Documente Academic
Documente Profesional
Documente Cultură
1 Wireline Component
Note Before using this information and the product it supports, read the information in Notices on page 219.
Copyright IBM Corporation 2006, 2010. US Government Users Restricted Rights Use, duplication or disclosure restricted by GSA ADP Schedule Contract with IBM Corp.
Contents
Preface . . . . . . . . . . . . . . vii
Audience . . . . . . Tivoli Netcool Performance Component . . . . . The Default UNIX Shell . . . . . . . . . . vii Manager - Wireline . . . . . . . . . vii . . . . . . . . . ix Install Oracle patches . . . . . . . . . . Relax permissions on ORACLE_HOME . . . . Set the ORACLE_SID variable . . . . . . . Set automatic startup of the database instance . . Configure the Oracle listener . . . . . . . Configure the Oracle net client . . . . . . . Shut down of performance impacting Oracle jobs Installing the Oracle client . . . . . . . . . Ensure that the Oracle User is not in NIS . . . Download the Oracle distribution and patch to disk . . . . . . . . . . . . . . . . Run the Oracle client configuration script . . . Set a password for the Oracle login name . . . Run the pre-installation script . . . . . . . Verify PATH and environment for the Oracle login name . . . . . . . . . . . . . Install the Oracle client . . . . . . . . . Run the root.sh script . . . . . . . . . . Install Oracle patches . . . . . . . . . . Relax permissions on ORACLE_HOME . . . . Configure the Oracle net client . . . . . . . Next steps . . . . . . . . . . . . . . . 43 44 45 45 46 48 50 50 50 51 52 54 54 55 56 58 58 60 60 62
Chapter 1. Introduction . . . . . . . . 1
Tivoli Netcool Performance Manager architecture . . 1 Co-location rules . . . . . . . . . . . . 2 Inheritance . . . . . . . . . . . . . . 4 Notable subcomponents and features . . . . . 5 Typical installation topology . . . . . . . . . 8 Basic topology scenario . . . . . . . . . . 8 Intermediate topology scenario . . . . . . . 9 Advanced topology scenario. . . . . . . . 10 Tivoli Netcool Performance Manager distribution . . 11
. . . . . . . . . . . . . . . .
13 14 14 14 15 18 20 21 23 23 24 24 25 25 25 26
. 26 27 . 27 . 28 . 28 . 29 . 30 . 30 . . . . . . 31 32 33 36 36 37
. 38 . 38 . 42
Overview . . . . . . . . Before you begin . . . . . Special consideration . . . Overriding default values . Installing a minimal deployment Download the MIB-II files . Starting the Launchpad . . Start the installation . . . The post-installation script . . Next steps . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
89 89 90 90 91 91 91 92 94 94
Restrictions and behavior . Performing the uninstall. . Uninstalling the topology editor Residual files . . . . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
Appendix B. DataChannels
. . . . . 141
. 141 . 141 142 . 143 . 144 . 145 . 146 . 146 . 147
Data collection . . . . . . . . . . . . Data aggregation . . . . . . . . . . Management programs and watchdog scripts DataChannel application programs . . . . Starting the DataLoad SNMP collector . . . . DataChannel management components in a distributed configuration . . . . . . . . Manually starting the Channel Manager programs . . . . . . . . . . . . . Adding DataChannels to an existing system . . DataChannel terminology . . . . . . . .
115 116 117 118 118 118 119 120 121 123 124 124 124 125 . 125 . 126 . 126
159
. 160
129
. 129 . 129 . 130 . 131 . 131
iv
Verifying the DataView installation . . . . . 176 Assigning Tivoli Netcool Performance Manager roles to LDAP users . . . . . . . . . . 176
. 206
183
. . . . . 183 183 183 184 184
187
. . . . . . . . 187 187 200 204 205 205 206 206
Contents
vi
Preface
The purpose of this manual IBM Tivoli Netcool Performance Manager 1.3.1 is a bundled product consisting of a wireline component and a wireless component. The purpose of this guide is to help you install theTivoli Netcool Performance Manager product suite and the Oracle database management system. This guide provides instructions for installing Tivoli Netcool Performance Manager components, but not necessarily for configuring the installed components into a finished system that produces management reports. After going through the steps in this guide, you will have a set of running Tivoli Netcool Performance Manager components ready to configure into a fully functional system. The goal of this guide is to get each component installed and running in its barest form. The running component does not necessarily have network statistical data flowing into and out of it yet. In particular, at the end of this installation procedure, there are no or few management reports that can be viewed in DataView. Configuring installed components into a working system is the subject of other manuals in the Tivoli Netcool Performance Manager documentation set.
Audience
The audience for this manual. The audience for this manual is the network administrator or operations specialist responsible for installing the Tivoli Netcool Performance Manager product suite on an enterprise network. To install Tivoli Netcool Performance Manager successfully, you should have a thorough understanding of the following subjects: v Basic principles of TCP/IP networks and network management v SNMP concepts v Administration of the Linux, Solaris or AIX operating environment v Administration of the Oracle database management system v Tivoli Netcool Performance Manager
vii
v DataLoad provides flexible, distributed data collection and data import of SNMP and non-SNMP data to a centralized database. v DataChannel aggregates the data collected through Tivoli Netcool Performance Manager DataLoad for use by the Tivoli Netcool Performance Manager DataView reporting functions. It also processes online calculations and detects real-time threshold violations. v DataView is a reliable application server for on-demand, web-based network reports. v Technology Packs extend the Tivoli Netcool Performance Manager system with service-ready reports for network operations, business development, and customer viewing. The following figure shows the different Tivoli Netcool Performance Manager modules.
Tivoli Netcool Performance Manager documentation consists of the following: v v v v v Release notes Configuration recommendations User guides Technical notes Online help
The documentation is available for viewing and downloading on the information center at http://publib.boulder.ibm.com/infocenter/tivihelp/v8r1/topic/ com.ibm.netcool_pm.doc/welcome_tnpm.html.
viii
Preface
ix
Chapter 1. Introduction
Introduction to Tivoli Netcool Performance Manager installation. This chapter provides an overview of the Tivoli Netcool Performance Manager product suite and provides important pre-installation setup information. In addition, this chapter provides an overview of the installation interface introduced in version 1.3.1.
The Tivoli Netcool Performance Manager system components are as follows: v Tivoli Netcool Performance Manager database - The Tivoli Netcool Performance Manager database is hosted on Oracle.
v Tivoli Netcool Performance Manager DataMart - Tivoli Netcool Performance Manager DataMart is the user and administrative interface to the Tivoli Netcool Performance Manager database and to other Tivoli Netcool Performance Manager components. v Tivoli Netcool Performance Manager DataLoad - Tivoli Netcool Performance Manager DataLoad consists of one or more components that collect network statistical raw data from network devices and from network management systems. v Tivoli Netcool Performance Manager DataChannel - Tivoli Netcool Performance Manager DataChannel is a collection of components that collect data from DataLoad collectors, aggregate and process the data, and load the data into the Tivoli Netcool Performance Manager database. DataChannel components also serve as the escalation point for collected data that is determined to be over threshold limits. v Tivoli Netcool Performance Manager DataView - Tivoli Netcool Performance Manager DataView is the Web server hosting and analysis platform. This platform is used to display Web-based management reports based on network data aggregated and placed in the Tivoli Netcool Performance Manager database. v Tivoli Netcool Performance Manager Technology Packs - Each technology pack is a set of components that describes the format and structure of network statistical data generated by network devices. Each technology pack is specific for a particular device, or class of devices; or for a particular company's devices; or for a protocol (such as standard SNMP values) common to many devices. v Tivoli Integrated Portal - The Tivoli Integrated Portal application provides a database-aware Web server foundation for the Web-based management reports displayed by Tivoli Netcool Performance Manager DataView. The Tivoli Integrated Portal application server is an essential component of each DataView installation.
Platform support
All components of the Tivoli Netcool Performance Manager - Wireline Component can run on a heterogenous mix of AIX, Linux, and Solaris operating systems.
Co-location rules
Allowed component deployment numbers and co-location rules. Table 1 lists how many of each component can be deployed per Tivoli Netcool Performance Manager system and whether multiple instances can be installed on the same server. In this table: v N - Depends on how many subchannels there are per channel, and how many channels there are per system. For example, if there are 40 subchannels per channel and 8 channels, theoretically N=320. However, the practical limit is probably much lower. v System - The entire Tivoli Netcool Performance Manager system. v Per host - A single physical host can be partitioned using zones, which effectively gives you multiple hosts. Note: All CME, DLDR, FTE, and LDR components within a channel must share the same filesystem.
Component AMGR
Number of Instances Allowed One per host that supports DataChannel components v N per system v One per corresponding subchannel
Co-Location Constraints
BCOL
Yes
One per subchannel One per system One per system One per DataChannel; maximum of 8 v N per system v One per corresponding subchannel v One per host
Filesystem
Yes
DataMart
Yes
DataView
v N per system v One per host Co-locate with corresponding DataMart Filesystem Filesystem
Discovery Server
Yes
One per channel One per subchannel N+M per system, where N is the number of collectors that HAM is monitoring, and M is the number of standby collectors One per channel One per system v N per system v One per corresponding subchannel
Filesystem
UBA (complex)
Pack-dependent
Pack-dependent
Pack-dependent
Chapter 1. Introduction
v In the Logical view of the Topology Editor, the DataChannel component contains the subchannels, LDR, and DLDR components, with a maximum of 8 channels per system. The subchannel contains the collector, FTE, and CME, with a maximum of 40 subchannels per channel.
Inheritance
Inheritance is the method by which a parent object propagates its property values to a child component. The following rules should be kept in mind when dealing with these properties. v A Child Property can be read only, but is not always. v If the Child Property is not read only, then it can be changed to a value different from the Parent Property. v If the Parent Property changes, and the Child and Parent properties were the same before the change, then the child property will be changed to reflect the new Parent Property value v If the Child Property changes, the Parent Property value will not be updated v The Default Value of the Child Property is always the current Parent Property value
Example
As an example of how a new component inherits property values: The Disk Usage Server (DUS) is a child component of the Host object. The DUS Remote User property inherits its value from the Host PV User Property on creation of the DUS. The DUS property value will be taken from the Host property value. Child properties that have been inherited are marked as inherited. As an example of what happens when you change inherited property values: If we change the Host PV User Property value, it gets pushed down to the DUS Remote User property value, updating it. The asscociated Default Value is also updated. If we change the DUS Remote User property value, that is the child value, it does not propogate up to the host; the parent Host PV User Property value remains unchanged. Now the child and parent properties are out of sync, and if we change the parent property value it is not reflected in the child property, though the default value continues to be updated.
Collectors
Collectors description. The DataLoad collector takes in the unrefined network data and stores it in a file that Tivoli Netcool Performance Manager can read. This file is known as a binary object format file (BOF). The following processes are employed in the DataLoad module: v SNMP Collector - The DataLoad SNMP Collector sends SNMP requests to network objects. Only the data requested by the configuration that was defined for those network objects is retrieved. v Bulk Collector - The Bulk Collector uses a Bulk Adaptor, which is individually written for specific network resources, to format the unrefined data into a file, called a PVline file, which is passed to the Bulk Collector. Installation or topology considerations: Installation and topology considerations for collectors. The DataLoad modules can be loaded on lightweight servers and placed as close to the network as possible (often inside the network firewall). Because a DataLoad module does not contain a database, the hardware can be relatively inexpensive and can still reliably handle high volumes of data. Up to 320 DataLoad modules can be supported per Tivoli Netcool Performance Manager installation. The number of collectors in your system will affect the topology configuration. You can have multiple BULK collectors, UBA or BCOL, on a single host, but you can only have one SNMP based collector per host. The number of collectors is in turn driven by the number of required Technology Packs.
Technology packs
Technology packs description. Tivoli Netcool Performance Manager Technology Packs are custom designed collections of MIBs, discovery formulas, collection formulas, complex formulas, grouping rules, reporters, and other functions. Technology packs provide all Tivoli Netcool Performance Manager needs to gather data for targeted devices. Technology packs make it possible for Tivoli Netcool Performance Manager to report on technology from multiple vendors. Installation or topology considerations: Installation and topology considerations for technology packs. If you are creating a UBA collector, you must associate it with a specific technology pack.
Chapter 1. Introduction
Note: General installation information for technology packs can be found in the IBM Tivoli Netcool Performance Manager: Technology Pack Installation Guide, pack-specific installation guides are also provided. Please consult both sets of documentation for important installation or topology information.
High Availability
High Availability description. High availability can be implemented for Tivoli Netcool Performance Manager in two forms: v High Availability Manager (HAM): This is a DataChannel component that can be configured to handle availability of SNMP collectors. v Veritas Cluster or Sun Cluster (referred to as HA within the documentation): This method of implementing high availability has a much broader scope and can cover all or a combination of the database, DataChannel, DataMart and DataView components. The following High Availability (HA) documents are available for download from the Tivoli Open Process Automation Library (OPAL), http://www-01.ibm.com/ software/brandcatalog/opal/. v TNPM High Availability Overview Describes high availability solutions for the Tivoli Netcool Performance Manager product. v Sun Cluster TNPM Agent Guide Describes how Sun Clusters can be used with TNPM to created a high availability Tivoli Netcool Performance Manager system. v High Availability Operations and Deployment Describes an example system that was configured to provide high availability. v TNPM High Availability Installation and Configuration Describes the steps necessary to install and configure components of Tivoli Netcool Performance Manager in a highly available configuration. For information covering the High Availability Manager, see Chapter 6, Using the High Availability Manager, on page 105 Installation or topology considerations: Installation and topology considerations for the High Availability Manager. The HAM must be put on the same machine as the channel manager.
v Disk space is running low: Disk space may be impacted by the addition of a new DataChannel component. In which case, the user may want to add a new file system managed by a new Disk Usage Server. v Separate disk quota management: The user may want to separately manage the quotas assigned to discrete DataChannel components. For more information, see Disk quota management. The user can assign the management of a new file system to a Disk Usage Server by editing the local_root_directory property of that Disk Usage Server using the Topology Editor. The user can then add DataChannel components to the host, and can assign the component to a Disk Usage Server, either in the creation wizard or by editing the DUS_NUMBER property inside the component. Disk quota management: Disk Quota Management description. The addition of a Disk Usage Server endeavors to make the process of assigning space to a component much easier than it has been previously. No longer is a user required to calculate the requirements of each component and assign that space individually, but components now work together to more effectively utilize the space they have under the Disk Usage Server. Also, the user is relieved of trying to figure out which component needs extra space and then changing the quota for that component. Now, the user can just change the quota of the DUS and all components on that Disk Usage Server will get the update and share the space on an as needed basis. Good judgement of space requirements is still needed. However, the estimating of space requirements is being made at a higher level; and should an estimate be incorrect, only one number needs to be changed instead of potentially updating the quota for each component separately. Flow control: Flow Control description. Optimized flow control further eliminates problems with component level quotas. Each component holds on to only a five hours of input and output, and once it has reached this limit, it stops processing until the downstream component picks up some of the data. This avoids the cascading scenario where one component stops processing and the components feeding it begin to stockpile files, which results in the quota being filled and causes all components to shut down because they have run out of file space. Installation or topology considerations: Installation or Topology considerations for flow control. DataChannel components can only be added to hosts that include a Disk Usage Server.
Chapter 1. Introduction
Notes Install the Topology Editor and primary deployer on this system.
Notes Install the Topology Editor and primary deployer on this system.
Chapter 1. Introduction
This scenario has an added copy of both collectors on corinth to a second machine, thessaloniki, for the purposes of failover. HAM only manages SNMP collectors; therefore, the HAM in this scenario will manage availability of the DataLoad SNMP collector and not the Bulk Load collector. The HAM must be put on the same machine as the channel manager.
Notes Install the Topology Editor and primary deployer on this system.
10
Table 4. Tivoli Netcool Performance Manager advanced topology scenario (continued) Tivoli Netcool Performance Manager Components Hosted v Oracle client v Tivoli Integrated Portal v Tivoli Netcool Performance Manager DataView
Notes You could install Tivoli Netcool Performance Manager components remotely on this system. Your configuration can use a pre-existing Tivoli Integrated Portal, or install and include a new instance. You could install Tivoli Netcool Performance Manager components remotely on this system. Your configuration can use a pre-existing Tivoli Integrated Portal, or install and include a new instance.
rhodes
v Oracle client v Tivoli Integrated Portal v Tivoli Netcool Performance Manager DataView
Chapter 1. Introduction
11
12
Overview
Before beginning the Tivoli Netcool Performance Manager installation, you must install the prerequisite software listed in the IBM Tivoli Netcool Performance Manager: Configuration Recommendations Guide. The required software includes: v Oracle server: To use Oracle with Tivoli Netcool Performance Manager, you must install Oracle as described in this chapter - do not use a separate Oracle installation method provided by Oracle Corporation. v Oracle client: You must install Oracle client software on each system where you plan to install a Tivoli Netcool Performance Manager component, except for the system where you installed the Oracle server. When you complete the steps in this chapter, the Oracle server and client will be installed and running, with tablespaces sized and ready to accept the installation of a Tivoli Netcool Performance Manager DataMart database. You can communicate with Oracle using the SQLPlus command-line utility. The steps in this chapter use IBM-provided installation scripts to install and configure the Oracle database from the Oracle distribution and patch. For use with Tivoli Netcool Performance Manager, you must install Oracle as described in this chapter. Do not use a separate Oracle installation method provided by Oracle Corporation. You should obtain the official Oracle distribution from your edelivery site (after purchase of an Oracle license). See the IBM Tivoli Netcool Performance Manager: Configuration Recommendations Guide for recommendations when purchasing a license from Oracle. Note: The Tivoli Netcool Performance Manager script used to install Oracle is platform-independent and can be used to install on Solaris, AIX, or Linux, regardless of the operating system distribution media. v OpenSSH: You must install and configure OpenSSH before installing Tivoli Netcool Performance Manager. For details, see Appendix E, Secure file transfer installation, on page 163. Linux systems require the installation of VSFTP (Very Secure FTP). v Web browser: The launchpad requires a Web browser. IBM recommends using Mozilla with the launchpad. For the complete list of supported browsers, see the IBM Tivoli Netcool Performance Manager: Configuration Recommendations Guide document. v Java: Java is used by DataMart, DataLoad, and the technology packs. You must ensure you are using the IBM JRE and not the RHEL JRE. The IBM JRE is supplied with the Topology Editor or with TIP. To ensure you are using the right JRE you can either:
13
Set the JRE path to conform to that used by the Topology Editor, do this using the following commands (using the default location for the primary deployer):
PATH=/opt/IBM/proviso/topologyEditor/jre/bin:$PATH export $PATH
For a remote server, that is one that does not host the primary deployer, you must download and install the required JRE, and set the correct JRE path. See the IBM Tivoli Netcool Performance Manager: Configuration Recommendations Guide document for JRE download details. Note: See the IBM Tivoli Netcool Performance Manager: Configuration Recommendations Guide document for the complete list of prerequisite software and their supported versions.
Supported platforms
The platforms supported by Tivoli Netcool Performance Manager. Refer to the following table for platform requirement information.
Tivoli Netcool Performance Manager Component All Tivoli Netcool Performance Manager Components: v Database v DataView v DataChannel v DataLoad v DataMart Required Oracle Software v Solaris 10 64-bit v AIX 6.1 64-bit v RHEL 5.5, 64-bit
14
Procedure
In general, set DISPLAY as follows:
$ DISPLAY=Host_IP_Address:0.0 $ export DISPLAY
To make sure the DISPLAY environment variable is set, use the echo command:
$ echo $DISPLAY
Procedure
1. Set the DISPLAY environment variable. 2. Enter the following command when logged in as root:
# /usr/openwin/bin/xhost +
Note: Disabling access control is what enables access to the current machine from X clients on other machines.
AIX systems
Changing ethernet characteristics on AIX.
15
Procedure
1. Using the System Management Interface Tool (SMIT), navigate to Devices > Communication > Ethernet Adapter > Change/Show Characteristics of an Ethernet Adapter. Select your ethernet adapter (the default is ent0). Change the Media Speed setting to 100_Full_Duplex. Change the setting Apply change to DATABASE only to yes. Set the port on the switch or router that the AIX node is plugged into to 100_Full_Duplex. 6. Reboot your system. 2. 3. 4. 5.
Solaris systems
This section describes how to set a network interface card (NIC) and a BGE network driver to full duplex mode. NIC: Change the NIC to full duplex mode on Solaris systems About this task To change the NIC to full duplex mode: Procedure 1. Determine which type of adapter you have by running the following command:
ifconfig -a
2. To determine the current settings of the NIC, run the command ndd -get /dev/hme with one of the following parameters:
Command Parameter link_status v 1 - Up v 0 - Down link_speed Determines the link speed v 0 - 10Mb/sec v 1 - 100Mb/sec link_mode Determines the duplex mode v 0 - Half duplex v 1 - Full duplex adv_autoneg_cap Determines whether auto negotiation is on v 0 - Off v 1 - On Description Determines whether the link is up
For example:
ndd -get /dev/hme link_status
In these commands, /dev/hme is your NIC; you might need to substitute your own /dev/xxx. 3. To set your NIC to 100Mb/s with full duplex for the current session, run the following commands:
16
ndd -set /dev/hme adv_100hdx_cap 0 ndd -set /dev/hme adv_100fdx_cap 1 ndd -set /dev/hme adv_autoneg_cap 0
However, these commands change the NIC settings for the current session only. If you reboot, the settings will be lost. To make the settings permanent, edit the /etc/system file and add the following entries:
set hme:hme_adv_autoneg_cap=0 set hme:hme_adv_100hdx_cap=0 set hme:hme_adv_100fdx_cap=1
4. Verify that your NIC is functioning as required by rerunning the commands listed in Step 2. BGE network driver: Change a BGE network driver to full duplex mode. About this task To change a BGE network driver to full duplex mode. Procedure 1. To determine the link speed and current duplex setting, run the following command:
% kstat bge:0 | egrep speed|duplex
2. Create a file namedbge.conf in the /platform/uname -i/kernel/drv directory (for example, /platform/SUNW,Sun-Fire-V210/kernel/drv/bge.conf). 3. Add the following lines to the file:
speed=100; full duplex=1;
Linux systems
Enabling 100 full duplex mode on Linux systems.
17
Procedure
1. Enter the following command:
# dmesg | grep -i duplex
If this is not contained within the output, you must enable full duplex mode. The example output resulting from the command executed in step 1: eth0: link up, 100Mbps, full-duplex, lpa 0x45E1 indicate that the primary network interface is eth0. The actions specified in the following process presume that your primary network interface is eth0. Enabling full duplex mode on Linux: To enable full duplex mode. Procedure 1. Open the file ifcfg-eth0, which is contained in:
/etc/sysconfig/network-scripts/
Note: The ETHTOOL_OPTS speed setting can be set to either 100 or 1000 depending on speed of connection available 100Mbit/s or 1000Mbit/s (1Gbit/s).
18
Procedure
1. Log in as root. 2. Set and export the DISPLAY environment variable. (seeSetting up a remote X Window display on page 14.) 3. If one does not already exist, create a group to which you can add pvuser. You can create a group with the name of your choice using the following command:
groupadd <group>
where: v <group> is the name of the new group, for example, staff. 4. At a shell prompt, run the following command:
# useradd -g <group> -m -d <home_dir>/<username> -k /etc/skel -s /bin/ksh <username>
Where: v <group> is the name of the group to which you want to add pvuser. v <home_dir> is the home directory for the new user, for example, /export/home/ can be used as the example home directory. v <username> is the name of the new user. This can be set to any string. Note: For the remainder of this document this user will be referred to as pvuser. 5. Set a password for pvuser:
# passwd pvuser
The system prompts you to specify a new password twice. The default pvuser password assumed by the Tivoli Netcool Performance Manager installer is pv. This can be set to a password conforming to your organization's standards. 6. Test logging in as pvuser, either by logging out and back in, or with the su command, such as:
# su - pvuser
These instructions create a pvuser login name with the following attributes:
Attribute login name member of group pvuser staff Value
19
Attribute home directory login shell copy skeleton setup files (.profile, and so on) from this directory
Note: The pvuser account must have write access to the /tmp directory.
When you create the pvuser login name on the next computer, add the -u option to the useradd command to specify the same user ID number:
# useradd -g <group> -m -d <home_dir>/pvuser -k /etc/skel -s /bin/ksh -u 1001 pvuser
Where: v <group> is the name of the group to which you want to add pvuser. v <home_dir> is the home directory for the new user, for example, /export/home/ can be used as the example home directory. v <username> is the name of the new user. This can be set to any string.
Procedure
1. Log in as root. 2. Change your working directory to /etc/security by entering the following command:
20
# cd /etc/security
3. Make a backup copy of the limits file by entering the following command:
# cp limits limits.ORIG
4. Using a text editor, open the limits file and set the following values:
default: fsize = -1 core = -1 cpu = -1 data = -1 rss = 65536 stack = 65536 nofiles = 2000 totalProcesses = 800
Note: Apply these settings to every AIX system running a Tivoli Netcool Performance Manager program: the database server, DataLoad servers, DataChannel servers, and DataMart servers. 5. Write and quit the file. 6. After modifying the settings, log off every Tivoli Netcool Performance Manager user and then log in again for the changes to take effect.
Procedure
1. Set the NOEXEC_USER_STACK parameter in the system file: a. Log in as root. b. Change to the /etc directory:
# cd /etc
c. Create a backup of the system file, then open the system file with a text editor. d. Set the parameter NOEXEC_USER_STACK to 1, by adding the following line at the bottom of the file:
set NOEXEC_USER_STACK=1
e. Save and exit the system file. 2. Set resource controls correctly. The parameters affected by the deployment size are project.max-sem-ids, process.max-sem-nsems, project.max-shm-memory, and project.max-shm-ids. These parameters define the maximum size of a semaphore set and the maximum number of semaphores in the system. a. In Solaris 10, kernel parameters are replaced by resource controls. See Oracle Database Installation Guide 10g Release 2 (10.2) for Solaris Operating System (SPARC 64-Bit) Part Number B15690-02
Chapter 2. Installing and configuring the prerequisite software
21
(http://download.oracle.com/docs/cd/B19306_01/install.102/b15690/ pre_install.htm#sthref259), section 2.6: Configuring Kernel Parameters. See also Oracle Metalink ID 169706.1, Oracle Database on Unix AIX, HP-UX, Linux, Mac OS X, Solaris, Tru64 Unix Operating Systems Installation and Configuration Requirements Quick Reference (8.0.5 to 11.2), which lists Solaris requirements. b. Oracle recommends the following values, noting that they are guidelines and should be tuned for production database systems. If you use a custom configuration, you must change the values of the parameters to the appropriate level.
Resource Control project.max-sem-ids process.max-sem-nsems project.max-shm-memory project.max-shm-ids 100 256 429496725 100 Recommended Value
c. Log in as the Oracle user (for example, oracle). d. To find the current kernel parameter settings, check the project id, and then check the resource control settings for that project id:
$ id -p uid=4074(oracle) gid=9999(dba) projid=3(default) $ prctl -n project.max-shm-memory -i project 3 project: 3: default NAME PRIVILEGE VALUE FLAG ACTION project.max-shm-memory privileged 1.95GB - deny system 16.0EB max deny $ prctl -n project.max-sem-ids -i project 3 project: 1: user.root NAME PRIVILEGE VALUE FLAG ACTION project.max-sem-ids privileged 128 - deny system 16.8M max deny $ prctl -n project.max-shm-ids -i project 3 project: 3: default NAME PRIVILEGE VALUE FLAG ACTION project.max-shm-ids privileged 128 - deny system 16.8M max deny $ prctl -n process.max-sem-nsems $$ process: 12134: bash NAME PRIVILEGE VALUE FLAG ACTION process.max-sem-nsems privileged 600 - deny system 32.8M max deny
RECIPIENT -
RECIPIENT -
RECIPIENT -
RECIPIENT -
e. To change values, check the Solaris documentation for complete information on projects. Here is one example, which sets the value of project.max-shm-memory to 4GB. Log in as root and add a project, attached to the dba group (assuming the oracle user is part of the dba group), and set the value:
# projadd -p 100 -G dba -c "Oracle Project" \ -K "project.max-shm-memory=(privileged,4G,deny)" group.dba
f. Check by logging back in as oracle, checking with id -p that the projid is now the new project number 100, and run prctl again to check that the max-shm-memory value has been updated.
22
$ id -p uid=4074(oracle) gid=9999(dba) projid=100(group.dba) bash-3.00$ prctl -n project.max-shm-memory -i project 100 project: 100: group.dba NAME PRIVILEGE VALUE FLAG ACTION project.max-shm-memory privileged 4.00GB - deny system 16.0EB max deny
RECIPIENT -
Procedure
1. Log in as root: 2. Change to the following directory:
# /etc/init.d
Procedure
1. Open the SELinux config file for editing:
$ cat /etc/selinux/config
To:
SELINUX=disabled
Note: You can also set the SELINUX setting to permissive. Setting SELINUX to permissive will result in a number of warnings at install time, but it will allow the installation code to run.
23
Procedure
1. Add the following the lines in the file /etc/sysctl.conf v kernel.shmall = physical RAM size / pagesize For most systems, this will be the value 2097152. See Note 301830.1, which is available from the Oracle website, for more information. v kernel.shmmax = 1/2 of physical RAM, but not greater than 4GB. This would be the value 2147483648 for a system with 4Gb of physical RAM. v kernel.shmmni = 4096 v kernel.sem = 250 32000 100 128 v fs.file-max = 512 x processes (for example 65536 for 128 processes) v net.ipv4.ip_local_port_range =9000 65500 v net.core.rmem_default = 262144 v net.core.rmem_max = 2097152 v net.core.wmem_default = 262144 v net.core.wmem_max = 1048576 v fs.aio-max-nr = 1048576 2. To effect these changes, execute the command:
# sysctl -p
Install a libcrypto.so
For full SNMPv3 support, SNMP DataLoad must have access to the libcrypto.so.
Procedure
1. Install the OpenSSL package. This package can be downloaded from http://www.openssl.org/. 2. As root, extract and install the libcrypto.so file using the following code:
# cd /usr/lib # ar -xv ./libcrypto.a # ln -s libcrypto.so.0.9.8 libcrypto.so
3. Update the dataload.env file so that the LD_LIBRARY_PATH (on Solaris & Linux) or LIBPATH (on AIX) environment variables include the path:
/ProvisoAutomation/proviso/thirdparty/openssl.org/openssl-0.9.8o/{YourOS}/lib
24
What to do next
Check the variable has been set by doing the following: 1. Open a fresh shell 2. Check the dataload.env file. 3. Bounce the SNMP DL Upon startup, with a valid library, the collector will log the following log messages:
INFO:CRYPTOLIB_LOADED Library libcrypto.so (OpenSSL 0.9.8e-fips-rhel5 01 Jul 2008, 0x90802f) has been loaded. INFO:SNMPV3_SUPPORT_OK Full SNMPv3 support Auth(None,MD5,SHA-1) x Priv(None,DES,AES) is available.
Deployer pre-requisites
Minimum filesystem specification and pre-requisites for the Deployer script. The Deployer will check the for the items described under the following headings. You should ensure that all elements are installed before running the deployer.
25
Procedure
1. On the target host, log in as the Tivoli Netcool Performance Manager user, such as pvuser. 2. Create a directory to hold the contents of your Tivoli Netcool Performance Manager distribution. For example:
$ mkdir /var/tmp/cdproviso
Note: Any further references to this directory within the install will be made using the token <DIST_DIR>. You will run a variety of scripts and programs from directories residing in the directory created on the hard drive, including: v Oracle configuration script v Pre-installation script v Installation script v Tivoli Netcool Performance Manager setup program 3. Download the Tivoli Netcool Performance Manager distribution to the host directory created in the previous step and expand the contents of the distribution package. 4. If the current host is the primary deployer do the following: a. Change to the following directory: On Solaris systems:
$ cd <DIST_DIR>/proviso/SOLARIS/DataChannel/SOL10/sparc
On AIX systems:
$ cd <DIST_DIR>/proviso/AIX/DataChannel/AIX/powerpc
On Linux systems:
$ cd <DIST_DIR>/proviso/RHEL/DataChannel/RHEL5/i686/
26
Procedure
1. Create a folder named TCR as a peer to the other Tivoli Netcool Performance Manager Components, that is, DataView, DataChannel, etc. For example:
<DIST_DIR>/proviso/SOLARIS/TCR
2. Extract the TCR 2.1 inside this folder. Should the user decide not to extract the tar as a peer to the other components, a TCR folder must still be created having a path to the TCR install.sh the same as: ./TCR/TCRInstaller/install.sh Note: If the user extracts the tar directly into the same root location as the Tivoli Netcool Performance Manager Components then the TCR launchpad.sh will overwrite the Tivoli Netcool Performance Manager Installer launchpad.sh, meaning the launchpad cannot be started for the installer.
