Server goes "unresponsive" every day

Seeing a lot of OOM errors, this is generally not lack of RAM but RAM fragmentation, you can try limiting your containers/VMs or using for example the swap file plugin.

5 minutes ago, JorgeB said:

Seeing a lot of OOM errors, this is generally not lack of RAM but RAM fragmentation, you can try limiting your containers/VMs or using for example the swap file plugin.

Where do you see this in the logging? And not sure what the cause for this would be. VMs are limited, the dockers aren't but as far as I can see aren't consuming a lot of RAM.

e.g.:

 

Jun  7 10:11:01 MassiveDump kernel: syncthing invoked oom-killer: gfp_mask=0x140cca(GFP_HIGHUSER_MOVABLE|__GFP_COMP), order=0, oom_score_adj=0

 

Are there (m)any downsides to downgrading unraid?
The server used to run for several months and only became unstable after upgrading.

Usually no, but see the release notes about docker.

On 7/6/2023 at 6:46 PM, JorgeB said:

e.g.:

 

Jun  7 10:11:01 MassiveDump kernel: syncthing invoked oom-killer: gfp_mask=0x140cca(GFP_HIGHUSER_MOVABLE|__GFP_COMP), order=0, oom_score_adj=0

 

On one of the latest crashes I was still logged on to ssh and noticed high cpu usage from "kswapd0" which kind of matches your indication of that the system was swapping memory too much.

Still there's no obvious reason for it to do so. The memory isn't over-allocated at the moment.

 

Since my last post, I downgraded back to 6.11.5 with minor issues. Just as the release note states you have to force update all of your dockers and manually start them. Besides that no specials.

 

The system has been running great ever since. Indicating to me that there's ether something wrong with my combination of plugins/ dockers and 6.12.2 or the software i.c.m. with my hardware. 

I'm leaning towards ruling the last one out since there's a lot of complaint about the stability of this version.