Problem when backing up to disk

Question

Hi,

we have a dedicatedstorage node with Networker 7.3.2 using a backup-to-disk device (mounted as drive s:\b2d and connected via fibre channel). When backing up data we get the following error:

Jan 9 13:32:00 < > root: [ID 702911 daemon.notice] NetWorker media: (emergency) Cannot write to s:\b2d\07\14\d3f0b333-00000006-50a3205d-45a3205d-12b10000-0a018832 - errno=22

The error occurs always after backing up about 188GB of data so it doesn't seem to be a connectivity issue. Can anyone explain what the error number means? There are no more informations in any Logfile on nw server and storage node.

PS: Have the networker 7.3.2 release notes been updated after jumbo release or are there no jumbo release notes? I would be interested in a list of fixed issues in jumbo before applying the patch to our environement. The latest version I found on powerlink is from Dec 4th.

Thanks.

ble1 · Answer

PS: Have the networker 7.3.2 release notes been
updated after jumbo release or are there no jumbo
release notes? I would be interested in a list of
fixed issues in jumbo before applying the patch to
our environement. The latest version I found on
powerlink is from Dec 4th.

Yes they were. If you check LGTpa list of fixed issues for 7.3.2 you will noticed that new LGTpas addressed by jumbo have been added on top of the list. For full list contact your support. I must say I do not support it in a way as it was done which I also stressed out to some EMC people recently.

dk3 · Answer

Well, as I posted in my first message the backup was broken after around 188 GB. This is much more than the 32 GB limit you are describing... Full backups of the same dedicated storage node to the same backup to disk device have been working for several month (full and incremental). I tried staging all savesets, unmounting the tapes, formatting the device and re-mounting the (virutal) device tape. The restarted backup operation is still running so I can't tell if the problem is solved now. I'll keep you up to date.

xvan_egmond · Answer

Hi, have you set the filesystem blocksize (or whatever windows calls it ) large enough to create the saveset? We have found that the standard blocksize will not be usable for networker as that stops at about 32GB filesize and usually saveset (especially full backups) are significantly larger than that.

hsinjwm · Answer

I got the same problem as you described.Do you find the solution?

ble1 · Answer

What disk do you use?

DeaconZ28-2015 · Answer

I also have experienced this problem with Backup to disk. It is almost like it hits a bad disk sector or something. I have seen it resolved only by rebooting. Unfortunately the server in question is the storage node and also the client, so I don't know if its a Backup to Disk problem or Disk Array problem.

I have had nothing but headaches since Backup to Disk. And in my case it writes at the same speed as tape even though its connected via Fibre.

DeaconZ28-2015 · Answer

Clariion CX300 with a dedicated RAID 3 LUN. Drives are 500GB ATA in a DAE2.

ble1 · Answer

I was never able to get good speed with ATA disks... they are fine for slow clients, but if you are looking for solution which would give you speeds above tape then forget ATA.

DeaconZ28-2015 · Answer

What about 300GB Fibre Channel drives? I have an array of those. Only problem is production directories reside currently on the FC's. Would require some downtime to switch'em.

ble1 · Answer

FC should give you better results, especially if you have multiple streams towards or from those devices.

DeaconZ28-2015 · Answer

I just got the same error again last night. I'm backing up a 660GB saveset and it aborts at 65GB each time.

Error Message:
Cannot write to G:\Disk0\91\85\ee661c74-00000006-a714d42f-005a0000-0a680426 - errno=22

ble1 · Answer

If it's each time can you make a test and do copy with explorer of that saveset to disk (eg. same area which is in this case G:\Disk0\91\85) and see if that breaks. The error itself stands for I/O error 'Invalid argument' which may indicate issue with disk or layer between application and disk.

DeaconZ28-2015 · Answer

I ran tests with large amounts of data to see if its bad disk. No problems.

Last night, a different saveset (500GB) failed too, but in a different directory on the device. I have even gone so far as to destroy the LUN and start over, still no luck.
04/05/07 23:37:24 nsrd: media emergency: Cannot write to G:\Disk0\57\67\95e4211f-00000006-d715b84e-4615b84e-002a0000-0a680426 - errno=22

What I don't understand is that when I back it up to another LUN (same array) that is attached to the NW Server itself, it has no problems (except speed is still too slow).

Message was edited by:
Zguy28

ble1 · Answer

What I don't understand is that when I back it up to
another LUN (same array) that is attached to the NW
Server itself, it has no problems (except speed is
still too slow).

Perhaps LUN is not backup friendly

On more serious note, do you have problem with always the same LUN?

lpamintuan · Answer

we are having the same problem too. I have a DXi7500 Express NAS. Applies all what is recommended in Quantum and EMC Best Practice Guide regarding NIC bonding NFS mount, RPC tuning and also created new shares and new devices, still having the same issues. The backups will work halfway, then will stop with the following error message:

'Media 38758 03/28/2012 07:28:00 PM nsrd media emergency: Cannot write to /mnt/DXdisk4/46/23/0a3acfbc-00000006-a173a3a4-4f73a3a4-00601b00-78d03621 - Stale NFS file handle"

Backup Environment:

Red Hat v5.7 x64 O/S

Networker v7.6.2.4 (back rev. from v7.6.3.1, hoping there was bug in this version, but still the same)

Any resolution/fix on this yet?

NetWorker

Problem when backing up to disk

Was this post helpful?