SQL VM regularly freezes.

[BiteMyElbow]

Greetings.

I have PVE 8.0.3 with several identical virtual machines, each of which has its own role. All VMs are Server 2019. One of them (Microsoft SQL Server) is constantly hanging up one time in two-three days. Black screen in Console and the only thing I can do is hard stop it.

I can't find anything useful in windows logs. Please advise me how to find the cause of the virtual machine hanging from Proxmox itself, which logs should I look at for diagnostics?

upd: i have virtio-win-guest-tools 0.1.229 in this VM

 

[Moayad]

Hi,

The Windows VM with VirtIO have a bug, see the below link[0]. To see if that related to the below bug, you can check the `Event Log` in Windows VM looking for `vioscsi`.


[0] https://forum.proxmox.com/threads/r...system-unresponsive.139160/page-3#post-691337

 

[crea]

We have same problem with server 2022. But we don`t have any message about vioscsi in event log.

It can be new update in windows?

 

[BiteMyElbow]

Hi,

The Windows VM with VirtIO have a bug, see the below link[0]. To see if that related to the below bug, you can check the `Event Log` in Windows VM looking for `vioscsi`.


[0] https://forum.proxmox.com/threads/r...system-unresponsive.139160/page-3#post-691337

I changed VIRTIO driver to 1.208 version but it does not resolve my problem. Is there any other options?

 

[Fantu]

Obvious checks like chkdsk and sfc I hope have already been done.
Make sure there is no trace of problems in the windows event logs (in most cases there are traces, perhaps of apparently unrelated problems, checking warning error and critical events), you are not using problematic driver versions (and maybe this thing is confirmed), there have been no significant changes in software and configurations (since the problem happens).

If it's not something listed above then it could be if windows starts having anomalies due to full ram (there are however few cases where it leaves no traces in the Windows events) or there are some cases of hardware problems on the host, but to not see them in other vm are quite rare cases, for example for using disks not in raid (or at least not a raid with redundancy, which can mitigate problems and is more likely to display errors) that have errors in parts used by that vm, but should be also without a host filesystem in between (vm disks on lvm) otherwise you would see errors in the host logs.

 

[BiteMyElbow]

The problem still exists, i do not know how to solve it.

From time to time, the VM just freezes, the Console screen is black and I can't do anything with it, just reboot. The only clue is that the hang-up occurs mainly on weekends (when some kind of maintenance is possible). In the computer resource monitor, the processor is loaded by more than 100%, although in general there are more than enough resources. Other VMs on this hypervisor are working correctly, there is only one problem, I suspect that there is some kind of incompatibility between the QEMU and MS SQL drivers.

Can you please tell me how to diagnose and solve the problem?

 

[Deepen Dhulla]

Check your Windows Server version and drivers. Plus, from your screenshots, it seems you are not even using the QEMU guest agent. Do you have it installed? Also, check your storage type – is it Ceph or external storage, etc.? This is to check out the IOPS.