Tower bricking after a few days runtime.

[Jamfin]

Hi people!

 

Fairly new to unraid so forgive ingorance or being clueless. I've had my server setup now for the best part of a month; all of which has been perfect and no issues at all.

 

However; since upgrading to 6.12.8 (could be completely unrelated) my server seems to just completely brick after a few days of runtime. Initially I noticed that my server was running for 9 days until I went into my Homarr container and went to save something then the server completely bricked and I had to power cycle it, so I had assumed it was something to do with Homarr.

 

However about an hour ago, it happened again (2 days after the first one) when I was out the house so it couldn't be related. I do have a few plugins, containers, a windows VM (which literally just runs an idle game on it for when i'm out the house and don't want my computer on also) which is low usage and a rclone mount.

 

I've attached my diagnosis for those that know what to look for but unfortunately i'm kind of in the blue. I've not had syslog logging to my flash so I've turned that on to be able to catch the culprit if it happens again.

 

I'd appreciate any insight into the matter, i'm kind of a bit blind and this server now runs quite a few public services so i'd like to be able to get to the bottom of it before I do any damage to the drives by having to power cycle the machine.

home-diagnostics-20240308-1301.zip

[itimpi]

Your syslog contains:

Mar  8 12:48:02 Home root: Fix Common Problems: Error: Invalid folder disk1 contained within /mnt

but you do not seem to have a disk1.    This means that you have something (probably a docker container) configured to write to disk1 which will be in RAM so could gradually fill RAM up and result in the server crashing.

[Jamfin]

14 minutes ago, itimpi said:

Your syslog contains:

Mar  8 12:48:02 Home root: Fix Common Problems: Error: Invalid folder disk1 contained within /mnt

but you do not seem to have a disk1.    This means that you have something (probably a docker container) configured to write to disk1 which will be in RAM so could gradually fill RAM up and result in the server crashing.

Silly me, forgot to fix that a while back... I had old SSD's in my array whilst I waited for my NAS drive so disk1 doesn't exist anymore correct.

 

Although; i've found the offending container and done a rescan of common problems but it still remains so it must be elsewhere on the system.

 

Do you have any recommendations of where I can check? All my docker containers no longer contain disk1 checks.

[JonathanM]

2 minutes ago, LubricantJam said:

Although; i've found the offending container and done a rescan of common problems but it still remains

Was that after a reboot?

[Jamfin]

Just now, JonathanM said:

Was that after a reboot?

No, it was just after restarting and rebuilding the offending container. Should I be restarting then? Is that /mnt/disk1 cached in RAM now?

Edited March 8 by LubricantJam

[itimpi]

Just now, LubricantJam said:

No, it was just after restarting and rebuilding the offending container. Should I be restarting then? Is that `/mnt/disk1` cached in RAM now?

You need to restart to clear this.   

[Jamfin]

Just now, itimpi said:

You need to restart to clear this.   

Ah! Apologies for the misunderstanding, i'll give the system a reboot now. Thank you for the clarification.