Out of memory errors

[Maginos]

Hi guys,

 

the Fix Common Problems plugin reported, that I have several Out of memory errors on my server. For resolving this issue I should post my diagnostics here in the forum.

 

I think I might know from where these errors came from, so here some background information:

- My server has 32 GB of RAM, around 7 GB are used in idle

 

- Yesterday I created a Ubuntu server vm and assigned first 6 GB to the vm (Initial and Max memory = 6 GB). On this vm I installed Nextcloud All In One with all the available containers. On this vm, I think I ran out of memory, when the NC did some scans (build up the photo library, etc.).

- Then I added 2 GB more Ram to the vm, so that it has a total of 8 GB.

 

- Suddenly I noticed, that my unraid server used over 90% of its Ram permanently and the CPU usage went up to around 100%, so I checked the Ram usage of the Ubuntu VM. This was around 5 GB. 

- Since CPU and memory usage didn't go down, I shut down the Ubuntu server vm and after some minutes, CPU and memory were at a normal level again. 

 

I also have a screenshot of the Memory utilization recorded by my Zabbix Server. The time in the graph, where the gap is, I restarted the server and before that, the zabbix agent on the unraid server was not available.

 

I hope, you can help me find out, what process caused this OOM error.

 

Thank you for your help!

Maginos

 

tilda-diagnostics-20231229-1031.zip

[Maginos]

Some further information concerning the Ubuntu vm:

Since yesterday, I had network connectivity errors coming from nowhere. Today I obverved, that the vm doesn't get an IP address from the DHCP server. I was able to resolve this by executing the command "dhclient enp1s0". After that, the vm got a valid IP address from the DHCP server. To make this persistent, I edited the file in the /etc/netplan/ folder and changed "enp11s0" to "enp1s0" (for whatever reason the interface name is wrong...). After a reboot, the vm got its IP from the DHCP.

 

While I was working today on this problem, the unraid host ran into high memory and cpu utilization again (screenshot below). I checked the cpu utilization of the vm again and it was around 50% (4 GB). I shut down the vm and after around 5 minutes, CPU and memory on the unraid host went back to normal. 

So I think it's very likely, that the OOM error is related with the ubuntu vm. 

[Maginos]

After deactivating Talk in the Mastercontainer, the OOM errors were gone, so this seemed to be the fix.