27
Procedure
Provide a basename, which the installation retains as the variable DB_USER_ROOT. Note: This is not an operating system environment variable, but a variable used internally by the installer. The default DB_USER_ROOT value is PV. IBM strongly encourages you to retain the default value.
Results
Oracle login names are generated from the DB_USER_ROOT basename by appending a function or subsystem identifier to the basename, as in the following examples: v PV_ADMIN v PV_INSTALL v PV_LDR v PV_CHANNEL v PV_COLL v PV_CHNL_MANAGER v PV_GUI In addition, separate Oracle login names are generated for each Tivoli Netcool Performance Manager DataChannel and subsystem, identified by an appended channel number, as in the following examples: v PV_CHANNEL_01 v PV_CHANNEL_02 v PV_LDR_01 v PV_LDR_02
28
Procedure
You can retain the default password, or enter passwords of your own according to your site password standards. You should use the same password for all Tivoli Netcool Performance Manager subsystem Oracle login names. If you use different passwords for each login name, keep a record of the passwords you assign to each login name.
Results
The Tivoli Netcool Performance Manager installer uses PV for three default values, as described in Table 5. Table 5: Uses of PV as Default Values
Installer Default Value PV Used As Default value of the DB_USER_ROOT variable, the basename on which Oracle login names are generated Default password for all Oracle login names Default Oracle database name, also called the Oracle TNS name Recommendation In all instances, use the default value PV, unless your site has an explicit naming standard or an explicit password policy.
PV or pv PV
What to do next
Note: If you use a non-default value, you must remember to use the same value in all installation stages. For example, if you set your Oracle TNS name to PROV instead of PV, you must override the default PV entry in all subsequent steps that call for the TNS name.
Assumed values
The steps in this chapter assume the following default values:
Setting Hostname of the Oracle server Oracle server program files installed in ORACLE_BASE = Value Assumed in this Chapter delphi /opt/oracle /opt/oracle
Operating system login name for Oracle user oracle Note: The default name created is oracle. However, you can set another name for the Oracle user. Password for Oracle user ORACLE_SID = TNS name for Tivoli Netcool Performance Manager database instance oracle PV PV
29
Value Assumed in this Chapter /opt/oracle/product/n Note: The value of ORACLE_HOME cannot contain soft links to other directories or filesystems. Be sure to specify the entire absolute path to Oracle. Tivoli Netcool Performance Manager expects an Optimal Flexible Architecture (OFA) structure where ORACLE_HOME is a sub-directory to ORACLE_BASE.
Oracle login name for database administrator (DBA) Password for Oracle DBA login name DB_USER_ROOT = Path for Oracle data, mount point 1 Path for Oracle data, mount point 2
Note: If your site has established naming or password conventions, you can substitute site-specific values for these settings. However, IBM strongly recommends using the default values the first time you install Tivoli Netcool Performance Manager. See Specifying a basename for DB_USER_ROOT on page 28 for more information.
Procedure
1. On the system where the Oracle database is to be installed, disable NIS. For more information, see your operating system documentation.
30
2. Run the configure_ora script to create a local oracle account (see Run the Oracle server configuration script on page 33). 3. Re-enable NIS. Note: The local Oracle account should be used before the NIS Oracle user.
Procedure
1. Log in as root. 2. Create a directory to hold the contents of the Oracle distribution. For example:
# mkdir /var/tmp/oracle10201 # mkdir /var/tmp/oracle10204
3. Download the Oracle files to the /var/tmp/oracle10201 directory. 4. Locate the appropriate upgrade patch file for your version of Oracle on the Oracle Web site and download it to your /var/tmp/oracle10204 directory. 5. Unzip the oracle distribution files that now reside in the /var/tmp/oracle10201 and /var/tmp/oracle10204 directories. Before you proceed to the next step, make sure that you obtain the upgrade instructions provided by Oracle for the patch. The instructions contain information on performing steps required for the upgrade that are not documented in this guide. If you are installing the Oracle patch on an AIX system, you might need to run the sbinclean command and execute the catpatch.sql and utlrp.sql SQL scripts as part of the upgrade procedure. Additional requirements might exist for other platforms. See your database administrator to determine whether there are any company-specific requirements for installing Oracle in your environment.
Procedure
1. Log in as root. 2. Create a directory to contain your Oracle 10.2.0.1.0 files. For example:
# mkdir /var/tmp/oracle10201
3. Download your gzipped 'cpio' Oracle files to this directory. 4. Oracle Server folder setup: a. Within the oracle 10.2.0.1.0 distribution directory you have created, create a subdirectory called 'database'. b. Gunzip the '10gr2_db_sol.cpio.gz' file that was downloaded from Oracle. c. Cpio extract '10gr2_db_sol.cpio' into the 'database' subdirectory
Chapter 2. Installing and configuring the prerequisite software
31
5. Oracle Client folder setup: a. Within the oracle 10.2.0.1.0 distribution directory you have created, create a subdirectory called 'client'. b. Gunzip the '10gr2_client_sol.cpio.gz ' file that was downloaded from Oracle. c. Cpio extract '10gr2_client_sol.cpio' into the 'client' subdirectory:
cd client cpio -idmv < 10gr2_client_sol.cpio)
Procedure
Make sure all the required Solaris packages and patches are installed on your system. All required packages and patches are specified in the IBM Tivoli Netcool Performance Manager: Configuration Recommendations Guide.
Procedure
Make sure all the required Solaris packages and patches are installed on your system. All required packages and patches are specified in the IBM Tivoli Netcool Performance Manager: Configuration Recommendations Guide.
On AIX systems
Verify that your AIX system has all prerequisite packages and patches. Before you install Oracle on AIX systems, verify that your system meets the required release level, as specified in the IBM Tivoli Netcool Performance Manager: Configuration Recommendations Guide document. Before installing the Oracle server, make sure all the required AIX packages and patches are installed on your system. All required packages and patches are specified in the IBM Tivoli Netcool Performance Manager: Configuration Recommendations Guide document.
On Linux systems
Verify that your Linux system has all prerequisite packages and patches. Before you install Oracle on Linux systems, verify that your system meets the required release level, and make sure all required Linux packages and patches are installed on your system. Release level and all required packages and patches are specified in the IBM Tivoli Netcool Performance Manager: Configuration
32
Recommendations Guide.
Procedure
1. Log in as root. 2. Set the ORACLE_BASE environment variable to point to the top-level directory where you want the Oracle server files installed. The default installation directory is /opt/oracle. This can be set to any directory required by your organization. For example:
# ORACLE_BASE=/opt/oracle # export ORACLE_BASE
Note: The configure_ora script places this variable into the oracle login account's .profile file. To check that the variable is set correctly, enter the following command:
# env | grep ORA
AIX systems:
# cd <DIST_DIR>/proviso/AIX/DataBase/AIX/instance
Linux systems:
# cd <DIST_DIR>/proviso/RHEL/DataBase/RHEL5/instance
where:
Chapter 2. Installing and configuring the prerequisite software
33
<DIST_DIR> is the directory on the hard drive where you copied the contents of the Tivoli Netcool Performance Manager distribution in Downloading the Tivoli Netcool Performance Manager distribution to disk on page 26. 4. Run the Oracle configuration script by entering the following command:
# ./configure_ora The following screen is displayed:
-------------------------------------------------configure_ora Setting the Oracle environment <Current Date> -------------------------------------------------OS ........... : [ SunOS 5.10 Generic ] Host ......... : [ delphi ] Logname ...... : [ root ] ORACLE_BASE .. : [ /opt/oracle ] DBA group ................. : [ dba ] OUI Inventory group ....... : [ oinstall ] Oracle Software owner ..... : [ oracle ] Configure Oracle release .. : [ 10.2.0 ] Menu : 1. Modify Oracle software owner. 2. Next supported release 3. Check environment. 0. Exit Choice:
5. (Optional) To create a different name for the Oracle user other than the default oracle, type 1 and follow the instructions. Note: The Oracle user will be referred to as "oracle" for the remainder of this document. If you have set the username to be a non-default string, that is, something other than "oracle" , make sure to substitute your non- default string for each occurrence of "oracle" within any task description. 6. Type 3 at the Choice prompt and press Enter. The script creates the dba and oinstall groups and the ORACLE_BASE directory, unless they already exist:
Checking environment... Checking for group [ dba ] --> Created. Checking for group [ oinstall ] --> Created. Checking ORACLE_BASE ** WARNING ** ORACLE_BASE directory does not exist. ** [ /opt/oracle ] ** ** Create it ? (n/y) y
7. Type y and press Enter. The script creates the /opt/oracle directory and continues as follows:
Checking for user [ oracle ] ** WARNING ** User [ oracle ] does not exist. ** ** Create it locally ? (n/y) y
8. Type y and press Enter. The script creates the oracle user and continues as follows:
34
--> Created. Checking for oracle directory tree : [ /opt/oracle/product ] --> Created. [ /opt/oracle/product/10.2.0 ] --> Created. [ /opt/oracle/product/10.2.0/dbs ] --> Created. [ /opt/oracle/admin ] --> Created. [ /opt/oracle/admin/skeleton ] --> Created. [ /opt/oracle/admin/skeleton/lib ] --> Ok. [ /opt/oracle/admin/skeleton/lib/libpvmextc.so ] --> Created. [ /opt/oracle/admin/skeleton/lib/libmultiTask.so ] --> Created. [ /opt/oracle/admin/skeleton/lib/libcmu.so ] --> Created. [ /opt/oracle/admin/skeleton/bin ] --> Ok. [ /opt/oracle/admin/skeleton/bin/snmptrap ] --> Created. [ /opt/oracle/local ] --> Created. Checking for oracle .profile file --> Created. Checking for dbora file --> Created. /etc/rc0.d/K10dbora link --> Created. /etc/rc1.d/K10dbora link --> Created. /etc/rc2.d/S99dbora link --> Created. Checking for dbora configuration files : /var/opt/oracle/oratab --> Created. /var/opt/oracle/lsnrtab --> Created. Press Enter to continue...
9. Press the Enter key to continue. The configure_ora main screen is refreshed. 10. Type 0 and press Enter to exit the configure_ora script. Note: You must set a password for the oracle login name (see Set a password for the Oracle login name on page 36).
The script creates the following setup files: Solaris specific files: v /etc/init.d/dbora, which starts the Oracle Listener and database server automatically on each system boot v Symbolic links to /etc/init.d/dbora in /etc/rc0.d, /etc/rc1.d, and /etc/rc2.d v Oracle configuration files /var/opt/oracle/oratab and lsnrtab. AIX specific files: v /etc/inittab is modified to contain the dbstart and lsnrctl startup calls. v /etc/rc.shutdown is modified to contain the dbshut and lsnrctl stop commands. v Oracle configuration files /etc/oratab and /etc/lsnrtab. Note: AIX does not use init.d. Common files:
Chapter 2. Installing and configuring the prerequisite software
35
v A .profile file for the oracle user containing the following lines:
# -- Begin Oracle Settings -umask 022 ORACLE_BASE=/opt/oracle ORACLE_HOME=$ORACLE_BASE/product/10.2.0 NLS_LANG=AMERICAN_AMERICA.WE8ISO8859P1 ORA_NLS33=$ORACLE_HOME/ocommon/nls/admin/data LD_LIBRARY_PATH=$ORACLE_HOME/lib TNS_ADMIN=$ORACLE_HOME/network/admin PATH=$PATH:$ORACLE_HOME/bin:/usr/ccs/bin EXTPROC_DLLS=ONLY:${LD_LIBRARY_PATH}/libpvmextc.so export PATH ORACLE_BASE ORACLE_HOME NLS_LANG export ORA_NLS33 LD_LIBRARY_PATH TNS_ADMIN export EXTPROC_DLLS # -- End Oracle Settings --
Note the following: v The value of ORACLE_HOME cannot contain soft links to other directories or filesystems. Be sure to specify the entire absolute path to Oracle. v You will add the ORACLE_SID variable to this file later, in Set the ORACLE_SID variable on page 45.
Procedure
1. Log in as root. 2. Enter the following command:
# passwd oracle
3. Enter and re-enter the password (oracle, by default) as prompted. The password is set.
Procedure
1. Log in as root. 2. Change to the following directory:
36
Solaris systems:
# cd <DIST_DIR>/proviso/SOLARIS/DataBase/SOL10/instance/ora_installer
AIX systems:
# cd <DIST_DIR>/proviso/AIX/DataBase/AIX/instance/ora_installer
Linux systems:
# cd <DIST_DIR>/proviso/RHEL/DataBase/RHEL5/instance/ora_installer
where: v <DIST_DIR> is the directory on the hard drive where you copied the contents of the Tivoli Netcool Performance Manager distribution in Downloading the Tivoli Netcool Performance Manager distribution to disk on page 26. 3. Set the ORACLE_BASE environment variable. For example:
# ORACLE_BASE=/opt/oracle # export ORACLE_BASE
You must use the same ORACLE_BASE setting that you specified in Run the Oracle server configuration script on page 33 4. Enter the following command:
# ./pre_install_as_root
If the script shows an error, correct the situation causing the error before proceeding to the next step.
Procedure
1. Log in as root or become superuser. 2. Set the DISPLAY environment variable. 3. Change to the directory /var/tmp/oracle10201/database/rootpre. Note: The Oracle server distribution is downloaded to /var/tmp/oracle10201 as per the instructions in the section Downloading the Tivoli Netcool Performance Manager distribution to disk on page 26. 4. Run the following command:
./rootpre.sh
This error can safely be ignored. Note: For more information on this Oracle error, see Oracle Metalink Article 282036.1.
37
Procedure
1. Log in as oracle. 2. Set and export the DISPLAY environment variable. If you are using the su command to become oracle, use a hyphen as the second argument so the oracle name's login environment is loaded:
# su - oracle
3. Verify that the environment variable ORACLE_BASE has been set by entering the following command:
$ env | grep ORA
If the response does not include ORACLE_BASE=/opt/oracle, stop and make sure the .profile file was set for the oracle user as described in Run the Oracle server configuration script on page 33. 4. To verify the path, enter the following command:
$ echo $PATH
The output should show that /usr/ccs/bin is part of the search path. For example:
/usr/bin:/opt/oracle/product/10.2.0/bin:/usr/ccs/bin
a. If the directory does not appear in the path, add it by entering the following commands:
$ PATH=$PATH:/usr/ccs/bin $ export PATH
Install Oracle
There are two ways to install the Oracle database files. v Using the menu-based script: The Oracle installation script provided by IBM is used to install Oracle server, Oracle client, and to install upgrade patches. You should follow this step if you are not familiar with the Oracle installation process. v Using batch mode on page 40: If you understand and are very familiar with the Oracle installation process, use batch mode. Note: Choose one method, either the menu based script or the batch mode, do not implement both.
38
Procedure
1. Log in as oracle. 2. Change to the following directory: Solaris systems:
# cd <DIST_DIR>/proviso/SOLARIS/DataBase/SOL10/instance/ora_installer
AIX systems:
# cd <DIST_DIR>/proviso/AIX/DataBase/AIX/instance/ora_installer
Linux systems:
# cd <DIST_DIR>/proviso/RHEL/DataBase/RHEL5/instance/ora_installer
where: v <DIST_DIR> is the directory on the hard drive where you copied the contents of the Tivoli Netcool Performance Manager distribution in Downloading the Tivoli Netcool Performance Manager distribution to disk on page 26. 3. Enter the following command to start the installer:
$ ./perform_oracle_inst
Menu : 1. Next supported release 2. Set install type to: Client 3. Perform install 0. Exit Choice :
4. Verify the following settings: v The Oracle release number should be 10.2.0. v The Installation type field should be set to Server. This field cycles between three settings: Server, Client, and Patch. Type 2 at the Choice prompt and press Enter until Server is displayed. 5. Type f at the Choice prompt and press Enter. 6. At the Choice prompt, enter the full path to the directory containing the installation files. For example:
Choice: f Enter new value for CD directory: /var/tmp/oracle10201
39
For example, if you used non-default values for ORACLE_BASE or ORACLE_HOME, enter your settings until the menu shows the correct information. 8. To begin the Oracle installation, type 3 at the Choice prompt and press Enter. The installation script checks the environment, then asks whether you want to perform the installation. 9. Type Y at the Choice prompt and press Enter. The installation script starts installing Oracle and displays a series of status messages. Note: You can safely ignore: any "font.properties not found" messages in the output, or any Linux operating system prerequisite failure, as this relates to a problem with the Oracle Universal Installer. For more information on this problem with the Oracle Universal Installer, see the Oracle technote "Prerequisite Checks Fail When Installing 10.2 On Red Hat 5 (RHEL5) [ID 456634.1]". When the installation reaches the In Summary Page stage, the installation slows down significantly while Oracle files are copied and linked. 10. When the installation is complete, messages similar to the following are displayed:
In End of Installation Page The installation of Oracle10 Database was successful. Please check /opt/oracle/oraInventory/logs/silentInstall2004-09-28_04-23-53PM.log for more details. The Oracle installation has completed. Please check the messages above to determine if the install completed successfully. If you do not see successful completion messages, consult the install log at: /opt/oracle/oraInventory/logs Press C to continue...
Note: Write down the log file location to aid in troubleshooting if there is an installation error. 11. Type C and press Enter to return to the installation menu. 12. Type 0 and press Enter to exit the installation menu.
40
Procedure
1. Log in as oracle. 2. Change to the following directory: Solaris systems:
# cd <DIST_DIR>/proviso/SOLARIS/DataBase/SOL10/instance/ora_installer
AIX systems:
# cd <DIST_DIR>/proviso/AIX/DataBase/AIX/instance/ora_installer
Linux systems:
# cd <DIST_DIR>/proviso/RHEL/DataBase/RHEL5/instance/ora_installer
where: v <DIST_DIR> is the directory on the hard drive where you copied the contents of the Tivoli Netcool Performance Manager distribution in Downloading the Tivoli Netcool Performance Manager distribution to disk on page 26. 3. Open the perform_oracle_inst.ini file with a text editor. This file contains a number of Oracle installation settings similar to the following:
# # # # You can install either a server install, a client install or a patchset. To install in batch mode you must enter a value. Uncomment the appropriate value.
# #ORA_TYPE=Server #ORA_TYPE=Client #ORA_TYPE=Patch # # What is the directory where you copied the # Oracle install CDs. # INSTALL_DIRECTORY= # # Script will automatically use the # ORACLE_BASE defined in you environment # or you can define it below. # #ORACLE_BASE= # # Script will automatically use the # ORACLE_HOME defined in you environment # or you can define it below. # #ORACLE_HOME= # # What unix account is the oracle software # owner? A typical default is supplied. # ORA_USR_OWN=oracle # # What is the oracle dba unix group name? # A typical default is supplied. # ORA_GRP_DBA=dba # # What is the oracle oui unix group name? # A typical default is supplied. # ORA_GRP_OUI=oinstall
4. Make the necessary edits to the installation values. You can uncomment lines already included in the file to use those values for the installation. For example, uncomment the following line to install Oracle server:
ORA_TYPE=Server
Chapter 2. Installing and configuring the prerequisite software
41
5. Save your edits and close the file. 6. Run the installation script with the b flag:
$ ./perform_oracle_inst b
The installation program verifies the values and performs the installation as shown in the section for the menu-based installation (see Install Oracle on page 38).
Procedure
1. Log in as root or become superuser. 2. Change to the directory where Oracle files were installed. (This is the value of the ORACLE_HOME environment variable as seen by the oracle login name.) For example:
# cd /opt/oracle/product/10.2.0
4. If the default entry, /usr/local/bin, is writable by root, press Enter to accept the default value. The default entry might be NFS-mounted at your site so it can be shared among several workstations and therefore might be write-protected. If so, enter the location of a machine-specific alternate bin directory. (You might need to create this alternate directory at a shell prompt first.) For example, enter /usr/delphi/bin. 5. The script continues as follows:
... Adding entry to /var/opt/oracle/oratab file... Entries will be added to the /var/opt/oracle/oratab file as needed by Database Configuration Assistant when a database is created Finished running generic part of root.sh script. Now product-specific root actions will be performed. #
42
Procedure
1. Log in as oracle. 2. Change to the following directory: Solaris systems:
# cd <DIST_DIR>/proviso/SOLARIS/DataBase/SOL10/instance/ora_installer
AIX systems:
# cd <DIST_DIR>/proviso/AIX/DataBase/AIX/instance/ora_installer
Linux systems:
# cd <DIST_DIR>/proviso/RHEL/DataBase/RHEL5/instance/ora_installer
where: v <DIST_DIR> is the directory on the hard drive where you copied the contents of the Tivoli Netcool Performance Manager distribution in Downloading the Tivoli Netcool Performance Manager distribution to disk on page 26. 3. Enter the following command to start the installer:
$ ./perform_oracle_inst
4. Verify the following settings: v The Oracle release number should be 10.2.0. v The Installation type field should be set to Patch. This field cycles between three settings: Server, Client, and Patch. Type 2 at the Choice prompt and press Enter until Patch is displayed. 5. To specify the path, type f at the Choice prompt and press Enter.
43
6. At the Choice prompt, enter the full path to the patch directory you created to hold the contents of the patch release (see Download the Oracle distribution and patch to disk on page 31). For example:
Choice: f Enter new value for CD directory: /var/tmp/oracle10204
7. To begin the upgrade installation, type 3 at the Choice prompt and press Enter. The installation script checks the environment, then asks whether you want to perform the installation. 8. Type y at the Choice prompt and press Enter. The script proceeds to install the patch release files, showing a long series of messages much like the ones shown during installation of the base release. Note: You can safely ignore any "font.properties not found" messages in the output. In the messages, look for "success" notifications. If the script shows unsuccessful installation messages, check the installation log file specified in the message. Correct the error conditions identified in the log file, then rerun the patch installation and double-check all menu settings. 9. When the installation is complete, type C and press Enter. Note: The message telling you to press C might have already scrolled past, because messages from several installation threads are written to the same screen. 10. Exit from the menu. Note: If you try to start Oracle as the database administrator (DBA) at this point, the startup will fail because the file initPV.ora does not yet exist. This file is created during the Tivoli Netcool Performance Manager installation. 11. Carry out the steps as described in Run the root.sh script on page 42. This step is also required after an Oracle patch installation.
Procedure
1. Log in as oracle. 2. Change to the following directory:
$ cd ORACLE_HOME/install.
44
Procedure
1. Log in as oracle. 2. Open the .profile file with a text editor. 3. Add the following line anywhere between the Begin and End Oracle Settings comment lines:
ORACLE_SID=PV; export ORACLE_SID
For example:
# -- Begin Oracle Settings -umask 022 ORACLE_BASE=/opt/oracle ORACLE_HOME=$ORACLE_BASE/product/10.2.0 ORACLE_SID=PV; export ORACLE_SID NLS_LANG=AMERICAN_AMERICA.WE8ISO8859P1 ORA_NLS33=$ORACLE_HOME/ocommon/nls/admin/data LD_LIBRARY_PATH=$ORACLE_HOME/lib TNS_ADMIN=$ORACLE_HOME/network/admin PATH=$PATH:$ORACLE_HOME/bin:/usr/ccs/bin:/usr/delphi/bin EXTPROC_DLLS=ONLY:${LD_LIBRARY_PATH}/libpvmextc.so export PATH ORACLE_BASE ORACLE_HOME NLS_LANG export ORA_NLS33 LD_LIBRARY_PATH TNS_ADMIN export EXTPROC_DLLS # -- End Oracle Settings --
4. Save and exit the .profile file. 5. Enter the following shell command to activate the change to your profile:
$ . ./.profile
6. Make sure the variable was set by entering the following command:
$ env | grep ORACLE_SID
Procedure
1. Log in as oracle. 2. Depending on your operating system, change to the following directory:
Chapter 2. Installing and configuring the prerequisite software
45
Solaris systems:
$ cd /var/opt/oracle
AIX systems:
$ cd /etc
Linux Systems:
$ cd /etc
3. Edit the oratab file with a text editor. The last line of this file looks like this example:
*:/opt/oracle/product/10.2.0:N
4. Make the following edits to this line: v Replace * with $ORACLE_SID (PV by default). v Replace N with Y. The last line should now be:
PV:/opt/oracle/product/10.2.0:Y
Procedure
1. Log in as oracle. 2. Change to one of the following directories:
$ cd $TNS_ADMIN
or
$ cd /opt/oracle/product/10.2.0/network/admin
Note: By Oracle convention, the keywords in this file are in uppercase but uppercase is not required.
# listener.ora network configuration file in directory # /opt/oracle/product/10.2.0/network/admin LISTENER= (DESCRIPTION_LIST = (DESCRIPTION = (ADDRESS_LIST = (ADDRESS = (PROTOCOL = TCP) (HOST = yourhost) (PORT = 1521)) )
46
(ADDRESS_LIST = (ADDRESS = (PROTOCOL = IPC) (KEY = EXTPROC)) ) ) ) SID_LIST_LISTENER = (SID_LIST = (SID_DESC = (SID_NAME = PLSExtProc) (ORACLE_HOME = /opt/oracle/product/10.2.0) (PROGRAM = extproc) ) (SID_DESC = (GLOBAL_DBNAME = PV.WORLD) (SID_NAME = PV) (ORACLE_HOME = /opt/oracle/product/10.2.0) ) )
4. Using a text editor, change the following: a. Replace the string yourhost in the line (HOST = yourhost) with the name of your Oracle server. Note: Specify the host using the hostname only, do not use the IP address. b. (optional) Replace the default port number 1521 in the line (PORT = 1521) with your required port number. c. Write and quit the file. 5. Depending on your operating system, change to the following directory: Solaris systems:
$ cd /var/opt/oracle
AIX systems:
$ cd /etc
Linux Systems:
$ cd /etc
6. Edit the lsnrtab file and add a line in the following format to the end of the file (after the initial comments):
LISTENER:value_of_ORACLE-HOME:Y
For example:
LISTENER:/opt/oracle/product/10.2.0:Y
In this syntax, LISTENER is the name of the listener process. 7. Write and quit the file. 8. Test that the listener process works correctly by starting it manually using the following command:
lsnrctl start
(The lsnrctl command also accepts the stop and status arguments.) Look for a successful completion message.
47
Procedure
1. Log in as oracle. 2. Change to one of the following directories:
$ cd $TNS_ADMIN
or
$ cd /opt/oracle/product/10.2.0/network/admin
3. Create the sqlnet.ora file, which will manage Oracle network operations. You must create an sqlnet.ora file for both Oracle server and Oracle client installations. Follow these steps: a. Copy the sample sqlnet.ora file, template.example_tnpm.sqlnet.ora, contained in the opt/oracle/admin/skeleton/bin/ directory:
$ cp /opt/oracle/admin/skeleton/bin/template.example_tnpm.sqlnet.ora sqlnet.ora
For example:
# sqlnet.ora network configuration file in # /opt/oracle/product/10.2.0/network/admin NAMES.DIRECTORY_PATH=(TNSNAMES) NAMES.DEFAULT_DOMAIN=WORLD
Note: If you do not use WORLD as the DEFAULT_DOMAIN value, make sure you enter the same value for DEFAULT_DOMAIN in both sqlnet.ora and tnsnames.ora. c. Write and quit the file. 4. Create the tnsnames.ora file, which maintains the relationships between logical node names and physical locations of Oracle Servers in the network. You can do this by copying the existing sample file:
cp /opt/oracle/admin/skeleton/bin/template.example_tnpm.tnsnames.ora tnsnames.ora
Follow these steps: a. Enter lines similar to the following example, using the actual name of your Oracle server in the HOST=delphi line and replacing {SID} with PV or your Oracle SID.
# tnsnames.ora network configuration file in # /opt/oracle/product/10.2.0/network/admin # # The EXTPROC entry only needs to exist in the # tnsnames.ora file on the Oracle server. # For Oracle client installations, tnsnames.ora # only needs the PV.WORLD entry. EXTPROC_CONNECTION_DATA.WORLD = (DESCRIPTION = (ADDRESS_LIST =
48
(ADDRESS = (PROTOCOL = IPC) (KEY = EXTPROC) ) ) (CONNECT_DATA = (SID = PLSExtProc) (PRESENTATION = RO) ) ) PV.WORLD = (DESCRIPTION = (ENABLE=BROKEN) (ADDRESS_LIST = (ADDRESS = (PROTOCOL = TCP) (HOST = delphi) (PORT = 1521) ) ) (CONNECT_DATA = (SERVICE_NAME = PV.WORLD) (INSTANCE_NAME = PV) ) )
Note: Indents in this file must be preserved. Note the following: v You will use the value in the INSTANCE_NAME field as the TNS entry when installing Tivoli Netcool Performance Manager DataMart. v IBM strongly recommends that you include the line (ENABLE=BROKEN) in the PV.WORLD entry, as shown in the example. This parameter setting prevents CME processes from hanging in the event that the CME is disconnected from the database before results are returned to the CME. v If configuring tnsnames.ora for a server installation, be sure to append the same domain suffix to all entries including EXTPROC_CONNECTION_DATA that you specified for the NAMES.DEFAULT_DOMAIN entry in the sqlnet.ora file. That is, append.WORLD to each entry. b. Write and quit the file. 5. Test the Oracle Net configuration by entering a command with the following syntax:
tnsping Net_service_name.domain 10
For example:
$ tnsping PV.WORLD 10
Look for successful completion messages (OK). To test without using the domain suffix, enter a command with the following syntax:
tnsping Net_service_name 10
For example:
$ tnsping PV 10
Note: If either test is not successful, check your configuration and retest.
49
Procedure
Should you discern any impact to performance, please confirm that these jobs are not in operation.
Procedure
1. On the system where the Oracle database is to be installed, disable NIS. For more information, see your operating system documentation. 2. Run the configure_ora script to create a local oracle account (see Run the Oracle server configuration script on page 33). 3. Re-enable NIS. Note: The local Oracle account should be used before the NIS Oracle user.
50
Procedure
1. Log in as root. 2. Create a directory to hold the contents of the Oracle distribution. For example:
# mkdir /var/tmp/oracle10201 # mkdir /var/tmp/oracle10204
3. Download the Oracle files to the /var/tmp/oracle10201 directory. 4. Locate the appropriate upgrade patch file for your version of Oracle on the Oracle Web site and download it to your /var/tmp/oracle10204 directory. 5. Unzip the oracle distribution files that now reside in the /var/tmp/oracle10201 and /var/tmp/oracle10204 directories. Before you proceed to the next step, make sure that you obtain the upgrade instructions provided by Oracle for the patch. The instructions contain information on performing steps required for the upgrade that are not documented in this guide. If you are installing the Oracle patch on an AIX system, you might need to run the sbinclean command and execute the catpatch.sql and utlrp.sql SQL scripts as part of the upgrade procedure. Additional requirements might exist for other platforms. See your database administrator to determine whether there are any company-specific requirements for installing Oracle in your environment.
Procedure
1. Log in as root. 2. Create a directory to contain your Oracle 10.2.0.1.0 files. For example:
# mkdir /var/tmp/oracle10201
3. Download your gzipped 'cpio' Oracle files to this directory. 4. Oracle Server folder setup: a. Within the oracle 10.2.0.1.0 distribution directory you have created, create a subdirectory called 'database'. b. Gunzip the '10gr2_db_sol.cpio.gz' file that was downloaded from Oracle. c. Cpio extract '10gr2_db_sol.cpio' into the 'database' subdirectory
cd database cpio -idmv < 10gr2_db_sol.cpio
5. Oracle Client folder setup: a. Within the oracle 10.2.0.1.0 distribution directory you have created, create a subdirectory called 'client'.
