Solved

ddb disks run out of space, what can I do

  • 12 December 2023
  • 4 replies
  • 102 views

Badge +5

Hello Team,

my customer has a 400TB backend storage and 4TB ddb space (two partitions). Both ddb disks are full.

I see 1.062.831.168 unique blocks on files db. Fact is he needs to increase his ddb space but it takes time to organize disks and install them. 

What can we do to bridge this time. Is there a “space reclamation” for the ddb disks too. May be a seal is the solution because from ddb view we start from scratch. I know the obvious consequences such as baseline and network load. Empty the ddb paths and do a restore was a thought.

 

Do you have other suggestions

Thanks

Joerg

icon

Best answer by Chris Hollis 14 December 2023, 05:50

View original

4 replies

Userlevel 6
Badge +15

Hi @Joerg 

Yes, if sparse support is disabled you can try space rec. 

https://kb.commvault.com/article/81665 outlines some things to try (first two are likely not possible for you at this stage), third thing to try is a compact: 

https://documentation.commvault.com/v11/expert/compacting_deduplication_database.html

This may help?

Otherwise besides that, the other options are things you’d probably not want to try or do… involving drastically lowering retention settings (at least 75%).. 

Regards,

Chris 

Badge +5

Hi Chris,

this helps👍

Thank you

Joerg

Badge +3

@Joerg, sealing option could be taken only in combination with marking the DDB as corrupted. Since otherwise we still need DDB for pruning.

 

Badge +5

@Abhishek Narulkar :  Sealing would be a very drastic step. I didn't suggest it to my customer, but now I'm curious.

I had a look at my lab machine because I did not remember where it is possible to mark  the ddb as corrupted. I found a checkbox “enable physical pruning” but nothing to mark the ddb as corrupted. Where can I do this?  

Reply