WSUS cleanup crashes VM

[shallax]

Tried three times so far with 3 VMs running the cryptsetup stuff in parallel. It's not reproduced the issue yet.

I might do this one more time or so as I'm adding quite a bit of mileage to my SSDs when I do this

 

[shallax]

I've done this 5 times now, going for a 6th. I have a feeling this has worked around the issue. If I understand correctly, io_uring is only marginally faster than using native, so if I'm okay with losing a tiny bit of performance, I guess this fixes it for me.

 

[showiproute]

Hello @fabian , in general my WSUS cleanup script now does not break my VM.
But I do have another issue where I do not know if this is a WIndows issue or linked somehow to this crash issue:

I robocopied a couple of full VM backups during the night and today morning I was unable to open my SMB share via network - on the VM itself I was able to open the corresponding disks.

Within the VM I have done a reboot and now it is stuck at the "Wird neu gestartet" screen.


Not sure if this is linked somehow but on the same physical host I ran PBS which does a daily backup at 0700 & 1900.
What I have seen is that the IO delay spiked up to ~ 80 but went down after the backup:

 

[showiproute]

Additional information: The reboot via the Windows GUI was not enought - also the Proxmox reset didn't do the job.
I had to stop the VM and reboot it.

 

[eider]

Some time has passed now and I can confirm that setting aio=native does solve the issue with VMs behaving as they were previously without any crashes. For VM 100 I have applied this only to OS drive while keeping second attached one at new defaults; it would seem that RAID5 backed array is either too slow to trigger the issue even under extensive I/O operations (such as heavy random read/write when Veeam B&R merges incremental backups), or the issue shows up only under combination with discard=on and/or ssd=1.