Skip to main content

I got during the last full backup the following message:
“Index cache eD:\IndexCache] free space on MediaAgent thost name] is i51197] MB. It is considered critically low. Backups will start failing if the cache runs out of space. Take one of the following actions: 1. Increase the index cache total disk space. 2. Reduce the retention setting on the index cache.”

Question:
How can I reduce the index cache size?
 

Hi Marius

Thank you for the question.

The simplest way to resolve this situation is to add more disk space to the index cache volume.

If this isn’t possible, there are some other things to check, but these involve different approaches depending on whether this is Indexing V1 or V2.

Index Cache Cleanup (V1)

Index Cleanup and Compaction Operations (V2)

 

Another potential operation, for Indexing V2 is to run index server load balancing to spread the load across all Indexing V2 servers in the environment:

http://kb.commvault.com/article/IND0012

 

We would need a little more detail, Commcell version, agents involved, etc. to understand if V1 or V2 indexing is being used here.

 

Thanks,

Stuart


We both in use: backup clients with indexing version 2 and backup clients with indexing version 1:

As first action point I just reduced the retention time for the index 365 → 180 days

Do you mean it will be sufficient?


Please see the attached files


CommVault version 11.21
Windows server 2016 (incl. indexing 2) file data agents
Windows server 2012 R2 (incl. indexing 1) file data agents


Hi Marius,

 

We recommend users to take a close look at the new “Index Cache Location” health report to check the current index cache status for their CommCells. You can check and follow-up the index cache space usage and prediction (a.k.a Date to be Full) at Cloud > Reports > CommCells (find and select your commcell) > Health.

We have a new version for “Index Cache Location” and “Index State” reports, with more details and guides to help solve issues.

 

 

Regarding the index cache usage, some important things to notice:

  1. Changing the pruning options won’t affect indexing V2, just V1, so you may not see a big difference immediately after changing it.
  2. We also recommend customers to run the “Load Balance Index Servers” workflow once in a while to distribute the indices between available caches, as mentioned by Stuart. This is only for Indexing V2.
  3. It’s important to have at least 2 Media Agents as data-path options in the primary copy of your Storage Policy’s configuration. V2 is smart and will try always to use the least loaded index server as option.
  4. When load balance does not help - for example, when there is no option to move index due to configuration or other index being overloaded as well - you’ll be asked to act by adding more disk to the cache or adding another Media Agent to share the load.
  5. Index load is also very volatile, during some jobs the load increases a lot and after some time some of the files related to the backup/restore will be deleted from the cache. So, you’ll see the cache changing size a lot during operation. Just keep checking everyday how it’s progressing and act as recommended by the health reports and you should be fine.

I’m glad to help or answer any further questions you have.


When I select  Cloud > Reports > CommCells (find and select your commcell) > Health

I see the message:

“This commcell environment is not sending data to the cloud metrics reporting server”

 


Yesterday I changed the retention time for the index 365 → 90 days (media agent → properties → catalog)
I see now there is reducation in the index cache by approx 50 GByte.

Please note I use only one media agent and no dedicated index server. The index cache is located on one physical drive of the media agent.
 


@Marius , regarding the error about not sending data, you can configure this in Command Center or the Commcell Console under Diagnostics and usage.

Regarding your space savings, is that adequate for your needs?


After the retention time reduction to 90 days the free storage space has been increased by 20%
I think that should be sufficient for the next 1 year. Notwithstanding the above I will try to activate th Index Cache Location health report that you mentioned. Thanks!


Reply about running the Load Balancing workflow was split here:

 

 


Reply