Unfortunately I had to learn this one the hard way – but in hopes that you don’t have to, here’s the gist of my situation. I’ve had Veeam running and backing up utilizing a Windows Server 2012 R2 box as a target for a while now! I’m getting some awesome dedup ratios utilizing both Veeam and Windows built in deduplication. That said, just last week I began to see this error occurring on one of my backups jobs.
“The requested operation could not be completed due to a file system limitation. Failed to write data to the file [path_to_backup_file]”
After a bit of Google Fu one can conclude that from here and here my problems were mainly due to the way my volumes were formatted – more specifically the support of for large file records. I, like most people went ahead and simply used the GUI to format and label all of my volumes. The problem being, utilizing the GUI also utilizes the default settings for the format operation, which in turn support only small size file records.
Therefore, after time, after some data is laid down to the disk, after dedup has been doing it’s thing for a while you might start to see the same error I was. The solution – well, unfortunately it is to reformat that volume with large size file records. The command to do so, pretty simple, listed below
Format <DRIVE> /NTFS /L
The key here being /L, which specifies the support for large size file records. Also, this process can take quite some time. From the time I ran the command to the time I had to get to this point in this blog I’m still sitting at 2%.
In my case, I removed all the backups from Veeam that lived on this disk. I’m comfortable that I also have replication running so I wasn’t worried about losing any data. If you are though, you could always simple disable dedup and copy your backups files to another location, run the format command and then copy them back. Honestly, I knew in my case it would be easier and quicker to simply just reseed those VMs.
Also, it’s not to say that Veeam won’t work without large size file records. I have 8 other volumes on this Veeam instance which have all been formatted with the same default settings and haven’t seen any issues whatsoever with them – just this one volume is throwing the error! For now, I plan on just leaving these other volumes the way they are. Am I just delaying the inevitable? Only time will tell!
Thanks for reading!