Proxmox VMs closing randomly.

[Script20]

Hello, I have proxmox. fully default, not done anything to it. But my VMs close without me actually doing it.

I have 4 VMs.
1. TrueNAS, not really doing anything.
2. Ubuntu, just a ubuntu server with a discord bot running, nothing else.
3. Ubuntu, minecraft server but the server has been offline so it's been idle.
4. Portainer, Homer, and tiny other things, hardly doing anything.

I turned it on maybe 24 hours ago, I haven't been on it apart from checking sensors command to make sure it's not overheating ( It's not )
I thought it was a SSD issue where it had 99% space taken, now I have a 500GB and using hardly any but still closing VMs randomly.
The Server I have on it has 2, 8 core CPUS so 32 threads, 64GB ram, 500GB Boot SSD, and 2 4TB HDDs for trueNAS.
On the summary tab on the server where it show CPU usage and the rest CPU can reach 2% max, Memory can reach 24GB MAX, and server load gets like 0.5 MAX
My VMs are not shutdown correctly as my MC server can be rolled back from last save, so it's not shut down correctly, I can't find any reason why.

I'm new to proxmox so I'm not sure why or can't find out why.

I have attached my syslog file just in case.
If I can't get it working I'm going to have to go to TrueNAS Scale for my server.

 

[Dunuin]

Is the syslog/journald showing any messages containing "oom"? Because usually VMs get killed when you run out of RAM.

 

[Script20]

Is the syslog/journald showing any messages containing "oom"? Because usually VMs get killed when you run out of RAM.

in the top main post, I stated it maxes to 24GB when I have 64GB

 

[Dunuin]

in the top main post, I stated it maxes to 24GB when I have 64GB

You could still run out of usable RAM if the RAM is fragmented and on. Wouldn't hurt to run a cat /var/log/syslog | grep oom to be sure.

 

[Script20]

You could still run out of usable RAM if the RAM is fragmented and on.

The server was off for a couple of week because of this issue, it was on for 24 hours and then VMs closed.
The ram can't be the issue, I also looked for OOM and in the syslog and it did nothing.

 

[HellrazorX]

In my case I had (so far) one VM that kept crashing within less than 5 minutes after each boot.

VM has allocation of 16 gb but I saw ''kill'' logs stating it went up to 17.8 to 18 before getting killed

I've been using this VM 40 hours a week, running unattended the remaining hours for more than 4 months and it never crashed even once before.
The PVE uptime at the moment the issue started was about 27 days.

This system has 96gb of ram running ZFS1 or 2 (not sure)

I rebooted the PVE and it seems to work beautifully again... so far.
It ''feels'' like it's a Mem fragmentation problem but it's a bit scary

Does, adding swap to memory, like suggested other posts, could actually solve any kind of headaches magically?