Skip to main content
Question

Restoring VMs from Tape / Index

  • April 25, 2025
  • 3 replies
  • 144 views

Forum|alt.badge.img+1

Hello all,

 

We have a typical Hyper-V environment with about 50 VMs. We're performing full backups daily.. For offsite copies, we use a GFS (Grandfather-Father-Son) backup setup with a tape library with daily/weekly and monthly/yearly retention policies.

Everything runs fine overall—we just need to manually delete daily tapes after a week due to retention issues with big data apps  not syncing correctly with the planned retention schedule. This is manageable on our end.

The real issue is with restores. We have a backup site that is fully offline and is a clone of the primary site. When we try to restore VMs, most restore fine and grab the index from the daily tape. However, some VMs require indexes from other tapes. To restore all VMs, we often need 2–3 additional tapes from within the 7-day window. In other words, not all indexes are stored on a single tape—this is evident from the index size per tape.

It seems the problem is related to the IndexCheckpoint and how the index database is written every 7 days. Since we’re performing full backups daily, it appears the system skips creating a new checkpoint or consolidated index, which might be causing this issue.

Commvault Support hasn’t been able to understand the problem so far. Has anyone else experienced something similar?

is there a way to force more frequent index checkpoints ?

3 replies

Forum|alt.badge.img+10

Hi ​@DTO ,

Good day!

In general for any client we will always keep 3 checkpoints (In V2 indexing) | For index backups there will be separate client Storage policy name _ index server (Bigdata apps) which is pointed by default to same storage policy or to other if manually change

During index reconstruction we will try to restore the indexDB from index backup job and then play the action logs from consecutive backup jobs required

Thats the reason why some time you may see 2-3 additional tapes required sometime.

During backup the index action logs only be created and they will be played back to index DB asynchronously.

Regards,

Sureshkumar S

 


Forum|alt.badge.img+1
  • Author
  • Byte
  • April 28, 2025

Hallo ​@Sureshkumar S 

Thanks for the explanation. Given our restore environment (offline DR clone with limited tape access), we would really benefit from having self-contained indexes on each daily tape.

Is there a way to change the default checkpoint interval and force a new index checkpoint to occur more frequently—ideally daily instead of every 7 days?

 

Thank you


Forum|alt.badge.img
  • Byte
  • August 28, 2025

Hi There,

Im also having the same issue regarding the tapes. Is there a workaround?

I was going through the Commvault documentation and came across media refresh. Can this help for these situation?

https://documentation.commvault.com/v11/commcell-console/media_refresh.html