user shares keep disappearing until unraid reboot

[papnikol]

  

Hi everyone,

I am using unraid v6.11.3. Suddenly (without me having made any significant changes), the user shares have started disappearing and they do not reappear until I reboot. Initially it happened once in a while but it keeps happening more and more often (3 times today). Here is what I tried (having also searched the forum for suggestions):

1. Tried stopping and restarting the array - didn't work

2. Changed Settings > Global Share Settings > Tunable (support Hard Links) to "No" - didn't work

3. I have some Docker Containers but they are not running so I dont think they could be the problem

4. Updated all apps - didn't work

5. Checked the XFS filesystem in all disks

 

EDIT: I just noticed an error in the logs:

Quote

kernel: Out of memory: Killed process 8995 (shfs) total-vm:569000kB, anon-rss:269704kB, file-rss:0kB, shmem-rss:524kB, UID:0 pgtables:588kB oom_score_adj:0

 

Could this be the problem? My server has 6GB and I am running cache_dirs (folder caching). I also have a LOT of files in the server. I always had but they have been slowly increasing. Could I have reached a limit where the memory is somehow filled?

 

Any help would be very welcome as my server has been rendered unusable.

Edited May 23, 2023 by papnikol

[itimpi]

The ‘shfs’ process is what supports User Shares so that getting killed is why the User Shares disappear.  The question is why?
 

You are likely to get better informed feedback if you attach your system’s diagnostics zip file to your next post in this thread taken after the problem occurs.

 

Have you tried booting in Safe Mode with Docker disabled to see if the problem still occurs?

[papnikol]

31 minutes ago, itimpi said:

The ‘shfs’ process is what supports User Shares so that getting killed is why the User Shares disappear.  The question is why?
 

You are likely to get better informed feedback if you attach your system’s diagnostics zip file to your next post in this thread taken after the problem occurs.

Thanks a lot for the input.

I will boot normally and when the problem occurs again (I am guessing soon), I will post the diagnostics.

 

31 minutes ago, itimpi said:

Have you tried booting in Safe Mode with Docker disabled to see if the problem still occurs?

I might try that too, afterwards. A plugin might be the problem but I seriously doubt it has something to do with Docker since I am nor running automatically any of the installed plugins.

[JonathanM]

11 hours ago, papnikol said:

My server has 6GB and I am running cache_dirs (folder caching).

Turn that off and see if the situation improves.

cache dirs should be set to operate on the bare minimum of folders, it can easily become counterproductive and keep the array needlessly spinning if your RAM can't contain the full list.

[papnikol]

So I tried various proposed stuff (and was away for a bit, sorry for taking too many days to get back to you after your prompt inputs).

 

What I found out is by disabling docker, the problem was fixed. Additionally, I noticed in the dashboard that docker was using quite a lot of memory, when enabled. More particularly 67% of my 6GBs, which is almost 4GBs. I noticed that in my other unraid machine docker was also using ~4GB. I do not understand why that would be since all my containers were stopped. Maybe the memory is reserved by docker?