ONTAP Discussions

Re: Broken? snapmirrors - how to delete them ?

Hi Hardy,

Have you already tried snapmirror release from aushafs2b?

Re: Broken? snapmirrors - how to delete them ?

No,  a reboot doesn't help.   I think the problem is on Backup Server side,  not filer.

But..  you are set the mirror from aushafs1b  to aushafs2b  .

What about the snapmirror status from  aushafs1b  ?   and  the  "snap list admin"   by aushafs1b  ??

Re: Broken? snapmirrors - how to delete them ?


aushafs2b:admin        aushafs2b:NDMP_LOCAL_005614  Source         -          Idle

aushafs2b:images       aushafs2b:NDMP_LOCAL_005617  Source         -          Idle

aushafs2b:domino2_vol  aushafs2b:nrst1a             Source         -          Idle

this should be source. please follow my steps in designation end.

can please share the below output from dst end.

snapmirror status

df -Vh nrst1a

lun show -l nrst1a

rdfile /etc/snapmirror.conf

lun show unmapped

one more Question this replication within filer or across the filer?

Re: Broken? snapmirrors - how to delete them ?

Yes I have - Thanks.  No good.

Re: Broken? snapmirrors - how to delete them ?

We are having the same issue as Max Hardy.  Our problem is related to the use of TSM to create NFS volume backups.  TSM is creating and starting the smtape jobs.  We currently have more than 220 orphaned smtape jobs on the controllers where the destination is an NDMP_local output like in Max Hardy's original post.  If there is no snapmirror destination to "snapmirror release", then it appears that there is no way to remove these orphans in the snapmirror status list.  They are not found in the snapmirror.conf file because they are supposed to be temporary associations and should go away when the TSM backup has completed.  We currently have a case open because one of our controller pairs is experiencing filer panics, crashing the controllers.  We have been told that root cause is smtape exhausting resources.  The crash and subsequent reboot did not remove the 220 orphaned entries from the snapmirror status list.   I am left believing that the orphaned NDMP_local snapmirrors that cannot be removed, released or otherwise deleted are the root cause of the filer crashes.  Perhaps we have all found a bug in the smtape engine.  We are still waiting for an answer from support on how to remove the orphans. 

Re: Broken? snapmirrors - how to delete them ?

Hi pburcin

I look forward to hearing Netapp's reply.  I havn't gone that far yet - I stopped using the smtape backups partly because of this issue.

Re: Broken? snapmirrors - how to delete them ?

The official response from support is that these "orphaned" smtape jobs that show up in the snapmirror status list can be removed by using "snapmirror release <volname>  <destination controller:NDMP_local_xxxxxx>".  You WILL receive an error that indicates there is "no release-able destination".  However, the orphaned job that you released DOES get removed from the snapmirror status list.  In my case, I have hundreds of these orphaned smtape jobs to remove across numerous controllers running Data ONTAP 8.1.2x.  So the manual removal process will be tedious.  Basically, that was my reply to support.  There's gotta be a better way to fix this permanently!!!  We should not have to watch this and remove the orphans manually on a regular basis.

So this is not a fix.  It is the way to remove the orphaned smtape jobs from the snapmirror status list only.  I received no admission of a bug found or a fix in process.

Re: Broken? snapmirrors - how to delete them ?

"snapmirror release <volname>  <destination controller:NDMP_local_xxxxxx>"    --->>> Worked a charm, cheers.

Review Banner
All Community Forums