Chapter 2. Installing and configuring the prerequisite software
51
b. Gunzip the '10gr2_client_sol.cpio.gz ' file that was downloaded from Oracle. c. Cpio extract '10gr2_client_sol.cpio' into the 'client' subdirectory:
cd client cpio -idmv < 10gr2_client_sol.cpio)
Procedure
1. Log in as root. 2. Set the ORACLE_BASE environment variable to point to the top-level directory where you want the Oracle client files installed. The default installation directory is /opt/oracle. For example:
# ORACLE_BASE=/opt/oracle # export ORACLE_BASE
Note: The configure_client script places this variable into the oracle login name's .profile file. To check that the variable is set correctly, enter the following command:
# env | grep ORA
AIX systems:
# cd <DIST_DIR>/proviso/AIX/DataBase/AIX/instance
Linux systems:
# cd <DIST_DIR>/proviso/RHEL/DataBase/RHEL5/instance
where:
52
<DIST_DIR> is the directory on the hard drive where you copied the contents of the Tivoli Netcool Performance Manager distribution in Downloading the Tivoli Netcool Performance Manager distribution to disk on page 26. 4. Run the Oracle configuration script using the following command:
# ./configure_client
5. (Optional) To create a different name for the Oracle user other than the default oracle, type 1 and follow the instructions. Note: The Oracle user will be referred to as "oracle" for the remainder of this document. If you have set the username to be a non-default string, that is, something other than "oracle" , make sure to substitute your non- default string for each occurance of "oracle" within any task description. 6. Type 3 at the Choice prompt and press Enter. The script creates the dba and oinstall groups, and the ORACLE_BASE directory, unless they already exist.
Checking environment... Checking for group [ dba ] --> Created. Checking for group [ oinstall ] --> Created. Checking ORACLE_BASE ** WARNING ** ORACLE_BASE directory does not exist. ** [ /opt/oracle ] ** ** Create it ? (n/y) y
If prompted, type y and press Enter. The script creates the /opt/oracle directory and continues as follows:
Checking for user [ oracle ] ** WARNING ** User [ oracle ] does not exist. ** ** Create it locally ? (n/y) y
If prompted, type y to create the oracle user and press Enter. The script creates the oracle user and continues as follows:
Chapter 2. Installing and configuring the prerequisite software
53
--> Created. Checking for oracle directory tree : [ /opt/oracle/product ] --> Created. [ /opt/oracle/product ] --> Created. [ /opt/oracle/product/9.2.0 ] --> Created. Checking for oracle .profile file --> Created. Press Enter to continue...
7. Press Enter to continue. The configure_client main screen is displayed. 8. Type 0 and press Enter to exit the script.
Procedure
1. Log in as root. 2. Enter the following command:
# passwd oracle
3. Enter and re-enter the password (oracle, by default) as prompted. The password is set.
Procedure
1. Log in as root. 2. Change to the following directory: Solaris systems:
# cd <DIST_DIR>/proviso/SOLARIS/DataBase/SOL10/instance/ora_installer
AIX systems:
# cd <DIST_DIR>/proviso/AIX/DataBase/AIX/instance/ora_installer
Linux systems:
# cd <DIST_DIR>/proviso/RHEL/DataBase/RHEL5/instance/ora_installer
where:
54
v <DIST_DIR> is the directory on the hard drive where you copied the contents of the Tivoli Netcool Performance Manager distribution in Downloading the Tivoli Netcool Performance Manager distribution to disk on page 26. 3. Set the ORACLE_BASE environment variable. For example:
# ORACLE_BASE=/opt/oracle # export ORACLE_BASE
You must use the same ORACLE_BASE setting that you specified in Run the Oracle server configuration script on page 33 4. Enter the following command:
# ./pre_install_as_root
If the script shows an error, correct the situation causing the error before proceeding to the next step.
Procedure
1. Log in as oracle. 2. Set and export the DISPLAY environment variable. If you are using the su command to become oracle, use a hyphen as the second argument so the oracle name's login environment is loaded:
# su - oracle
3. Verify that the environment variable ORACLE_BASE has been set by entering the following command:
$ env | grep ORA
If the response does not include ORACLE_BASE=/opt/oracle, stop and make sure the .profile file was set for the oracle user as described in Run the Oracle server configuration script on page 33. 4. To verify the path, enter the following command:
$ echo $PATH
The output should show that /usr/ccs/bin is part of the search path. For example:
/usr/bin:/opt/oracle/product/10.2.0/bin:/usr/ccs/bin
a. If the directory does not appear in the path, add it by entering the following commands:
$ PATH=$PATH:/usr/ccs/bin $ export PATH
55
Procedure
1. Log in as oracle. 2. Change to the following directory: Solaris systems:
# cd <DIST_DIR>/proviso/SOLARIS/DataBase/SOL10/instance/ora_installer
AIX systems:
# cd <DIST_DIR>/proviso/AIX/DataBase/AIX/instance/ora_installer
Linux systems:
# cd <DIST_DIR>/proviso/RHEL/DataBase/RHEL5/instance/ora_installer
where: v <DIST_DIR> is the directory on the hard drive where you copied the contents of the Tivoli Netcool Performance Manager distribution in Downloading the Tivoli Netcool Performance Manager distribution to disk on page 26. 3. Enter the following command to start the installer:
$ ./perform_oracle_inst
56
-------------------------------------------------perform_oracle_inst Installation of oracle binaries <Current Date> -------------------------------------------------OS ........... : [ SunOS 5.10 Generic ] Host ......... : [ delphi ] Logname ...... : [ oracle ] Install Oracle release .... : [ 10.2.0 ] Installation type.......... : [ Client ] Enter the appropriate letter to modify the entries below: a) b) c) d) e) f) ORACLE_BASE .. : [ /opt/oracle ] ORACLE_HOME .. : [ /opt/oracle/product/10.2.0 ] DBA group ..................... : [ dba ] OUI Inventory group ........... : [ oinstall ] Oracle Software owner ......... : [ oracle ] Directory where CDs were copied: [ ]
Menu : 1. Next supported release 2. Set install type to: Client 3. Perform install 0. Exit Choice :
4. Enter f at the Choice prompt and press Enter. 5. Enter the full path to the directory you created to hold the Oracle distribution in Download the Oracle distribution and patch to disk on page 31. For example:
Choice: f Enter new value for CD directory: /var/tmp/oracle10201
6. Edit any other menu settings as necessary. Make sure that the values for ORACLE_BASE and ORACLE_HOME correspond to the locations you specified when you ran the Oracle client configuration script. 7. To start the Oracle installation, type 3 at the Choice prompt and press Enter. 8. The installation script checks the environment, then asks whether you want to perform the installation. Type Y at the Choice prompt and press Enter. The installation script starts installing Oracle and displays a series of status messages. Note: You can safely ignore any "font.properties not found" messages in the output. When the installation reaches the In Summary Page stage, the installation slows down significantly while Oracle files are copied and linked. When the installation process completes, the installation displays a success message. Write down the log file location to aid in troubleshooting if there is an installation error. Type C and press Enter to return to the installation menu. Type 0 and press Enter to exit the installation menu. Perform the steps in Run the root.sh script on page 42.
9.
57
Procedure
1. Log in as root or become superuser. 2. Change to the directory where Oracle files were installed. (This is the value of the ORACLE_HOME environment variable as seen by the oracle login name.) For example:
# cd /opt/oracle/product/10.2.0
4. If the default entry, /usr/local/bin, is writable by root, press Enter to accept the default value. The default entry might be NFS-mounted at your site so it can be shared among several workstations and therefore might be write-protected. If so, enter the location of a machine-specific alternate bin directory. (You might need to create this alternate directory at a shell prompt first.) For example, enter /usr/delphi/bin. 5. The script continues as follows:
... Adding entry to /var/opt/oracle/oratab file... Entries will be added to the /var/opt/oracle/oratab file as needed by Database Configuration Assistant when a database is created Finished running generic part of root.sh script. Now product-specific root actions will be performed. #
58
Your specific system may require supplementary steps not covered in this document, but which the Oracle instructions cover in detail. For example, if you are installing the Oracle patch on an AIX system, it is possible you may need to run the sbinclean command and execute the catpatch.sql and utlrp.sql SQL scripts as part of the upgrade procedure. Additional requirements might exist for other platforms. Note: Do not skip this step. Upgrade the Oracle version before you install the Tivoli Netcool Performance Manager database structure. To install the Oracle patchset:
Procedure
1. Follow the instructions to run the perform_oracle_inst script, as shown in Install Oracle patches on page 43. Go through the menus as described in that section, with the following changes: a. Enter 2 at the Choice prompt and press Enter until the Installation type field is set to Patch. b. Type f at the Choice prompt and press Enter. c. At the Choice prompt, type the full path to the patch directory you created in Download the Oracle distribution and patch to disk on page 31 to hold the client patch 10.2.0.4 contents. For example:
Choice: f Enter new value for CD directory: /var/tmp/oracle10204
2. Type 3 at the Choice prompt and press Enter. The script checks the environment and displays a series of messages. 3. Type y at the Choice prompt and press Enter to begin the patch installation. The script proceeds to install the patch release files, showing a series of messages much like the ones shown during installation of the base release. Note: You can safely ignore any "font.properties not found" messages in the output. 4. Watch the script's messages for success entries. Note: If the script shows unsuccessful installation messages, check the installation log file named in the message. Correct the error conditions identified in the log file, then rerun the patch installation and double-check all menu settings. 5. At the Choice prompt, type C and press Enter to continue. The message telling you to press C may have already scrolled past, because messages from several installation threads are written to the same screen. If the installation has paused, type C to continue.Relax Permissions on ORACLE_HOME
59
Procedure
1. Log in as oracle. 2. Change to the following directory:
$ cd ORACLE_HOME/install.
Procedure
v You must configure sqlnet.ora and tnsnames.ora files for both Oracle server and Oracle client installations. However, the tnsnames.ora file for client installations should not have the EXTPROC_CONNECTION_DATA section. v If you are installing DataView and one or more other Tivoli Netcool Performance Manager components on the same system, you must make sure that the tnsnames.ora and sqlnet.ora files for each set of client software are identical. The easiest way to do this is to create these files when you are configuring the first client instance for Net and then to copy it to the corresponding directory when you configure the second instance.
Procedure
1. Log in as oracle. 2. Change to the following directory:
$ cd $ORACLE_HOME/network/admin
3. To create the sqlnet.ora file, FTP the following file from your Oracle server:
60
/opt/oracle/admin/skeleton/bin/template.example_tnpm.sqlnet.ora
For example:
# sqlnet.ora network configuration file in # /opt/oracle/product/10.2.0/network/admin NAMES.DIRECTORY_PATH=(TNSNAMES) NAMES.DEFAULT_DOMAIN=WORLD
Note: If you do not use WORLD as the DEFAULT_DOMAIN value, make sure you enter the same value for DEFAULT_DOMAIN in both sqlnet.ora and tnsnames.ora. 5. Write and quit the sqlnet.ora file.
Procedure
1. , FTP the following file from your Oracle server:
/opt/oracle/admin/skeleton/bin/template.example_tnpm.tnsnames.ora
Note: Indents in this file must be preserved. 3. Replace the string yourhost in the line (HOST = yourhost) with the name of your Oracle server. Note the following: v You will use the value in the INSTANCE_NAME field as the TNS entry when installing DataMart.
61
v If you reconfigure the Oracle client to connect to a different Oracle database in another Tivoli Netcool Performance Manager installation, be sure you update the HOST entry in the tnsnames.ora file, then restart the Oracle client. v Specify the host using the hostname only, do not use the IP address. 4. (optional) Replace the default port number 1521 in the line (PORT = 1521) with your required port number. 5. Write and quit the file.
Procedure
1. Log in as oracle. 2. Enter a command with the following syntax:
tnsping Net_service_name 10
For example: tnsping PV.WORLD 10 3. Test again, using the same Net instance name without the domain suffix:
tnsping PV 10
Next steps
The steps that follow installation of the prerequisite software. Once you have installed the prerequisite software, you are ready to begin the actual installation of Tivoli Netcool Performance Manager. Depending on the type of installation you require, follow the directions in the appropriate chapter: v Chapter 4, Installing as a minimal deployment, on page 89 - Describes how to install Tivoli Netcool Performance Manager in a distributed production environment. v Chapter 5, Modifying the current deployment, on page 95 - Describes how to install Tivoli Netcool Performance Manager as a minimal deployment, which is used primarily for demonstration or evaluation purposes. v If you are planning on Installing Tivoli Netcool Performance Manager as a distributed environment that uses clustering for high availability, please review the Tivoli Netcool Performance Manager HA (High Availability) documentation, which is available for download by going to http://www-01.ibm.com/software/ brandcatalog/opal/ and searching for "Netcool Proviso HA Documentation".
62
63
Before installing Tivoli Netcool Performance Manager, you must have installed the prerequisite software. For detailed information, see Chapter 2, Installing and configuring the prerequisite software, on page 13. In addition, you must have decided how you want to configure your system. Refer to the following sections: v Co-location rules on page 2 v Typical installation topology on page 8 v Appendix A, Remote installation issues, on page 137 The general steps used to install Tivoli Netcool Performance Manager are as follows: v Start the launchpad. v Install the Topology Editor. v Start the Topology Editor. v Create the topology.
64
v v v v
Add the Tivoli Netcool Performance Manager components. Save the topology to an XML file. Start the deployer. Install Tivoli Netcool Performance Manager using the deployer.
The following sections describe each of these steps in detail. Note: Before you start the installation, verify that all the database tests have been performed. Otherwise, the installation might fail. See Chapter 2, Installing and configuring the prerequisite software, on page 13 for information about tnsping.
Procedure
1. Log in as root. 2. Set and export the DISPLAY variable (see Setting up a remote X Window display on page 14). 3. Set and export the BROWSER variable to point to your Web browser. For example: On Solaris systems:
# BROWSER=/opt/mozilla/mozilla # export BROWSER
On AIX systems:
# BROWSER=/usr/mozilla/firefox/firefox # export BROWSER
On Linux systems:
# BROWSER=/usr/bin/firefox # export BROWSER
Note: The BROWSER command cannot include any spaces around the equal sign. 4. Change directory to the directory where the launchpad resides. On Solaris systems:
# cd <DIST_DIR>/proviso/SOLARIS
On AIX systems:
# cd <DIST_DIR>/proviso/AIX
On Linux systems:
# cd <DIST_DIR>/proviso/RHEL
<DIST_DIR> is the directory on the hard drive where you copied the contents of the Tivoli Netcool Performance Manager distribution in Downloading the Tivoli Netcool Performance Manager distribution to disk on page 26. 5. Enter the following command to start the launchpad:
# ./launchpad.sh
65
Procedure
1. You can begin the Topology Editor installation procedure from the command line or from the Launchpad. From the launchpad: a. On the launchpad, click the Install Topology Editor option in the list of tasks. b. On the Install Topology Editor page, click the Install Topology Editor link. From the command line: a. Log in as root. b. Change directory to the directory that contains the Topology Editor installation script: On Solaris systems:
# cd <DIST_DIR>/proviso/SOLARIS/Install/SOL10/topologyEditor/Disk1/InstData/VM
On AIX systems:
# cd <DIST_DIR>/proviso/AIX/Install/topologyEditor/Disk1/InstData/VM
On Linux systems:
# cd <DIST_DIR>/proviso/RHEL/Install/topologyEditor/Disk1/InstData/VM
<DIST_DIR> is the directory on the hard drive where you copied the contents of the Tivoli Netcool Performance Manager distribution in Downloading the Tivoli Netcool Performance Manager distribution to disk on page 26 c. Enter the following command:
# ./installer.bin
2. The installation wizard opens in a separate window, displaying a welcome page. Click Next. 3. Review and accept the license agreement, then click Next. 4. Confirm the wizard is pointing to the correct directory. The default is /opt/IBM/proviso. If you have previously installed the Topology Editor on this system, the installer does not prompt you for an installation directory and instead uses the directory where you last installed the application. 5. Click Next to continue. 6. Confirm the wizard is pointing to the correct base installation directory of the Oracle JDBC driver (/opt/oracle/product/version/jdbc/lib), or click Choose to navigate to another directory. 7. Click Next to continue. 8. Review the installation information, then click Run. 9. When the installation is complete, click Done to close the wizard.
66
The installation wizard installs the Topology Editor and an instance of the deployer in the following directories:
Interface Topology Editor Directory install_dir/topologyEditor For example: /opt/IBM/proviso/topologyEditor Deployer install_dir/deployer For example: /opt/IBM/proviso/deployer
The combination of the Topology Editor and the deployer is referred to as the primary deployer. For more information, see Resuming a partially successful first-time installation on page 87. Note: To uninstall the Topology Editor, follow the instructions in Uninstalling the topology editor on page 133. Do not delete the /opt/IBM directory. Doing so will cause problems when you try to reinstall the Topology Editor. If the /opt/IBM directory is accidentally deleted, perform the following steps: a. Change to the /var directory. b. Rename the hidden file .com.zerog.registry.xml (for example, rename it to .com.zerog.registry.xml.backup). c. Reinstall the Topology Editor. d. Rename the backup file to the original name (.com.zerog.registry.xml).
Procedure
v To start the Topology Editor from the launchpad: 1. If the Install Topology Editor page is not already open, click the Install Topology Editor option in the list of tasks to open it. 2. On the Install Topology Editor page, click the Start Topology Editor link. v To start the Topology Editor from the command line: 1. Log in as root. 2. Change directory to the directory in which you installed the Topology Editor. For example:
# cd /opt/IBM/proviso/topologyEditor
Note: If your DISPLAY environment variable is not set, the Topology Editor will fail with a Java assertion message (core dump).
67
Procedure
1. In the Topology Editor, select Topology > Create new topology. The New Topology window is displayed. 2. Enter the Number of resources to be managed by Tivoli Netcool Performance Manager. The default value is 10000. The size of your deployment affects the database sizing. 3. Click Finish. The Topology Editor creates the following entities: v In the Logical view, five items are listed: Tivoli Netcool Performance Manager Topology, Cross Collector CMEs, DataChannels, DataMarts and Tivoli Integrated Portals. v In the Physical view, there is a new Hosts folder.
Procedure
1. In the Physical view, right-click the Hosts folder and select Add Host from the menu. The Add Host window opens. 2. Specify the details for the host machine. The fields are as follows:
68
v Host name - Enter the name of the host (for example, delphi). v Operating system - Specifies the operating system (for example, SOLARIS). This field is filled in for you. v Oracle home - Specifies the default ORACLE_HOME directory for all Tivoli Netcool Performance Manager components installed on the system (by default, /opt/oracle/product/10.2.0). v PV User - Specifies the default Tivoli Netcool Performance Manager Unix user (for example, pvuser) for all Tivoli Netcool Performance Manager components installed on the system. v PV user password - Specifies the password for the default Tivoli Netcool Performance Manager user (for example, PV). v Create Disk Usage Server for this Host? - Selecting this check box creates a DataChannel subcomponent to handle disk quota and flow control. If you have not chosen to create a Disk Usage Server, Click Finish to create the host. The Topology Editor adds the host under the Hosts folder in the Physical view. If you have chosen to create a Disk Usage Server, click Next and the Add Host window will allow you to add details for your Disk Usage Server. 3. Specify the details for the Disk Usage Server. The fields are as follows: v Local Root Directory - The local DataChannel root directory. This property allows you to differentiate between a local directory and a remote directory mounted to allow for FTP access. v Remote Root Directory - Remote directory mounted for FTP access. This property allows you to differentiate between a local directory and a remote directory mounted to allow for FTP access. v FC FSLL - This is the Flow Control Free Space Low Limit property. When this set limit is reached the Disk Usage Server will contact all components who reside in this root directory and tell them to free up all space possible. v FC QUOTA - This is the Flow Control Quota property. This property allows you to set the amount of disk space in bytes available to Tivoli Netcool Performance Manager components on this file system. v Remote User - User account used when attempting to access this Disk Usage Server remotely. v Remote User Password - User account password used when attempting to access this Disk Usage Server remotely. v Secure file transfer to be used - Boolean indicator identifying if ssh should be used when attempting to access this directory remotely. v Port Number - Port number to use for remote access (sftp) in case it is a non default port. Click Finishto create the host. The Topology Editor adds the host under the Hosts folder in the Physical view. Note: The DataChannel properties will be filled in automatically at a later stage.
69
Procedure
1. In the Physical view, right-click the Hosts folder and select Add Multiple Host from the menu. The Add Multiple Hosts window opens. 2. Add new hosts by typing their names into the Host Name field as a comma separated list. 3. Click Next. 4. Configure all added hosts. The Configure hosts dialog allows you to enter configuration settings and apply these settings to one or more of the specified host set. To apply configuration settings to one or more of the specified host set: a. Enter the appropriate host configuration values. All configuration options are described in Steps 2, and 3 of the previous process, Add the hosts on page 68. b. Select the check box opposite each of the hosts to which you want to apply the entered values. c. Click Next. The hosts for which all configuration settings have been specified disappear from the set of selectable hosts. d. Repeat steps a, b and c till all hosts are configured. 5. Click Finish.
Procedure
1. In the Logical view, right-click the Tivoli Netcool Performance Manager Topology component and select Add Database Configurations from the menu. The host selection window opens. 2. You must add the Database Configuration component to the same server that hosts the Oracle server (for example, delphi). Select the appropriate host using the drop-down list. Note: The operating system of the host where you configure the Database Configurations component determines the main platform for the entire Tivoli Netcool Performance Manager deployment. This means that you should install all the other main components (DataView, DataMart, DataChannel and Tivoli Integrated Portal) on machines with the same operating system. The Topology Editor enforces this constraint by listing only those hosts with the same operating system used to host the database. 3. Click Next to configure the mount points for the database. 4. Add the correct number of mount points. To add a new mount point, click Add Mount Point. A new, blank row is added to the window. Fill in the fields as appropriate for the new mount point. 5. Enter the required configuration information for each mount point. a. Enter the mount point location:
70
v Mount Point Directory Name (for example, /raid_2/oradata) Note: The mount point directories can be named using any string as required by your organizations naming standards. v Used for Metadata Tablespaces? (A check mark indicates True.) v Used for Temporary Tablespaces? (A check mark indicates True.) v Used for Metric Tablespaces? (A check mark indicates True.) v Used for System Tablespaces and Redo? (A check mark indicates True.) b. Click Back to return to the original page. c. Click Finish to create the component. The Topology Editor adds the new Database Configurations component to the Logical view. 6. Highlight the Database Configurations component to display its properties. Review the property values to make sure they are valid. For the complete list of properties for this component, see the IBM Tivoli Netcool Performance Manager: Property Reference Guide. The Database Configurations component has the following subelements: v Channel tablespace configurations v Database Channels v Database Clients configurations v Tablespace configurations v Temporary tablespace configurations Note: Before you actually install Tivoli Netcool Performance Manager, verify that both the raid_2/oradata and raid_3/oradata directory structures have been created, and that the oradata subdirectories are owned by oracle:dba.
Add a DataMart
The steps required to add a DataMart component to your topology.
For a remote server, that is one that does not host the primary deployer, you must download and install the required JRE, and set the correct JRE path. See
Chapter 3. Installing in a distributed environment
71
the IBM Tivoli Netcool Performance Manager: Configuration Recommendations Guide document for JRE download details. To add a DataMart component:
Procedure
1. In the Logical view, right-click the DataMarts folder and select Add DataMart from the menu. The host selection host window is displayed. 2. Using the drop-down list of available hosts, select the machine on which DataMart should be installed (for example, delphi). 3. Click Finish. The Topology Editor adds the new DataMart x component (for example, DataMart 1) under the DataMarts folder in the Logical view. 4. Highlight the DataMart x component to display its properties. Review the property values to make sure they are valid. You can specify an alternate installation user for the DataMart component by changing the values of the USER_LOGIN and USER_PASSWORD properties in the Advanced Properties tab. For the complete list of properties for this component, see the IBM Tivoli Netcool Performance Manager: Property Reference Guide.
72
Procedure
In the Logical view, right-click the DataMart x folder and select Add Discovery server from the menu. The Topology Editor displays the new Discovery Server under the DataMart n folder in the Logical view.
Procedure
v Install the primary instance of DataMart and the Discovery Server on one target host system. v Install and configure any required technology packs on the primary host. You modify the contents of the inventory files during this step. v Install secondary instances of DataMart and the Discovery Server on corresponding target host systems. v Replicate the inventory files from the system where the primary instance of DataMart is running to the $PVMHOME/conf directory on the secondary hosts. You must also replicate the InventoryHook.sh script that is located in the $PVMHOME/bin directory and any other files that this script requires.
73
Procedure
1. In the Logical view, right-click on the Tivoli Integrated Portals folder and select Add TIP from the menu. The Configure TIP Wizard is displayed. 2. The Topology Editor gives you the choice of adding an already existing Tivoli Integrated Portal to the topology or to create a new Tivoli Integrated Portal. To create a new TIP, select the Create a new TIP radio button. To import an already existing Tivoli Integrated Portal into the topology, select the Import existing TIPs from host radio button. 3. Using the drop-down list of available hosts, select the host on which Tivoli Integrated Portal should be installed (for example, delphi). Note: The hostname of the host selected for the TCR install must not contain underscores. Underscores in the hostname will cause the installation of TCR to fail. 4. Click Finish. The Topology Editor adds the new Tivoli Integrated Portal component to the Logical view. 5. Highlight the Tivoli Integrated Portal component to display its properties. 6. Review the other property values to make sure they are valid. For the complete list of properties for this component, see the IBM Tivoli Netcool Performance Manager: Property Reference Guide.
Procedure
1. In the Logical view, right-click on the Tivoli Integrated Portals folder and select Import existing TIPs from host from the menu. The Run TIP Discovery Wizard Page is displayed. 2. Select the check box for each host on which you would like to perform Tivoli Integrated Portal discovery. 3. Click Add TIP. If the discovered Tivoli Integrated Portal is an old version, it is flagged within the topology for upgrade. Any DataView without a Tivoli Integrated Portal is flagged within the topology for Tivoli Integrated Portal installation on that host. The deployer will take the appropriate action when run.
74
4. Click Next. 5. Configure Tivoli Integrated Portal properties. a. Enter the appropriate host configuration values. v TCR_INSTALLATION_DIRECTORY: This is the directory in which Tivoli Common Reporting is installed. v TIP_INSTALLATION_DIRECTORY: This is the directory in which Tivoli Integrated Portal is installed. v WAS_USER_NAME: This is the WAS user name. v WAS_PASSWORD: This is the WAS password. If you would like to configure LDAP for Tivoli Integrated Portal, select the LDAP check box. b. Select the check box opposite each of the Tivoli Integrated Portal hosts to which you want to apply the entered values. c. Click Next. The hosts for which all configuration settings have been specified disappear from the set of selectable hosts. d. Repeat steps a, b and c till all hosts are configured. 6. Click Next to add the discovered Tivoli Intergrated Portals to the topology. Note: If you discover a Tivoli Common Reporting/Tivoli Intergrated Portal of version 2.1 that was installed using the Tivoli Common Reporting installer and not the Tivoli Netcool Performance Manager installer, the port will not align with a Technology Pack automatically. To align the port numbers you must specify the Tivoli Intergrated Portal port when performing the Technology Pack installation.
Add a DataView
How to add a DataView.
Procedure
In the Logical view, right-click on a Tivoli Integrated Portal and select Add DataView from the menu. The DataView is automatically added inheriting its properties from the Tivoli Integrated Portal instance.
75
Procedure
1. In the Logical view, right-click the DataChannels folder and select Add Administrative Components from the menu. The host selection window opens. 2. Using the drop-down list of available hosts, select the machine that you want to be the Channel Manager host for your DataChannel configuration (for example, corinth). 3. Click Finish. The Topology Editor adds a set of new components to the Logical view: v Channel Manager - Enables you to start and stop individual DataChannels and monitor the state of various DataChannel programs. There is one Channel Manager for the entire DataChannel configuration. The Channel Manager components are installed on the first host you specify v Corba Naming Server - Provides near real-time data to DataView. v High Availability Managers - This is mainly used for large installations that want to use redundant SNMP collection paths. The HAM constantly monitors the availability of one or more SNMP collection hosts, and switches collection to a backup host (called a spare) if a primary host becomes unavailable. v Log Server - Used to store user, debug, and error information. v Plan Builder - Creates the metric data routing and processing plan for the other components in the DataChannel. v Custom DataChannel properties - These are the custom property values that apply to all DataChannel components. v Global DataChannel properties - These are the global property values that apply to all DataChannel components.
Add a DataChannel
A DataChannel is a software module that receives and processes network statistical information from both SNMP and non-SNMP (BULK) sources.
Procedure
1. In the Logical view, right-click the DataChannels folder and select Add DataChannel from the menu. The Configure the DataChannel window is displayed.
76
2. Using the drop-down list of available hosts, select the machine that will host the DataChannel (for example, corinth). 3. Accept the default channel number (for example, 1). 4. Click Finish. The Topology Editor adds the new DataChannel (for example, DataChannel 1) to the Logical view. 5. Highlight the DataChannel to display its properties. Note that the DataChannel always installs and runs as the default user for the host (the Tivoli Netcool Performance Manager Unix username, pvuser). Review the other property values to make sure they are valid. For the complete list of properties for this component, see the IBM Tivoli Netcool Performance Manager: Property Reference Guide. The DataChannel has the following subelements: v Daily Loader x - Processes 24 hours of raw data every day, merges it together, then loads it into the database. The loader process provides statistics on metric channel tables and metric tablespaces. v Hourly Loader x - Reads files output by the Complex Metric Engine (CME) and loads the data into the database every hour. The loader process provides statistics on metric channel tables and metric tablespaces. The Topology Editor includes the channel number in the element names. For example, DataChannel 1 would have Daily Loader 1 and File Transfer Engine 1. Note: When you add DataChannel x, the Problems view shows that the Input_Components property for the Hourly Loader is blank. This missing value will automatically be filled in when you add a DataLoad collector (as described in the next section) and the error will be resolved.
Procedure
1. Create two directories on the DataChannel host, for example, DATA_DIR to hold the data and EXE_DIR to hold the executable. 2. Change the LOCAL_ROOT_DIRECTORY value on that host's Disk Usage Server to the data root folder DATA_DIR. In the Host advanced properties you will see the DATA_DIR value propagated to all DC folder values for the host. 3. Change DC_ROOT_EXE_DIRECTORY to the executable directory EXE_DIR. This change will propagate to the DC conf directory, the DataChannel Bin Directory and the Datachannel executable file name. Note: For advanced information about DataChannels, see Appendix B, DataChannels, on page 141.
77
Add a Collector
Collectors collect and process raw statistical data about network devices obtained from various network resources. The collectors send the received data through a DataChannel for loading into the Tivoli Netcool Performance Manager database. Note that collectors do not need to be on the same machine as the Oracle server and DataMart.
Collector types
Collector types and their description, plus the steps required to associate a collector with a Technology Pack.
Procedure
1. Install Tivoli Netcool Performance Manager, without creating the UBA collector. 2. Download and install the technology pack. 3. Open the deployed topology file to load the technology pack and add the UBA collector for it. Note: For detailed information about UBA technology packs and the installation process, see the Technology Pack Installation Guide. Configure the installed pack by following the instructions in the pack-specific user's guide.
Restrictions
There are a number of collector restrictions that must be noted. Note the following restrictions:
78
v The maximum collector identification number is 999. v There is no relationship between the channel number and the collector number (that is, there is no predefined range for collector numbers based on channel number). Therefore, collector 555 could be attached to DataChannel 7. v Each database channel can have a maximum of 40 subchannels (and therefore, 40 collectors).
Procedure
1. In the Logical view, right-click the DataChannel x folder. The pop-up menu lists the following options: v Add Collector SNMP - Creates an SNMP collector. v Add Collector UBA - Creates a UBA collector. v Add Collector BCOL - Creates a BCOL collector. This collector type is used in custom technology packs. DataMart must be added to the topology before a BCOL collector can be added. Select Add Collector SNMP. The Configure Collector window opens. 2. Using the drop-down list of available hosts on the Configure Collector window, select the machine that will host the collector (for example, corinth). 3. Accept the default collector number (for example, 1). 4. Click Finish. The Topology Editor displays the new collector under the DataChannel x folder in the Logical view. 5. Highlight the collector to view its properties. The Topology Editor displays both the SNMP collector core parameters and the SNMP technology pack-specific parameters. The core parameters are configured with all SNMP technology packs. You can specify an alternate installation user for the SNMP collector by changing the values of the pv_user, pv_user_group and pv_user_password properties in the Advanced Properties tab. Review the values for the parameters to make sure they are valid. Note: For information about the core parameters, see the IBM Tivoli Netcool Performance Manager: Property Reference Guide.
