cancel
Showing results for 
Search instead for 
Did you mean: 

Subdisk move fails with undocumented error

oschistad
Level 3

We have a 30TB volume striped across two 15TB sub-disks. We need to move one of those subdisks to a different storage system for performance reasons.

Since there is no good way to move the data on the underlying storage we need to do this through VxVM.

So we presented a 15.2TB LUN from the destination storage system, and attempted a subdisk move from VxVM.

Sadly, it always stops after a few hours with the following rather useless error:

 ID:
   V-76-58645-7950
Received:
   8. mars 2012 kl 21.58 CET
Severity:
   Error
Source:
   vxvm
Classification:
   dynamic disk and volume class
Brief Description:
   Failed to move a subdisk that is part of volume \Device\HarddiskDmVolumes\s9-DeDupDG\s9-DeDupDATA.
Details:
   Failed to move subdisk from one disk to another disk.
Recommended Action:
   Make sure that you have enough unallocated space on the destination disk.
Other Properties:
   \Device\HarddiskDmVolumes\s9-DeDupDG\s9-DeDupDATA

In the voldlog we see that throughout the entire time the following errors are logged:

 3/8/2012 6:42:10 PM:5232: ERROR: SDMoveStripe: atomic_copy failed with error c000000d
3/8/2012 6:42:10 PM:5232: ERROR: SDMoveStripe: atomic_copy failed with error c000000d
3/8/2012 6:42:10 PM:5232: ERROR: SDMoveStripe: atomic_copy failed with error c000000d
3/8/2012 6:42:10 PM:5232: ERROR: SDMoveStripe: atomic_copy failed with error c000000d
3/8/2012 6:42:10 PM:5232: ERROR: SDMoveStripe: atomic_copy failed with error c000000d
3/8/2012 6:42:10 PM:5232: ERROR: SDMoveStripe: atomic_copy failed with error c000000d
3/8/2012 6:42:10 PM:5232: ERROR: SDMoveStripe: atomic_copy failed with error c000000d
3/8/2012 6:42:10 PM:5232: ERROR: SDMoveStripe: atomic_copy failed with error c000000d 

until finally it stops:

 3/8/2012 9:58:12 PM:5232: ERROR: SDMoveStripe: atomic_copy failed with error c000000d
3/8/2012 9:58:12 PM:5232: ERROR: SDMoveStripe: atomic_copy failed with error c000000d
3/8/2012 9:58:27 PM:5196: VOLD_REQUEST req_plex_dis started
3/8/2012 9:58:27 PM:5196: VOLD_REQUEST req_plex_dis SUCCESS: client 1, size 0
3/8/2012 9:58:28 PM:5196: ERROR: ior_config_change_start: dg not found for messaging or version 50 not supported for ff
3/8/2012 9:58:28 PM:5196: ERROR: ior_config_change_start: dg not found for messaging or version 50 not supported for ff
3/8/2012 9:58:28 PM:5196: ERROR: ior_config_change_start: dg not found for messaging or version 50 not supported for ff
3/8/2012 9:58:28 PM:5196: ERROR: ior_config_change_start: dg not found for messaging or version 50 not supported for ff 

We upgraded from 5.1SP1 to 6.0 to see if that resolved the issue, but no go. So we seem stuck with a subdisk we are unable to move even though we are using arguably the best volume manager in the business..

 

Help..?

2 REPLIES 2

mikebounds
Level 6
Partner Accredited

I believe Subdisk moves uses temporary storage when it moves data, athough I see you have made destination disk a bit bigger (by 0.2TB) and 200gb is a LOT of storage so you would think this would be enough, but perhaps VxVm allocates stroage by percentage of volume size and then < 2% is not much, so you could perhaps try adding a 1TB lun to diskgroup so there is plenty of temp storage and then this temp LUN can be removed later.

Mike

 

oschistad
Level 3

Hi Mike,

on the matter of free space: Yes I have had a suspicion as well that this may be relevant. Maybe the problem here is that the source volume contains almost no free space? This is a NetBackup MSDP pool ,and the container files have pretty much spread across the entire disk at this point - about 2TB of free space in the filesystem.

I guess the disk move needs to log changes (copy-on-write) while the move is happening, which I suspect means that the source volume needs some free space.

Could it be that dirty region logging is required in such a scenario?