commvault.com commvault.com
Solved

Index Cache

  • 9 February 2022
  • 4 replies
  • 3237 views
Henke
Rank
Userlevel 4
Badge +13

Index Cache (V2).

Yesterday I moved an IndexCache to another drive due to space issues. The index was at the time 293GB in size on a 300GB drive. I moved it with the Workflow “ChangeIndexCacheConfig” to a new destination with 1TB of free space.

Today when I check the index today it is 440GB in size, so it growed almost 50% over night.

Why would it grow this much? Just normal incremental backups was run over night? No additional clients added or any changes made to clients being backed up by the MA.

Index cache settings as displayed below.

 

 

BR 

Henke

icon

Best answer by Niall 9 February 2022, 13:35

View original
If you have a question or comment, please create a topic

4 replies

Niall
Rank
Userlevel 3
Badge +8

Hi Henke, 

When you only had 7GB of free space on the Index Cache, the Index clean-up process would have been running constantly trying to clear down space as you will have hit capacity thresholds. And now you have a much larger disk you aren't hitting those capacity thresholds anymore. 

HTH

Henke
Rank
Userlevel 4
Badge +13

@Niall thank you for the answer.

So you think keeping only 2 days of backups in IndexCache using 440GB of disk seems to be accurate for an Index with  1782273823 objects? 

Or isn’t the “days” used when the index is V2?

//Henke

Niall
Rank
Userlevel 3
Badge +8

Hi Henke

With V2 Indexing, the Index is a database and action logs - whilst the action logs are cleaned up following an index backup the database never actually gets deleted unless you delete the associated backupset.  I think I am right in saying that your 2 day retention would only apply to v1 indexes (Even then I don't think that would work unless you were running full backups every 2 days). 

It always used to be the case that Index compaction and backups happened every 30 days or x number of objects. Given how tight you were on space previously I think this would been happening pretty much continuously and that right now it is waiting on the 30 days or x number objects thresholds to be reached. (Keep in mind that is per DB and there is a DB per backupset). Given the situation I am not at all surprised the size of the Index Cache has grown. 

That said 440GB does strike me as alot for the number of objects you mention but it is going to depend on what the objects are

 

Henke
Rank
Userlevel 4
Badge +13

Thanks for the input @Niall 

I’ll open a support ticket to have a sanity check done on the index cache of the systems in question.

//Henke

Terms & ConditionsCookie settings