Results
The collector has two components: v Complex Metric Engine x - Perform calculations on the collected data. v File Transfer Engine (FTE) x - Transfers files from the collector's output directories and places them in the input directory of the CME. The FTE writes data to the file /var/adm/wtmpx on each system that hosts a collector. As part of routine maintenance, check the size of this file to prevent it from growing too large.
79
Note: Your Solaris version can be configured with strict access default settings for secure environments. Strict FTP access settings might interfere with automatic transfers between a DataChannel subchannel and the DataLoad server. Check for FTP lockouts in /etc/ftpd/ftpusers, and check for strict FTP rules in /etc/ftpd/ftpaccess. Note: The Topology Editor includes the channel and collector numbers in the element names. For example, DataChannel 1 could have Collector SNMP 1.1, with Complex Metric Engine 1.1. and File Transfer Engine 1.1.
Procedure
1. In the Logical view, right-click the Cross Collector CME folder and select Add Cross Collector CME from the menu. The Specify the Cross Collector CME details window is displayed. 2. Using the drop-down list of available hosts, select the machine that will host the Cross-Collector CME (for example, corinth). 3. Select the desired Disk Usage Server on the selected host. 4. Select the desired channel number (for example, 1). 5. Click Finish. The Topology Editor adds the new Cross-Collector CME (for example, Cross-Collector CME 2000) to the Logical view. 6. Highlight the Cross-Collector CME to display its properties. Note: The Cross-Collector CME always installs and runs as the default user for the host (the Tivoli Netcool Performance Manager Unix username, pvuser). 7. Review the other property values to make sure they are valid. For the complete list of properties for this component, see the IBM Tivoli Netcool Performance Manager: Property Reference Guide 8. After running the deployer to install the Cross-Collector CME you will need to restart the CMGR process. Note: You will notice that dccmd start all will not start the Cross-Collector CME at this point. 9. You must first deploy a formula against the Cross-Collector CME using the DataChannel frmi tool. Run the frmi tool. The following is an example command:
frmi ecma_formula.js -labels formula_labels.txt
Where: v The format of formula_labels.txt is 2 columns separated by an "=" sign. v First column is Full Path to formula. v Second is the number of the Cross-Collector CME. v The file formula_labels.txt is of the format:
Path_to_ECMA_formulas~Formula1Name=2000 Path_to_ECMA_formulas~Formula2Name=2001
Note: When a Cross-Collector CME (CC-CME) is installed on the system and formulas are applied against it, the removal of collectors that the CC-CME
80
depends on is not supported. This is an exceptional case, that is, if you have not installed a CC-CME, collectors can be removed.
Procedure
1. In the Logical view, right-click the Cross Collector CME folder and select Add multiple Cross Collectors from the menu. The Add Cross Collector CME window is displayed. 2. (Optional) Click Add Hosts to add to the set of Cross Collector hosts. Only hosts that have a DUS can be added. Note: It is recommended that you have 20 Cross Collector CMEs spread across the set of toplolgy hosts. 3. Set the number of Cross Collector CMEs for the set of hosts, there are two ways you can do this: v Click Calculate Defaults to use the wizard to calculate the recommended spread across the added hosts. This will set the number of Cross Collector CMEs to the default value. v To manually set the number of cross collector for each host, use the drop-down menu opposite each host name. 4. Click Finish.
Procedure
1. In the Topology Editor, select Topology then either Save Topology As or Save Topology. Click Browse to navigate to the directory in which to save the file. By default, the topology is saved as topology.xml file in the topologyEditor directory. 2. Accept the default value or choose another name or location, then click OK to close the file browser window. 3. The file name and path is displayed in the original window. Click Finish to save the file and close the window. You are now ready to deploy the topology file (see Starting the Deployer on page 82). Note: Until you actually deploy the topology file, you can continue making changes to it as needed by following the directions in Opening an existing topology file on page 82. See Chapter 5, Modifying the current deployment, on page 95 for more information about making changes to a deployed topology file.
Chapter 3. Installing in a distributed environment
81
Note: Only when you begin the process of deploying a topology is it saved to the database. For more information, see the section Deploying the Topology.
Procedure
1. If it is not already open, open the Topology Editor (see Starting the Topology Editor on page 67). 2. In the Topology Editor, select Topology > Open existing topology. The Open Topology window is displayed. 3. For the topology source, click local then use Browse to navigate to the correct directory and file. Once you have selected the file, click OK. The selected file is displayed in the Open Topology window. Click Finish. The topology is displayed in the Topology Editor. 4. Change the topology as needed.
Primary Deployer
The steps required to run the primary deployer from the Topology Editor
Procedure
Click Run > Run Deployer for Installation. Note: When you use the Run menu options (install or uninstall), the deployer uses the last saved topology file, not the current one. Be sure to save the topology file before using a Run command.
82
Secondary Deployers
A secondary deployer is only required if remote installation using the primary deployer is not possible.
Procedure
v To 1. 2. v To run a secondary deployer from the launchpad: On the launchpad, click Start the Deployer. On the Start Deployer page, click the Start Deployer link. run a secondary deployer from the command line:
1. Log in as root. 2. Change to the directory containing the deployer within the downloaded Tivoli Netcool Performance Manager distribution: On Solaris systems:
# cd <DIST_DIR>/proviso/SOLARIS/Install/SOL10/deployer/
On AIX systems:
# cd <DIST_DIR>/proviso/AIX/Install/deployer/
On Linux systems:
# cd <DIST_DIR>/proviso/RHEL/Install/deployer/
<DIST_DIR> is the directory on the hard drive where you copied the contents of the Tivoli Netcool Performance Manager distribution in Downloading the Tivoli Netcool Performance Manager distribution to disk on page 26. 3. Enter the following command:
# ./deployer.bin
Note: See Appendix D, Deployer CLI options, on page 159 for the list of supported command-line options.
Pre-deployment check
The Deployer will fail if the required patches listed in this file are not installed.
83
Procedure
v To 1. 2. 3. v To 1. 2. 3. check if the required packages are installed: Click Run > Run Deployer for Installation to start the Deployer. Select the Check prerequisites check box. Click Next. The check will return a failure if any of the required files are missing. repair a failure: Log in as root. Install the packages listed as missing. (Linux only) If any openmotif package is listed as missing: Install the missing openmotif package and update the package DB using the command:
# updatedb
Procedure
1. The deployer opens, displaying a welcome page. Click Next to continue. 2. If you started the deployer from the launchpad or from the command line, enter the full path to your topology file, or click Choose to navigate to the correct location. Click Next to continue. Note: If you start the deployer from within the Topology Editor, this step is skipped. The database access window prompts for the security credentials.
84
3. Enter the host name (for example, delphi) and database administrator password (for example, PV), and verify the other values (port number, SID, and user name). Note that if the database does not yet exist, these parameters must match the values you specified when you created the database configuration component (see Add a database configurations component on page 70). Click Next to continue. 4. The node selection window shows the target systems and how the files will be transferred (see Secondary Deployers on page 83 for an explanation of this window). The table has one row for each machine where at least one Tivoli Netcool Performance Manager component will be installed. The default settings are as follows: v The Enable checkbox is selected. If this option is not selected, no actions will be performed on that machine. v The Check prerequisites checkbox is not selected, if selected scripts are run to verify that the prerequisite software has been installed. v Remote execution is enabled, using both RSH and SSH. If remote execution cannot be enabled, perhaps due to a particular customer's security protocols, see Appendix A, Remote installation issues, on page 137 and Resuming a partially successful first-time installation on page 87. v File transfer using FTP is enabled. If desired, reset the values as appropriate for your deployment. Click Next to continue. 5. Provide media location details. The Tivoli Netcool Performance Manager Media Location for components window is displayed, listing component and component platform. a. Click on the Choose the Proviso Media button. You will be asked to provide location of the media for each component. b. Enter the base directory in which your media is located. If any of the component media is not within the directory specified, you will be asked to provide media location detail for that component. 6. The deployer displays summary information about the installation. Review the information, then click Next. The deployer displays the table of installation steps (see Pre-deployment check on page 83 for an overview of the steps table). Note the following: v Regardless of whether the steps are run, or if they pass or fail, closing the wizard will result in the topology being posted to the Tivoli Netcool Performance Manager Database, assuming it exists. v If an installation step fails, see Resuming a partially successful first-time installation on page 87 for debugging information. Continue the installation by following the instructions in Resuming a partially successful first-time installation on page 87. v If the TCR installation step fails, which can happen when there is not enough space available in /usr and /tmp or directory cleanup has not been carried out, run the tcrClean.sh script. To run this script: a. Copy the tcrClean.sh script from the Primary Deployer (host where the Topology Editor is installed) to the server where the TCR installation step fails. The tcrClean.sh script can be found on the Primary Deployer in the directory:
/opt/IBM/proviso/deployer/proviso/bin/Util/
Chapter 3. Installing in a distributed environment
85
b. Run tcrClean.sh. c. When prompted, enter the install location of TCR. d. Continue the installation by following the instructions in Resuming a partially successful first-time installation on page 87 7. Click Run All to run all the steps in sequence. 8. The deployer prompts you for the location of the setup files. Use the file selection window to navigate to the top-level directory for your operating system to avoid further prompts. For example:
<DIST_DIR>/RHEL/
<DIST_DIR> is the directory on the hard drive where you copied the contents of the Tivoli Netcool Performance Manager distribution in Downloading the Tivoli Netcool Performance Manager distribution to disk on page 26. Note: This assumes that the Tivoli Netcool Performance Manager distribution was downloaded to the folder /var/tmp/cdproviso as per the instructions in Downloading the Tivoli Netcool Performance Manager distribution to disk on page 26. If Tivoli Integrated Portal is configured to install on a remote host, the Run Remote TIP Install step is included. This step will prompt the user to enter the root password. The deployer requires this information in order to run as root on the remote host and perform the Tivoli Integrated Portal installation. 9. When all the steps have completed successfully, click Done to close the wizard. 10. Stop and start TCR: a. Navigate to the /tip_install_dir/products/tcr/bin/ directory. b. Set the ORACLE_HOME environment variable. For example:
ORACLE_HOME=/opt/oracle/product/10.2.0 export ORACLE_HOME
d. Run the following scripts: v stopTCRserver.sh <username> <password> v startTCRserver.sh Note: These scripts must be run every time Tivoli Integrated Portal is restarted. Note: The Topology Editor must be closed after every deployment.
Next steps
The steps to perform after deployement. The next step is to install the technology packs, as described in Technology Pack Installation Guide. Once you have created the topology and installed Tivoli Netcool Performance Manager, it is very easy to make changes to the environment. Simply open the deployed topology file (loading it from the database), make your changes, and run
86
the deployer with the updated topology file as input. For more information about performing incremental installations, see Chapter 5, Modifying the current deployment, on page 95. Note: After your initial deployment, always load the topology file from the database to make any additional changes (such as adding or removing a component), because it reflects the current status of your environment. Once you have made your changes, you must deploy the updated topology so that it is propagated to the database. To make any subsequent changes following this deployment, you must load the topology file from the database again. To improve performance, IBM recommends that you regularly compute the statistics on metadata tables. You can compute these statistics by creating a cron entry that executes the dbMgr (Database Manager Utility) analyzeMetaDataTables command at intervals. The following example shows a cron entry that checks statistics every hour at 30 minutes past the hour. Note that the ForceCollection option is set to N, so that statistics will only be calculated when the internal calendar determines that it is necessary, and not every hour:
0 5 * * * [ -f /opt/DM/dataMart.env ] && [ -x /opt/DM/bin/dbMgr ] && . /opt/DM/dataMart.env && dbMgr analyzeMetaDataTables A N
For more information on dbMgr and the analyzeMetaDataTables command, see the Tivoli Netcool Performance Manager dbMgr Reference Guide. For each new SNMP DataLoad, change the env file of the TNPM user to add the directory with the openssh libcrypto.so to the LD_LIBRARY_PATH (or LIBPATH).
Procedure
1. After correcting the problem, restart the deployer from the command line using the following command:
Chapter 3. Installing in a distributed environment
87
./deployer.bin -Daction=resume
Using the resume switch enables you to resume the installation exactly where you left off. Note: If you are asked to select a topology file in order to resume your installation, select the topology file you saved before beginning the install. 2. The deployer opens, displaying a welcome page. Click Next to continue. 3. Accept the default location of the base installation directory of the Oracle JDBC driver (/opt/oracle/product/version/jdbc/lib), or click Choose to navigate to another directory. Click Next to continue. 4. The steps page shows the installation steps in the very same state they were in when you stopped the installation (with the completed steps marked Success, the failed step marked Error, and the remaining steps marked Held). 5. Select the step that previously failed, reset it to Ready, then click Run Next. Verify that this installation step now completes successfully. 6. Run any remaining installation steps, verifying that they complete successfully. 7. At the end of the installation, the deployer loads the updated topology information into the database.
88
Overview
A minimal deployment installation is used primarily for demonstration or evaluation purposes, and installs the product on the smallest number of machines possible, with minimal user input. This installation type installs all the Tivoli Netcool Performance Manager components on the local host using a predefined topology file to define the infrastructure. The minimal deployment installation also installs the MIB-II SNMP technology pack. When you perform a minimal deployment installation, the Tivoli Netcool Performance Manager components are installed on the server you are running the deployer from.
89
Special consideration
By default, Tivoli Netcool Performance Manager uses Monday to determine when a new week begins. If you wish to specify a different day, you must change the FIRST_WEEK_DAY parameter in the Database Registry using the dbRegEdit utility. This parameter can only be changed when you first deploy the topology that installs your Tivoli Netcool Performance Manager environment, and it must be changed BEFORE the Database Channel is installed. For more information, see the Tivoli Netcool Performance Manager Database Administration Guide.
90
Procedure
v The product distribution site: https://www-112.ibm.com/software/howtobuy/ softwareandservices Located on the product distribution site are the ProvisoPackInstaller.jar file, the bundled jar file, and individual stand-alone technology pack jar files. v (Optional) The Tivoli Netcool Performance Manager CD distribution, which contains the ProvisoPackInstaller.jar file and the jar files for the Starter Kit components. See your IBM customer representative for more information about obtaining software. Note: Technology Pack Installer and the MIB-II jar files must be in the same directory (for example, AP), and no other application jar files should be present, if there are any more jars in that folder the installation step will fail due to "too many jars" in the specified folder. In addition, you must add the AP directory to the Tivoli Netcool Performance Manager distribution's directory structure.
Procedure
1. Log in as root. 2. Set and export the DISPLAY variable (see Setting up a remote X Window display on page 14). 3. Set and export the BROWSER variable to point to your Web browser. For example: On Solaris systems:
# BROWSER=/opt/mozilla/mozilla # export BROWSER
On AIX systems:
# BROWSER=/usr/mozilla/firefox/firefox # export BROWSER
On Linux systems:
Chapter 4. Installing as a minimal deployment
91
Note: The BROWSER command cannot include any spaces around the equal sign. 4. Change directory to the directory where the launchpad resides. On Solaris systems:
# cd <DIST_DIR>/proviso/SOLARIS
On AIX systems:
# cd <DIST_DIR>/proviso/AIX
On Linux systems:
# cd <DIST_DIR>/proviso/RHEL
<DIST_DIR> is the directory on the hard drive where you copied the contents of the Tivoli Netcool Performance Manager distribution in Downloading the Tivoli Netcool Performance Manager distribution to disk on page 26. 5. Enter the following command to start the launchpad:
# ./launchpad.sh
Procedure
1. On the launchpad, click the Install Tivoli Netcool Performance Manager 1.3.1 for Minimal Deployment option in the list of tasks, then click the Install Tivoli Netcool Performance Manager 1.3.1 for Minimal Deployment link to start the deployer. Alternatively, you can start the deployer from the command line, as follows: a. Log in as root. b. Set and export your DISPLAY variable (see Setting up a remote X Window display on page 14). c. Change directory to the directory that contains the deployer: On Solaris systems:
# cd <DIST_DIR>/proviso/SOLARIS/Install/SOL10/deployer
On AIX systems:
# cd <DIST_DIR>/proviso/AIX/Install/deployer
On Linux systems:
# cd <DIST_DIR>/proviso/RHEL/Install/deployer
2. The deployer opens, displaying a welcome page. Click Next to continue. 3. Accept the terms of the license agreement, then click Next. 4. Accept the default location of the base installation directory of the Oracle JDBC driver (/opt/oracle/product/version/jdbc/lib), or click Choose to navigate to another directory. Click Next to continue.
92
5. The deployer prompts for the directory in which to install Tivoli Netcool Performance Manager. Accept the default value (/opt/proviso) or click Choose to navigate to another directory. Click Next to continue. 6. Verify the following additional information about the Oracle database: v Oracle Base. The base directory for the Oracle installation (for example, /opt/oracle). Accept the provided path or click Choose to navigate to another directory. v Oracle Home. The root directory of the Oracle database (for example, /opt/oracle/product/10.2.0). Accept the provided path or click Choose to navigate to another directory. v Oracle Port. The port used for Oracle communications. The default value is 1521. Click Next to continue. 7. The node selection window shows the target system and how the files will be transferred. These settings are ignored for a minimal deployment installation because all the components are installed on a single server. Click Next to continue. 8. Provide media location details. The Tivoli Netcool Performance Manager Media Location for components window is displayed, listing component and component platform. a. Click on the Choose the Proviso Media button. You will be asked to provide location of the media for each component. b. Enter the base directory in which your media is located. If any of the component media is not within the directory specified, you will be asked to provide media location detail for that component. 9. The deployer displays summary information about the installation. Review the information, then click Next to begin the installation. The deployer displays the table of installation steps (see Pre-deployment check on page 83 for an overview of the steps table). Note the following: v If an installation step fails, see Appendix I, Error codes and log files, on page 187 for debugging information. Continue the installation by following the instructions in Resuming a partially successful first-time installation on page 87 v Some of the installation steps can take a long time to complete. However, if an installation step fails, it will fail in a short amount of time. 10. Click Run All to run all the steps in sequence. 11. When all the steps have completed successfully, click Done to close the wizard. 12. Run chmod -R 777 on /opt/IBM/tivoli in order to make all files in the TIP directory structure accessible. Your installation is complete. See The post-installation script on page 94 for information about the post-installation script, or Next steps on page 94 for what to do next.
93
Procedure
1. 2. 3. 4. Starts the DataChannel. Starts the DataLoad SNMP Collector, if it is not already running. Creates a DataView user named tnpm. Gives the poc user permission to view reports under the NOC Reporting group, with the default password of tnpm.
Results
The script writes a detailed log to the file /var/tmp/poc-postinstall.${TIMESTAMP}.log.
Next steps
The steps to be performed following the deployment of your system. When the installation is complete, you are ready to perform the final configuration tasks that enable you to view reports on the health of your network. These steps are documented in detail in the Tivoli Netcool Performance Manager documentation set. For information about the MIB-II Technology Pack, see the MIB-II Technology Pack User's Guide. For information about installing additional technology packs, see the Technology Pack Installation Guide For each new SNMP DataLoad, change the env file of the TNPM user to add the directory with the openssh libcrypto.so to the LD_LIBRARY_PATH (or LIBPATH).
94
Procedure
1. If it is not already open, open the Topology Editor (see Starting the Topology Editor on page 67). 2. In the Topology Editor, select Topology > Open existing topology. The Open Topology window is displayed. 3. For the topology source, select From database (v. 443) and click Next. 4. Verify that all of the fields for the database connection are filled in with the correct values:
95
v Database hostname - The name of the database host. The default value is localhost. v Port - The port number used for communication with the database. The default value is 1521. v Database user - The user name used to access the database. The default value is PV_INSTALL. v Database Password - The password for the database user account. For example, PV. v SID - The SID for the database. The default value is PV. If desired, click Save as defaults to save these values for future incremental installations. Click Finish.
Results
The topology is retrieved from the database and is displayed in the Topology Editor.
Procedure
1. If it is not already open, open the Topology Editor (see Starting the Topology Editor on page 67). 2. Open the existing topology (see Opening a deployed topology on page 95). 3. In the Logical view of the Topology Editor, right-click the folder for the component you want to add. 4. Select Add XXX from the pop-up menu, where XXX is the name of the component you want to add. 5. The Topology Editor prompts for whatever information is needed to create the component. See the appropriate section for the component you want to add: v Add the hosts on page 68 v Add a database configurations component on page 70 v Add a DataMart on page 71 v Add a Discovery Server on page 73 v Add a Tivoli Integrated Portal on page 74 v Add a DataView on page 75 v Add the DataChannel administrative components on page 76 v Add a DataChannel on page 76 v Add a Collector on page 78 Note: that if you add a collector to a topology that has already been deployed, you must manually bounce the DataChannel management
96
components (cnsw, logw, cmgrw, amgrw). For more information, see Manually starting the Channel Manager programs on page 146. v Add a Discovery Server on page 73 6. The new component is displayed in the Logical view of the Topology Editor. 7. Save the updated topology. You must save the topology after you add the component and before you run the deployer. This step is not optional. 8. Run the deployer (see Starting the Deployer on page 82), passing the updated topology as input. The deployer can determine that most of the components described in the topology are already installed, and installs only the new component. 9. When the installation ends successfully, the deployer uploads the updated topology into the database. For information about removing a component from the Tivoli Netcool Performance Manager environment, see Removing a component from the topology on page 129.
Example
In this example, you update the installed version of Tivoli Netcool Performance Manager to add a new DataChannel and two SNMP DataLoaders to the existing system. To update the Tivoli Netcool Performance Manager installation: 1. If it is not already open, open the Topology Editor (see Starting the Topology Editor on page 67). 2. Open the existing topology (see Opening a deployed topology on page 95). 3. In the Logical view of the Topology Editor, right-click the DataChannels folder. 4. Select Add Data Channel from the pop-up menu. Following the directions in Add a DataChannel on page 76, add the following components: a. Add a new DataChannel (Data Channel 2) with two different SNMP DataLoaders to the topology. The Topology Editor creates the new DataChannel. b. Add two SNMP collectors to the channel structure created by the Topology Editor. The editor automatically creates a Daily Loader component, an Hourly Loader component, and two Sub Channels with an FTE component and a CME component. 5. Save the updated topology. 6. Run the deployer (see Starting the Deployer on page 82), passing the updated topology as input. The deployer can determine that most of the components described in the topology are already installed, and installs only the new components (in the example, DataChannel 5 with two new subchannels and DataLoaders). 7. When the installation ends, successful or not, the deployer uploads the updated topology into the database.
97
Procedure
1. tart the Topology Editor (if it is not already running) and open the topology that includes the component's current host (see Starting the Topology Editor on page 67 and Opening a deployed topology on page 95). 2. In the Logical view, navigate to the name of the component to move. 3. Right-click the component name, then click Change Host from the pop-up menu.
98
The Migrate Component dialog appears, containing a drop-down list of hosts where you can move the component. 4. Select the name of the new host from the list, then click Finish. The name of the new host appears in the Properties tab.
Procedure
1. Start the Topology Editor (if it is not already running) and open the topology that includes the collector's current host (see Starting the Topology Editor on page 67 and Opening a deployed topology on page 95). 2. In the Logical view, navigate to the name of the collector to move. For example if moving SNMP 1.1, navigate as follows: DataChannels > DataChannel 1 > Collector 1.1 > Collector SNMP.1.1 3. Right-click the collector name (for example, Collector SNMP 1.1), then click Migrate from the pop-up menu. The Migrate Collector dialog appears, containing a drop-down list of hosts where you can move the collector. Note: If you are moving a collector that has not been deployed, select Change host from the pop-up menu (Migrate is grayed out). After the Migrate Collector dialog appears, continue with the steps below. 4. Select the name of the new host from the list, then click Finish. In the Physical view, the status of the collector on the new host is Configured. The status of the collector on the original host is To be uninstalled. You will remove the collector from the original host in Step 9. Note: If you are migrating a collector that has not been deployed, the name of the original host is automatically removed from the Physical view. 5. Click Topology > Save Topology to save the topology data. 6. Click Run > Run Deployer for Installation to run the deployer, passing the updated topology as input. For more information on running the deployer, see Starting the Deployer on page 82. The deployer installs the collector on the new host and starts it.
99
Note: Both collectors are now collecting data - the original collector on the original host, and the new collector on the new host. 7. Before continuing with the steps below, note the current time, and wait until a time period equivalent to two of the collector's collection periods elapses. Doing so guards against data loss between collections on the original host and the start of collections on the new host. Because data collection on the new host is likely to begin sometime after the first collection period begins, the data collected during the first collection period will likely be incomplete. By waiting for two collection time periods to elapse, you can be confident that data for one full collection period will be collected. The default collection period is 15 minutes. You can find the collection period for the sub-element, sub-element group, or collection formula associated with the collector in the DataMart Request Editor. For information on viewing and setting a collection period, see the Tivoli Netcool Performance Manager DataMart Configuration Guide. 8. Bounce the FTE for the collector on the collector's new host, as in the following example:
./dccmd bounce FTE.1.1
The FTE now recognizes the collector's configuration on the new host, and will begin retrieving data from the collector's output directory on the new host. 9. In the current Topology Editor session, click Run > Run Deployer for Uninstallation to remove the collector from the original host, passing the updated topology as input. For more information, see Removing a component from the topology on page 129. Note: This step is not necessary if you are moving a collector that has not been deployed.
Procedure
1. Move the collector as described in Moving a deployed SNMP collector on page 99. Note: If you are moving a spare collector out of the HAM environment, the navigation path is different than the path shown in Step 2 of the above instructions. For example, suppose you have a single HAM environment with a cluster MyCluster on host MyHost, and you are moving the second SNMP spare out of the HAM. The navigation path to the spare would be as follows: DataChannels > Administrative Components > High Availability Managers > HAM MyServer.1 > MyCluster > Collector Processes > Collection Process SNMP Spare 2. 2. Log in as Tivoli Netcool Performance Manager Unix user, pvuser, on the collector's new host. 3. Change to the directory where DataLoad is installed. For example:
100
cd /opt/dataload
6. Edit the file dataLoad.env and set the field DL_HA_MODE as follows: v Set DL_HA_MODE=true if you moved the collector onto a HAM host. v Set DL_HA_MODE=false if you moved the collector off of a HAM host. 7. Source the DataLoad environment again:
. ./dataLoad.env
Note: If you move an SNMP collector to or from a HAM host, you must bounce the HAM. For information, see Stopping and restarting modified components on page 125.
Procedure
1. Log in as pvuser to the DataChannel host where the UBA collector is running. 2. Change to the directory where DataChannel is installed. For example:
cd /opt/datachannel
4. Stop the collector's UBA and FTE components. For example, to stop these components for UBA collector 1.1, run the following commands: dccmd stop UBA.1.1 and... dccmd stop FTE.1.1 For information on the dccmd command, see the Tivoli Netcool Performance Manager Command Line Interface Guide. Note: Some technology packs have additional pack-specific components that must be shut down - namely, BLB (bulk load balancer) and IF (inventory file) components. IF component names have the format xxxIF, where xxx is a pack-specific name. For example, Cisco CWM packs have a CWMIF component, Alcatel 5620 SAM packs have a SAMIF component, and Alcatel 5620 NM packs have a QCIF component. Other packs do not use these technology-specific components.
101
5. Tar up the UBA collector's UBA directory. You will copy this directory to the collector's new host later in the procedure (Step 13). For example, to tar up a UBA directory for UBA collector 1.1, run the following command: Note: This step is not necessary if the collector's current host and the new host share a file system.
tar -cvf UBA_1_1.tar ./UBA.1.1/*
Note: Some technology packs have additional pack-specific directories that need to be moved. These directories have the same names as the corresponding pack-specific components described in Step 4. 6. Start the Topology Editor (if it is not already running) and open the topology that includes the collector's current host (see Starting the Topology Editor on page 67 and Opening a deployed topology on page 95). 7. In the Logical view, navigate to the name of the collector to move - for example, Collector UBA.1.1. 8. Right-click the collector name and select Migrate from the pop-up menu. The Migrate Collector dialog appears, containing a drop-down list of hosts where you can move the collector. 9. Select the name of the new host from the list, then click Finish. In the Physical view, the status of the collector on the new host is Configured. The collector is no longer listed under the original host. Note: If the UBA collector was the only DataChannel component on the original host, the collector will be listed under that host, and its status will be "To be uninstalled." You can remove the DataChannel installation from the original host after you finish the steps below. For information on removing DataChannel from the host, see Removing a component from the topology on page 129. 10. Click Topology > Save Topology to save the topology. 11. Click Run > Run Deployer for Installation to run the deployer, passing the updated topology as input. For more information on running the deployer, see Starting the Deployer on page 82. If DataChannel is not already installed on the new host, this step installs it. 12. Click Run > Run Deployer for Uninstallation to remove the collector from the original host, passing the updated topology as input. For more information, see Removing a component from the topology on page 129. 13. Copy any directory you tarred in Step 5 and the associated JavaScript files to the new host. Note: This step is not necessary if the collector's original host and the new host share a file system. For example, to copy UBA_1_1.tar and the JavaScript files from the collector's original host: a. Log in as pvuser to the UBA collector's new host. b. Change to the directory where DataChannel is installed. For example:
cd /opt/datachannel
c. FTP to the collector's original host. d. Run the following commands to copy the tar file to the new host. For example:
102
e. Change to the directory where the JavaScript files for the technology pack associated with the collector are located:
cd /opt/datachannel/scripts
f. FTP the JavaScript files from the /opt/datachannel/scripts directory on the original host to the /opt/datachannel/scripts directory on the new host. 14. Log in as pvuser to the Channel Manager host where the Administrator Components (including CMGR) are running. 15. Stop and restart the Channel Manager by performing the following steps: a. Change to the $DC_HOME directory (typically, /opt/datachannel). b. Source the DataChannel environment:
. dataChannel.env
The process ID appears in the output immediately after the user ID, as shown below in bold:
pvuser 6561 6560 0 Aug 21 ?
d. Stop the CMGR process. For example, if 6561 is the CMGR process ID: kill -9 6561 e. Change to the $DC_HOME/bin directory (typically, /opt/datachannel/bin). f. Restart CMGR by running the following command:
./cmgrw
16. Log in as pvuser to the UBA collector's new host and change to the $DC_HOME/bin directory (typically, /opt/datachannel/bin). 17. Run the following command to verify that Application Manager (AMGR) is running on the new host:
./findvisual
If the AMGR process is running, you will see output that includes an entry like the following:
pvuser 6684 6683 0 Aug 21 ?
Note: If AMGR is not running on the new host, do not continue. Verify that you have performed the preceding steps correctly. 18. Start the collector's UBA and FTE components on the new host. For example, to start these components for collector 1.1, run the following commands: ./dccmd start UBA.1.1 and... ./dccmd start FTE.1.1 Note: If any pack-specific components were shut down on the old host (see Step 4), you must also start those components on the new host.
103
104
Overview
The High Availability Manager (HAM) is an optional component for large installations that want to use redundant SNMP collection paths. The HAM constantly monitors the availability of one or more SNMP collection hosts, and switches collection to a backup host (called a spare) if a primary host becomes unavailable. The following figure shows a simple HAM configuration with one primary host and one spare. In the panel on the left, the primary host is operating normally. SNMP data is being collected from the network and channeled to the primary host. In the panel on the right, the HAM has detected that the primary host is unavailable, so it dynamically unbinds the collection path from the primary host and binds it to the spare.
HAM basics
An SNMP collector collects data from a specific set of network resources according to a set of configuration properties. A collector has two basic parts: the collector process running on the host computer, and the collector profile that defines the collector's properties. Note: Do not confuse a "collector profile" with an "inventory profile." A collector profile contains properties used in the collection of data from network resources properties such as collector number, polling interval, and output directory for the collected data. An inventory profile contains information used to discover network resources - properties such as the addresses of the resources to look for and the mode of discovery. A collector that is not part of a HAM environment is static - that is, the collector process and the collector profile are inseparable. But in a HAM environment, the
Copyright IBM Corp. 2006, 2010
105
collector process and collector profile are managed as separate entities. This means that if a collector process is unavailable (due to a collector process crash or a host machine outage), the HAM can dynamically reconfigure the collector, allowing data collection to continue. The HAM does so by unbinding the collector profile from the unavailable collector process on the primary host, and then binding the collector profile to a collector process on a backup (spare) host. Note: It may take several minutes for the HAM to reconfigure a collector, depending on the amount of data being collected.
