V9 backup errors with "Failed to flush file buffers"

[Tarook]

Hi,
I have a number of backup & backup copy jobs that are failing with the error,

Code: Select all

The device is not ready. Failed to flush file buffers.

The backup destination repository is a volume from an EMC VNX5300 SAN that is configured on the backup server as a windows volume of 16 Tb. The permissions on the volume are set correctly and the repository permissions are set also to allow everybody. I have submitted a ticket to support on this as Case # 01702188. Their initial analysis is from the job logs is that veeam is having issues writing to it.
Any one seen anything like this as all my searching is not turning up very much on it. Any help would be appreciated.

[veremin]

We've seen similar problems reported on these forums before. Though, no universal answer regarding the root cause was found. So, kindly, keep working with our support team on addressing that. Thanks.

[Tarook]

I have been trying to sort out this problem with Veeam support (case 01702188) and we are no closer to a resolution than I was back in mid Feb when I raised it with them. Surely someone out there has come across this issue. Any help or guidance would be appreciated.

[Gostev]

It does look like your backup storage is the issue. Have you tried using any other backup target for a test? Do you get the same error there?

[Tarook]

I realise this but I have to disagree with that diagnosis. In fact Veeam is able at the start of any backup job to create a folder in the repository and create a file in that folder with I think a .vbm file extension. It appears that when Veeam starts to write the backup file the whole thing goes to mud and we get the Failed to flush file buffers error.
If veeam can create a folder and then a file within that folder it isn't having trouble talking to the storage. I am also able through windows and a normal copy command I am able to create folders and copy large file (2Tb) to it without an y issues.

[Tarook]

I just ran another test job after configuring the VNX storage as a server 2012R2 storage space(I am getting desperate to find a solution which I hate to say is more than Veeam support is doing) and the job creates a folder for the job files, creates a .vbm file for the job name ( that's twice it's written to the disk) and then created a .vbk file of 4Kb (that's a third time to write to it) and that's as far as it goes. The job fails with the same error message and deletes the .vbk file (able to delete from it now) and leave the folder and a 1kb .vbm file behind. Obviously it's had enough now and takes it's bat & ball and goes home. Sorry a bit of Aussie humor or sarcasm there.

How about kicking this up the line as it seem to be too hard to solve for the guys that are looking at it so far?

[foggy]

Chris, you can ask for your case escalation to a higher support tier.

[Gostev]

Chris, based on the above I strongly suggest that you try a test backup to any other target (really, anything that is not backed by said VNX - even locally to the backup server itself). This would be the first test I personally would do to rule out storage specific issues here. Thanks!

[DaveWatkins]

Do you by chance have Windows dedup running on the server?

I've seen that error before when the dedup volume was too fragmented

[mkartono]

Has there been any resolution to this? I'm having the same issue but on a NAS directly attached by iSCSI.

[Vitaliy S.]

There was no solution in the OP case, everything started to work once the support case was opened.

[insearchof]

I run VMWare ESXi Host 6.0
Built many vm's
I installed Veeam Backup and Replicate v9 on a Windows 2008 Server which is a VM

I added an external USB Drive 1 TB to the VM Windows 2008 Server as E Drive

Selected one of the VM's and started a backup and selected the USB drive as the location

Code: Select all

Processing VM001 Error The system cannot find the file specified Failed to flush the file buffers

I saw this format E: /FS:NTFS /L

Trying that now but it is taking a long time to format the drive

Any thoughts

Thanks

Tom

[veremin]

The said error might illustrate underlying problem with a backup storage. Can you try to point jobs to a different location just to limit a scope of potential suspects?

Also, do you run default Windows Server 2008 or it has SP 2 installed? I'm wondering because basic 2008 is not supported.

Thanks.

[insearchof]

My Windows 2008 is Windows 2008 R2 Standard SP1

Yes I tried to run the same zip backup to my Local C Drive and received the same error

This error when I run the backup to my local C Drive

Code: Select all

8/4/2016 11:05:34 PM :: Processing TGCS002 Error: Current license or ESXi version prohibits execution of the requested operation. 

This error when I try to run the backup on the USB drive

Code: Select all

8/4/2016 8:40:27 PM :: Processing TGCS014 Error: The system cannot find the file specified.
Failed to flush file buffers. File: [E:\Veeam Backup\TGCS014\TGCS0142016-08-04T204025.vbk].

Thoughts

Thank you

[veremin]

8/4/2016 11:05:34 PM :: Processing TGCS002 Error: Current license or ESXi version prohibits execution of the requested operation.

This means you leverage Free ESXi which is not supported, you need a paid VMware license. Thanks.

[insearchof]

Hi

Forgot to reply

This issue can be locked as resolved I now have licensed versions of ESXi hosts running

Veeam now working

[veremin]

Correct, paid VMware edition doesn't have certain APIs blocked, therefore, backups become possible now. Thanks.