Hi,
After maintenance a couple days ago, one of the Windows VM on our cluster has decided to stop completing back-ups.
Whilst the back-up job is running, the qemu-guest-agent for Windows seems to crash, which also results in the console not functioning and displaying an vnc proxy error.
This only seems to happen when there are more than 2 scsi disks attached to this VM, any of the 5 disks successfully back-up one their own. Once a third disk is attached to the VM, the back-up job freezes midway at ~40 / 50% completion. Displaying the following increase in read speed just before freezing the VM:
We have tried all the usual troubleshooting steps, such as:
> rebooting the machine, reinstalling virtio & qemu drivers, basically testing every single disk option there is. To no avail.
We've cloned all the disks over to another VM, which displays the exact same behavior with those disks.
In Windows they do report as error-free, successfully completing a chkdsk, dism, sfc scan.
Has anyone seen behavior like this before?
Attached is the output of pveversion -v
After maintenance a couple days ago, one of the Windows VM on our cluster has decided to stop completing back-ups.
Whilst the back-up job is running, the qemu-guest-agent for Windows seems to crash, which also results in the console not functioning and displaying an vnc proxy error.
This only seems to happen when there are more than 2 scsi disks attached to this VM, any of the 5 disks successfully back-up one their own. Once a third disk is attached to the VM, the back-up job freezes midway at ~40 / 50% completion. Displaying the following increase in read speed just before freezing the VM:
Code:
NFO: 40% (341.3 GiB of 853.0 GiB) in 42m 22s, read: 141.4 MiB/s, write: 138.2 MiB/s
INFO: 41% (349.8 GiB of 853.0 GiB) in 43m 28s, read: 132.1 MiB/s, write: 132.1 MiB/s
INFO: 42% (358.3 GiB of 853.0 GiB) in 43m 31s, read: 2.8 GiB/s, write: 222.7 MiB/s
INFO: 47% (400.9 GiB of 853.0 GiB) in 43m 34s, read: 14.2 GiB/s, write: 0 B/sWe have tried all the usual troubleshooting steps, such as:
> rebooting the machine, reinstalling virtio & qemu drivers, basically testing every single disk option there is. To no avail.
We've cloned all the disks over to another VM, which displays the exact same behavior with those disks.
In Windows they do report as error-free, successfully completing a chkdsk, dism, sfc scan.
Has anyone seen behavior like this before?
Attached is the output of pveversion -v