106
Clusters
A HAM environment can consist of a single set of hosts or multiple sets of hosts. Each set of hosts in a HAM environment is called a cluster. A cluster is a logical grouping of hosts and collector processes that are managed by a HAM. The use of multiple clusters is optional. Whether you use multiple clusters or just one has no affect on the operation of the HAM. Clusters simply give you a way to separate one group of collectors from another, so that you can better deploy and manage your primary and spare collectors in a way that is appropriate for your needs. Multiple clusters may be useful if you have a large number of SNMP collector hosts to manage, or if the hosts are located in various geographic areas. The clusters in a given HAM environment are distinct from one another. In other words, the HAM cannot bind a managed definition in one cluster to a collector process in another.
107
1 + 1, fixed spare
A fixed spare cluster with one primary host and one designated spare. The figure below shows a fixed spare cluster with one primary host and one designated spare: v In the panel on the left, Primary1 is functioning normally. The designated spare is idle.
108
v In the panel on the right, Primary1 experiences an outage. The HAM unbinds the collector from Primary1 and binds it to the designated spare. v With the spare in use and no other spares in the HAM cluster, failover can no longer occur - even after Primary1 returns to service. For failover to be possible again, you must reassign Collector 1 to Primary1. This idles the collector process on the spare, making it available for the next failover operation if Primary 1 fails again.
Note: When a designated spare serves as the only spare for a single primary, as in a 1+1 fixed spare cluster, the HAM pre-loads the primary's collector definition on the spare. This results in a fast failover with a likely loss of no more than one collection cycle. The following table shows the bindings that the HAM can and cannot make in this cluster:
Collector Collector 1 Possible Host Bindings Primary1 (default binding) Designated spare Host Bindings Not Possible -
2 + 1, fixed spare
A fixed spare cluster with two primary hosts and one designated spare The figure below shows a fixed spare cluster with two primary hosts and one designated spare: v In the panel on the left, Primary1 and Primary2 are functioning normally. The designated spare is idle. v In the panel on the right, Primary2 experiences an outage. The HAM unbinds the collector from Primary2 and binds it to the designated spare. v With the spare in use and no other spares in the HAM cluster, failover can no longer occur - even after Primary2 returns to service. For failover to be possible again, you must reassign Collector 2 to Primary2. This idles the collector process on the spare, making it available for the next failover operation.
109
The following table shows the bindings that the HAM can and cannot make in this cluster:
Collector Collector 1 Possible Host Bindings Primary1 (default binding) Designated spare Collector 2 Primary2 (default binding) Designated spare Primary1 Host Bindings Not Possible Primary2
v The following figure shows the same cluster after Primary2 has returned to service. In the panel on the left, Primary2 is idle, prepared to act as backup if needed.
110
v In the panel on the right, Primary1 experiences an outage. The HAM unbinds the collector from Primary1 and binds it to the floating spare, Primary2.
The following table shows the bindings that the HAM can and cannot make in this cluster:
Collector Collector 1 Possible Host Bindings Primary1 (default binding) Primary2 Designated spare Collector 2 Primary1 Primary2 (default binding) Designated spare Host Bindings Not Possible -
3+ 2, fixed spares
A fixed spare cluster with three primary hosts and two designated spares. The figure below shows a fixed spare cluster with three primary hosts and two designated spares: v In the panel on the left, all three primaries are functioning normally. The designated spares are idle. v In the panel on the right, Primary3 experiences an outage. The HAM unbinds the collector from Primary3 and binds it to Designated Spare 2. The HAM chose Designated Spare 2 over Designated Spare 1 because the managed definition for Collector 3 set the failover priority in that order. Note: Each managed definition sets its own failover priority. Failover priority can be defined differently in different managed definitions. v With one spare in use and one other spare available (Designated Spare 1), failover is now limited to the one available spare - even after Primary3 returns to service. For dual failover to be possible again, you must reassign Collector 3 to Primary3.
111
The following table shows the bindings that the HAM can and cannot make in this cluster:
Collector Collector 1 Possible Host Bindings Primary1 (default binding) Designated Spare 1 Designated Spare 2 Collector 2 Primary2 (default binding) Designated Spare 1 Designated Spare 2 Collector 3 Primary3 (default binding) Designated Spare 1 Designated Spare 2 Primary1 Primary2 Primary1 Primary3 Host Bindings Not Possible Primary2 Primary3
112
The following table shows the bindings that the HAM can and cannot make in this cluster:
Collector Collector 1 Possible Host Bindings Primary1 (default binding) Primary2 Primary3 Designated Spare 1 Designated Spare 2 Collector 2 Primary1 Primary2 (default binding) Primary3 Designated Spare 1 Designated Spare 2 Collector 3 Primary1 Primary2 Primary3 (default binding) Designated Spare 1 Designated Spare 2 Host Bindings Not Possible -
113
Resource pools
When you configure a managed definition in the Topology Editor, you specify the hosts that the HAM can bind to the managed definition, and also the priority order in which the hosts are to be bound. This list of hosts is called the resource pool for the managed definition. A resource pool includes: v The managed definition's primary host and collector process (that is, the host and collector process that are bound to the managed definition by default). v Zero or more other primary hosts in the cluster. If you add a primary host to a managed definition's resource pool, that primary host becomes a floating spare for the managed definition. v Zero or more designated spares in the cluster. Typically, each managed definition includes one or more designated spares in its resource pool. Note: If no managed definitions include a designated spare in their resource pools, there will be no available spares in the cluster, and therefore failover cannot occur in the cluster.
The SNMP collector can reside in one of the following states, as shown in the following table:
114
Event
Description Initial state; a collector number may or may not be assigned; the collection profile has not been loaded. Intermediate state between Idle and Ready. Occurs after a Load event. Collector number is assigned, and the collection profile is being loaded. Collector number assigned, profile loaded, but not scheduling requests or performing collections. Intermediate state between Idle and Running. Occurs after a Start event. Collector number assigned, and profile is being loaded. Actively performing requests and collections. Intermediate state between Running and Idle.
Loading
Load
Ready
N/A
Starting
Start
Running Stopping
N/A Stop/Pause
The following state diagram shows how the SNMP collector transitions through its various states depending upon events or time-outs:
How failover works with the HAM and the SNMP collector
How Failover Works With the HAM and the SNMP Collector The following tables illustrate how the HAM communicates with the SNMP collectors during failover for a 1+1 cluster and a 2+1 cluster.
115
Table 5. HAM and SNMP Collector in a 1+1 Cluster State of Primary Running Idle State of Spare Events and Actions The HAM sends the spare the Load event for the specified collection profile. The HAM sends a Pause event to the spare to extend the timeout. Note: If the timeout expires, the spare will perform start actions and transition to a Running state. The HAM sends a Pause event to the collector process that has been in a Running state for a shorter amount of time. The HAM sends a Start event to the spare.
Running
Ready
Running
Running
No response
Ready
Table 6. HAM and SNMP Collector in a 2+1 Cluster State of Primary Running Running Running Idle Ready Running State of Spare Events and Actions No action No action The HAM sends a Stop event to the collector process that has been in Running state for the shorter amount of time. The HAM sends a Start event to the spare. The HAM sends a Start event to the spare.
No Response No Response
Idle Ready
Because more than one physical system may produce SNMP collections, the File Transfer Engine (FTE) must check every capable system for a specific profile. The FTE retrieves all output for the specific profile. Any duplicated collections are reconciled by the Complex Metrics Engine (CME).
116
The following list describes EXTENDED STATUS information: v 1.1 - Load # Collection profile 1.1 v Ok: - Status of the load. Ok means it is properly collected, Not Running indicates a severe problem (data losses) v (box1:3012 -> Running 1.1 for 5h2m26s)- The collector that is currently performing the load, with its status and uptime. v No avail spare - List of possible spare, if something happens to the collector currently working. In this example there is no spare available, a failover would fail. A list of host:port would indicate the possible spare machines. v Check: box4:3002, box5:3002 - Indicates what is currently wrong with the system/configuration. Machines box4:3002 and box5:3002 should be spare but are either not running, or not reachable. The user is instructed to check these machines. For a 1-to-1 failover configuration, the dccmd command might return output like the following:
$ dccmd status HAM.SERVER.1 COMPONENT APPLICATION HOST STATUS ES DURATION EXTENDED STATUS HAM.SERVER.1 HAM SERVER running 10010 1.1 Ok: (box1:3002 -> Running 1.1 for 5h2m26s); 1 avail spare: (box2:3002 -> Ready 1.1)
This preceding output shows that Collector 1.1 is in a Running state on Box1, and that the Collector on Box2 is in a Ready state, with the profile for Collector 1.1 loaded.
Procedure
1. Install all collectors. 2. Configure and start the HAM. 3. Install all technology packs. 4. Perform the discovery.
117
Topology prerequisites
The minimum component prerequisite. A 3+1 HAM cluster requires that you have a topology with the following minimum components: v Three hosts, each bound to an SNMP collector. These will act as the primary hosts. You will create a managed definition for each of the primary hosts. v One additional host that is not bound to an SNMP collector. This will act as the designated spare. For information on installing these components, see Adding a new component on page 96.
Procedures
The general procedures for creating a single-cluster HAM with one designated spare and three floating spares.
Procedure
1. Start the Topology Editor (if it is not already running) and open the topology where you want to add the HAM (see Starting the Topology Editor on page 67 and Opening a deployed topology on page 95). 2. In the Logical view, right-click High Availability Managers, located at DataChannels > Administrative Components. 3. Select Add High Availability Manager from the pop-up menu. The Add High Availability Manager Wizard appears. 4. In the Available hosts field, select the host where you want to add the HAM. Note: You can install the HAM on a host where a collector process is installed, but you cannot install more than one HAM on a host. 5. In the Identifier field, accept the default identifier. The identifier has the following format: HAM.<HostName>.<n> where HostName is the name of the host you selected in Step 4, and n is a HAM-assigned sequential number, beginning with 1, that uniquely identifies this HAM from others that may be defined on other hosts. 6. Click Finish. The HAM identifier appears under the High Availability Managers folder. 7. Right-click the identifier of the HAM you just created. 8. Select Add Cluster from the pop-up menu. The Add Cluster Monitor Wizard appears. 9. In the Identifier field, type a name for the cluster and click Finish. The cluster name appears under the HAM identifier folder you added in Step 6. The following folders appear under the cluster name: v Collector Processes v Managed Definitions
118
Note: To add additional clusters to the environment, repeat Step 7 through Step 9.
Procedure
1. In the Logical view, right-click the Collector Processes folder that you created in Step 9 of the previous section, Create the HAM and a HAM cluster on page 118. 2. Select Add Collection Process SNMP Spare from the pop-up menu. The Add Collection Process SNMP Spare - Configure Collector Process SNMP Spare dialog appears. 3. In the Available hosts field, select the host that you want to make the designated spare. This field contains the names of hosts in the Physical view that do not have SNMP collectors assigned to them. 4. In the Port field, specify the default port number, 3002, for the spare's collector process, then click Finish. Under the cluster's Collector Processes folder, the entry Collection Process SNMP Spare <n> appears, where n is a HAM-assigned sequential number, beginning with 1, that uniquely identifies this designated spare from others that may be defined in this cluster. Note: Repeat Step 1 through Step 4 to add an additional designated spare to the cluster.
What to do next
Should you be making changes to an already existing configuration, please make sure the dataLoad.env file contains all the right settings: 1. Change to the directory where DataLoad is installed. For example:
cd /opt/dataload
3. Make sure that DL_HA_MODE field in the dataLoad.env file and set to DL_HA_MODE=true. 4. Source the DataLoad environment again:
. ./dataLoad.env
119
Procedure
1. In the Logical view, right-click the Managed Definitions folder that you created in Create the HAM and a HAM cluster on page 118. 2. Select Add Managed Definition from the pop-up menu. The Add Managed Definition - Choose Managed Definition dialog appears. 3. In the Collector number field, select the unique collector number to associate with this managed definition. 4. Click Finish. The following entries now appear for the cluster: v Under the cluster's Managed Definitions folder, the entry Managed Definition <n> appears, where n is the collector number you selected in Step 3. v Under the cluster's Collector Processes folder, the entry Collector Process [HostName] appears, where HostName is the host that will be bound to the SNMP collector you selected in Step 3. This host is the managed definition's primary host. Note: Repeat Step 1 though to Step 4 to add another managed definition to the cluster.
Example
When you finish adding managed definitions for a 3+1 HAM cluster, the Logical and Physical views might look like the following:
120
In this example, the hosts dcsol1a, dcsol1b, and docserver1 are the primaries, and docserver2 is the designated spare.
Procedure
1. Right-click a managed definition in the cluster's Managed Definitions folder. 2. Select Configure Managed Definition from the pop-up menu. The Configure Managed Definition - Collector Process Selection dialog appears, as shown below. In this example, the resource pool being configured is for Managed Definition 1 (that is, the managed definition associated with Collector 1).
121
3. In the Additional Collector Processes list, check the box next to each host to add to the managed definition's resource pool. Typically, you will add at least the designated spare (in this example, docserver2) to the resource pool. If you add a primary host to the resource pool, that host becomes a floating spare for the managed definition. Note: You must add at least one of the hosts in the Additional Collector Processes list to the resource pool. Since the goal in this example is to configure all primaries as floating spares, the designated spare and the two primaries (docserver1 and dcsol1a) will be added to the resource pool. 4. When finished checking the hosts to add to the resource pool, click Next. Note: If you add just one host to the resource pool, the Next button is not enabled. Click Finish to complete the definition of this resource pool. Return to Step 1 to define a resource pool for the next managed definition in the cluster, or skip to Save and start the HAM on page 123 if you are finished defining resource pools. The Configure Managed Definition - Collector Process Order dialog appears, as shown below:
122
5. Specify the failover priority order for this managed definition. To do so: a. Select a host to move up or down in the priority list, then click the Up or Down button until the host is positioned where you want. b. Continue moving hosts until the priority list is ordered as you want. c. Click Finish. In this example, if the primary associated with Managed Definition 1 fails, the HAM will attempt to bind the managed definition to the floating spare dcsol1a. If dcsol1a is in use or otherwise unavailable, the HAM attempts to bind the managed definition to docserver1. The designated spare docserver2 is last in priority. 6. Return to Step 1 to define a resource pool for the next managed definition in the cluster, or continue with the next section if you are finished defining resource pools.
Procedure
1. Click Topology > Save Topology to save the topology file containing the HAM configuration. 2. Run the deployer (see Starting the Deployer on page 82), passing the updated topology file as input. 3. Open a terminal window on the DataChannel host. 4. Log in as pvuser. 5. Change your working directory to the DataChannel bin directory (/opt/datachannel/bin by default), as follows:
cd /opt/datachannel/bin
6. Bounce (stop and restart) the Channel Manager. For instructions, see Step 15 on page 111. 7. Run the following command:
dccmd start ham
Chapter 6. Using the High Availability Manager
123
Monitoring of the HAM environment begins. For information on using dccmd, see the Tivoli Netcool Performance Manager Command Line Interface Guide.
124
Procedure
1. Open a terminal window on the DataChannel host. 2. Log in as pvuser. 3. Change your working directory to the DataChannel bin directory (/opt/datachannel/bin by default), as follows:
cd /opt/datachannel/bin
For example: v To bounce the HAM with the identifier HAM.dcsol1b.1, run:
dccmd bounce ham.dcsol1b.1
v To bounce the FTE for collector 1.1 that is managed by a HAM, run:
dccmd bounce fte.1.1
You do not need to bounce the HAM that the FTE and collector are in. For information on using dccmd, see the Tivoli Netcool Performance Manager Command Line Interface Guide. 5. Bounce the Channel Manager. For instructions, see Step 15.
Procedure
1. Right-click the collector process or managed definition to view. 2. Select Show from the pop-up menu. The Show Collector Process... or Show Managed Definition... dialog appears. The following sections describe the contents of these dialogs.
125
The configuration values are described as follows: v dcsol1a. The primary host where this collector process runs. v 3002. The port through which the collector process receives SNMP data. v 3 2 (Primary) 1. The managed definitions that the HAM can bind to this collector process. The values have the following meanings: 3. The managed definition for Collector 3. 2 (Primary). The managed definition for Collector 2. This is the default managed definition for the collector process. 1. The managed definition for Collector 1.
126
Note the following about this managed definition's resource pool: v The priority order of the hosts is from top to bottom - therefore, the first collector process that the HAM will attempt to bind to this managed definition is the one on host dcsol1a. The collector process on host docserver2 is last in the priority list. v The first three hosts are floating spares. They are flagged as such by each having a primary managed definition. v The host docserver2 is the only designated spare in the resource pool. It is flagged as such by not having a primary managed definition.
127
128
129
v If you are uninstalling a DataChannel component, the component should first be stopped. If you are uninstalling all DataChannel components on a host, then you should remove the DataChannel entries from the crontab. v If you delete a DataChannel or collector, the working directories (such as the FTE and CME) are not removed; you must delete these directories manually. v When a Cross-Collector CME (CC-CME) is installed on the system and formulas are applied against it, the removal of collectors that the CC-CME depends on is not supported. This is an exceptional case, that is, if you have not installed a CC-CME, collectors can be removed. DataView restrictions: Uninstall DataView manually if other products are installed in the same Tivoli Integrated Portal instance. If other products are installed in the same Tivoli Integrated Portal instance, you must use the following procedure to uninstall a DataView component: 1. Run the uninstall command:
<tip_location>/products/tnpm/dataview/bin/uninstall.sh <tip_location> <tip_administrator_username> <tip_administrator_password>
3. In the Topology Editor: 4. Remove the DataView component. 5. 6. 7. 8. Save the topology. Run the deployer for uninstallation. Mark the DataView step successful. Run the unregister DataView step.
Note: Once this manual un-install is completed, the DataView instance will remain in the topology after the un-install operation completes, this is not usually the case for un-installed components.
Removing a component
To remove component from the topology.
Procedure
1. If it is not already open, open the Topology Editor (see Starting the Topology Editor on page 67). 2. Open the existing topology (see Opening a deployed topology on page 95). 3. In the Logical view of the Topology Editor, right-click the component you want to delete and select Remove from the pop-up menu. 4. The editor marks the component as "To Be Removed" and removes it from the display. 5. Save the updated topology. 6. Run the deployer (see Starting the Deployer on page 82). Note: If you forgot to save the modified topology, the deployer will prompt you to save it first. The deployer can determine that most of the components described in the topology file are already installed, and removes the component that is no longer part of the topology.
130
7. The deployer displays the installation steps page, which lists the steps required to remove the component. Note that the name of the component to be removed includes the suffix "R" (for "Remove"). For example, if you are deleting a DataChannel, the listed component is DCR. 8. Click Run All to run the steps needed to delete the component. 9. When the installation ends successfully, the deployer uploads the updated topology file into the database. Click Done to close the wizard. Note: If you remove a component and redeploy the file, the Topology Editor view is not refreshed automatically. Reload the topology file from the database to view the updated topology.
What to do next
If you have uninstalled DataChannel Components, you will need to bounce CMGR after you have run the deployer, so it will pick up the updated configuration and realize the components have been removed. If you do not bounce CMGR after the deployer runs then you may get errors when you the components are restarted.
Order of uninstall
The order in which you must uninstall components.
Procedure
1. DataLoad and DataChannel When uninstalling DataChannel from a host, you must run ./dccmd stop all, disable or delete the dataload cron processes and manually stop (kill -9) any running channel processes (identified by running findvisual). See Appendix B, DataChannels, on page 141 for more information about the findvisual command. 2. DataMart 3. DataChannel Administrative Components and any remaining DataChannel components. 4. DataView Also remove Tivoli Integrated Portal/Tivoli Common Reporting. 5. Tivoli Netcool Performance Manager Database (remove only after all the other components have been removed). The database determines the operating platform of the Tivoli Netcool Performance Manager environment.
131
Procedure
1. You can start the uninstaller from within the Topology Editor or from the command line. To start the uninstaller from the Topology Editor: v Select Run > Run Deployer for Uninstallation. To start the uninstaller from the command line: a. Log in as root. b. Set and export your DISPLAY variable (see Setting up a remote X Window display on page 14). c. Change directory to the directory that contains the deployer. For example:
# cd /opt/IBM/proviso/deployer
2. The uninstaller opens, displaying a welcome page. Click Next to continue. 3. Accept the default location of the base installation directory of the Oracle JDBC driver (/opt/oracle/product/version/jdbc/lib), or click Choose to navigate to another directory. Click Next to continue. 4. A pop-up opens, asking whether you want to download the topology from the database. Click Yes. 5. The database access window prompts for the security credentials. Enter the host name (for example, delphi) and database administrator password (for example, PV), and verify the other values (port number, SID, and user name). Click Next to continue.
132
6. The uninstaller displays a message stating that the topology download was successful and saved to the file /tmp/ProvisoConsumer/Topology.xml. Click Next to continue. 7. The uninstaller displays several status messages, then displays a message stating that the environment status was successfully downloaded and saved to the file /tmp/ProvisoConsumer/Discovery.xml. Click Next to continue. 8. A pop-up opens, stating that no operations need to be executed on your nodes. The uninstaller closes. 9. Repeat the process on each machine in the deployment. Note: After the removal of each Component using the Topology Editor, the Topology Editor should reload the topology from the Database.
Procedure
1. Log in as root. 2. Set and export your DISPLAY variable (see Setting up a remote X Window display on page 14). 3. Change directory to the install_dir/uninstall directory. For example:
# cd /opt/IBM/proviso/uninstall
5. The Uninstall wizard opens. Click Uninstall to uninstall the Topology Editor. 6. When the script is finished, click Done.
Residual files
How to remove the possible residual files that may exist after the uninstall process.
133
Procedure
1. Log in as oracle. 2. Enter the following commands to stop Oracle:
sqlplus "/ as sysdba" shutdown abort exit lsnrctl stop
3. As root, enter the following commands to delete these files and directories:
rm rm rm rm rm rm rm rm rm rm rm rm rm rm rm -fR -fR -fR -fR -fR -fR -fR -fR -fR -fR -fR -fR -fR -fR -fR /tmp/PvInstall /var/tmp/PvInstall /opt/Proviso /opt/proviso $ORACLE_BASE/admin/PV $ORACLE_BASE/admin/skeleton $ORACLE_HOME/dbs/initPV.ora $ORACLE_HOME/dbs/lkPV $ORACLE_HOME/dbs/orapwPV $ORACLE_HOME/lib/libpvmextc.so $ORACLE_HOME/lib/libmultiTask.so $ORACLE_HOME/lib/libcmu.so $ORACLE_HOME/bin/snmptrap $ORACLE_HOME/bin/notifyDBSpace $ORACLE_HOME/bin/notifyConnection
where $ORACLE_BASE is /opt/oracle and $ORACLE_HOME is /opt/oracle/product/10.2.0. 4. Enter the following commands to clear your Oracle mount points and remove any files in those directories:
rm -r /raid_2/oradata/* rm -r /raid_3/oradata/*
5. Enter the following command to delete the temporary area used by the deployer:
rm -fr /tmp/ProvisoConsumer
7. Delete the startup file, netpvmd. v For Solaris, use the command:
rm /etc/init.d/netpvmd
What to do next
Following TCR uninstallation: To prevent any possible system instability caused by residual processes post-uninstall of TCR, run the tcrClean.sh script on all systems where TCR has been uninstalled: 1. On the host where the TCR installation failed, change to the directory containing tcrClean.sh:
cd /opt/IBM/proviso/deployer/proviso/bin/Util/
2. Run tcrClean.sh 3. When prompted, enter the location where TCR was installed.
134
Note: If you have uninstalled TCR on a remote host, the tcrClean.sh file will need to be sent using ftp to the remote host for execution.
135
136
Procedure
v Option 1: 1. Unselect the Remote Command Execution option during the installation. The deployer creates and transfers the directory with the required component package in it. 2. As root, log in to the remote system and manually run the run.sh script. v Option 2: Follow the directions outlined in Installing on a remote host using a secondary deployer on page 138.
137
Procedure
v Option 1: 1. Unselect the FTP option during the installation. The deployer creates a directory containing the required component package. 2. Copy the required component directory to the target system. 3. As root, log in to the remote system and manually run the run.sh script. v Option 2: Follow the directions outlined in Installing on a remote host using a secondary deployer.
Procedure
1. Copy the Tivoli Netcool Performance Manager distribution to the server on which you would like to set up the secondary deployer, that is, copy the distribution to corinth. For more information on copying the Tivoli Netcool Performance Manager distribution to a server, see Downloading the Tivoli Netcool Performance Manager distribution to disk on page 26. 2. Open the Topology Editor on the primary deployer host, that is, on delphi, and add the remote component to the topology definition. You may have completed this task already when creating your original topology definition. If you have already added the remote component to your topology definition, skip to the next step. 3. Deploy the new topology containing the added component using the Topology Editor. This is done by clicking Run > Run Deployer for Installation. This will push the edited topology to the database. 4. Open the Deployer on corinth by doing the following:
138
a. Connect to corinth, change to the directory where you have downloaded the product distribution, and launch the deployer either in graphical mode (by starting the Launchpad and clicking Start Deployer) or CLI mode (by navigating to the directory containing the deployer and entering the command ./deployer.bin). b. Enter the database credentials when prompted. The deployer connects to the database. For more information on how to run a secondary deployer, see Secondary Deployers on page 83. Note: Due to Step 3, the secondary deployer sees the topology data and knows that the required component is still to be installed on corinth. 5. Follow the on screen instructions to install the desired component. Note: You cannot launch the deployer simultaneously from two different hosts. Only one deployer can be active at any given time.
139
140
Appendix B. DataChannels
This section provides detailed information about the DataChannel architecture.
Data collection
DataChannel data collection. A Tivoli Netcool Performance Manager DataChannel consists of a number of components, including the following: v File Transfer Engine (FTE) v Complex Metric Engine (CME) v Daily Database Loader (DLDR) v Hourly Database Loader (LDR) v Plan Builder (PBL) v Channel Manager The FTE, DLDR, LDR, and PBL components are assigned to each configured DataChannel. The FTE and CME components are assigned to one or more Collector subchannels. Data is produced by Tivoli Netcool Performance Manager DataLoad Collectors. Both SNMP and BULK Collectors are fed into a subchannel's channel processor. Data moves through the CME and is synchronized in the Hourly Loader. The Hourly Loader computes group aggregations from resource aggregation records. The Daily Loader provides statistics on metric channel tables and metric tablespaces and inserts data into the database. Data is moved from one channel component to another as files. These files are written to and read from staging directories between each component. Within each staging directory there are subdirectories named do, output, and done. The do subdirectory contains files that are waiting to be processed by a channel component. The output subdirectory stores data for the next channel component to work on. After files are processed, they are moved to the done directory. All file movement is accomplished by the FTE component.
Data aggregation
A DataChannel aggregates data collected by collectors for eventual use by DataView reports. The DataChannel provides online statistical calculations of raw collected data, and detects real-time threshold violations. Aggregations include: v Resource aggregation for every metric and resource v Group aggregation for every group v User-defined aggregation computed from raw data Threshold detections in real time include: v Raw data violating configured thresholds
Copyright IBM Corp. 2006, 2010
141
v Raw data violating configured thresholds and exceeding the threshold during a specific duration of time v Averaged data violating configured thresholds
* The actual component's executable file seen in the output of ps -ef is named XXX_visual, where XXX is an entry in this column. For example, the file running for CMGR is seen as CMGR_visual.
The watchdog scripts run every few minutes from cron. Their function is to monitor their corresponding management component, and to restart it, if necessary. You can add watchdog scripts for the Channel Manager programs to the crontab for the pvuser on each host on which you installed a DataChannel component. To add watchdog scripts to the crontab: 1. Log in as pvuser. Make sure this login occurs on the server running the Channel Manager components. 2. At a shell prompt, go to the DataChannel conf subdirectory. For example:
$ cd /opt/datachannel/conf
3. Open the file dc.cron with a text editor. (The dc.cron files differ for different hosts running different DataChannel programs. The following example shows the dc.cron file for the host running the Channel Manager programs.)
0,5,10,15,20,25,30,35,40,45,50,55 1,6,11,16,21.26,31,36,41,46,51,56 2,7,12,17,22.27,32,37,42,47,52,57 3,8,13,18,23.28,33,38,43,48,53,58 1-31 1-31 1-31 1-31 1-12 1-12 1-12 1-12 0-6 0-6 0-6 0-6 /opt/datachannel/bin/cnsw > /dev/null 2>&1 /opt/datachannel/bin/logw > /dev/null 2>&1 /opt/datachannel/bin/cmgrw > /dev/null 2>&1 /opt/datachannel/bin/amgrw > /dev/null 2>&1
4. Copy the lines in the dc.cron file to the clipboard. 5. At another shell prompt, edit the crontab for the current user.
crontab -e
A text editor session opens, showing the current crontab settings. 6. Paste the lines from the dc.cron tab into the crontab file. For example:
142
0 * * * * [ -f /opt/datamart/dataMart.env ] 0,5,10,15,20,25,30,35,40,45,50,55 1-31 1-12 1,6,11,16,21,26,31,36,41,46,51,56 1-31 1-12 2,7,12,17,22,27,32,37,42,47,52,57 1-31 1-12 3,8,13,18,23,28,33,38,43,48,53,58 1-31 1-12
&& [ -x /opt/datamart/bin/pollinv ] && .... 0-6 /opt/datachannel/bin/cnsw > /dev/null 2>&1 0-6 /opt/datachannel/bin/logw > /dev/null 2>&1 0-6 /opt/datachannel/bin/cmgrw > /dev/null 2>&1 0-6 /opt/datachannel/bin/amgrw > /dev/null 2>&1
7. Save and exit the crontab file. 8. Repeat steps 1 to 8 on each DataChannel host, with this difference: The dc.cron file on collector and loader hosts will have only one line, like this example:
0,5,10,15,20,25,30,35,40,45,50,55 1-31 1-12 0-6 /opt/datachannel/bin/amgrw > /dev/null 2>&1
On such hosts, this is the only line you need to add to the pvuser crontab.
UBA.n.c
UBA.1.100
CME.n.s
CME.2.1
* The actual application's executable file visible in the output of ps -ef is named XXX_visual, where XXX is an entry in this column.
Note: For historical reasons, the SNMP DataLoad collector is managed by Tivoli Netcool Performance Manager DataMart, and does not appear in Table 11.
Appendix B. DataChannels
143
Procedure
v Verify that the DataChannel management programs are running: 1. Log in as pvuser on each DataChannel host. 2. Change to the DataChannel installation's bin subdirectory. For example:
$ cd /opt/datachannel/bin
In the resulting output, look for: The AMGR process on every DataChannel host The CNS, CMGR, LOG, and AMGR processes on the Channel Manager host v If the DataChannel management programs are running on all DataChannel hosts, start the application programs on all DataChannel hosts by following these steps: 1. Log in as pvuser. Make sure this login occurs on the host running the Channel Manager programs. 2. Change to the DataChannel installation's bin subdirectory. For example:
$ cd /opt/datachannel/bin
3. Run the following command to start all DataChannel applications on all configured DataChannel hosts:
./dccmd start all
See the Tivoli Netcool Performance Manager Command Line Interface Guide for information about the dccmd command.
If the collector is running, you will see output similar to the following:
pvuser 27118 1 15 10:03:27 pts/4 0:06 /opt/dataload/bin/pvmd -nologo
-noherald /opt/dataloa
Procedure
1. Log into the server that is running Tivoli Netcool Performance Manager SNMP DataLoad by entering the username and password you specified when installing SNMP DataLoad. 2. Source the DataLoad environment file by entering the following command:
./$DLHOME/dataLoad.env
144
where $DLHOME is the location where SNMP DataLoad is installed on the system (/opt/dataload, by default). Note: If DataLoad shares the same server as DataMart, make sure you unset the environment variable by issuing the following command from a BASH shell command line:
unset PV_PRODUCT
The command displays the following message when the SNMP collector has been successfully started:
PVM Collecting Daemon is running.
Results
The script controlling the starting and stopping of SNMP collectors, pvmdmgr, prevents the possibility that multiple collector instances can be running simultaneously. If a user starts a second instance, that second instance will die by itself in under two minutes without ever contacting or confusing the relevant watchdog script.
