ONTAP Discussions

WALF and general block reclamation

darraghos
5,738 Views

Guys, whilst reading these forums, I see a lot of deference to storage reclamation when snapshots are deleted from aggregates/volumes or flexclones are deleted (not LUN space reclamation where files are deleted from an OS as that I understand). It seems like that when a vol snapshot,aggregate snapshot or flexclone is deleted a background process runs to:

  1. Mark the blocks locked by the snapshot as reusable ?
  2. Zero or clean those blocks for reuse ?

Is there any info on this process and how to monitor? Maybe the WAFL Scan status command?

2 REPLIES 2

MCARDINA221975
5,654 Views

You are correct on both counts.

Keep in mind that these commands should be run from a diag promt <priv set diag>

WAFL Scan status <volume name> will show you the status of block reclimation on that particular volume.

Example:

Volume Test_Volume:

Scan id                   Type of scan                         progress

     313              active bitmap rearrangement     fbn 235 of 314 w/ max_chain_len 3

Active bitmap rearrangement is always running on the volume.  If a reclaim is kicked off...either by the system or by the user a WAFL Scan status of the volume will look like this:

Example:

Volume Test_Volume:

Scan id                   Type of scan     progress

     244    active bitmap rearrangement     fbn 16 of 314 w/ max_chain_len 3

   43244    container block reclamation     block 0 of 315 (fbn 0)

You could run the status command and watch the reclaimed blocks incriment until it completes.

With regards to the process itself, I do not know that we, as admins, have visibility into the exact process itself.

You could try to run the ps command and locate the process ID.  That may show you how much CPU and stack % it is using:

netapp*> ps

Process statistics over 82302.705 seconds...

   ID      State      Domain      %CPU      StackUsed      %StackUsed      Name

    1        RR              k             0%              0                      100%           dummy_1

In the above example, PID 1 is a dummy process running in the Kahuna domain.

Hope that helps?

darraghos
5,654 Views

Thanks Michael, much appreciated.

Public