I/O device error

[fsseremba]

Normally get the error below when backing to disk using CIFS;

Error: The request could not be performed because of an I/O device error.
Failed to read data from the file [\\158.113.12.3\3UN-PRD-CIFS-001\3UN-PRD-CIFS-LUN-005\3UN-PRD-CIFS-LUN-0052015-12-28T003042.vib].

Veeam Support gave me a work around of cloning the failed backup job and rerun the job which works but it created another backup folder which requires clean up later. But it keeps on re-occurring later.

I would like to know the cause of this issue and if there a permanent solution other than using the work around

[foggy]

Francis, this kind of errors points to the issue with your backup storage and that's why it tends to re-occur after some time after starting a new chain (that the cloned job effectively does). I recommend to take a closer look at the storage device/disks.

[Vitaliy S.]

What kind of storage is that? Can you mount it to any Windows/Linux box to avoid using CIFS?

[fsseremba]

Alexander,

I have already engaged HP Support to have a closer look at the storage device/disks but did not find any hardware issues and no failed disks.

Vitaliy,
I do have HP StoreOnce B6200 storage system and
I was using Linux box (NFS) mounts but the backup speeds were too slow, so I was advise to use CIFS with v8.0 of which the backup speeds are better compared to NFS.

[Vitaliy S.]

Since this is an I/O related issue, then job cloning as well as running active full backup can help you with having valid backups, but the only solution to this problem would be backup storage performance investigation. What were events/error messages logged on this device when your backup job failed? Is there any way to check that?

[fsseremba]

Vitaliy, they are no event/error messages logged on the device. I will try to engage the storage vendor and see if they can find something.

[Vitaliy S.]

The simplest experiment that could nail down the issue would be to switch to another repository type. If this issue doesn't occur any longer, then it seems like CIFS implementation might be the reason for this error message.

[fsseremba]

Vitaliy, i will try to switch to another repository type and see how it performances

[fsseremba]

The problem got resolved after upgrading to Veeam v9.