Skip to main content

Good afternoon,

I wanted to check with the community before generating a case with support, regarding the number of outstanding prunable blocks. 

On a library, which has started to increase 1TB per day, we have applied a 'run space reclamation' and have recovered 30TB of 65TB it had in size. Although this library has only 6TB in use. It is strange. 

We are seeing a large number of outstanding prunable blocks but performing the 'run space reclamation' does not remove them. 

 

 

 
 
​​​​

Does anyone know the reason?

Thanks

Good morning.  Are sparse files supported on the mount path?  Space Reclamation is for mount paths that do not support sparse files


Hey @Johana ,

 

Firstly, as @Orazan mentioned, space reclamation is really only if your storage does not support sparse and thus needs this process to reclaim some of that extra usage back.

 

Can’t tell from your first screenshot if that is the case as that information on the right hand side there.

 

The dedupe store is consuming 32TB on disk (total data size on disk) and this value will match the size on media value when viewing from mount path level. You however cannot use the Data Written figures at mount path level since these are only counting the currently in retention jobs and their respective data written figures, and does NOT include any deduplication baseline blocks that are being referenced by these current jobs in retention.

 

Lastly, regarding your pending prunable records, they are essentially the jobs (and job streams) that have met retention and purged by Data Aging process however have yet to have a chance to be processed by the DDB.

 

If you are running space reclamation, that DDB level pruning process will be paused to allow space reclamation to run so if you are constantly running space reclamation, then you are not giving normal day to day data aging and pruning process and time to work. Best course of action right now is to just kill space reclamation jobs for a few days and see if that pending prunable record count to start dropping. 

 

Thanks


Reply