Appendix B. DataChannels
145
Procedure
v To start the Channel Manager programs manually: 1. Log in as pvuser on the host running the Channel Manager programs. 2. At a shell prompt, change to the DataChannel bin subdirectory. For example:
$ cd /opt/datachannel/bin
3. Enter the following commands at a shell prompt, in this order: For the Channel Name Server, enter:
./cnsw
v To manually start the DataChannel programs on all hosts in your DataChannel configuration: 1. Start the Channel Manager programs, as described in the previous section. 2. On each DataChannel host, start the amgrw script. 3. On the Channel Manager host, start the application programs as described in Starting the DataChannel management programs on page 143.
Procedure
1. On the DataChannel host, log in as the component user, such as pvuser.
146
2. Change your working directory to the DataChannel bin directory (/opt/datachannel/bin by default) using the following command: $ cd /opt/datachannel/bin 3. Shut down the DataChannel FTE. Prior to shutting down all DataChannel components, some DataChannel work queues must be emptied. To shut down the DataChannel FTE and empty the work queues:
$ ./dccmd stop FTE.*
4. Let all DataChannel components continue to process until the .../do directories for the FTE and CME components contain no files. The .../do directories are located in the subdirectories of $DCHOME (typically, /opt/datachannel) that contain the DataChannel components - for example, FTE.1.1, CME.1.1. 5. Shut down all CMEs on the same hour (So the operator state files will be in synch with each other). To accomplish this: a. Identify the leading CME by either looking at the do and done directories in each CME and the DAT files inside there; or using dccmd status all to see which CME is reporting the latest hour in its processing status. b. All CMEs on that hour must be stopped and then continue using the same approach to finding the hour being processed and stop each CME as it reaches the same hour until all CMEs are stopped. CMEs are stopped using the command:
$ ./dccmd stop CME
Note: For details on how to restart a DataChannel, see Manually starting the Channel Manager programs on page 146.
DataChannel terminology
Terms used throughout the Tivoli Netcool Performance Manager DataChannel installation procedure. v Collector Subchannel: A subdivision of a DataChannel, with each Collector subchannel associated with a single Collector and CME. The division into Collector subchannels helps eliminate latency or loss of data caused by delayed Collectors. If a Collector subchannel disconnects for a period of time, only that Collector is affected, and all other Collector subchannels continue processing. The number of Collector subchannels per DataChannel differs according to the needs of a particular deployment. See the Tivoli Netcool Performance Manager Configuration Recommendations for information related to system configuration requirements. The terms Collector and Collectors are used to refer to Collector subchannel and Collector subchannels. v Complex Metric Engine (CME): A DataChannel program that performs on-the-fly calculations on raw metrics data for a DataChannel. These calculations include time aggregations for resources, as well as statistical calculations using raw data, thresholds, properties, and constants as inputs. If CME formulas are defined for the incoming metrics data, the values are processed by those
Appendix B. DataChannels
147
formulas. The CME synchronizes metadata at the beginning of each hour, and only processes the metadata that exists for the hour. CORBA (Common Object Request Broker Architecture): An industry-standard programming architecture that enables pieces of programs, called objects, to communicate with one another regardless of the programming language that they were written in or the operating system they are running on. Daily Database Loader (DLDR): A DataChannel program that gathers statistical data processed by a DataChannel's CME and inserts it into the Tivoli Netcool Performance Manager database. There is one Daily Loader for each DataChannel; it is part of the channel processor component of the DataChannel. DataChannel Remote (DCR): A DataChannel installation configuration in which the subchannel, CME and FTE components are installed and run on one host, while the Loader components are installed and run on another host. In this configuration, the subchannel hosts can continue processing data and detecting threshold violations, even while disconnected from the Channel Manager server. DataChannel Standard: A DataChannel installation configuration in which all component programs of each subchannel are installed and run on the same server. DataChannel Standard installation is described in this chapter. DataLoad Bulk Collector: A DataLoad program that processes different data formats and resource files supplied by network devices, network management platforms, network probes, and other types of resources such as BMC Patrol. The Bulk Collector translates bulk statistics provided in flat files into Tivoli Netcool Performance Manager metadata and data. If operating in synchronized inventory mode, the Bulk Collector passes the resources and properties to the Tivoli Netcool Performance Manager DataMart Inventory application.
v DataLoad SNMP Collector: A DataLoad program that collects data from network resources via SNMP polling. The SNMP Collector provides raw data files to the DataChannel for processing by the CME. v DataLoad UBA Bulk Collector: A DataLoad program that imports data from files (referred to as Bulk input files) generated by non-SNMP devices, including Alcatel 5620 NM, Alcatel SAM, and Cisco CWM. These Bulk input files contain formats that the Bulk Collector is unable to handle. v Discovery Server (DISC): A DataChannel program that runs as a daemon to perform an inventory of SNMP network devices from which to gather statistical data. v Hourly Database Loader (LDR): A DataChannel program that serves as the point of data synchronization and merging, and of late data processing, for each DataChannel. The Hourly Loader gathers files generated by the CME, computes group aggregations from the individual resource aggregation records, and loads the data into the Tivoli Netcool Performance Manager database. v File Transfer Engine (FTE): A DataChannel program that periodically scans the Collector output directories, examines the global execution plan to determine which computation requests require the data, then sorts the incoming data for storage. v Next-Hour Policy: Specifies the number of seconds to wait past the hour for files to arrive before the next hourly directory is created. The default value causes the DataChannel to wait until 15 minutes after the hour before it starts processing data for the next hour. To avoid losing data, you need to set a percentage and a time-out period during the configuration of the CME. v Plan Builder (PBL): A DataChannel program that creates the metric data routing and processing plan for the other components in the DataChannel.
148
Overview
An aggregation set is a grouping of network management raw data and computed statistical information stored in the Tivoli Netcool Performance Manager database for a single timezone. For example, if your company provides network services to customers in both the Eastern and Central US timezones, you must configure two aggregation sets. Because each aggregation set is closely linked with a timezone, aggregation sets are sometimes referred to as timezones in the in Tivoli Netcool Performance Manager documentation. However, the two concepts are separate. Note: "Aggregation set" is abbreviated to "Aggset" in some setup program menus.
Procedure
1. Log in as root. (Remain logged in as root for the remainder of this appendix.) 2. At a shell prompt, change to the directory where Tivoli Netcool Performance Manager DataMart program files are installed. For example:
# cd /opt/datamart
3. Load the DataMart environment variables into your current shell's environment using the following command:
# . ./dataMart.env
149
# cd bin
6. Type 1 at the Choice prompt and press Enter to enter the password for PV_ADMIN. The script prompts twice for the password you set up for PV_ADMIN.
==> Enter password for PV_ADMIN : PV ==> Re-enter password : PV
Note: The script obtains the DB_USER_ROOT setting from the Tivoli Netcool Performance Manager database configured in previous chapters, and constructs the name of the Tivoli Netcool Performance Manager database administrative login name, PV_ADMIN, from that base. If you set a different DB_USER_ROOT setting, the "Database user" entry reflects your settings. For example, if you previously set DB_USER_ROOT=PROV, this script would generate the administrative login name PROV_ADMIN. 7. To configure the first aggregation set, type 2 at the Choice prompt and press Enter twice. The script shows the current settings for the aggregation set with ID 0 (configured by default):
The following Time Zones are defined into the Database : ___________________________________________________________________________________ id | Date (in GMT) | offset in | Name | Aggset status | | seconds | | ___________________________________________________________________________________ 0 | 1970/01/01 00:00:00 | 0 | GMT | Aggset created ==> Press <Enter> to continue ....
You can use this aggregation set as-is, or modify it to create a new timezone. 8. Press Enter. A list of predefined timezones and their timezone numbers is displayed:
150
Num | OffSet | Time zone Name | Short | Long | Hours | | Description | Description ___________________________________________________________________________________ [ 1] : -10:00 | America/Adak | HAST | Hawaii-Aleutian Standard Time [ 2] : -10:00 | Pacific/Rarotonga | CKT | Cook Is. Time [ 3] : -09:00 | America/Anchorage | AKST | Alaska Standard Time [ 4] : -09:00 | AST | AKST | Alaska Standard Time [ 5] : -08:00 | PST | PST | Pacific Standard Time [ 6] : -07:00 | MST | MST | Mountain Standard Time [ 7] : -06:00 | America/Mexico_City| CST | Central Standard Time [ 8] : -06:00 | CST | CST | Central Standard Time [ 9] : -05:00 | EST | EST | Eastern Standard Time [10] : -04:00 | America/Santiago | CLT | Chile Time [11] : -03:00 | America/Sao_Paulo | BRT | Brazil Time [12] : -01:00 | Atlantic/Azores | AZOT | Azores Time [13] : 000:00 | Europe/London | GMT | Greenwich Mean Time [14] : +01:00 | Europe/Paris | CET | Central European Time [15] : +01:00 | ECT | CET | Central European Time [16] : +02:00 | Africa/Cairo | EET | Eastern European Time [17] : +02:00 | Europe/Helsinki | EET | Eastern European Time [18] : +02:00 | Europe/Bucharest | EET | Eastern European Time [19] : +03:00 | Asia/Baghdad | AST | Arabia Standard Time [20] : +03:00 | Europe/Moscow | MSK | Moscow Standard Time [21] : +04:00 | Asia/Baku | AZT | Azerbaijan Time [22] : +05:00 | Asia/Yekaterinburg | YEKT | Yekaterinburg Time [23] : +06:00 | Asia/Novosibirsk | NOVT | Novosibirsk Time [24] : +07:00 | Asia/Krasnoyarsk | KRAT | Krasnoyarsk Time [25] : +08:00 | Asia/Irkutsk | IRKT | Irkutsk Time [26] : +09:00 | Asia/Yakutsk | YAKT | Yakutsk Time [27] : +10:00 | Australia/Sydney | EST | Eastern Standard Time (New South Wales) [28] : +11:00 | Pacific/Noumea | NCT | New Caledonia Time [29] : +12:00 | Pacific/Auckland | NZST | New Zealand Standard Time [30] : +12:00 | Asia/Anadyr | ANAT | Anadyr Time ==> Select Time Zone number [1-30 ] (E : Exit) : 9
9. Type the number of the timezone you want to associate with aggregation set 0. For example, type 9 for Eastern Standard Time. The script prompts:
==> Select an Aggset ID to add/modify (E: Exit) : 0
To associate the specified timezone, EST, with the database's default aggregation set, type 0. 10. The script asks whether you want your aggregation set to include Daylight Saving Time (DST) transition dates:
Does your Time Zone manage DST [Y/N] : Y
For most time zones, type Y and press Enter. 11. The script displays the results:
151
Complete with Success ... The following Time Zone has been modified: ___________________________________________________________________________________ id | Date (in GMT) | offset in | Name | Aggset status | | seconds | | ___________________________________________________________________________________ 0 | 1970/01/01 00:00:00 | 0 | GMT | Aggset created 0 | 2004/09/29 22:00:00 | -14400 | EST_2004_DST | Aggset created 0 | 2004/10/31 06:00:00 | -18000 | EST_2004 | Aggset created 0 | 2005/04/03 07:00:00 | -14400 | EST_2005_DST | Aggset created 0 | 2005/10/30 06:00:00 | -18000 | EST_2005 | Aggset created 0 | 2006/04/02 07:00:00 | -14400 | EST_2006_DST | Aggset created 0 | 2006/10/29 06:00:00 | -18000 | EST_2006 | Aggset created 0 | 2007/04/01 07:00:00 | -14400 | EST_2007_DST | Aggset created 0 | 2007/10/28 06:00:00 | -18000 | EST_2007 | Aggset created 0 | 2008/04/06 07:00:00 | -14400 | EST_2008_DST | Aggset created 0 | 2008/10/26 06:00:00 | -18000 | EST_2008 | Aggset created 0 | 2009/04/05 07:00:00 | -14400 | EST_2009_DST | Aggset created 0 | 2009/10/25 06:00:00 | -18000 | EST_2009 | Aggset created 0 | 2010/04/04 07:00:00 | -14400 | EST_2010_DST | Aggset created 0 | 2010/10/31 06:00:00 | -18000 | EST_2010 | Aggset created ==> Press <Enter> to continue ....
Note: The dates that appear in your output will most likely be different from the dates that appear in the example. 12. Press Enter to return to the script's main menu. 13. To configure a second aggregation set, type 2 at the Choice prompt and press Enter three times. 14. Specify the timezone number of your second timezone. For example, type 8 to specify Central Standard Time. The script prompts:
==> Select an Aggset ID to add/modify (E: Exit) : 1
If you enter a set number that does not exist in the database, the script creates a new aggregation set with that number. Type the next available set number, 1. 15. Respond Y to the timezone management query. The script shows the results of creating the second aggregation set:
____________ The following Time Zone has been modified : ___________________________________________________________________________________ ______________ id | Date (in GMT) | offset in | Name | Aggset status | | seconds | | ___________________________________________________________________________________ ______________ 1 | 2004/09/29 23:00:00 | -18000 | CST_2004_DST | Aggset created 1 | 2004/10/31 07:00:00 | -21600 | CST_2004 | Aggset created 1 | 2005/04/03 08:00:00 | -18000 | CST_2005_DST | Aggset created 1 | 2005/10/30 07:00:00 | -21600 | CST_2005 | Aggset created 1 | 2006/04/02 08:00:00 | -18000 | CST_2006_DST | Aggset created 1 | 2006/10/29 07:00:00 | -21600 | CST_2006 | Aggset created 1 | 2007/04/01 08:00:00 | -18000 | CST_2007_DST | Aggset created 1 | 2007/10/28 07:00:00 | -21600 | CST_2007 | Aggset created 1 | 2008/04/06 08:00:00 | -18000 | CST_2008_DST | Aggset created 1 | 2008/10/26 07:00:00 | -21600 | CST_2008 | Aggset created 1 | 2009/04/05 08:00:00 | -18000 | CST_2009_DST | Aggset created 1 | 2009/10/25 07:00:00 | -21600 | CST_2009 | Aggset created 1 | 2010/04/04 08:00:00 | -18000 | CST_2010_DST | Aggset created 1 | 2010/10/31 07:00:00 | -21600 | CST_2010 | Aggset created ==> Press <Enter> to continue ....
16. Press Enter to return to the main menu, where you can add more aggregation sets, or type 0 to exit.
152
The next step is to install the aggregation sets on the server on which you installed Tivoli Netcool Performance Manager DataMart.
Procedure
1. Make sure your EDITOR environment variable is set. 2. Change to the /opt/Proviso directory:
cd /opt/Proviso
4. Type 1 at the Choice prompt and press Enter. The Install menu is displayed:
Tivoli Netcool Performance Manager <version number> - [Install] 1. Tivoli Netcool Performance Manager Database Configuration 0. Previous Menu Choice [1]> 1
Procedure
1. Type 1 at the Choice prompt and press Enter. Setup displays the installation environment menu:
153
Tivoli Netcool Performance 1. PROVISO_HOME 2. DATABASE_DEF_HOME 3. CHANNELS_DEF_HOME 4. AGGRSETS_DEF_HOME 5. Continue 0. Exit Choice [5]> 5
[installation environment]
Note: Menu options 2, 3, and 4 are used later in the installation process. 2. Make sure the value for PROVISO_HOME is the same one you used when you installed the database configuration. If it is not, type 1 at the Choice prompt and correct the directory location. 3. The script displays the component installation menu:
Tivoli 1. 2. 3. 0. Choice Netcool Performance Manager Database Configuration Database Channel Aggregation set Exit [1]> 3 <version number> [component installation]
4. Type 3 at the Choice prompt and press Enter. The script displays the installation environment menu:
Tivoli Netcool Performance Manager Aggregation Set <version number> - [installation 1. PROVISO_HOME : /opt/Proviso 2. ORACLE_HOME : /opt/oracle/product/10.2.0 3. ORACLE_SID : PV 4. DB_USER_ROOT : 5. Continue 0. Previous Menu Choice [5]> 4 environment]
5. Type 4 at the Choice prompt and press Enter to specify the same value for DB_USER_ROOT that you specified in previous chapters. This manual's default value is PV.
Enter value for DB_USER_ROOT [] : PV
6. Make sure that the values for PROVISO_HOME, ORACLE_HOME, and ORACLE_SID are the same ones you entered in previous chapters. Correct the values if necessary. 7. Type 5 at the Choice prompt and press Enter. Setup displays the Aggregation Set installation options menu:
Tivoli Netcool Performance Manager Aggregation Set 1. List of configured aggregation sets 2. List of installed aggregation sets 3. Number of the aggregation set to install : 4. Channel where to install aggregation set : 5. Start date of aggregation set : 6. Continue 0. Back to options menu Choice [6]> <version number> - [installation options]
Note: Do not change the value for option 4. Retain the default value, "all." 8. The first time you use any menu option, the script prompts for the password for PV_ADMIN:
Enter password for PV_ADMIN : PV
154
9. Use menu option 1 to list the aggregation sets you configured in Configuring aggregation sets on page 149. The script displays a list similar to the following:
============= LIST OF CONFIGURED AGGREGATION SETS ============ Num Effect Time Name ---- --------------------- ------------------------------------------0 01-01-1970 00:00:00 GMT 04-01-2007 07:00:00 EST_2007_DST 04-02-2006 07:00:00 EST_2006_DST 04-03-2005 07:00:00 EST_2005_DST 04-04-2010 07:00:00 EST_2010_DST 04-05-2009 07:00:00 EST_2009_DST 04-06-2008 07:00:00 EST_2008_DST 09-29-2004 22:00:00 EST_2004_DST 10-25-2009 06:00:00 EST_2009 10-26-2008 06:00:00 EST_2008 10-28-2007 06:00:00 EST_2007 10-29-2006 06:00:00 EST_2006 10-30-2005 06:00:00 EST_2005 10-31-2004 06:00:00 EST_2004 10-31-2010 06:00:00 EST_2010 1 04-01-2007 08:00:00 CST_2007_DST 04-02-2006 08:00:00 CST_2006_DST 04-03-2005 08:00:00 CST_2005_DST 04-04-2010 08:00:00 CST_2010_DST 04-05-2009 08:00:00 CST_2009_DST 04-06-2008 08:00:00 CST_2008_DST 09-29-2004 23:00:00 CST_2004_DST 10-25-2009 07:00:00 CST_2009 10-26-2008 07:00:00 CST_2008 10-28-2007 07:00:00 CST_2007 10-29-2006 07:00:00 CST_2006 10-30-2005 07:00:00 CST_2005 10-31-2004 07:00:00 CST_2004 10-31-2010 07:00:00 CST_2010 2 aggregation sets configured Press enter... Time lag -------+0h -4h -4h -4h -4h -4h -4h -4h -5h -5h -5h -5h -5h -5h -5h -5h -5h -5h -5h -5h -5h -5h -6h -6h -6h -6h -6h -6h -6h
10. Select option 2 to list the aggregation sets already installed. The output is similar to the following:
============== LIST OF CREATED AGGREGATION SETS ============== ============ X: created ==== #: partially created ============ Channels 0 | 1 AggSets ----------------------------------------------------------------------| 0 X Press enter...
Remember that aggregation set 0 is automatically installed when you install the database channel, and continues to be installed even if you modified set 0 by assigning a different timezone. 11. Select option 3 to designate the aggregation set to install. In the examples above, set 0 is already installed, but set 1 is waiting to be installed. Thus, enter 1 at the prompt:
Enter Aggregation Set number between 1 and 998 : 1
12. By default, the date to start collecting data on the designated aggregation set is today's date. You can instead use menu option 5 to designate a future date to start collecting data. Set an appropriate future date for your installation.
Enter start date (GMT) using Oracle format yyyy.mm.dd-hh24 : 2009.08.31-00 WARNING! Start date is set in the future. No loading is allowed until start date (GMT) is reached. Do you confirm the start date (Y/N) [N] ? y
155
13. When all menu parameters are set, type 6 at the Choice prompt and press Enter.
2. Do not make changes to this file unless you have explicit instructions from Professional Services. Only if you have guidelines from Professional Services for advanced configuration of your aggregation sets, make the suggested edits. Save and close the file. 3. When you close the configuration file, the script checks the file parameters and starts installing the aggregation set. The installation takes three to ten minutes, depending on the speed of your server. A message like the following is displayed when the installation completes:
P R O V I S O A g g r e g a t i o n S e t <version number> |||||||||||||||||||||||| AggregationSet installed Tivoli Netcool Performance Manager Aggregation Set 1 on Channel
1 successfully inst
Press Enter...
156
Procedure
v Best practice: Use a separate calendar for each timezone. If you link multiple timezones to the same calendar, a change to one timezone calendar setting will affect all the timezones linked to that calendar. v To link a group to a timezone: 1. Create a calendar with the DataView GUI, or use the default CME Permanent calendar. 2. Create a text file (for example, linkGroupTZ.txt) with the following format: Each line has three fields separated by |_|. The first field is a DataView group name. The second field is a timezone name from the Tivoli Netcool Performance Manager internal timezone list. See Configuring aggregation sets on page 149 for a list of timezone names. The third field is the name of the calendar you create, or CME Permanent. The following example line demonstrates the file format:
~Group~USEast|_|EST_2005_DST|_|CME Permanent|_|
Enter as many lines as you have timezone entries in your aggregation set configuration. 3. At a shell prompt, enter a command similar to the following, which uses the Resource Manager's CLI to link the group to the timezone:
resmgr -import segp -colNames "npath tz.name cal.name" -file linkGroupTZ.txt
v To review timezone to group associations: Use the resmgr command. For example:
resmgr -export segp -colNames "name tz.name cal.name" -file link.txt
157
158
-Daction=patch
-Daction=poc
-Daction=resume
-Daction=uninstall
-DCheckUser
-DOracleClient=oracle_client_home
-DOracleServerHost=hostname -DOracleServerPort=port
159
Description Specifies the Oracle server ID. Default is PV. Specifies the administrator username for the Oracle server. Default is PV_INSTALL. Specifies the administrator password for the Oracle server. Default is PV. Indicates that the deployer is running on the primary server. This option is used by the Topology Editor to invoke the deployer. Use this option to force a channel configuration update in the database. Instructs the deployer to install or uninstall the component specified using the id parameter, regardless of the current status of the component in the topology. Use this option to force an install or uninstall of a component in a high-availability (HA) environment, or when fixing an incomplete or damaged installation. Table 12 contains a list of possible values for the id parameter. Tells the deployer to use the specified topology file instead of prompting for the file. Causes the deployer to log additional diagnostic information. Enables you to override the hostname that the deployer uses to define where it is running. This option is useful when hostname aliasing is used and none of the hostnames listed in the topology.xml file match the hostname of the machine where the deployer is running.
-DTarget=id
-DTopologyFile=topology_file_path
-DTrace=true -DUsehostname=hostname
where id is a supported target identifier code. If you are using the -DTarget option to force the uninstall of a component, you must also specify the -Daction=uninstall option when you run the deployer application. The following example shows how to force the uninstallation of DataMart on the local system:
deployer.bin -Daction=uninstall -DTarget=DMR
160
DL TIP DBR
DMR
DVR
DCR
DLR
TIPR
DBU
DLU
161
When you run the deployer using the -DTarget option, note the following: v The deployer does not perform component registration in the versioning tables of the database. v The deployer does not upload modified topology information to the database. v The deployer does not allow you to you select other nodes besides the local node in the Node Selection panel. v In the case of an uninstall, the deployer does not remove the component from the topology.
162
Overview
How to install OpenSSH for Secure File Transfer (SFTP) among Tivoli Netcool Performance Manager components. This document explains how to install OpenSSH for Secure File Transfer (SFTP) among Tivoli Netcool Performance Manager components. You must be proficient in your operating system and have a basic understanding of public/private key encryption when working with SFTP. For the purposes of this document, an SFTP "client" is the node that initiates the SFTP connection and login attempt, while the SFTP "server" is the node that accepts the connection and permits the login attempt. This distinction is important for generating public/private keys and authorization, as the SFTP server should have the public key of the SFTP client in its authorized hosts file. This process is described in more detail later. For Tivoli Netcool Performance Manager to use SFTP for the remote execution of components and file transfer, OpenSSH must be configured for key-based authentication when connecting from the Tivoli Netcool Performance Manager account on the client (the host running the Tivoli Netcool Performance Manager process that needs to use SFTP) to the account on the server. In addition, the host keys must be established such that the host key confirmation prompt is not displayed during the connection.
Enabling SFTP
The use of SFTP is supported in Tivoli Netcool Performance Manager. Tivoli Netcool Performance Manager SFTP can be enabled for a single component, set of components, or all components as needed. This table shows the Tivoli Netcool Performance Manager components that support SFTP:
Client Server Description
Node on which DataChannel All other DataChannel nodes Installer can use SFTP to resides. to be installed. install Tivoli Netcool Performance Manager software to remote locations. Bulk Collector FTE FTE CME/LDR DataMart Inventory Bulk Collector DataLoad SNMP collector Remote CME Transfer of inventory files. FTE transfers files from BCOL to CME. Transfer of SNMP data. Downstream CME and LDR both transfer files from remote CME.
Note: This document is intended only as a guideline to installing OpenSSH. Tivoli Netcool Performance Manager calls the ssh binary file directly and uses the SFTP
Copyright IBM Corp. 2006, 2010
163
protocol to transfer files, so the essential Tivoli Netcool Performance Manager requirement is that OpenSSH is used and public key authentication is enabled. The following procedures are examples of one method of installing and configuring OpenSSH. The precise method and final configuration for your site should be decided by your local operating system security administrator. For detailed information about OpenSSH and its command syntax, visit the following URL:
http://www.openssh.com/manual.html
Installing OpenSSH
This section describes the steps necessary to install OpenSSH on AIX, Solaris and Linux. Note: The following sections refer to the earliest supported version of the required packages. Refer to the OpenSSH documentation for information about updated versions.
AIX systems
To install OpenSSH on AIX systems you must follow all steps described in this section.
Procedure
1. In your browser, enter the following URL:
http://www-03.ibm.com/servers/aix/products/aixos/linux/download.html
2. From the AIX Toolbox for Linux Applications page, download the following files according to the instructions to each Tivoli Netcool Performance Manager system where SFTP is to be used: v prngd - Pseudo Random Number Generation Daemon (prngd-0.9.291.aix5.1.ppc.rpm or later). v zlib - zlib compression and decompression library (zlib-1.2.24.aix5.1.ppc.rpm or later). 3. From the AIX Toolbox for Linux Applications page, click the AIX Toolbox Cryptographic Content link. 4. Download the following files to each Tivoli Netcool Performance Manager system where SFTP is to be used: openssl-0.9.7g-1.aix5.1.ppc.rpm or later 5. In your browser, enter the following URL:
http://sourceforge.net/projects/openssh-aix
6. From the OpenSSH on AIX page, search for and download the following files according to the instructions to each Tivoli Netcool Performance Manager system where SFTP is to be used: openssh-4.1p1_53.tar.Z or later
164
Procedure
1. Log in to the system as root. 2. Change your working directory to the location where the software packages have been downloaded by using the following command:
# cd /download/location
3. Run the RPM Packaging Manager for each package, in the specified order, using the following commands:
# rpm -i zlib # rpm -i prndg # rpm -i openssl
4. Uncompress and untar the openssh tar file by entering the following commands:
$ uncompress openssh-4.1p1_53.tar.Z $ tar xvf openssh-4.1p1_53.tar
5. Using the System Management Interface Tool (SMIT), install the openssh package. 6. Exit from SMIT.
Procedure
To configure the server to start on system boot, modify the /etc/rc.d/rc2.d/Ssshd init script as follows:
#! /usr/bin/sh # # start/stop the secure shell daemon case "$1" in start) # Start the ssh daemon if [ -x /usr/local/sbin/sshd ]; then echo "starting SSHD daemon" /usr/local/sbin/sshd & fi ;; stop) # Stop the ssh daemon kill -9 `ps -eaf | grep /usr/local/sbin/sshd | grep -v grep | awk {print $2} | xargs` ;; *) echo "usage: sshd {start|stop}" ;;
165
Solaris systems
OpenSSH is required for SFTP to work with Tivoli Netcool Performance Manager on Solaris systems. The version of SSH installed with the Solaris 10 operating system is not supported. Note: The following sections refer to the current version of the required packages. Refer to the OpenSSH documentation for information about updated versions. To install OpenSSH on Solaris systems, follow all steps described in this section.
Procedure
1. In your browser, enter the following URL: http://www.sunfreeware.com 2. From the Freeware for Solaris page, follow the instructions to download the following files to each Tivoli Netcool Performance Manager system where SFTP is to be used. Ensure that you download the correct files for your version of Solaris. v gcc - Compiler. Ensure that you download the full Solaris package and not just the source code (gcc-3.2.3-sol9-sparc-local.gz or later). v openssh - SSH client (openssh-4.5p1-sol-sparc-local.gz or later). v openssl - SSL executable files and libraries (openssl-0.9.8d-sol9-sparc-local.gz or later). v zlib - zlib compression and decompression library (zlib-1.2.3-sol9-sparclocal.gz or later).
What to do next
Note: The user should ensure they have the libcrypto.so.0.9.8 instead of libcrypto.so.1.0.0. to use OpenSSH on Solaris.
Procedure
1. Log in to the system as root. 2. Change your working directory to the location where the software packages have been downloaded using the following command:
# cd /download/location
3. Copy the downloaded software packages to /usr/local/src, or a similar location, using the following commands:
# # # # cp cp cp cp gcc-version-sparc-local.gz /usr/local/src zlib-version-sparc-local.gz /usr/local/src openssl-version-sparc-local.gz /usr/local/src openssh-version-sparc-local.gz /usr/local/src
166
5. Install the gcc compiler: a. Uncompress gcc using the following command:
gunzip gcc-version-sparc-local.gz
6. Install the zlib compression library: a. Uncompress zlib using the following command:
gunzip zlib-version-sparc-local.gz
7. Install the OpenSSL executable and binary files: a. Uncompress OpenSSL using the following command:
gunzip openssl-version-sparc-local.gz
8. Install the OpenSSH client: a. Uncompress OpenSSH using the following command:
gunzip openssh-version-sparc-local.gz
c. Create a group and user for sshd using the following commands:
groupadd sshd useradd -g sshd sshd
9. Optional: Remove Sun SSH from the path and link OpenSSH: a. Change your working directory to /usr/bin using the following command:
cd /usr/bin
b. Move the Sun SSH files and link the OpenSSH files using the following commands:
# mv ssh ssh.sun # mv ssh-add ssh-add.sun # mv ssh-agent ssh-agent.sun # mv ssh-keygen ssh-keygen.sun # mv ssh-keyscan ssh-keyscan.sun # ln -s /usr/local/bin/ssh ssh # ln -s /usr/local/bin/ssh-add ssh-add # ln -s /usr/local/bin/ssh-agent ssh-agent # ln -s /usr/local/bin/ssh-keygen ssh-keygen # ln -s /usr/local/bin/ssh-keyscan ssh-keyscan
Procedure
1. Create or modify the /etc/init.d/sshd init script as follows:
Appendix E. Secure file transfer installation
167
#! /bin/sh # # start/stop the secure shell daemon case "$1" in start) # Start the ssh daemon if [ -x /usr/sbin/sshd ]; then echo "starting SSHD daemon" /usr/sbin/sshd & fi ;; stop) # Stop the ssh daemon /usr/bin/pkill -x sshd ;; *) echo "usage: /etc/init.d/sshd {start|stop}" ;;
2. Check that /etc/rc3.d/S89sshd exists (or any sshd startup script exists) and is a soft link to /etc/init.d/sshd. If not, create it using the following command:
ln -s /etc/init.d/sshd /etc/rc3.d/S89sshd
Linux systems
OpenSSH is required for VSFTP to work with Tivoli Netcool Performance Manager.OpenSSH is installed by default on any RHEL system.
Configuring OpenSSH
This section describes how to configure the OpenSSH server and client.
Procedure
1. Log in to the system as root. 2. Change your working directory to the location where the OpenSSH Server was installed (/usr/local/etc/sshd_config by default) using the following command:
# cd /usr/local/etc
3. Using the text editor of your choice, open the sshd_config file. This is an example of a sshd_config file:
#*************************************************************************** # sshd_config # This is the sshd server system-wide configuration file. See sshd(8) # for more information. # The strategy used for options in the default sshd_config shipped with # OpenSSH is to specify options with their default value where # possible, but leave them commented. Uncommented options change a # default value.
