Sunteți pe pagina 1din 2

1. MUST run "pxehttp" in addition to "pxetftp" to enable PXE installation entirely from LAN. 2.

on the 2nd stage of the PXE installation, when asked about the rPath location, enter: HTTP --> server: 192.168.16.1:8000 rPath: openfiler/mnt/ 3. it appears that when installing to a new USB drive, the uninitialized MBR will prevent openfiler from recognizing it as a valid installation device. therefore we MUST initialize the MBR, with either the following: a) in linux: dd if=/dev/zero of=/dev/<sd# of USB> count=1 bs=512 cfdisk /dev/<sd# of USB> # create a new partition. OR: b) write a gPXE usb image on it: dd if=/path/to/gpxe.usb of=/dev/<sd# of USB> 4. default web GUI's user name & password: username: password: openfiler password

5. httpd SSL certificate and configurations: /opt/openfiler/etc/httpd/conf/ 6. creating groups and users with LDAP: in "Accounts" tab, select "Use LDAP". then: "Use Local LDAP Server", check. "Use TLS", DO NOT CHECK. SMB has problem with this in OF 2.99. "Server" 127.0.0.1 "Base DN": dc=<your-domain>,dc=<your-net> "Root bind DN:" cn=manager,dc=<your-domain>,dc=<your-net> (appears MUST be "manager"). "Root bind password:" use root password. "Login SMB server to root DN", check. "Allow user to change password", check. <Submit>, then goto "Administration" on the right-panel to add groups and users. 7. OF 2.99's volume management part is quite buggy. the 1st problem is that when creating partitions using the GUI only about 95% capacity is used. this can be done by 'parted' from the shell. run: parted /dev/<disk> print rm <#> mklabel mkpart primary 0 -0 set <#> raid on select /dev/<disk> quit # # # # # # # show the current disk/partition info. remove the given partition. create label (should be MBR), usually choose GPT. use the whole disk as partition. set the given partition as RAID partition. select another disk. exit parted.

the above can be used to create a partition of the entire capacity. but the GUI still cannot add the partitions to RAID-1 (mirror). this is

because that the 'mdadm' comes with OF 2.99 is broken. to fix that: conary update mdadm=openfiler.rpath.org@rpl:devel/2.6.4-0.2-1 this downgrade mdadm from version 3.2-2-1 to 2.6.4-0.2-1. now everything should be working. [https://forums.openfiler.com/viewtopic.php?id=6252] 8. (LVM) volume group, logical volume commands: vgdisplay, vgcreate, vgremove lvdisplay, lvcreate, lvremove [http://tldp.org/HOWTO/LVM-HOWTO/commontask.html] [http://jreypo.wordpress.com/tag/mkfs-ext3/] 9. to use full size of the volumes, seems we have to create an XFS filesystem volume. 10. MUST SET "NETWORK ACCESS CONFIGURATION" to allow client access. 11. a good reference: [http://www.maykinmedia.nl/blog/2009/apr/28/installing-openfiler-part-two/] 12. FAILURE RECOVERY: just doc here from the forum posts. may be useful when things go wrong with a disk in the RAID array: a) b) c) d) e) f) Power down OpenFiler. Remove failed drive, and install new drive. Restart OpenFiler. Partition the new drive to match the failed drive. Add new volume into the RAID set. Be patient. Initially the system will show the drive added as a spare. Eventually, the status will change to show that it is initializing. g) As with any RAID, the system is usable during initialization, although performance might be a bit diminished. TO IDENTIFY THE FAILED DRIVE: we can use "dd" command to read a block of the drives to see which one is failing.

S-ar putea să vă placă și