Backup from storage snapshot gives iSCSI errors in eventlog

[FrancWest]

Hi,

case #03937924.

When using storage integration with our Dell/EMC Unity array every time a backup from storage snapshot is finsihed the following error and warning is being recorded in the eventlog:

Unexpected provider failure. Restarting the service may fix the problem. Error code: 80070037@02000014

and

Disk <n> has been surprise removed.

Is there a way to get rid of these errors, since they clutter the eventlog and hide possible real storage issues when there actually is a problem with a production iSCSI lun.

Franc.

[PetrM]

Hello Franc,

I don't think there is an option to get rid of the specific event except event log filtering.

I'd recommend to continue working with our support team to find the solution.
Actually, you could contact Microsoft support team as well, may be they have some ideas about the event and how to suppress it.

Thanks!

[FrancWest]

Hi,

This specific event also occurs when there's a real issue with iSCSI luns, so that's why I ask why these events are occurring when backing up from storage snapshot. In case of backup from storage snapshot I can ignore these events, but that way it's really easy to miss actual severe problems.

The support team suggested me to use the forum and the case was closed.

I quote:

'The best way to proceed with this is opening a thread on Veeam Community Forums, it is the best since this is a product design level question. Moreover, I have already gone in'n'out to a storage integration specialist. He shared the position that the problem is in between an OS and a storage appliance.'

Franc.

[PetrM]

Thanks, I've reviewed the case.

Veeam unmounts storage snapshot from backup proxy once data processing is finished and I believe no changes are possible in this algorithm.
Moreover, I don't think it's technically possible to say to OS that some disk will be removed soon and the corresponding event must be suppressed. (at least, the research is required)

May be you could try to include ID of the event and use numbers of permanently mounted disks as keywords in the event log filter if you would like to be sure that severe issues will not be missed?
This filter allows you to ignore events only for storage snapshots being mounted/unmounted during backup.

I'm not sure that the test with iSCSI initiator repeats precisely unmount operation performed during backup because you did the test at the level of OS itself (as far as I understand).
In our case, Veeam instructs storage system to remove a storage snapshot and OS doesn't know about it, as result it generates the mentioned event.

I think exporting/mounting LUN snapshot on proxy and then unmounting this snapshot would be a good test in order to reproduce the issue "manually".
(I suppose you may ask for assistance of storage vendor support team)

Please let us know if you have some questions.

Thanks!