Data Domain Cleaning

Hi,

Are you talking about DD VTL - so it is the tapes that are recyclable?

If so, Networker has only marked them as expired - the DD will not know this and you could still technically restore the data from the tapes.

Not until the tape is relabelled will the data be essentially removed from the DD and then be eligible for DD cleaning.

Rewind followed by write is how the DD will know to dump the data.

Apologies if you're talking about AFTD backups.

Regards,

Jonathan

The DD cleaning by default runs every Tuesday @ 6am and should reclaim that space once completed.

Hi Kenneth,

Thanks for your reply.

The cleaning runs as scheduled but does not free up as much space as I would expect.

Even though there is a couple of TB of save sets which are recyclable there is only about 100GB of space cleanable according to the show space command on the DD.

If using tapes, data get's cleaned when removed from tapes - that means even if you tape is eligible for recycling, data is still there and you can, for example, get it back with scanner command.  To remove it, you would need to relabel tape.  I believe recent version of NW even had a such option to automate that (since I no longer use tapes I wouldn't know exactly where such setting is).  Once it is removed from tape, DD will see is as something that can be removed and will update its own stats and data will be removed upon next DD scheduled (or manual) cleaning.  It is only with disk devices (eg. AFTD, DDBoost) that data is removed as soon as it is expired in NW and thus immediately visible on DD as well.

Hi Jbrooksuk,

Thank you for your reply also.

I am using DDBoost devices.

The DD should free up the space during the cleaning cycle, as stated by Kenneth, but it looks like this is not happening (or at least as expected).

The nsrim command has been run manually this morning by DD still reports only 100GB cleanable. I'm wondering when this value gets updated? Should this change when nsrim has been run or does it talk a while for the DD to realize that save sets have expired?

In that case do following:

mminfo -avot -q scan -r volume

I had couple of months ago strange situation when server was stopped in a way that storage nodes were disabled by operator and that happened during nsrim cleanup.  As result, volume got scan flag set and that also keeps away nsrim from doing cleanup on such devices.  It is just an idea, nothing more.

You can also do something like nsr_render_log -S '2 days ago' /nsr/res/daemon.raw | grep Recove  to see how much has been recovered from devices.  Not that when it says 100TB that doesn't mean 100TB free on DD; NW uses native sizes and on DD you see de-duped one.  Ratio, very very roughly, can be calculated over what is your average de-dupe ratio.

Hi Hrvoje, thanks for your reply,

I am using DDBoost devices.

NetWorker shows the save sets as recyclable, I have run nsrim on networker server.

The save sets still show as recyclable in NetWorker and the cleanable capacity still shows as only 100GB on Data Domain.