shfs killed due to OOM

[megagator]

Since upgrading to 6.4.1, I've had a frequent issue with shfs being killed due to low memory. This causes all sorts of issues, namely the loss of /mnt/user/:

/bin/ls: cannot access '/mnt/user/': Transport endpoint is not connected

(I also have a docker container with a known memory leak, but I capped it's usage with Docker so it doesn't exceed 3GB)

 

The system I'm on has 16GB of RAM, but unraid has always seemed pretty light on it, rarely consuming a tenth of that in the past.

 

I would love to get to the bottom of why shfs is taking up so much memory, but in the mean time, is there a simple way to get the shares working again without a reboot?

syslog

[Frank1940]

You can try this as it has eliminated the problem for a lot of folks:

 

        https://lime-technology.com/forums/topic/58855-regular-out-of-memory-problems/#comment-577424

 

 

I am not sure about the restarting of shfs  but I, personally would be reboot.  Whenever I get an error that cause me an issue (which has been very seldom with unRAID but I only run a basic NAS), I would reboot to guarantee that the system is working properly again. 

[megagator]

Ah yes, great idea, but I've had that in place for a while. My current settings are at:

Disk Cache 'vm.dirty_background_ratio' (%): 1%
Disk Cache 'vm.dirty_ratio' (%): 2%

Resolved an issue for me in the past on 6.3.5, don't recall what exactly it was though...

[Frank1940]

My next suggestion would be to post your diagnostics file in a new post.   Tools   >>>    Diagnostics

[Frank1940]

You also might want to look at these two threads:

 

        https://lime-technology.com/forums/topic/65720-shfs-memory-leak/

and

        https://lime-technology.com/forums/topic/69047-640-641-shfs-taking-a-lot-of-memory-~8gb-now/

 

Normally, I don't recommend 'hijacking' another user's thread but, at least, you should be aware of them and monitor them to see if anything fits your situation.