Proactive Support - Unexpected SQL TempDB Database and Log growth

  • 15 June 2021
  • 2 replies
  • 336 views

Userlevel 5
Badge +9

After 12th of June 2021, you may encounter a situation on your Commserve whereby the SQL TempDB (tempdb.mdf) and Log (templog.ldf) are increasing in size unexpectedly.


This may continue until the volume on the Commserve holding these files becomes full, which can lead to performance issues on the Commserve and observed in the Commcell console.

Commvault released a reporting enhancement on the 11th June 2021 that has inadvertently caused tempdb/templog file growth.

This enhancement has since been fixed however your environment may still be showing high space cosumption from the tempdb and templog files.

If you are witnessing this behaviour and have low disk space on the drive hosting these SQL database files, please carry out the following on the Commserve server:

1. Disable All Activity in the Console Activity Control (commvault.com)

(Right click Commserve > Properties > Activity Control > Uncheck "Enable All Job Activity" > Set "Enable after a Delay" for 30 minutes)

2. Suspend all jobs running in the job controller Suspending a Job in the Job Controller (commvault.com)

3. Stop all Commvault services on the Commserve instance:

(Launch Commvault Process Manager for Instance001 > Services Tab > Highlight "All Services" > Stop)

4. Launch Services.msc from the Run dialog box on the Commserve and Restart the SQL Server (COMMVAULT) service Start, stop, pause, resume, restart SQL Server services - SQL Server | Microsoft Docs

5. The tempDB and templog files should have now reduced back to their startup size
 

6. Start the Commvault services on the Commserve and reactivate activity (if it hasn't already automatically enabled). You may then resume suspended jobs.

7. Should the issue remain after SQL server service restart, please notify the hotline/raise a case for further analysis.

Additionally, if you have sufficient disk space, the tempdb and templog files will shrink naturally.


2 replies

Userlevel 4
Badge +11

Hi there,

 

I had the case in july, and we provided disk expansion for the pending SQL transactions to be properly executed, before we would stop activity (though in fact when there were no more disk space, the CS Activity stopped). 

Then I created an alert so make sure that from the warning event of the CS tempDB usage growing beyond 80%, I would be notified, and would be able to pro-act instead, of react..

 

Last night, at I got that notification.


  • Event ID: 19956392
  • Monitoring Criteria: (Event Code equals to 59:66)
  • Severity: Major
  • Event Date: Thu Sep 23 23:04:37 2021
  • Program: Sql Agent
  • Client: myCommserve
  • Description: TempDb Usage is high

 

Yesterday I applied FP58 (and a hotfix), while I was on FP53 (v11FR11).

May we know in which FP or FR.FP this issue is solved, as I understood that it should now be?

Otherwise, of course I would raise a case. 

Userlevel 7
Badge +21

Hey @Laurent !  the enhancement/change was on our cloud reporting side, so any update you applied (if that was the cause) would have had no effect on the above issue.

Likely worth a case!

Reply