Corrupt files with no error

[lwarner]

Not sure where to start with this, so I though I would ask the community.
We have the latest B&R 12 and vSphere 7.0 U3.
Generally everything runs smooth as glass.

But, I have encountered two separate incidents of a single corrupt file inside a Windows Server VM, one yesterday, one about six months ago.

The latest one: I was using SureBackup to test one of our accounting application updates. So I made a simple backup of the server with the application; it was only 64 GB, and it was basically a file server. The backup job completed without error. I made my SureBackup app group and job, and fired it up. The application update failed, and after digging into it, I found a single file that was corrupt - checked with a file hash. I did a Guest Files restore on that same backup and verified that the file in question was corrupt.
Unsure about the cause, I just ran the same backup job again - zero changes - just hit Start. The single file now was ok, as tested by doing both Guest File recovery and also doing my SureBackup 'lab' again.
No errors were logged in Veeam anywhere.
No errors were logged anywhere else, including the Windows servers, VMware, disks, network, or hardware.
The data path from Veeam to Repository is all local, storage is all flash, everything is enterprise grade.

This is obviously very disconcerting, we can't have files just randomly becoming corrupt with absolutely no knowledge of the incidents.

I supposed it could be due to random chance; perhaps the 1 in 10^16 bit error lottery, cosmic rays, aliens....
Has anyone else encountered this? Could it just be bad luck? How can a "silent" error like this occur?

[Mildur]

Hi Leigh,

I recommend checking this with our customer support team. We can't speculate over a forum post about why those files were corrupted on one day and consistent on another. I suspect it could be related to the consistency level (crash consistent backup) of your backup, but I'll let our support team comment on that after analyzing the log files.

Please provide me with the case number or this topic may be deleted by a moderator.

Best,
Fabian

[lwarner]

Fabian,
I am not asking for tech support, I am simply asking if anyone else has encountered this issue. My audience is forum members, not tech support.
However, I have opened a support case, 07067378 at your request.
-Leigh

[lwarner]

Just to close the loop on this, support was unable to find any issue, nor did they have any similar cases or reports of similar behavior. Since nobody else replied to this thread, I guess I'll chalk it up as a 1 in 10E15 bit error. But if any readers do see anything like this, please chime in. Thanks.