memory should be ok, it memory has problem, windows guest should die like debian guest.
VM freezes irregularly
- Thread starter thimplicity
- Start date
Since I have been stable for 11 days (kernel 5.19.4-1), I wonder if installing the intel-microcode package (as I did) is playing some role as well...there are a lot of fixes in it related to this type of cpu...maybe who is still suffering crashes could give it a try...
Similar issue here, running Proxmox 7.2-9 off a Terramaster F2-423 with an Intel Celeron N5095.
Have one TrueNAS VM and another Ubuntu VM with Docker, the latter keeps randomly freezing.
Any pointers on how I can get the 5.19 kernel installed? I could only find the 5.15 package in the no-subscription repo.
Have one TrueNAS VM and another Ubuntu VM with Docker, the latter keeps randomly freezing.
Any pointers on how I can get the 5.19 kernel installed? I could only find the 5.15 package in the no-subscription repo.
A brief update on my NUC11ATKC4.
All the VMs are working fine so far, I just updated to 5.19.8 edge (thanks Fabian) and added a Debian VM (OMV).
I will report back in a few days, hoping it keeps running.
I just wanted to add to the chatter here. I have been using Proxmox VE on a micro PC I bought with an Intel N5095 CPU. Much like everyone else, Proxmox itself runs solid, but my VMs randomly freeze. I have 2 Debian 11 VMs (one running LibreNMS and one I'm experimenting with Kamailio), and a CentOS 7 VM running Asterisk. I find a random VM locks up at least once a week. It can be any of the VMs, and can be anywhere from 2 days to (a personal best) 13 days after boot, but eventually one always just locks up. I just installed the Edge kernel so I will also provide an update on how things go for me and if it improves things.
Hey all! Adding my experience here to help with debugging.
I'm also experiencing exactly the same issue this thread talks about. I've tried two boxes with the N5105.
Exhaustive details about my setup can be found in this thread:
https://forum.proxmox.com/threads/ubuntu-vm-periodically-stops-responding.114859/
I installed the edge kernel earlier today and left the house for a bit. I have an uptime robot monitoring the problematic VM. While I was gone everything seemed fine and the VM stayed up. When I got home I decided to do some work on the VM. I started some processes that used up all of the RAM of the VM and interestingly it seems to cause the freeze/crash. I've replicated this twice now.
Specifically I added some torrents to a torrent client and checking the files eats through the RAM allocated to the VM. It seems like this has triggered whatever issue causes the freeze to happen.
Proxmox continues to be responsive, and I have three cores allocated to the VM. After the freeze my CPU usage stays steady at 33% so it looks like the freeze causes one core to be pegged indefinitely. A single-threaded process maybe?
Anyway, just wanted to say I'm experiencing the same issue even after updating to the edge kernel. Interested to hear everyone else's experience on the new kernel.
I'm also experiencing exactly the same issue this thread talks about. I've tried two boxes with the N5105.
Exhaustive details about my setup can be found in this thread:
https://forum.proxmox.com/threads/ubuntu-vm-periodically-stops-responding.114859/
I installed the edge kernel earlier today and left the house for a bit. I have an uptime robot monitoring the problematic VM. While I was gone everything seemed fine and the VM stayed up. When I got home I decided to do some work on the VM. I started some processes that used up all of the RAM of the VM and interestingly it seems to cause the freeze/crash. I've replicated this twice now.
Specifically I added some torrents to a torrent client and checking the files eats through the RAM allocated to the VM. It seems like this has triggered whatever issue causes the freeze to happen.
Proxmox continues to be responsive, and I have three cores allocated to the VM. After the freeze my CPU usage stays steady at 33% so it looks like the freeze causes one core to be pegged indefinitely. A single-threaded process maybe?
Anyway, just wanted to say I'm experiencing the same issue even after updating to the edge kernel. Interested to hear everyone else's experience on the new kernel.
T
I ran CPU burn in on a VM for 24 hours. No issues.
The issue may not be with CPU but some kind of RAM allocation issue from hypervisor to VM.
That’s an interesting find.
I ran CPU burn in on a VM for 24 hours. No issues.
The issue may not be with CPU but some kind of RAM allocation issue from hypervisor to VM.
Thanks for the insight, I start simulating a mild load on the testing VMs.
stress-ng --cpu 0 -l 30 --io 2 --vm 2 --vm-bytes 80% -t 0 --metrics-briefI will report back anomalies.
Last edited:
Thanks, updated to 5.19.8 now. Originally held off going to the Edge kernel because it seemed to have run a bit more stable doing an apt upgrade, but after half a day it froze again. Will report back if it's working.
One symptom I did notice was, for the VM that was freezing, it would spike up to 50% CPU and stay there. Other VM running TrueNAS BSD works just fine.
I have replace my cluster of 3 x (Intel Apollo Lake N3450 8 Go ), with one N5105 with 32Go. So far i have no crash of the Hypervisor itself with stock kernel ( uptime 10 Hours), but i have perhaps one Vm ( container with bitwarden) with probably the same problem as bobbyrc.
The VM crashed, and the cpu inside the Vm was stuck at 100%. Force reboot needed.
I see tomorrow if i have another problem or not...
edit :
Well the same VM crash this night. i update everything in it this morning. I don't have hope... I see some people trouble with a pfsense VM. On my setting, it's running on another hypervisor
The VM crashed, and the cpu inside the Vm was stuck at 100%. Force reboot needed.
I see tomorrow if i have another problem or not...
edit :
Well the same VM crash this night. i update everything in it this morning. I don't have hope... I see some people trouble with a pfsense VM. On my setting, it's running on another hypervisor
Last edited:
A brief report of my NUC11ATKC4 running:
Now 3 days running the 7 VMs under stress-ng mild load, one issue with an Ubuntu VM which eventually filled up its memory and crashed (memory leak of the stress test? I am not sure), however the symptoms were totally different from the old freezes I had before installing the edge kernel.
Now I am testing one VM at time, starting with Fedora, under heavier load (for my little NUC).
stress-ng --cpu 0 -l 40 --io 2 --vm 2 --vm-bytes 40% -t 0 --metrics-briefI will report back in a few days.
qBittorrent v 4.4.0
Also just for context from my side, I have two VMs set up from the same template and both installed with the same packages on the guest OS. One does not have any docker containers running and has not yet frozen, the other with docker containers running (the one with qBittorrent) freezes.
I see, that's similar to what I have. 2 VMs, one is running TrueNAS core which never froze, but my Ubuntu VM with Docker and Deluge will freeze.
Knock on wood it hasn't froze since I updated to Edge though.
Knock on wood it hasn't froze since I updated to Edge though.
Setting the CPU type to host makes no difference. Unless fixes have been rolled into the 5.15.x kernel, you're unlikely to have your VMs run for a week (that was my record).