I recently upgraded from 6.11.5 to 6.12.10. All of my VMs have been working flawlessly for the last  year or more. After the upgrade, most of my VMs would have random freezes requiring the VM to be force stopped. The situation occurs multiple times during a 24h period. i have updated the templates to the latest virtiofs drivers iso - 0.1.248-1 and the machine to i440fx-7.2. I have tried to rebuild the template from scratch using the templates in the new Unraid version and just attaching the vdisk to that template. No go.

 

I am at a loss at where to even begin to start looking for root cause of the issue. I like to stay current on the unraid server version and finally took the leap. Do I or should I downgrade back to 6.11.5?

I had the same issue, I too upgraded unraid but thought the issue was because I changed from AMD to Intel CPU. Figured it would be more likely that the hardware change was the cause as the VM has been operating flawlessly for years before that. I finally got it fixed by just isolating a core for the VM, if you haven't done so maybe give it a try?

9 hours ago, Mokkisjeva said:

solating a core for the VM

isolate the core or pin the core in the xml? I have cores pinned in the xml already. I'm reluctant to isolate cores as I've seen references to the fact it harms the vm cpu performance. I'm not passing through any graphics adapters. However, I am running pretty consistently 4-5 VMs that are constantly on, none are tremendous cpu or memory intensive. I've got 32G memory in the server and Xeon® CPU E31240 (4c/8t). I am in the process of upgrading the server hardware throughout the rest of this year, but I'd rather get all the nits and niggles resolved before subjecting everything to newer, better hardware (cpu, memory, MB).

 

I'm also attempting to basically re-make the VM, using the W11 template in 6.12.10 using all the latest virtio drivers and changing machine type to Q35 (as I've heard that machine type is better suited to TPM and W11.

 

Thanks for the suggestion! I'll put it into my research to helping to stop this nagging issue of random vm freezes requiring force stops.

At the time of changing hardware i upgraded Unraid too so when I read your post I was optimistic that it could just be an unraid issue, but I was still tempted to going back to the AMD system as I too run several VMs and are dependent on them. It's been an absolute shitshow the past few weeks trying to solve this problem and whatever information I could find on the internet has been tried. Then I came across a gaming VM optimization thread where it was recommended to isolate a core for the VM I got the idea to try it. I mean, it's either try that last thing or go back to the old AMD system.

I hardly know what I'm doing most of the time when it comes to server stuff so finding a solution to me doesn't mean I know what the correct solution actually is. So if anyone see this and can give an educated guess to what the initial problem is and if there's a solution outside isolating cores then great, until then that's what I'm going with. I do run a i9-14900K (was supposed to be in my main rig but expensive mistakes were made) which has enough cores I can "waste".

Edited June 23, 20242 yr by Mokkisjeva

I did isolate and pin 2 cores to the VMs. They still randomly freeze, requiring a Force Stop. The last bit of detail I've come up with is the whole macvlan vs ipvlan mess. I've been with Unraid for quite some time, so I guess I started on macvlan. Now the default is ipvlan. And with bridging enabled for dockers and VMs this seems to cause issues with some hardware. I have yet to change the dockers to ipvlan. I have an ecosystem of *arrs that can get finicky with network settings, so I'll have to study up on the potential pitfalls. I've come to learn, it's hardly ever as simple as 'flip the switch' and no unintended consequences.

 

To summarize: Still getting VM freezes requiring Force Stop with 2 cores isolated AND pinned to the VMs only. My next steps are to weigh downgrading Unraid back to 6.11.5 to see if the VMs are still freezing (they did not before the upgrade to 6.12.10) OR take the plunge and change docker engine to ipvlan, leaving VMs with bridging, and deal with the *arrs potential issues.

On this continuing saga that is becoming much more of a thorn than a nagging issue. The random freezes are now happening more frequently. I have narrowed it down to highly likely something to do with how I set up my Win11 VMs. I have one that is running and has been running without issue since before i upgraded. I did update the virtio drivers, guest-agent, and machine type. This VM setup used the old Win10 template, so I did not use OVMF-TPM, just OVMF.

 

On all recent VMs with Win11, I have tried with and without TPM, from the initial setup/install. Continued random, but now quite frequent machine freezing. I used the Win10 template in the new Unraid VM section, still the same result. I even went so far as copying the xml for the one Win11 machine that is working well, changing the necessary parameters (network, mac address, VM name, vdisk location, etc.) and wound up with frequent freezes again.

 

I have allocated up to 8G memory. I have pinned two cpu/threads. All continue to randomly and frequently freeze. I turned docker services off. Same issue. I tried changing network adapter to e-1000. No joy. I made a new Win10 machine with W10 template. NO ISSUES thus far, and as many times as the other W11 VMs freeze (and frequency), this is certainly the fall back position. Other than the single W11 VM I have that's been working since before the upgrade, this most definitely seems to be some error in my W11 VM setup. My motherboard is too old to have TPM (upgrade planned but not yet executed). To this point I've just been using the "modify the registry" to bypass hardware checks for TPM on install.

 

Wit's end on troubleshooting! Frustration compounds.