168
Port 22 Protocol 2 ListenAddress 0.0.0.0 HostKey /usr/local/etc/ssh_host_dsa_key SyslogFacility AUTH LogLevel INFO PubkeyAuthentication yes AuthorizedKeysFile .ssh/authorized_keys RhostsAuthentication no RhostsRSAAuthentication no HostbasedAuthentication no PasswordAuthentication yes ChallengeResponseAuthentication no Subsystem sftp /usr/local/libexec/sftp-server #****************************************************************
4. Locate the Protocol parameter. For security purposes, it is recommended that this parameter is set to Protcol 2 as follows:
Protocol 2
5. Locate the HostKeys for protocol version 2 parameter and ensure that it is set as follows:
HostKey /usr/local/etc/ssh_host_dsa_key
8. Locate the Subsystem parameter and ensure that the SFTP subsystem and path are correct. Using defaults, the Subsystem parameter appears as follows:
Subsystem sftp /usr/local/libexec/sftp-server
Procedure
1. Log in as pvuser on the node that will be the SFTP client. This node is referred to as SFTPclient in these instructions, but you must replace SFTPclient with the name of your node. 2. Create an .ssh directory in the home directory of the Tivoli Netcool Performance Manager user, set permissions to x/r/w for owner (700), then change to the directory using the following commands:
$ mkdir ~/.ssh $ chmod 700 ~/.ssh $ cd ~/.ssh
169
3. Generate a DSA public and private key with no passphrase (DSA encryption is used as an example). The following example shows a UNIX server called SFTPclient:
$ /usr/local/bin/ssh-keygen -t dsa -f SFTPclient -P "" Generating public/private dsa key pair. Your identification has been saved in SFTPclient. Your public key has been saved in SFTPclient.pub. The key fingerprint is: 77:67:2f:34:d4:2c:66:db:9b:1f:9a:36:fe:c7:07:c6 pvuser@SFTPclient
4. The previous command generates two files, SFTPclient (the private key) and SFTPclient.pub (the public key). Copy the private key to id_dsa in the ~/.ssh directory by entering the following command:
$ cp -p ~/.ssh/SFTPclient ~/.ssh/id_dsa
id_dsa identifies the node when it contacts other nodes. 5. To permit Tivoli Netcool Performance Manager components on SFTPclient to communicate, you must append the contents of the SFTPclient.pub key file to the file authorized_keys in the ~/.ssh directory by using the following commands:
cd ~/.ssh cat SFTPclient.pub >> authorized_keys
6. Log on to the other node that will be the SFTP server. This node is referred to as SFTPserver in these instructions, but you must replace SFTPserver with the name of your node. 7. Repeat Step 1 through Step 5 on the SFTPserver node, replacing SFTPclient with SFTPserver. 8. Copy (with FTP, scp, or a similar utility) the public key from SFTPclient to SFTPserver and append the contents of the key file to the file authorized_keys in the ~/.ssh directory. If you cut and paste lines, be careful not to introduce carriage returns. Use the following FTP session as an example:
SFTPserver:~/.ssh> ftp SFTPclient Connected to SFTPclient. 220 SFTPclient FTP server (SunOS 5.8) ready. Name (SFTPclient:pvuser): pvuser 331 Password required for pvuser. Password: 230 User pvuser logged in. ftp> bin 200 Type set to I. ftp> get .ssh/SFTPclient.pub 200 PORT command successful. 150 Binary data connection for .ssh/SFTPclient.pub 226 Binary Transfer complete. local: .ssh/SFTPclient.pub remote: .ssh/SFTPclient.pub ftp> quit 221 Goodbye. SFTPserver:~/.ssh> cat SFTPclient.pub >> authorized_keys
9. Optional: If you want to set up bidirectional SFTP, repeat Step 8, but from the SFTserver node to the SFTPclient node. Note: This step is not needed for Tivoli Netcool Performance Manager. Use the following FTP session as an example:
SFTPclient:~/.ssh> ftp SFTPserver Connected to SFTPserver. 220 SFTPserver FTP server (SunOS 5.8) ready. Name (SFTPserver:pvuser): pvuser
170
331 Password required for pvuser. Password: 230 User pvuser logged in. ftp> bin 180 Tivoli Netcool Performance Manager Installation Guide, Version 1.3.1 200 Type set to I. ftp> get .ssh/SFTPserver.pub 200 PORT command successful. 150 Binary data connection for .ssh/SFTPserver.pub 226 Binary Transfer complete. local: .ssh/SFTPserver.pub remote: .ssh/SFTPserver.pub ftp> quit 221 Goodbye. SFTPclient:~/.ssh> cat SFTPserver.pub >> authorized_keys
10. When finished, the SFTPclient and SFTPserver should look similar to the following:
SFTPclient:~/.ssh> ls -al ~/.ssh total 10 drwx-----drwxr-xr-x -rw-------rw-r--r-2 pvuser pvuser 512 Nov 25 16:56 . 28 pvuser pvuser 1024 Nov 25 15:25 .. 1 pvuser pvuser 883 Nov 25 15:21 id_dsa 1 pvuser pvuser 836 Nov 25 16:33 known_hosts
SFTPserver:~/.ssh> ls -al ~/.ssh total 10 drwx-----drwxr-xr-x -rw-------rw-r--r-2 pvuser pvuser 512 Nov 25 16:56 . 28 pvuser pvuser 1024 Nov 25 15:25 .. 1 pvuser pvuser 883 Nov 25 15:21 id_dsa 1 pvuser pvuser 836 Nov 25 16:33 known_hosts
The important files are: v authorized_keys, which contains the public keys of the nodes that are authorized to connect to this node v id_dsa, which contains the private key of the node it is on v known_hosts, which contains the public keys of the node that you want to connect to For security, the private key (id_dsa) should be -rw------. Likewise, the public key Node<number>.pub, authorized_keys, and known_hosts should be -rw-r--r--. The directory itself should be -rwx-----. Note: The directory that contains the .ssh directory might also need to be writable by owner. 11. The first time you connect using SSH or SFTP to the other node, it will ask if the public key fingerprint is correct, and then save that fingerprint in known_hosts. Optionally, you can manually populate the client's known_hosts file with the server's public host key (by default, /usr/local/etc/ ssh_host_dsa_key.pub). For large-scale deployments, a more efficient and reliable procedure is: a. From one host, ssh to each SFTP server and accept the fingerprint. This builds a master known_hosts file with all the necessary hosts. b. Copy that master file to every other SFTP client. Note: If the known_hosts file has not been populated and secure file transfer (SFTP) is attempted through Tivoli Netcool Performance Manager, SFTP fails with vague errors.
Appendix E. Secure file transfer installation
171
Procedure
1. On both nodes, kill any existing sshd processes and start the sshd process from the packages you installed, by entering the following commands:
pkill -9 sshd /usr/local/sbin/sshd &
The path can be different depending on the installation. 2. From SFTPclient, run the following command:
/usr/local/bin/ssh SFTPserver
4. Optional: If you set up bidirectional SFTP, run the following command from SFTPserver:
/usr/local/bin/ssh SFTPclient
5. Optional: If you set up bidirectional SFTP, run the following command from SFTPserver:
/usr/local/bin/sftp SFTPclient
6. If all tests allow you to log in without specifying a password, follow the Tivoli Netcool Performance Manager instructions on how to enable SFTP in each Tivoli Netcool Performance Manager component. Make sure to specify the full path to SSH in the Tivoli Netcool Performance Manager configuration files. In addition, make sure the user that Tivoli Netcool Performance Manager is run as is the same as the user that you used to generate keys.
Troubleshooting
How to troubleshoot OpenSSH and its public keys.
Procedure
1. Check the ~/known_hosts file on the node acting as the SSH client and make sure the client host name and IP information is present and correct. 2. Check the ~/authorized_keys file on the node acting as the SSH server and make sure that the client public key is present and correct. Ensure that the permissions are -rw-r--r--.
172
3. Check the ~/id.dsa file on the node acting as the SSH client and make sure that the client's private key is present and correct. Ensure that the permissions are -rw-------. 4. Check the ~/.ssh directory on both nodes to ensure that the permissions on the directories are -rwx------. 5. Check for syntax errors (common ones are misspelling authorized_keys and known_hosts without the "s" at the end). In addition, if you copied and pasted keys into known hosts or authorized keys files, you probably have introduced carriage returns in the middle of a single, very long line. 6. Check the ~ (home directory) permissions to ensure that they are only writable by owner. 7. If the permissions are correct, kill the sshd process and restart in debug mode as follows:
pkill -9 sshd /usr/local/sbin/sshd -d
8. Test SSH again in verbose mode on the other node by entering the following command:
/usr/local/bin/ssh -v SFTPserver
9. Read the debugging information about both client and server and troubleshoot from there. 10. Check the log file /var/adm/messages for additional troubleshooting information.
173
174
LDAP configuration
The configuration of LDAP as a default authentication and authorization mechanism for Tivoli Netcool Performance Manager is achieved using the Topology Editor.
Procedure
1. In Logical View of Topology Editor, select the Tivoli Integrated Portal that you have added. 2. Open the Advanced Properties tab. 3. Click the checkbox opposite the IAGLOBAL_USER_REIGSTRY_LDAP_SELECTED property. This enables LDAP. 4. In Advanced Properties tab, enter the LDAP connection details. This requires that you populate the following fields: v WAS_USER_NAME: This is the name you have registered as the Tivoli Integrated Portal user. For example, "tipadmin". v IAGLOBAL_LDAP_BIND_DN: This username specified must have read and write permissions in LDAP 3. Typically this will be an LDAP administrator username. For example "cn=Directory Manager". v IALOCAL_LDAP_BIND_PASSWORD: This is the password for the Bind Distinguished Name specified.
175
v IAGLOBAL_LDAP_NAME: This is the LDAP server host name. Should this LDAP server exist behind firewall, make sure that this host has been authenticated. v IAGLOBAL_LDAP_PORT: For example, "1389". v IAGLOBAL_LDAP_REPOSITORY_ID: This is a string used to identify the LDAP repository, which can be set to the string of your choice. v IAGLOBAL_LDAP_BASE_ENTRY: The distinguished name of a base entry in LDAP. For example, for IBM the base entry is o=IBM, c=US.
Procedure
Verify from the UI that the users tnpm and tnpmScheduler are members of the tnpmAdministrators group.
Procedure
To successfully authenticate you LDAP user you need to assign them to one of the following roles: v tnpmUser v tnpmAdministrator This can be done by tipadmin user, by navigating to Users and Groups > Administrative User Roles, and assigning the correct roles. Alternatively tipcli.sh can be used for assigning roles to the user.
<tip_location>/profiles/TIPProfile/bin/tipcli.sh MapRolesToUser --username <tip_admin_user> --password <tip_admin_password> --userID <userUniqueId> --rolesList <roleName>
where <userUniqueId> is the concatenation of username and realm in which user information is stored. For example:
176
<tip_location>/profiles/TIPProfile/bin/tipcli.sh MapRolesToUser --username <tip_admin_user> --password <tip_admin_password> --userID uid=<user_name>,dc=<server>,dc=<server>,dc=<company>,dc=com --rolesList tnpmUser
Roles specific to an application, such as tnpmUser for Tivoli Netcool Performance Manager, are not stored in LDAP. Roles are stored in a flat XML file in the TIP directory. For example, if you assign a Tivoli Netcool Performance Manager role to an LDAP user on tip_instance1, you must also assign the same role to the user on tip_instance2. Otherwise, the user cannot authenticate on tip_instance2. Alternatively you can assign tnpmUser role to tnpmUsers group on tip_instance1. If the user is a member of this group then user can authenticate on tip_instance2.
177
178
The Deployer
How to run the deployer in silent mode.
Procedure
1. Log in as root. 2. Log in to the machine on which you want to run the silent installation. 3. In a text editor, open the Fresh.properties file and make the following edits: a. Set and verify that the Oracle client path is correct. b. Set the DownloadTopology flag to True (1) or False (0). c. If you set DownloadTopology flag to False, set the TopologyFilePath to the location of your topology.xml file. d. If you are running the deployer application on the same system where the Topology Editor is installed, set the Primary flag to true. e. Set and verify that the Database Access Information is correct. f. Set and verify the PACKAGE_PATH variable for the relevant system: On Solaris systems:
<DIST_DIR>/proviso/SOLARIS
On AIX systems:
<DIST_DIR>/proviso/AIX
On Linux systems:
<DIST_DIR>/proviso/RHEL
Copyright IBM Corp. 2006, 2010
179
<DIST_DIR> is the directory on the hard drive where you copied the contents of the Tivoli Netcool Performance Manager distribution in Downloading the Tivoli Netcool Performance Manager distribution to disk on page 26. Your edited file will look similar to the following:
#Oracle client JDBC driver path #-----------------------------OracleClient=/opt/oracle/product/10.2.0.4.0/jdbc/lib #Download Topology from Proviso database # 1 is true # 0 is false #------------DownloadTopology=0 #Primary # Specify if the configuration has to be updated # Specify true if running the deployer on the same # system where the Topology Editor is installed. # true or false #------------Primary=false #Topology file # If DownloadTopology=1 this parameter is ignored #------------TopologyFilePath=/tmp/ProvisoConsumer/Topology.xml #Database access information #--------------------------OracleServerHost=lab238053 OracleServerPort=1521 OracleSID=PV OracleAdminUser=PV_INSTALL OracleAdminPassword=PV #Check Prerequisites Flag(true/false) #Use true only for first time install #------------------------------------CHECK_PREREQ=true # Tivoli Netcool Performance Manager installation packages path #--------------------------PACKAGE_PATH=/cdrom/SOLARIS #Silver Stream installation packages path #------------------------------------SS_BUNDLE=/cdrom/exteNd40k
g. Write and quit the file. 4. Change to the /opt/IBM/proviso/deployer directory. 5. Run the following command:
./deployer.bin -i silent -f propertyFileWithPath
For example:
./deployer.bin -i silent -f /opt/IBM/proviso/deployer/proviso/data/Silent/Fresh.properties
180
if the install has not worked and one of the steps fails you will see one of the following errors: v Product images not found during silent installation v An installation step has failed during silent installation v Silent Installatoin Suspended because a reboot is needed v Engine main loop internal error
Restrictions
Deployer restrictions. Note the following restrictions: v The silent deployer does not support remote installations. You must manually invoke the script on each machine. v Silent resume is not supported. If you need to resume a partial silent installation, use the -Daction=resume option to complete the installation using graphical mode (the steps table). The step that originally failed might have been in the middle of a step sequence that cannot be re-created by a subsequent -i silent invocation.
On AIX systems:
# cd <DIST_DIR>/proviso/AIX/Install/topologyEditor/Disk1/InstData/VM
On Linux systems:
# cd <DIST_DIR>/proviso/RHEL/Install/topologyEditor/Disk1/InstData/VM
<DIST_DIR> is the directory on the hard drive where you copied the contents of the Tivoli Netcool Performance Manager distribution in Downloading the Tivoli Netcool Performance Manager distribution to disk on page 26. To install the Topology Editor in silent mode:
Procedure
1. Log in as root to the server on which you want to run the silent installation. 2. Change to the directory that contains the deployer.bin file (for example, /opt/IBM/proviso/deployer), then change to the /proviso/data/Silent subdirectory. 3. Using a text editor, open the topologyEditor.properties file and make the following edits: a. Set and verify that the Oracle client path is correct. b. Set the DownloadTopology flag to True (1) or False (0).
Appendix G. Using silent mode
181
c. If you set DownloadTopology flag to False, set the TopologyFilePath to the location of your topology.xml file. d. Set and verify that the Database Access Information is correct. e. Set and verify the PACKAGE_PATH variable. f. Write and quit the file. 4. Run the following command:
./installer.bin -i silent -f ..../silent/topologyEditor.properties
182
Overview
Interim fix installation overview. Unlike major, minor, and maintenance releases, which are planned, patch releases (interim fixes and fix packs) are unscheduled and are delivered under the following circumstances: v A customer is experiencing a "blocking" problem and cannot wait for a scheduled release for the fix. v The customer's support contract specifies a timeframe for delivering a fix for a blocking problem and that timeframe does not correspond with a scheduled release. v Development determines that a patch is necessary. Note: Patches are designed to be incorporated into the next scheduled release, assuming there is adequate time to integrate the code.
Installation rules
Rules that apply to the installation of patches. Note the following installation rules for patch installations: v Apply fix to Database before any other components. v Fixes for the Database and DataMart must be installed on that host. v Fixes for the DataChannel, DataLoad, DataMart and DataView can be installed remotely from the local host in a distributed system. v Fix packs are installed on general availability (GA) products. v Sequentially numbered fix packs can be installed on any fix pack with a lower number. v Interim fixes must be installed on the absolute fix pack. The patch installer verifies that your installation conforms to these rules.
183
Installing a patch
How to install a patch.
Procedure
1. You must have received or downloaded the maintenance package from IBM Support. The maintenance package contains the Maintenance Descriptor File, an XML file that describes the contents of the fix pack. Follow the instructions in the README for the fix pack release to obtain the maintenance package and unzip the files. Note: for each tar.gz file, you must unzip them, and then un-tar them. For example:
gunzip filename.tar.gz tar -xvf filename.tar
2. Log in as root. 3. Set and export your DISPLAY environment variable (see Setting up a remote X Window display on page 14). 4. Start the patch deployer using one of the following methods: From the launchpad: a. Click the Start Tivoli Tivoli Netcool Performance Manager Maintenance Deployer option in the list of tasks. b. Click the Start Tivoli Tivoli Netcool Performance Manager Maintenance Deployer link. From the command line: v Run the following command:
# ./deployer.bin -Daction=patch
5. The deployer displays a welcome page. Click Next to continue. 6. Accept the default location of the base installation directory of the Oracle JDBC driver (/opt/oracle/product/version/jdbc/lib), or click Choose to navigate to another directory. Click Next to continue. 7. On the patch folder page, click Choose to select the patch you want to install. 8. Navigate to the directory that contains the files for the fix pack, and click into the appropriate directory. Click Select to select that directory, then click Next to continue.
184
9. A pop-up window asks whether you want to download the topology file. Click Yes. 10. Verify that all of the fields for the database connection are filled in with the correct values: v Database hostname - Enter the name of the database host. v Port - Specifies the port number used for communication with the database. The default value is 1521. v Database user - Specifies the username used to access the database. The default value is PV_INSTALL. v Database Password - Enter the password for the database user account (for example, PV). v SID - Specifies the SID for the database. The default value is PV. Click Next. 11. When the topology has been downloaded from the database, click Next. 12. The node selection window shows the target systems and how the files will be transferred. The table has one row for each machine where at least one Tivoli Netcool Performance Manager component will be installed. Verify the settings, then click Next to continue. 13. The deployer displays summary information about the installation. Review the information, then click Next. The deployer displays the table of installation steps. 14. Run through each installation step just as you would for a normal installation. 15. When all the steps have completed successfully, click Done to close the wizard.
185
186
Error codes
The following sections describe the error messages generated by the Deployer, the Toplogy Editor and InstallAnywhere.
Deployer messages
The Deployer messages. Table 13 lists the error messages returned by the Tivoli Netcool Performance Manager deployer. Table 13: Deployer Messages
Error Code DataView Messages GYMCI5000E A system command failed. A See the installation log for standard UNIX system more details. command failed. These commands are used for standard system operations, such as creating directories, changing file permissions, and removing files. The operating system is not at the prerequisite patch level. Some required operating system patches are not installed. See the installation log for details. Install the required patches, then try the installation again. Description User Action
GYMCI5002E
GYMCI5003E
The Oracle configuration file, Add an entry for tnsnames.ora, does not SilverMaster to the include an entry for tnsnames.ora file, then try the installation again. SilverMaster. The Oracle configuration file, Ensure that the file exists in tnsnames.ora, was not found. the correct location. The tnsnames.ora file must be created and stored in the $TNS_ADMIN directory. Unable to connect to the Oracle database. It is possible that a specified connection parameter is incorrect, or the Oracle server might not be available. See the installation log for more details. Ensure that the connection parameters you are using are correct and that the Oracle server is up and running.
GYMCI5004E
GYMCI5005E
187
Description An error occurred while running the DVOptimizerToRule.sql script to initialize the database. It is possible that the Oracle database and listener are not running. An error occurred while trying to remove entries for a resource from a database table. It is possible that the Oracle database and listener are not running. An error occurred while trying to remove version information from a database table. It is possible that the Oracle database and listener are not running.
User Action See the installation log for more details. Check that the database and listener are running.
GYMCI5007E
See the installation log for more details. Check that the database and listener are running.
GYMCI5008E
See the installation log for more details. Check that the database and listener are running.
GYMCI5009E
An error occurred while Contact IBM Software Support. reading the configuration file. The name of a parameter or the format of the file is incorrect. The file system does not have sufficient free space to complete the installation. See the installation log for more details. Ensure that you have sufficient space on the file system before retrying the installation. Contact IBM Software Support.
GYMCI5010E
GYMCI5011E
The DataView license file is missing. The license file was not found, but this file should not be required. The installation log will contain more details of the error. A configuration file or directory is missing.
GYMCI5012E GYMCI5013E
An error occurred while See the installation log for creating a configuration file. more details. The file could not be created. The installer failed to create one of the required configuration files. See the installation log for An error occurred while updating a configuration file. more details. The file could not be modified. The installer failed to make a required modification to one of the configuration files.
GYMCI5014E
DataMart Messages GYMCI5101E The DataMart installation failed. See the DataMart installer logs for details.
188
Description
User Action
The database installation See the root_install_dir/ failed. See the installation log database/install/log/ for details. Oracle_SID/install.log file. Check the syntax and run The database uninstallation script failed because of a the script again. syntax error. This script must be run as oracle. For example: ./uninstall_db /var/tmp/PvInstall/ install.cfg.silent Check that all the required The database could not be Oracle variables are set and removed because some Oracle environment variables try again. are not correctly set. Some or all of the Oracle environment variables are not set (for example, ORACLE_HOME, ORACLE_SID, or ORACLE_BASE). An error occurred when trying to start the Oracle database. See the Oracle alert file for possible startup errors. Resolve any problems reported in the log and try again. See the Oracle alert file for possible shutdown errors. Resolve any problems reported in the log and try again. See the Oracle alert file for details of errors. You might need to manually delete Oracle data files using operating system commands.
GYMCI5202E
GYMCI5204E
GYMCI5205E
GYMCI5206E
GYMCI5207E
An error occurred while querying the database to determine the data files that are owned by the database.
DataChannel Messages GYMCI5301E The database channel installation failed. See the installation log for details. An error occurred while running a script. Unable to find an expected file. See the file root_install_dir/channel/ install/log/Oracle_SID/ install.log. See the message produced with the error code for more details. See the message produced with the error code for more details.
GYMCI5401E
GYMCI5402E
GYMCI5403E
The data in one of the files is See the message produced not valid. with the error code for more details. Unable to find an expected file or expected data. See the message produced with the error code for more details.
GYMCI5404E
189
Description Scripts cannot function correctly because the LD_ASSUME_KERNEL variable is set. An action parameter is missing. An error occurred while processing the tar command. The product version you are trying to install seems to be for a different operating system. Unable to locate installed package information for the operating system. A file has an unexpected owner, group, or permissions.
GYMCI5406E
See the message produced with the error code for more details. See the message produced with the error code for more details. See the message produced with the error code for more details. See the message produced with the error code for more details. See the message produced with the error code for more details.
GYMCI5407E
GYMCI5408E
GYMCI5409E
GYMCI5410E
GYMCI5411E
A problem was found by the See the message produced PvCheck module when with the error code for more checking the environment. details. The installation module failed. The patch installation failed. The remove action failed. An unrecoverable error occurred while running the script. See the messages in standard error for more details. See the messages in standard error for more details. See the messages in standard error for more details. See the message produced with the error code for more details.
Dataload Messages GYMCI5501E An error occurred when running the script. Unable to find an expected file. See the message produced with the error code for more details. See the message produced with the error code for more details.
GYMCI5502E
GYMCI5503E
The data in one of the files is See the message produced not valid. with the error code for more details. Unable to find an expected file or expected data. Scripts cannot function correctly because the LD_ASSUME_KERNEL variable is set. See the message produced with the error code for more details. Unset the variable and try again.
GYMCI5504E
GYMCI5505E
190
Description An action parameter is missing. An error occurred while processing the tar command. The product version you are trying to install seems to be for a different operating system. Unable to locate installed package information for the operating system.
User Action See the message produced with the error code for more details. See the message produced with the error code for more details. See the message produced with the error code for more details. See the message produced with the error code for more details.
GYMCI5507E
GYMCI5508E
GYMCI5509E
GYMCI5510E
A file has an unexpected See the message produced owner, group or permissions. with the error code for more details. A problem was found by the See the message produced PvCheck module when with the error code for more checking the environment. details. The installation module failed. The patch installation failed. The remove action failed. An unrecoverable error occurred while running the script. See the message produced with the error code for more details. See the messages in standard error for more details. See the messages in standard error for more details. See the message produced with the error code for more details.
GYMCI5511E
GYMCI5512E
Prerequisite Checkers: Operating System GYMCI6001E The syntax of the check_os script is not correct. The specified component does not exist.The syntax is: check_os PROVISO_COMPONENT where PROVISO_COMPONENT is DL, DC, DM, DB, or DV. Correct the syntax and try again.
GYMCI6002E
This version of IBM Tivoli See the check_os.ini file for a list of supported operating Tivoli Netcool Performance Manager is not supported on systems. the host operating system. The specified component does not exist or is not supported on this operating system. Ensure that you have specified the correct component. If you have, the operating system must be upgraded before the component can be installed.
GYMCI6003E
191
Description The operating system is not at the prerequisite patch level. Some required operating system patches are not installed. The host operating system is not supported for this installation. In the /etc/security/limits file, some values are missing or incorrect. Values must not be lower than specified in the check_os.ini file.
User Action Check the product documentation for a list of required patches. Apply any missing patches and try again. Perform the installation on a supported operating system. Check the values in the check_os.ini and edit the default stanza in the /etc/security/limits file so that valid values are specified for all required limits.
GYMCI6005E
GYMCI6006E
Prerequisite Checkers: Database GYMCI6101E The syntax of the check_db script is not correct. The syntax is: check_db [client server] [new - upgrade] [ORACLE_SID or tnsnames.ora entry] The host operating system is not supported for this installation. Correct the syntax and try again.
GYMCI6102E
GYMCI6103E
See the check_os.ini file for a This version of the IBM list of supported operating Tivoli Tivoli Netcool system versions. Performance Manager database is not supported on the current version of the host operating system. Some required Oracle variables are missing or undefined. An Oracle binary is missing or not valid. Check the Oracle users environment files (for example, .profile and .bash_profile). Ensure that Oracle is correctly installed.
GYMCI6104E
GYMCI6105E GYMCI6106E
The instance of Oracle Check the list of supported installed on the host is not at Oracle versions in the check_db.ini file. a supported version. Unable to contact the Oracle server using the tnsping utility with the specified ORACLE_SID. An Oracle instance is running on the host where you have requested a new server installation. Check that your Oracle listener is running on the database server. Start the listener if it is not running. Check whether you have selected the correct host for a new Oracle server installation. If the selected host is correct, remove the existing Oracle instance first.
GYMCI6107E
GYMCI6108E
192
Description
User Action
The number of bits (32 or 64) Check the list of supported for the Oracle binary does Oracle versions in the not match the values defined check_db.ini file. in the check_db.ini file. The installation method passed to the script is not valid. Valid installation methods are New and Upgrade. The installation type passed to the script is not valid. Valid installation methods are Client and Server. The script was run with options set for a new server installation, but an Oracle instance configuration file (init.ora) already exists for the specified SID. The presence of the init.ora file indicates the presence of an Oracle instance. Pass the New or Upgrade option to the script.
GYMCI6110E
GYMCI6111E
GYMCI6112E
Check that a new server installation is the correct action for this SID. If it is, remove the existing Oracle instance configuration files.
GYMCI6113E
Remove any symbolic links. A symbolic link was found in the Oracle home path. The Specify the Oracle home path using only real directories. Oracle home path cannot contain any symbolic links. Cannot contact the Oracle Listener. The tnsping utility was run to check the Oracle Listener status, but, the Listener could not be contacted. The Solaris semaphore and shared memory check failed. The sysdef command was used to check the values for semaphores and shared memory. The command did not report the minimum value for a particular semaphore or shared memory. Could not find the bos.adt.lib package in the COMMITTED state. The package might not be installed. The package is either not installed or not in a COMMITTED state. Check that the Oracle Listener is running. Start it if necessary.
GYMCI6114W
GYMCI6115E
Check that the required /etc/system parameters are set up for Oracle. Check that the values of these parameters meet the minimum values listed in the check_db.ini file.
GYMCI6116E
Ensure that the bos.adt.lib package is installed and committed and then try again.
193
Description Could not log in to the database. The verify base option was used. The option attempts to log into the database to ensure it is running. However, the script could not log in to the database. The checkextc script failed. The verify base option was used. The option runs the checkextc script to ensure external procedure calls can be performed. The tnsnames.ora file is missing. A tnsnames.ora file in should exist in ORACLE_HOME/network/ admin directory.
User Action Check that the database and Oracle Listener are up and running. If not, start them.
GYMCI6118E
Check that the Tivoli Netcool Performance Manager database was created properly.
GYMCI6119E
Check that the tnsnames.ora file exists in the ORACLE_HOME/network/ admin directory. If it does not, create it.
Minimal Deployment: Post-Installation Messages GYMCI7500E An internal processing error occurred in the script. Check the logs and the output from the script. Look for incorrect configuration or improper invocation.
GYMCI7501E
Check for errors that The required configuration occurred during the or messages files for the poc-post-install script are not installation steps. in the same directory as the script. These files should be unpacked by the installer together with the script. An environment file is missing or is in the wrong location. Check the poc-post-install configuration file. The missing environment file and expected path will be identified in the log file.
GYMCI7502E
GYMCI7503E
The SNMP DataLoad did not Check the SNMP DataLoad log for errors during startup. start. The SNMP DataLoad process (pvmd) failed to start. The network inventory failed. New devices cannot be discovered unless the inventory runs successfully. Check the inventory log for errors. Ensure the DISC server and SNMP DataLoad (Collector) processes are running.
GYMCI7504E
GYMCI7505E
The Report Grouping Check the inventory log file operation failed. This action for more details of the does not depend on any Report Grouping failure. external application processes. The database must be running, and correct DataMart grouping rule definitions are required.
194
Description The DataChannel command line failed. It is possible that the CNS, CMGR, and AMGR processes are not running. The Report User was not created. The Web user will not be able to view reports. The DataMart resmgr utility is used to add this configuration to the database. It is possible that the database is not running. Failed to associate a Report User to a group. The report user is associated with a group to allow the user to view reports. The DataMart resmgr utility is used to add this configuration to the database. It is possible that the database is not running. A report user could not be deleted from the database. Failed to create a Web User. The user will not be able to authenticate with the Web/application server. The Web group could not be created, and the Web user might not be properly configured to view reports. Failed to associate the Web User with a group. The Web user might not be properly configured to view reports unless successfully associated with a group. Failed to delete Web Users. Web user authentication was not removed.
User Action Ensure that the required processes are running. Check the proviso.log for details of the failure. Ensure that the database is running, and check for error logs in the DataMart logs directory.
GYMCI7507E
GYMCI7508E
Ensure that the database is running, and check for error logs in the DataMart logs directory. Ensure that the specified report group exists.
GYMCI7509E
Check for error and trace logs in the DataMart logs directory. Check the Web/application server log file for errors. Ensure that the Web/application server is running. Check the Web/application server log file for errors. Ensure that the Web/application server is running. Check the Web/application server log file for errors. Ensure that the Web/application server is running. This step relies on the database component only. Check the Web/application server logs.
GYMCI7510E
GYMCI7511E
GYMCI7512E
GYMCI7513E
GYMCI7514E
The Channel Naming Service Check for walkback or error failed to start. files in the DataChannel log or state directory. Cross-application communication cannot function. The central LOG server failed to start. Logging for DataChannel will be unavailable. Check for walkback or error files in the DataChannel log or state directory.
GYMCI7515E
195
Description The Channel Manager failed to start. DataChannel applications cannot be started or stopped. Application status will be unavailable. The Application Manager failed to start. DataChannel applications cannot be started or stopped. Application status will be unavailable. Failed to create the DV user group. The DV user will remain in the Orphans group. Failed to associate the DV user to the DV group. The DV user will remain in the Orphans group. The Web Application server is not running or took too long to start up. The MIB-II Technology Pack jar file was not found in the specified directory.
