Server is very slow [SOLVED]

[Gico]

Server is very slow in the last couple of weeks. I don't feel it in SMB but very much in GUI actions.

Containers updates takes very long and often doesn't end at all.

Diagnostics also sometimes doesn't end (not downloading the log).

It's a bit better when the containers are shut down, but still one container can take half hour to update,

and another doesn't finish at all.

I restarted the server this morning and problem still exists.

Downgraded from 6.7.0-RC7 and problem still continues.

Attached diagnostics and screenshot of CPU usage and top command.

 

 

juno-diagnostics-20190426-1601.zip

Edited April 27, 2019 by Gico

[Frank1940]

This has been a problem in the past.  I can't remember if it has a single or (most likely) multiple causes.  I would suggest that you google   shfs unraid  and see if any of the 'solutions' might address your problem.  

[Gico]

One of the proposed solutions was to change the containers mappings from "/mnt/user/appdata" to "/mnt/cache/appdata".

I'm doing that now, takes a lot of time,

however I got help and it seems like that cache_dirs is running:

find /mnt/cache/appdata -noleaf

and with a plex container, this takes a very long time.

 

Edit: Excluded appdata from cache_dirs, rebooted, and all seems ok now.

 

Edited April 26, 2019 by Gico

[NewDisplayName]

as far as i know it should be mnt user appdata

but yeah, cache dirs wrong configured can make much problems.

[itimpi]

You definitely want cache_dirs configured so that it is not trying to handle folders that might have a very large number (millions?) of tiny files (as it not atypical for the Plex appdata folder).