Sun System Handbook - ISO 3.4 June 2011 Internal/Partner Edition | |||
|
|
Solution Type Technical Instruction Sure Solution 1298177.1 : VTL - Single Instance Repository (SIR) Storage Usage Alert - Threshold Exceeded
What to do when Single Instance Repository (SIR) storage usage is above threshold. In this Document Created from <SR 3-3019953491> GoalThe objective of this document is to advise on what to do when the following alerts are encountered:Repository storage usage is 92%. Alert level is 90%.-Last Sunday on 20/02/2011 @18:52:14 SGT 2011 Repository storage usage is 91%. Alert level is 90%.-Last Saturday on 19/02/2011 @18:51:45 SGT 2011 Repository storage usage is 90%. Alert level is 90%.-Last Friday on 19/02/2011 @18:51:19 SGT 2011 SolutionThese alerts are generated by the SIRmonitor.sh script which checks the index repository and data repository to see if the usage is above the alert trigger level (by default, the trigger level is 90%).When these alerts are encountered, do the followings for troubleshooting: 1. On the VTL Console, highlight the VTL server and select the Deduplication Statistics tab in the right panel. This will display the Deduplication Statistics which identify if it is the Index Disk Capacity or Data Disk Capacity that is triggering the alerts.
2. On the VTL Console, right-click your VTL server and select: Deduplication --> Reclamation --> Reclamation Parameters This will display the Reclamation Configuration window. From this window, check if the "Reclamation Enabled" box is checked. If it is not, then automatic space reclamation is not enabled. Preferably, it should be enabled (by default it is enabled). 2.1 If automatic space reclamation is not enabled, then it will be necessary to perform manual reclamation. To manually run reclamation, right-click your VTL server and select: Deduplication --> Reclamation --> Start Space Reclamation (or Start Index Pruning). Notes: • Start Space Reclamation - Composed of index reclamation and storage reclamation, this option reclaims data disk space and memory. • Start Index Pruning - Reclaims space on the index disk. This process is CPU intensive and does not need to be run frequently. 2.2 If automatic space reclamation is enabled, check the Interval value in the Reclamation Configuration window which shows how frequently automatic reclamation will run in minutes. If this value is too high (frequency low), then the Interval value can be reduced so that space reclamation will run more often. The appropriate value will depend on how fast the spaces are being used up and vary from site to site. Observation of usage is required to determine an appropriate value. 3. There may be genuine situations whereby the SIR storage is really running out of space, perhaps due to new and additional backups being introduced. In this case, all Sun-STK VTL Prime configurations are pre-configured using all available disk capacity with no capacity upgrade available. So it is not possible to add additional disk trays to an exisitng configuration. Short term workaround is expire and remove any older backups from VTL Cache and SIR db. Checklist of items to be performed: 1. Expire/remove the backup tapes from backup application 2. Delete the virtual tapes associated with this (these) backups from VTL 3. Run Space Reclamation on the VTL cache 4. Run Space Reclamation on the Data portion of the SIR db 5. Run Index Pruning on the Index portion of othe SIR db 6. Re-create virtual tapes and assign to backup application for use Long term solution is to migrate to a larger VTL Prime configuration that fits the current backup requirements (VTL Prime models range from 2 TB to 65 TB). Please contact local Oracle Sales Representative for options. Attachments This solution has no attachment |
||||||||||||
|