User Action Check the proviso.log file for errors. Check for walkback or error files in the DataChannel log or state directory. Check the proviso.log file for errors. Check for walkback or error files in the DataChannel log or state directory. Check the poc-post-install log in /var/tmp for more details on the error condition. Check the poc-post-install log in /var/tmp for more details on the error condition. Start up the Web Application server as documented. Add the MIB2 Technology Pack jar to the directory. Remove other jar files and try again.
GYMCI7517E
GYMCI7518E
GYMCI7519E
GYMCI7520E
GYMCI7597E
GYMCI7598E
Too many jar files are present Remove the other jar files and try again. in the specified directory. Only two jar file can be present in the directory: the ProvisoPackInstaller.jar and the MIB-II Technology Pack jar. The Technology Pack installer failed. Check the Technology Pack installer logs for details.
GYMCI7599E
Installer Action Messages and IA Flow Messages GYMCI9998E Unable to find a message for See the installation log for more details. the key. The message was not retrieved from the message catalog. An unknown error occurred See the installation log for more details. for the component name with the error code code. The message could not be retrieved from the catalog. An error occurred during See the installation log for installation. An exception has more details. been generated during an installation step.
GYMCI9999E
GYMCI9001E
196
Description An unrecoverable error occurred when running the command command. An unrecoverable error occurred while running a command. An error occurred while connecting to the database. An error occurred while performing a database operation. Remote File Transfer has been disabled.
User Action See the installation log for more details. See the installation log for more details. See the installation log for more details. See the installation log for more details. To continue, change the step property to Allow Remote Execution and run the step again, or manually transfer the directory to the host. When the transfer is completed, change the step status to Success and continue the installation. See the installation log for more details.
GYMCI9003E
GYMCI9004E GYMCI9005E
GYMCI9006E
GYMCI9007E
An error occurred while remotely connecting to target. There are connection problems with the host.
GYMCI9008E
See the installation log for An error occurred while more details. connecting to target. There are connection problems with the host. An error occurred while copying install_dir. See the installation log for more details.
GYMCI9009E GYMCI9010E
Remote Command Execution To continue: 1. Change the has been disabled. step property to Set Allow Remote Execution. 2. Run the step again. Or, manually transfer the directory to the host. When the transfer is completed, change the step status to Success and continue the installation. An error occurred during file See the installation log for creation. more details. An error occurred while loading the discovered topology file. An error occurred while loading the topology file. The installation engine encountered an unrecoverable error. An error occurred while saving the topology file. See the installation log for more details. See the installation log for more details. See the installation log for more details. See the installation log for more details.
GYMCI9011E GYMCI9012E
GYMCI9013E GYMCI9014E
GYMCI9015E
197
Description The installer cannot proceed with the installation because there is insufficient disk space on the local host. The installer cannot download the topology from the specified database. Verify that the Tivoli Netcool Performance Manager database exists and that it has been started. If it does not exist, launch the installer, providing a topology file. The installer cannot connect to the specified database indicated because of incorrect credentials. The installer could not establish a connection to the specified database. Check that the Tivoli Netcool Performance Manager database can be contacted. Click Next to proceed without checking the current environment status. The database connection parameters do not match those in the topology file. An error occurred while loading the Oracle client jar. The configuration file name was not found. The step cannot run.
GYMCI9017E
Ensure that the correct host name, port, and SID were specified and that the database has been started.
GYMCI9018E
Ensure that you provide the correct user name and password. Check that the Tivoli Netcool Performance Manager database can be contacted.
GYMCI9019W
GYMCI9020E
Ensure that you provide the correct parameters. See the installation log for more details. See the installation log for more details.
GYMCI9021E GYMCI9022E
GYMCI9023W
See the installation log for There appear to be no more details. differences between the desired topology state and the current state of the Tivoli Netcool Performance Manager installation. The installer shows this message when it determines there is not work that it can do. Normally, this occurs when the Tivoli Netcool Performance Manager system is already at the desired state. However, it can also occur when there are component dependencies that are not satisfied. The operating system specified for this node in the topology file is not correct. Correct the topology file.
GYMCI9024E
198
Description The path is not valid or you do not have permissions to write to it. The path is not a valid Oracle path. The sqlplus command could not be found. The specified port is not valid. At least one parameter is null. The specified host name contains unsupported characters. The specified host cannot be contacted. The path not exists on the local system.
User Action Correct the parameter and try again. Correct the parameter and try again.
GYMCI9026E
Correct the parameter and try again. Specify values for the required parameters. Ensure that host names include only supported characters. Ensure that the host name is correct and check that the host is available. Correct the path and try again.
GYMCI9030E
GYMCI9031E GYMCI9032E
See the log file for further An error occurred while details. saving the topology. It has not been uploaded to the Tivoli Netcool Performance Manager database. This error occurs when there is a database connection error or when the Tivoli Netcool Performance Manager database has not yet been created One of the following parameters must be set to 1: param1 param2 An error occurred while creating mount point directories. An error occurred while changing the ownership or the group of mount point directories. The machine hostname was not found in the Tivoli Netcool Performance Manager model (topology.xml file). The machine where the installer is running is not part of the Tivoli Netcool Performance Manager topology. Check the log file for further details. Redefine the parameters and try again. See the log file for further details. See the log file for further details.
GYMCI9033E
GYMCI9034E
GYMCI9035E
GYMCI9036E
If a host name alias is used, make the machine host name match the host name in the model. Alternatively, use the option -DUsehostname=hostname to override the machine host name used by the installer.
199
Description The Deployer version you are using is not compatible with the component that you are trying to install. The XML file cannot be read or cannot be parsed. The deployment cannot proceed, because an error occurred the deployment plan was being generated.
User Action Use a Deployer at a version that supports the deployment of the component you are trying to install. Ensure the file is not corrupted. See the log file for more details. See the log file for more details. Check that there is sufficient disk space and that the Deployer images are not corrupted.
GYMCI9038E
GYMCI9039E
GYMCI9040E
The Deployer cannot manage See the log file for more the indicated component on details about the condition the specified node. that was detected. The user ID you specified is not defined on the target system. You specified a host that is running on an unsupported platform. The value you specified is not supported. Check that you have specified the correct user ID. Check that you have specified the correct host name. Specify one of the supported values.
GYMCI9041E
GYMCI9042E
GYMCI9043E
GYMCI0000E
Ensure that you have the correct location for the technology pack metadata files and try the operation again. Contact IBM Software Support.
GYMCI0001E
200
Description No item has been found that satisfies the filtering criteria.
User Action Ensure that you enter the correct filtering criteria and try the operation again.
GYMCI0003E
An error occurred when Ensure that you have selected the correct file and reading XML file name. The XML file might be corrupt or try the operation again. in an incorrect format. The input value must be an integer. An unexpected element was found when reading the XML file. A value must be specified. The value must represent a log filter matching regular expression expression. Correct the input value and try the operation again. Ensure that you have selected the correct file and try the operation again. Correct the input value and retry the operation. Correct the input value and try the operation again.
GYMCI0004E GYMCI0005E
GYMCI0006E GYMCI0007E
GYMCI0008E
Metadata file name was not Ensure that you have the found. The specified file does correct file name and path not exist. and retry the operation. Metadata file name is corrupted. Metadata file name was already imported. Do you want to replace it? Object name was not found in the repository. The specified object does not exist. Contact IBM Software Support. Click Yes to replace the file or No to cancel the operation. Ensure that you have the correct object name and try the operation again.
GYMCI0009E GYMCI0010E
GYMCI0011E
GYMCI0012E
Ensure that you have the The specified value must identify an existing directory. correct directory name and The specified directory does try the operation again. not exist. Removing object from host in No user action required. Physical View. File name does not exist. Ensure that you have the correct file name and try the operation again. Ensure that there is sufficient space to write the file in the file system where the Topology Editor is running. Correct the login credentials and try the operation again.
GYMCI0013E GYMCI0014E
GYMCI0015E
An unexpected error occurred writing file name. See the trace file for details. The user or password that you specified is wrong.
GYMCI0016E GYMCI0017E
The value specified for at Correct the input value or least one of the following values and try the operation fields is not valid: host name, again. port, or SID. The file name is corrupted. Select a valid XML file.
GYMCI0018E
201
Description An unexpected error occurred when retrieving data from the database. See the trace file for details. An unexpected error occurred when parsing file name. See the trace file for details. An unexpected error occurred. See the trace file for details. The input value must be a boolean. The specified value must be one of the following operating systems: AIX, SOLARIS, or Linux.
User Action Ensure that the database is up and running and that you can connect to it. Select a valid XML file.
GYMCI0020E
GYMCI0021E
Contact IBM Software Support. Correct the input value and try the operation again. Correct the input value and try the operation again.
GYMCI0022E GYMCI0023E
GYMCI0024E
The value must be a software Correct the input value and version number in the format try the operation again. n.n.n or n.n.n.n. For example 7.1.2, or 7.1.2.1. The value must be an integer Correct the input value and in the range minValue to try the operation again. maxValue, inclusive. The value must be a comma-separated list of strings. The value must be a file size expressed in kilobytes. For example, 1024K. The value must be a file size expressed in megabytes. For example, 512M. The value must be a file size expressed in kilobytes or megabytes. For example 1024K or 512M. Correct the input value and try the operation again. Correct the input value and try the operation again. Correct the input value and try the operation again. Correct the input value and try the operation again.
GYMCI0025E
GYMCI0026E
GYMCI0027E
GYMCI0028E
GYMCI0029E
GYMCI0030E
The value must be an FTP or Correct the input value and try the operation again. SFTP connection string. For example, ftp:// username:password@hostname/directory. The value must be a comma-separated list of directories. For example, /opt, /var/tmp, /home. Value cannot be a fully-qualified domain name, IP address, or name containing hyphen or period. Correct the input value and try the operation again.
GYMCI0031E
GYMCI0032E
Supply the unqualified host name without the domain. Do not use the IP address or a name that contains hyphens.
202
Description Metadata file name contains an technology pack with a wrong structure. Value should be in the format YYYY-MM-DD, cannot be a date prior than 1970-01-01, or later than the current date. The meta-data file contains an technology pack with the wrong structure. Value should be in the format YYYY-MM-DD, cannot be a date prior than 1970-01-01, or later than the current date. The operation failed because the specified file does not exist.
User Action Contact IBM Software Support. Specify a date that is within the range and in the correct format.
GYMCI0034E
GYMCI0035E
Obtain a valid meta-data file and try again. Correct the input value and retry the operation.
GYMCI0036E
GYMCI0037E
Ensure that the file name and path you specified is correct and retry the operation.
GYMCI0038E
The operation failed because See the trace file for more of an error while validating details. the host name mappings file. Correct the entry for the The host name retrieved by specified host name in the the upgrade process is not topology definition. valid. Fully qualified host names, IP addresses and names containing hyphens or periods are not supported. The upgrade process retrieved two entries for the specified host name. The fully qualified host name is not supported. The upgrade process did not retrieve a valid value for the specified property. A default value has been used. No component is present on the specified host. Remove the entry for the fully qualified host name.
GYMCI0039E
GYMCI0040E
GYMCI0040W
Check that the default assigned is appropriate and change it if necessary. Specify a host where at least one component is present.
GYMCI0041E GYMCI0042E
The operation failed because Correct the input value and the input value is not the retry the operation. correct data type. The correct data type is Long. The operation failed because the input value is not valid. The upgrade process did not retrieve a valid value for the specified property. A default value has been used. Correct the input value and retry the operation. Check that the default assigned is appropriate and change it if necessary.
GYMCI0043E GYMCI0044W
203
InstallAnywhere messages
The InstallAnywhere messages. Table 14 lists the InstallAnywhere error messages. These messages could be returned by either the deployer or the Topology Editor. See the InstallAnywhere documentation for more information about these error codes and how to resolve them. Table 15: Install Anywhere Messages
Error Code 0 1 Description Success: The installation completed successfully without any warnings or errors. The installation completed successfully, but one or more of the actions from the installation sequence caused a warning or a non-fatal error. The silent installation failed because of step Error errors. One or more of the actions from the installation sequence caused a fatal error. The installation was cancelled by the user. The installation includes an invalid command-line option. Unhandled error. The installation failed the authorization check, may indicate an expired version. The installation failed a rules check. A rule placed on the installer itself failed. An unresolved dependency in silent mode caused the installer to exit. The installation failed because not enough disk space was detected during the execution of the Install action. The installation failed while trying to install on a Windows 64-bit system, but installation did not include support for Windows 64-bit systems. The installation failed because it was launched in a UI mode that is not supported by this installer. Unhandled error specific to a launcher. The installation failed due to an error specific to the lax.main.class property. The installation failed due to an error specific to the lax.main.method property. The installation was unable to access the method specified in the lax.main.method property.
2005
2006
204
Description The installation failed due to an exception error caused by the lax.main.method property. The installation failed because no value was assigned to the lax.application.name property. The installation was unable to access the value assigned to the lax.nl.java.launcher.main.class property. The installation failed due to an error specific to the lax.nl.java.launcher.main.class property. The installation failed due to an error specific to the lax.nl.java.launcher.main.method property. The installation was unable to access the method specified in the lax.nl.launcher.java.main.method property. A Java executable could not be found at the directory specified by the java.home system property. An incorrect path to the installer jar caused the relauncher to launch incorrectly.
3005
3006
3007
3008
3009
4000
4001
Log files
A description of the files that are used to log errors for the Tivoli Netcool Performance Manager components and its underlying framework. Several files are used to log errors for the Tivoli Netcool Performance Manager components and its underlying framework. These log files include: v COI log files v Deployer log file on page 206 v Eclipse log file on page 206 v Trace log file on page 206 See the Technology Pack Installation Guide for information about the technology pack log files.
205
Contains detailed MachinePlan_machinename_ [INSTALL_mmdd_hh.mm].log information about the tasks executed by the For example: COI steps on the MachinePlan_delphi_[INSTALL _0610_10.37].log specified machine DeploymentPlan.log Contains high-level information about the COI Plan execution
tmp/ProvisoConsumer/Plan/ logs/INSTALL_mmdd_hh.mm
Procedure
1. In the Topology Editor, select Window > Preferences. The Log Preferences window opens. 2. Select the new trace level. If desired, change the name of the log file. 3. Click Apply to apply your changes. To revert back to the default values, click Restore Defaults. 4. Click OK to close the window.
206
Appendix J. Troubleshooting
This section lists problems that might occur during an installation and how to resolve them. The problems are grouped by the interface or component exhibiting the problem.
Deployment problems
A list of deployment problem descriptions and solutions.
Problem The deployer window does not automatically become the focus window after launching from it from the Topology Editor. When the user tries to launch the Firefox browser an error is displayed regarding the Cairo 1.4.10 package: In a fresh installation, the database installation step fails. Solution Cause: In some cases (for example, when you export the display on a VNC session on Linux systems), the deployer window does not get the focus. User action: Click on the deployer window or move other windows to make the deployer window the focus window. Cause: Cairo 1.4.10 may not support the requested image format. User action: Start VNC server using the following command: /usr/bin/X11/vncserver -depth 24 -geometry 1280x1024 Cause: You did not perform the necessary preparatory steps. User action: This step verifies that the Oracle Listener is working properly before actually creating the Tivoli Netcool Performance Manager database. If the step fails: 1. Complete the necessary manual steps (see Configure the Oracle listener on page 46). 2. Change the status of the step to Ready. 3. Resume the installation. The step should complete successfully. Cause: There are many possible causes. User action: 1. Make sure the installation step is really in a hung state. For example, the Tivoli Netcool Performance Manager database-related steps might take more than an hour to complete; other steps complete in far less time. 2. Determine which child process is causing the hang. First, find the installer process by entering the following command: ps -ef The installer process has an entry similar to this one: root 12899 7290 10 13:43:31 pts/7 0:10 /tmp/install.dir.12899/Solaris/resource/ jre/jre/bin/java -Djava.compiler=NONE 3. Find the process that has that process number (for example, 12899) as its father. Continue until you find the last process. 4. Kill the last process using the following command: kill -9 At this point, the status of the hung step will change to Error. 5. If you can determine the cause of the hang, fix the problem and resume the installation. Otherwise, collect the log files and contact IBM for support. The deployer hangs when displaying Cause: The NFS file system is not working properly. User action: Run the df the Preview page. (This step normally -k command and make sure that all NFS mounted file systems are working takes only a few seconds). properly. When the problem has been corrected, restart the deployer.
Copyright IBM Corp. 2006, 2010
207
Solution Cause: The deployer uses either RSH or OpenSSH to perform remote command execution. You must configure OpenSSH to make this connection possible. User action: After configuring OpenSSH, run the test program provided in deployer_root/proviso/data/Utils/testremote.sh to test your configuration, where deployer_root is the root directory for the deployer. For example: /export/home/pvuser/443/SOLARIS/Install/ SOL9/deployer
Installation messages report success, This is screen noise and can safely be ignored. but might include messages similar to the following: Fatal Error]:4:1: An invalid XML character (Unicode: 0x1b) was found in the element content of the document. When you click the Done button to complete a fresh installation, the deployer displays database access error messages. Cause: You stopped a fresh installation before the installing and configuring the Tivoli Netcool Performance Manager database. User action:If the Tivoli Netcool Performance Manager database has not been installed, complete the installation using the -Daction=resume option (see Resuming a partially successful first-time installation on page 87). If the database has been installed, there is another problem. Contact IBM Software Support. Cause: When it starts, the channel manager (CMGR) places information in the database that is needed for real-time reports to start correctly. During installation, a cron job is created that starts CMGR. A silent installation might run fast enough that the cron job does not run before DataView is started. In this case, CMGR does not add the required information to the database, and real-time reports do not start up correctly. User action: 1. Make sure that the CMGR process is running (see Management programs and watchdog scripts on page 142 and Starting the DataChannel management programs on page 143). 2. Restart DataView. During Tivoli Integrated Portal install, User action: the Deployment Engine failed to find 1. Log in as root. pre-installed Tivoli Integrated Portal. 2. Enter the following commands to restart the Deployment Engine: # cd /usr/ibm/common/acsi/bin # ./acsisrv.sh -start 3. Check DE is running with the following command # ./listIU.sh This will list all IUs in the system.
Data does not appear in real-time reports, and right-clicking on a real-time report does not display the option menu. This problem can occur with a silent installation or a minimal deployment installation on a Solaris system.
208
The following excerpt from the file shows the resulting XML element:
<equals arg1="${remove.temporary.files}" arg2="false"/>
When you contact IBM support about a Tivoli Netcool Performance Manager installation problem, the support staff might ask you for these files. You can create a tar file or zip archive that contains the entire contents of the /tmp/ProvisoConsumer directory and send it to the IBM support staff for assistance.
A new channel component was deployed, or Cause: The channel components need to be the channel configuration was changed, but bounced. User action: Bounce the the change has no effect. components, as described in Appendix B, DataChannels, on page 141.
Appendix J. Troubleshooting
209
Telnet problems
A list of Topology Editor problems and solutions.
Problem Telnet client fails at initial connection and reports the following error: Not enough room in buffer for display location option reply Can occur when you start Tivoli Netcool Performance Manager components from a Solaris 10 system where the user interface is displayed remotely on a Windows desktop using an X Window tool like Exceed. Solution Cause: Length of the DISPLAY variable passed via the telnet client is too long (for example, XYZ-DA03430B70B-009034197130.example.com:0.0). User action: Set the value of the DISPLAY variable using the IP address of the local system, or the hostname only without the domain name. Then, reconnect to the Solaris 10 machine using the telnet client.
210
Java problems
A list of Java problems and solutions.
Problem Installer reports a Java Not Found error during installation of technology packs. Solution Cause: The installer expected, but did not find, Java executables in the path reported in the error message. The technology pack installation requires the correct path in order to function. User action: Create a symbolic link from the reported directory to the directory on the system where the Java executables are installed, for example: ln -s bin_path $JAVA_HOME/bin/java where bin_path is the directory where the binaries are located. After you create the symbolic link, you must re-start the technology pack installation.
Procedure
1. Make sure you are logged in as oracle and that the DISPLAY environment variable is set. 2. Enter the following command:
$ sqlplus system/password@PV.WORLD
In this syntax: v password is the password you set for the Oracle system login name. (The default password is manager.) v PV is the TNS name for your Tivoli Netcool Performance Manager database defined in your Oracle Net configuration. For example:
$ sqlplus system/manager@PV.WORLD
Connected to: Oracle9i Enterprise Edition Release 9.2.0.8.0 - Production With the Partitioning option JServer Release 9.2.0.8.0 - Production SQL>
Appendix J. Troubleshooting
211
Procedure
1. Make sure you are logged in as oracle and that the DISPLAY environment variable is set. 2. At a shell prompt, change to the following directory path:
$ cd $ORACLE_BASE/admin/skeleton/bin
3. Run the checkextc script, using the system database login name and password as a parameter:
$ ./checkextc system/password
For example:
$ ./checkextc system/manager
Connecting to Oracle ... Creating library LibExtCall ... Creating function ExternalCall ... Calling function ExternalCall ... Check libpvmextc.so configuration succeeded. Creating function Version ... Calling function Version ... Check Version libpvmextc.so - Revision: 1.0.1.1 Creating function ExternalPipe ... Calling function ExternalPipe ... Check ExternalPipe - /var/opt/oracle Dropping function Version ... Dropping function ExternalCall ... Dropping function ExternalPipe ... Dropping library LibExtCall ...
212
Location
<tip_location>/products/tnpm/dataview/legacy/bin Where <tip_location> is the Tivoli Integrated Portal installation directory, by default /opt/IBM/tivoli/tipv2.
Required privileges
Adequate privileges are required to read and write files to the file system. You must run this command from the UNIX command line as the Tivoli Netcool Performance Manager UNIX user (by default, pvuser), or a user with similar or greater privileges.
Syntax
synchronize.sh -tipuser <tip_username -tippassword <tip_password> -sourceuser <source_username> -sourcepassword <source_password> -sourceurl <source_url> [-pattern <pattern>]
Parameters
<tip_username> A Tivoli Integrated Portal user name for the local Tivoli Integrated Portal. <tip_password> The Tivoli Integrated Portal user password for the local Tivoli Integrated Portal. <source_username> A Tivoli Integrated Portal user name for the remote Tivoli Integrated Portal.
213
<source_password> The Tivoli Integrated Portal user password for the remote Tivoli Integrated Portal. <source_url> The URL of the remote server, including the port and DataView context.
Optional parameter
<pattern> The name pattern that identifies the types of files to filter for the synchronization. Wildcards * and ? are supported. To synchronize all files, omit the pattern; do not use * on its own to synchronize all files.
Example
The following command synchronizes the DataView custom .jsp file content from a remote Tivoli Integrated Portal to a local Tivoli Integrated Portal: synchronize.sh -tipuser <tip_username> -tippassword <tip_password> -sourceuser <source_username> -sourcepassword <source_password> https://server.ibm.com:16711:PV *.jsp
214
The HTML and Java code sections are read from the page, processed and combined to generate a new JSP.
HTML
When processing HTML sections, SilverStream page controls are replaced with custom DataView JSP tags. For example, consider the following HTML from a SilverStream page:
<div> <h1><AGCONTROL name="stylesheetNameLabel"></h1> </div>
That HTML is converted to the following JSP. The className attribute in the new JSP tag comes from the SilverStream page control metadata section of the SilverStream page.
<<div> <h1><proviso:pageControl className="com.sssw.shr.page.AgpLabel" name="stylesheetNameLabel"></h1> </div>
Java code
When processing Java code sections, the code is embedded into the JSP directly using the standard JSP statement tag. For example, consider the following Java code from a SilverStream page:
<class SomePage extends AgpPage { private AgpLabel label; public SomePage() { this.label.setText("foo") } }
All JSP fragments generated from the HTML and Java code conversion are combined into one JSP page. When executed, this JSP page generates the same HTML as the SilverStream page. The resulting JSP can only be run by using the SilverStream emulation layer. If you are using unsupported SilverStream page controls (AgpTabPane, AgpImageHotSpot and AgpButtonRadio) then the resulting JSP will not work correctly.
215
Location
<tip_location>/products/tnpm/dataview/legacy/bin Where <tip_location> is the Tivoli Integrated Portal installation directory, by default /opt/IBM/tivoli/tipv2.
Required privileges
Adequate privileges are required to read and write files to the file system. You must run this command from the UNIX command line as the Tivoli Netcool Performance Manager UNIX user (by default, pvuser), or a user with similar or greater privileges.
Syntax
migrate.sh -tipuser <tip_username> -tippassword <tip_password> -ssurl <silverstream_URL> -ssuser <silverstream_username> -sspassword <silverstream_password> -target <content|users|all> [-import ] [-verbose ]
Parameters
<tip_username> A Tivoli Integrated Portal user name for the local Tivoli Integrated Portal. <tip_password> The Tivoli Integrated Portal user password for the local Tivoli Integrated Portal. <silverstream_URL> The URL of the DataView SilverStream server. <silverstream_username> The SilverStream administrator user name. <silverstream_password> The SilverStream administrator password. <content|users|all> Indicates the type of data to be migrated: content Migrates all SilverStream content. All content is copied to the content directory in the Tivoli Integrated Portal installation <tip_location>/products/tnpm/dataview/legacy/content. Depending on the type of content, it is copied to the following directories: v <tip_location>/products/tnpm/dataview/legacy/content/ SilverStream/Pages
216
v <tip_location>/products/tnpm/dataview/legacy/content/ SilverStream/Objectstore/Images v <tip_location>/products/tnpm/dataview/legacy/content/ SilverStream/Objectstore/General users Migrates all SilverStream users. Each user in SilverStream is either an administrator or user. In the Tivoli Integrated Portal, the administrator role is mapped to tnpmAdministrator, and the user role is mapped to tnpmUser. The migration tool creates two user groups, tnpmAdministrators and tnpmUsers, that contain all the users with the corresponding roles. Password information is not migrated. Under Tivoli Integrated Portal, the password is set to be the same as the user name. For example, the SilverStream user pvuser with password pv becomes user pvuser with password pvuser when migrated to the Tivoli Integrated Portal. all Exports all SilverStream content and users.
Optional parameters
-import Indicates data should be imported into the Tivoli Integrated Portal. -verbose Indicates additional migration messages should be displayed.
Example
Run the following command as user root. Assuming a default installation, this command copies all SilverStream content and users in the SilverStream server silverstreamserver to the /opt/IBM/tivoli/tipv2/products/tnpm/dataview/ legacy/content directory in a Tivoli Integrated Portal installation: migrate.sh -tipuser <tip_username> -tippassword <tip_password> http://silverstreamserver:8080/PV -ssuser <silverstream_username> -sspassword <silverstream_password> all
217
218
Notices
This information was developed for products and services offered in the U.S.A. IBM may not offer the products, services, or features discussed in this document in other countries. Consult your local IBM representative for information on the products and services currently available in your area. Any reference to an IBM product, program, or service is not intended to state or imply that only that IBM product, program, or service may be used. Any functionally equivalent product, program, or service that does not infringe any IBM intellectual property right may be used instead. However, it is the user's responsibility to evaluate and verify the operation of any non-IBM product, program, or service. IBM may have patents or pending patent applications covering subject matter described in this document. The furnishing of this document does not give you any license to these patents. You can send license inquiries, in writing, to: IBM Director of Licensing IBM Corporation North Castle Drive Armonk, NY 10504-1785 U.S.A. For license inquiries regarding double-byte (DBCS) information, contact the IBM Intellectual Property Department in your country or send inquiries, in writing, to: Intellectual Property Licensing Legal and Intellectual Property Law IBM Japan, Ltd. 1623-14, Shimotsuruma, Yamato-shi Kanagawa 242-8502 Japan The following paragraph does not apply to the United Kingdom or any other country where such provisions are inconsistent with local law: INTERNATIONAL BUSINESS MACHINES CORPORATION PROVIDES THIS PUBLICATION "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESS OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF NON-INFRINGEMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Some states do not allow disclaimer of express or implied warranties in certain transactions, therefore, this statement might not apply to you. This information could include technical inaccuracies or typographical errors. Changes are periodically made to the information herein; these changes will be incorporated in new editions of the publication. IBM may make improvements and/or changes in the product(s) and/or the program(s) described in this publication at any time without notice. Any references in this information to non-IBM Web sites are provided for convenience only and do not in any manner serve as an endorsement of those Web sites. The materials at those Web sites are not part of the materials for this IBM product and use of those Web sites is at your own risk.
219
IBM may use or distribute any of the information you supply in any way it believes appropriate without incurring any obligation to you. Licensees of this program who wish to have information about it for the purpose of enabling: (i) the exchange of information between independently created programs and other programs (including this one) and (ii) the mutual use of the information which has been exchanged, should contact: IBM Corporation 2Z4A/101 11400 Burnet Road Austin, TX 78758 U.S.A. Such information may be available, subject to appropriate terms and conditions, including in some cases payment of a fee. The licensed program described in this document and all licensed material available for it are provided by IBM under terms of the IBM Customer Agreement, IBM International Program License Agreement or any equivalent agreement between us. Any performance data contained herein was determined in a controlled environment. Therefore, the results obtained in other operating environments may vary significantly. Some measurements may have been made on development-level systems and there is no guarantee that these measurements will be the same on generally available systems. Furthermore, some measurement may have been estimated through extrapolation. Actual results may vary. Users of this document should verify the applicable data for their specific environment. Information concerning non-IBM products was obtained from the suppliers of those products, their published announcements or other publicly available sources. IBM has not tested those products and cannot confirm the accuracy of performance, compatibility or any other claims related to non-IBM products. Questions on the capabilities of non-IBM products should be addressed to the suppliers of those products. This information contains examples of data and reports used in daily business operations. To illustrate them as completely as possible, the examples include the names of individuals, companies, brands, and products. All of these names are fictitious and any similarity to the names and addresses used by an actual business enterprise is entirely coincidental. COPYRIGHT LICENSE: This information contains sample application programs in source language, which illustrate programming techniques on various operating platforms. You may copy, modify, and distribute these sample programs in any form without payment to IBM, for the purposes of developing, using, marketing or distributing application programs conforming to the application programming interface for the operating platform for which the sample programs are written. These examples have not been thoroughly tested under all conditions. IBM, therefore, cannot guarantee or imply reliability, serviceability, or function of these programs. You may copy, modify, and distribute these sample programs in any form without payment to IBM for the purposes of developing, using, marketing, or distributing application programs conforming to IBMs application programming interfaces.
220
If you are viewing this information in softcopy form, the photographs and color illustrations might not be displayed.
Notices
221
222
Trademarks
IBM, the IBM logo, and ibm.com are trademarks or registered trademarks of International Business Machines Corp., registered in many jurisdictions worldwide. Other product and service names might be trademarks of IBM or other companies. A current list of IBM trademarks is available on the Web at Copyright and trademark information at http://www.ibm.com/legal/copytrade.shtml. Adobe, Acrobat, PostScript and all Adobe-based trademarks are either registered trademarks or trademarks of Adobe Systems Incorporated in the United States, other countries, or both. Cell Broadband Engine and Cell/B.E. are trademarks of Sony Computer Entertainment, Inc., in the United States, other countries, or both and is used under license therefrom. Intel, Intel logo, Intel Inside, Intel Inside logo, Intel Centrino, Intel Centrino logo, Celeron, Intel Xeon, Intel SpeedStep, Itanium, and Pentium are trademarks or registered trademarks of Intel Corporation or its subsidiaries in the United States and other countries. IT Infrastructure Library is a registered trademark of the Central Computer and Telecommunications Agency which is now part of the Office of Government Commerce. ITIL is a registered trademark, and a registered community trademark of the Office of Government Commerce, and is registered in the U.S. Patent and Trademark Office.
Java and all Java-based trademarks and logos are trademarks or registered trademarks of Sun Microsystems, Inc. in the United States, other countries, or both.
Linux is a trademark of Linus Torvalds in the United States, other countries, or both. Microsoft, Windows, Windows NT, and the Windows logo are trademarks of Microsoft Corporation in the United States, other countries, or both. UNIX is a registered trademark of The Open Group in the United States and other countries. Other company, product, and service names may be trademarks or service marks of others. For trademark attribution, visit the IBM Terms of Use Web site (http://www.ibm.com/legal/us/).
223
224
Printed in USA