VM filesystem regular fails while running backup (snapshot)

[A.M.]

Hello,

We have Proxmox 4.4 installed on VDS with ZFS and several VM running on it.
After running regular nightly VM-snapshot, sometimes VM filesystem becomes to mount in readonly (all VM running Debian ext4 with SCSI VirtIO and no cashing mode), some resent failures got system in non-bootable state with super-block errors.

Maybe there are some mistakes in configuration? but we have no idea what was that.
Thanks.

 

[e100]

What is the IO Wait on the Proxmox host during the backup?

If its high then you are starving the VM of disk IO causing the VM to think its disks are bad because they are not responding.

I've only started using ZFS a few months ago so I am far from an expert.
It seems that ZFS has its own internal IO scheduler separate from the kernel.
ZFS will allow a single process, like the backup, to suck up all of the IO and happily starve everything else.
At least thats what I observe.