Issues with ndmpcopy | NetApp | Page 1

primal oasis Aug 10, 2024, 6:51 AM

#

Hi all,
trying to do an ndmpcopy (volume with millions of files) to a new flexgroup. The first one took quite some time.
Started a level 1 incremental, controller crashed after it finished (A400 running 9.12.1P10).

No I believe I am stuck, as a level one states the wrong start date and level two says previous was level 0.

Ndmpcopy: Starting copy [ 0 ] ...
Ndmpcopy: 10.106.97.151: Log: DUMP: Using inowalk incremental dump for Full Volume
Ndmpcopy: 10.106.97.151: Log: DUMP: Using snapshot_for_backup.1 snapshot
Ndmpcopy: 10.106.97.151: Log: DUMP: Date of this level 1 dump snapshot: Fri Aug 9 21:19:10 2024.
Ndmpcopy: 10.106.97.151: Log: DUMP: Date of last level 0 dump: Thu Jul 11 11:39:42 2024.
Ndmpcopy: 10.106.97.151: Log: DUMP: Dumping /c1122p1006/V_cifs_p_8142_024 to NDMP connection
Ndmpcopy: 10.106.97.151: Log: RESTORE: Error: This incremental dump does not follow the previous dump that you have restored.
Ndmpcopy: 10.106.97.151: Log: RESTORE: Error: This dump is based on Thu Jul 11 11:39:42 2024.
Ndmpcopy: 10.106.97.151: Log: RESTORE: Error: It should be based on Thu Aug 8 11:03:31 2024.
Ndmpcopy: 10.106.97.151: Log: DATA: Operation terminated: EVENT: INTERNAL ERROR (for V_cifs_p_8142_025)
Ndmpcopy: 10.106.97.151: Connection halted: HALT: Internal error!
Ndmpcopy: 10.106.97.151: Log: DUMP: Error: DUMP IS ABORTED
Ndmpcopy: Notify: Transfer failed
Ndmpcopy: Done

fi100300::> node run -node node-01 ndmpcopy -l 2 -sa ndmpuser:xxxxx -da ndmpuser:xxxxx 10.106.97.151:/c1122p1006/V_cifs_p_8142_024 10.106.97.151:/c1122p1006/V_cifs_p_8142_025
Ndmpcopy: Starting copy [ 1 ] ...
Ndmpcopy: Previous level was 0. Next valid level is 1
Ndmpcopy: 10.106.97.151: Log: Session identifier: 42149
Ndmpcopy: 10.106.97.151: Log: Session identifier: 42155
Ndmpcopy: Done

Any advice or help?

reef phoenix Aug 16, 2024, 5:26 PM

#

have you tried to do this with XCP ?? its quite good to do that and I have used xcp to copy multimillion files and had no issues.
but running XCP for a physical might be worth, as it needs more cores memory to work better and faster.

pastel python Aug 20, 2024, 2:32 PM

#

Or just upgrade volume to a flexgroup, and rebalance. That's what I did. OG vol was only 10TB tho.

versed crest Aug 21, 2024, 1:24 PM

#

pastel python Or just upgrade volume to a flexgroup, and rebalance. That's what I did. OG vol ...

Rebalance isn't recommended for large amounts of files. It works better when datasets are single files in the GB-TB range each.

primal oasis Aug 21, 2024, 2:41 PM

#

NetAppstrongly adviced against conversion and rebalance.

versed crest Aug 21, 2024, 2:46 PM

#

Conversion is fine if the single volume isn't too busy, it's the rebalance that is a problem.

pastel python Aug 21, 2024, 6:04 PM

#

Yeah, I had a low # of user ops on the volume I did.

versed crest Aug 22, 2024, 12:55 PM

#

pastel python Yeah, I had a low # of user ops on the volume I did.

Ok. Just keep in mind if you have perf problems on there, the conversion may be why.

primal oasis Aug 25, 2024, 6:44 PM

#

We use the volume for user profiles, average around 15k Io/sec

versed crest Aug 29, 2024, 2:41 PM

#

Yeah...bad idea...glad you didn't.

#Issues with ndmpcopy