Out fo memory errors with 32 GB RAM

[L0rdRaiden]

I  am getting this errors every 10-15 days or so

 

Mar 6 04:30:05 Unraid root: Fix Common Problems: Error: Invalid folder cache contained within /mnt
Mar 6 04:30:08 Unraid root: Fix Common Problems: Error: Out Of Memory errors detected on your server

 

 

Other related settings

Disk Cache 'vm.dirty_background_ratio' (%): 2

Disk Cache 'vm.dirty_ratio' (%): 3

 

Any Idea how to fix it?

Edited March 6, 2022 by L0rdRaiden

[itimpi]

You are likely to get better informed feedback if you post your systems diagnostics zip file as without those we are just guessing.

 

The messages suggest you have something mapped to /mnt/cache (probably a docker container) and you do not have a pool called 'cache'.  This would mean that the /mnt/cache location is in RAM so writing to this that will fill up your RAM.

[L0rdRaiden]

19 minutes ago, itimpi said:

You are likely to get better informed feedback if you post your systems diagnostics zip file as without those we are just guessing.

 

The messages suggest you have something mapped to /mnt/cache (probably a docker container) and you do not have a pool called 'cache'.  This would mean that the /mnt/cache location is in RAM so writing to this that will fill up your RAM.

 

It might be that I have plex transcode on RAM bust last time plex was used, was like 10 hours before that out of memory error happened.

I have attached the diagnostic file

 

unraid-diagnostics-20220306-1050.zip

Edited March 6, 2022 by L0rdRaiden

[itimpi]

Your original post mentioned getting:

Mar 6 04:30:05 Unraid root: Fix Common Problems: Error: Invalid folder cache contained within /mnt

which was my big concern.  If this repeats then it is almost certainly a misconfigured docker container referencing that path.  I do not see this repeated recently so you might want to run Fix Common Problems again to see if that path still exists and if it does look at what that path contains to give you a clue as to what creates it.

[Squid]

Additionally, you do NOT want to use a host path of /tmp for Plex's transcoding  Use /tmp/Plex instead  Without the subfolder, Plex *may* wind up deleting files at any given point in /tmp which will adversely affect the OS

[L0rdRaiden]

41 minutes ago, itimpi said:

Your original post mentioned getting:

Mar 6 04:30:05 Unraid root: Fix Common Problems: Error: Invalid folder cache contained within /mnt

which was my big concern.  If this repeats then it is almost certainly a misconfigured docker container referencing that path.  I do not see this repeated recently so you might want to run Fix Common Problems again to see if that path still exists and if it does look at what that path contains to give you a clue as to what creates it.

 

I have re run fix common problems and still appears, I think is because this "Remote" folder that I think I created when I had rclone. But now I can't delete it.

 

 

And regarding the out of memory errors could be related with this settings? what are the recommended values if I have 32 GB of ram and usually half of it is no being used.

Disk Cache 'vm.dirty_background_ratio' (%): 2

Disk Cache 'vm.dirty_ratio' (%): 3

[L0rdRaiden]

Regarding OOM errors I am going to follow this advice and see how it goes

 

 

[Squid]

55 minutes ago, L0rdRaiden said:

I have re run fix common problems and still appears, I think is because this "Remote" folder that I think I created when I had rclone. But now I can't delete it.

 

No.  The message from FCP says invalid folder "cache" within /mnt

 

In your screenshot it shows "cache".  You do not have a pool named "cache", so something is creating it.  Look first at any docker apps (hit show more settings) for something that has a path mapping directly referencing /mnt/cache  Fix that and reboot and then see what FCP says again

[L0rdRaiden]

14 minutes ago, Squid said:

No.  The message from FCP says invalid folder "cache" within /mnt

 

In your screenshot it shows "cache".  You do not have a pool named "cache", so something is creating it.  Look first at any docker apps (hit show more settings) for something that has a path mapping directly referencing /mnt/cache  Fix that and reboot and then see what FCP says again

 

I think this will end up being a mistery, I rebooted because I change the vm.dirty values and now Fix common errors doesn't find the cache folder error

 

In addition I have reviewed the mounting points of all the dockers as you mentioned and not a single docker is referencing to /mnt/cache

 

 

Edited March 6, 2022 by L0rdRaiden

[L0rdRaiden]

On 3/6/2022 at 1:00 PM, Squid said:

No.  The message from FCP says invalid folder "cache" within /mnt

 

In your screenshot it shows "cache".  You do not have a pool named "cache", so something is creating it.  Look first at any docker apps (hit show more settings) for something that has a path mapping directly referencing /mnt/cache  Fix that and reboot and then see what FCP says again

 I am getting again similar errors

 

Mar 15 23:54:42 Unraid smbd[6677]: make_connection_snum: '/mnt/user/appdata' does not exist or permission denied when connecting to [appdata] Error was No such file or directory
Mar 15 23:54:42 Unraid smbd[6677]: [2022/03/15 23:54:42.398776, 0] ../../source3/smbd/service.c:841(make_connection_snum)
Mar 15 23:54:42 Unraid smbd[6677]: make_connection_snum: '/mnt/user/appdata' does not exist or permission denied when connecting to [appdata] Error was No such file or directory
Mar 15 23:54:42 Unraid smbd[6677]: [2022/03/15 23:54:42.399123, 0] ../../source3/smbd/service.c:841(make_connection_snum)
Mar 15 23:54:42 Unraid smbd[6677]: make_connection_snum: '/mnt/user/appdata' does not exist or permission denied when connecting to [appdata] Error was No such file or directory
Mar 15 23:54:42 Unraid smbd[6677]: [2022/03/15 23:54:42.399546, 0] ../../source3/smbd/service.c:841(make_connection_snum)
Mar 15 23:54:42 Unraid smbd[6677]: make_connection_snum: '/mnt/user/appdata' does not exist or permission denied when connecting to [appdata] Error was No such file or directory
Mar 15 23:54:42 Unraid smbd[6677]: [2022/03/15 23:54:42.400016, 0] ../../source3/smbd/service.c:841(make_connection_snum)
Mar 15 23:54:42 Unraid smbd[6677]: make_connection_snum: '/mnt/user/appdata' does not exist or permission denied when connecting to [appdata] Error was No such file or directory
Mar 15 23:54:42 Unraid smbd[6677]: [2022/03/15 23:54:42.400375, 0] ../../source3/smbd/service.c:841(make_connection_snum)
Mar 15 23:54:42 Unraid smbd[6677]: make_connection_snum: '/mnt/user/appdata' does not exist or permission denied when connecting to [appdata] Error was No such file or directory
Mar 15 23:54:42 Unraid smbd[6677]: [2022/03/15 23:54:42.400734, 0] ../../source3/smbd/service.c:841(make_connection_snum)
Mar 15 23:54:42 Unraid smbd[6677]: make_connection_snum: '/mnt/user/appdata' does not exist or permission denied when connecting to [appdata] Error was No such file or directory
Mar 15 23:54:42 Unraid smbd[6677]: [2022/03/15 23:54:42.401102, 0] ../../source3/smbd/service.c:841(make_connection_snum)
Mar 15 23:54:42 Unraid smbd[6677]: make_connection_snum: '/mnt/user/appdata' does not exist or permission denied when connecting to [appdata] Error was No such file or directory

unraid-diagnostics-20220316-0015.zip

I have no docker using appdata, in fact appdata doesn't exist, but this looks like related with samba, any ideas?