Hello. I ask for help. Raised Proxmox Virtual Environment 6.3-3. Node on SSD. Two virtual machines on NVME disks. Each VM has its own video card and one for the node. The system works for two or three days and then just freezes. There are no hang-time events in kern.log. Only hard restart of the server helps. What to do? Where to dig?
video card forwarding VE 6.4-4
- Thread starter Pir0man
- Start date
Hey,
could you check if the RAM is filling up, this looks like some process might be leaking memory?
could you check if the RAM is filling up, this looks like some process might be leaking memory?
So the RAM usage reported in the Web UI does not match the one reported in the VM? What OS do you run in the VM?
The problem node has hard disks: 120 GB for a node and two 512 GB for two VMs. On the node where the memory is displayed normally, HDD 120 has both a node and a VM on it, but the VM is small, only 20 GB. A small node is ordinary without forwarding a video card.
Try https://docs.microsoft.com/en-us/sysinternals/downloads/procmon, this might help to figure out if something is using up resources. BTW Do you have Ballooning [1] enabled?
[1] https://pve.proxmox.com/wiki/Dynamic_Memory_Management#Ballooning
[1] https://pve.proxmox.com/wiki/Dynamic_Memory_Management#Ballooning
Just that I understand correctly, your PVE node crashed, not the VM, right? Does the node boot now?
Disabled backup of both VMs. The weekend node did not fall. Question: Why does the node have a memory usage percentage of 96% and this value is slightly floating? Can this value jump dramatically? How can VM backups affect memory usage? Maybe I should reduce the memory reserve for the VM?
