Dat e: April 27, 2007 From: J im Ridenour - CSG RE: Suggested Pro-Watch Maintenance Tasks Pre-requisit es: This bulletin describes various tasks that are recommended to help achieve the best performance and availability of the Pro-Watch software suite. These tasks will help ensure Pro-Watch is operating at its best performance. Some of these tasks may be able to be performed by Pro-Watch administrators; others may require some SQL or Windows knowledge. Required media: For ProWatch Professional (PWPE) systems, it may be of value to install SQL Enterprise Manager on the server machine. Inst allat ion st eps: The tasks described are to be performed on the Pro-Watch server. 1. Monthly Tasks a. Archive and purge the Pro-Watch Event Log i. It is recommended to archive and purge one month of log activity at a time. In some cases for extremely busy systems, this may need to be broken into smaller increments in order for the purge to complete successfully.
Note: PWPE installations employ MSDE, which is limited to a 2GB database file size. If this size is reached, the system may exhibit reduced performance attitudes. It is important to monitor this and schedule Archive/Purges in order that the maximum file size is not reached. Recovering from the instance where the database has reached its maximum size may result in the loss of all transaction log data that is in the database.
ii. Purging data is resource intensive. It is recommended that the purge process be run during off hours. Archive/Purge jobs can be scheduled using the Desired Start Time field in the Archive Purge screen
b. Check the Windows Event Logs (System and Application) i. You should take a look through the Application Log and the System log of the server to see if there are any Errors or Warnings. Many times these errors will tell if there are any potential hardware or Windows problems before they cause serious issues. You should also see messages from MSSQLServer telling you that your backups are succeeding. Some errors in the System log may be normal, depending on your environment. CSG Technical Support Bulletin 4/27/2007 Confidential 2
c. Check your drive space. i. You should check the amount of free space on your system drive (typically C) and your data drive (wherever the PwntData.mdf file is located) to make sure neither one is getting low. If the data drive runs out of space, it may cause database corruption and will definitely shutdown Pro-Watch. Ideally, half the drive should remain as free space, especially if backups are sent to an external destination.
2. Annual or Semi-Annual Tasks Note: Highly active systems, those with more than 100,000 events per day, may need the following maintenance twice a year or perhaps even more frequently. More moderately active systems may require this maintenance only once a year.
PWPE installations may require more maintenance to achieve the database size limits and may require these activities more often. This can only be gauged over time and should be monitored closely, especially if the system configuration or number of cardholders changes or if there is a significant change in daily activity.
a. Archive and purge the Audit log i. The Audit Log stores information about what changes were made to specific tables in the Pro-Watch database. In general, the data stored in this table takes less space than the Event Log, but it will build up over time. Audit log Archiving and Purging is executed at the same screen as Event Log Archiving and Purging.
b. Check and clean the CHANNEL_DOWNLOAD table i. This will need to be done by running scripts through Query Analyzer or using SQL Enterprise Manager. Generally speaking, there should be no records in this table, unless something is actively being downloaded. However, overtime records can occasionally be placed in this table, and they need to be deleted.
c. Check and clean the HI_QUEUE table i. This will need to be done by running scripts through Query Analyzer or using SQL Enterprise Manager. Generally speaking, there should be no records in this table, unless something is actively being downloaded. However, overtime records can occasionally be placed in this table, and they need to be deleted.
d. Check the UNACK_AL table i. The UNACK_AL table contains all of the alarms that are being displayed on the Alarm Monitor. This table should show about twice the number of rows that are on the Alarm Monitoring screen for a user that should see all of the channels. If there are many rows showing in the table that are not on the screen, then some channels in the system are not being properly included in the Routing Groups. A large number of rows in this table negatively affect system performance.
e. Confirm the number rows in the EV_LOG table is where it is expected to be i. This will confirm that the Archive and Purges have been running properly and that no large amount of data has been accidentally missed. CSG Technical Support Bulletin 4/27/2007 Confidential 3
f. Re-Index all of the tables in the PWNT database i. Re-Indexing the tables will improve performance and reduce database size. This should be scheduled during a period when the Pro-Watch system can be stopped. Because re-indexing locks tables, it causes Pro-Watch to stop if you do not manually shut it down first. With this in mind, it is generally not recommended that the Re-Index tool be run as part of a regularly-scheduled Database Maintenance plan, because it might cause a high number of unexpected shutdowns of the Pro-Watch Server service.
g. Clean out spool files. i. As part of the download process, Pro-Watch creates temporary files in the ProWatch\Spool\ directory. These files have long hexadecimal names, and most likely will not have an extension. On occasion, Pro-Watch may leave these files in the NonPri or Pri folders under the individual channel folders. These files will not typically cause problems, however it is recommended that they are cleaned to reclaim drive space and prevent unexpected downloads from occurring (Pro- Watch tries to resend any spool files when the Pro-Watch Server service is restarted.)