commvault.com commvault.com
Solved

Critical Event - DDB reaching QI time threshold...

  • 2 July 2021
  • 5 replies
  • 3672 views
B
Rank
Userlevel 2
Badge +8

Still a noice commvault user..  Noticed that we have been getting the following Event in the evvent viewer..

 

DDB(s) reaching QI time threshold for adding new Subclients [
{DDB [XXXXXX-GlobalDedup_Databases_5] MediaAgent [XXXX-cv1a] Path [D:\XXXX-DDP] AvgQITime [1153] Free Space Percentage [94] }
]. New DDB(s) may gets created if it crosses threshold. Please review the DDB configuration, for more details refer Deduplication Building Block Guide on Books Online.

 

Can anybody help me understand this and how to correct this.   Our small environment is all on AWS.

 

Thanks BC

 

icon

Best answer by Stuart Painter 5 July 2021, 08:33

View original

5 replies

S
Rank
Userlevel 7
Badge +15

Hi @bc1410 

Thanks for the question.

Query and Insert (Q&I) times are a measure of DDB performance with the expectation that Q&I remains below 1000μs.

Since the average Q&I time for your DDB has cross over this threshold, it’s being highlighted as a potential issue and can become a factor in job performance.

This may be because the sub/clients associated with the DDB have increased and the DDB can no longer keep up with the volume or the MA/disk performance was ok, but now the number of sub/clients has pushed this above a threshold where the DDB can cope with volume.

Please take a look at the DDB Sizing Guide to ensure the MA and DDB disk meet the expected minimum standards for the size of your environment.

Thanks,

Stuart 

Customer Support
B
Rank
Userlevel 2
Badge +8

Thanks Stuart.

I will look at the guide..  But just to note - since we started the commvault environment on AWS which was back in late March early april 2021 there has been no additional CLIENTs added.   Its a fairly small environment.  We dont even have the license capabilty to add more clients.  

Thanks

Brian

S
Rank
Userlevel 7
Badge +15

Hi @bc1410 

Thanks for the update.

The graph above shows a considerable increase in average Q&I times on or around 13th June, indicating that something has changed.

This might be caused by:

  • an increase in clients (ruled out now with your update)
  • A change in hardware, ie. DDB relocated from SSD to spinning disk or reduction in MA RAM/CPU.
  • A change in schedule, e.g. if DDB backup or pruning is happening during backup window, this can push up average Q&I. Try and separate these schedules, so these tasks are performed separately.
  • A change in client workload/scheduling. Has more content been added or have client schedules been moved so more backup jobs are run at a similar time with overlapping schedules
  • A change in dedupe workload, e.g. more clients enabled for dedupe or maybe a shift from client-side to MA-side dedupe processing, causing overall load increase on MA

There’s a few more things to consider and check that might impact DDB performance.

Thanks,

Stuart

Customer Support
G
Rank
Userlevel 4
Badge +9

https://documentation.commvault.com/commvault/v11_sp20/article?p=111985.htm

M
Rank
Userlevel 2
Badge +4

Hello,

let me add here that one common issue with Q&I times increasing may be caused by Antivirus software exclusions not being in place for deduplication engine path. If you confirm that from sizing perspective machine is all okay, I’d suggest to also take a look at recommended AV exclusions:

Antivirus Exclusions

 

Best Regards,

Marcin

Reply


Terms & ConditionsCookie settings