Inotify Resources Exhausted

[Spritzup]

/usr/bin/tail: inotify resources exhausted
/usr/bin/tail: inotify cannot be used, reverting to polling

 

I'm often seeing this in my server log, which seems to coincide with Plex having a pausing/stuttering issue.  Has anybody else seen this, and/or have suggestions on how to fix it?

 

Thanks!

 

~Spritz

[Spritzup]

I'm running the latest version (6.3.2), and this is happening when Fix Common Problems is not running... I only have it scheduled to run once a week.  This error is showing in unraid's log file (though FCP also picked it up).  Further, I have none of the apps installed that are identified as causing this issue by FCP installed.

 

~Spritz

 

EDIT - Apparently I somehow deleted someone's reply to my post... apologies, it had looked like I had made a duplicate post, I simply meant to delete my duplicate post.

[Squid]

I'm running the latest version (6.3.2), and this is happening when Fix Common Problems is not running... I only have it scheduled to run once a week.  This error is showing in unraid's log file (though FCP also picked it up).  Further, I have none of the apps installed that are identified as causing this issue by FCP installed.
 
~Spritz
 
EDIT - Apparently I somehow deleted someone's reply to my post... apologies, it had looked like I had made a duplicate post, I simply meant to delete my duplicate post.

What is your list of plugins?

Sent from my SM-T560NU using Tapatalk

[Spritzup]

15 hours ago, Squid said:

What is your list of plugins?

Sent from my SM-T560NU using Tapatalk
 

Here you go -->

 

CA Auto Update Applications

CA Backup/Restore Appdata

CA Cleanup Appdata

Community Applications

Dynamix Cache Directories

Dynamix Local Master

Dynamix SSD Trim

Dynamix System Statistics

Dynamix WebGui

Fix Common Problems

IPMI Support

Nerd Tools

Open Files

Preclear Disks

Recycle Bin

Tips and Tweaks

Unassigned Devices

unBalance

unRaid Server OS

 

Thanks!

 

~Spritz

[Squid]

Is plex running as a docker app or is it the old semi official plugin

Sent from my LG-D852 using Tapatalk

[Spritzup]

It's running as a docker app, along with -->

 

Plexpy

Sonarr

NZBGet

Radarr

Couchpotato (soon to be gone)

Crashplan

Mylar

Headphones

NZBHydra

Playpy

Deluge

 

Thanks again for your help.

 

~Spritz

[Squid]

Strange...  Only the plugins should be able to max it all out.  And I don't think any of your listed plugins have the ability to do that.

 

What is the output of 

cat /proc/sys/fs/inotify/max_user_watches

You may have to create a script composed of 

echo someNumber > /proc/sys/fs/inotify/max_user_watches

and either have it run via the user_scripts plugin at array start or by modifying /boot/config/go

[Spritzup]

As requested -->

 

root@Brain:/# cat /proc/sys/fs/inotify/max_user_watches
524288
 

[Squid]

4 hours ago, Spritzup said:

As requested -->

 

root@Brain:/# cat /proc/sys/fs/inotify/max_user_watches
524288
 

Not sure what's using it, but dlandon was going to pump out an update to Tips&Tweaks plugin to allow you to increase it via a GUI.  Barring that, you'll have to add the command to your go file as I posted above.

[Spritzup]

It is a mystery for the ages.  I appreciate the help you provided, thank you.  Unless dlandon rolls out the update prior to Sunday night, I'll just modify my go script.

 

Again, thank you.

 

~Spritz

[Squid]

I think its available now

 

[Spritzup]

2 minutes ago, Squid said:

I think its available now

 

 

So it is.  Is their a recommended setting to bump it to for unRaid?  Or should I just look to see what people are doing in the various flavors of Linux?

 

~Spritz

[Squid]

There is no recommended setting for any flavour of linux.  Its all 100% dependent upon what you are monitoring.  Each folder and each file monitored each take up one watch.  Increasing the number of available watches does not consume any resources.  Each utilized watch takes up maybe 50-100 bytes of RAM.  So basically, you just set it to whatever works for you.

 

I guess the output of

ps -aux | grep inotify

would be helpful to determine what is consuming the watches and help you come up with a decent number (should've thought of that before)

[Spritzup]

Nothing appears to be found... is it possible that inotify has not restarted?

 

 

root@Brain:/# ps -aux | grep inotify
root     116116  0.0  0.0   9640  1832 pts/1    S+   13:19   0:00 grep inotify
root@Brain:/#
 

[Squid]

No idea.  inotify doesn't need to be started / restarted to make the change.  However, if whatever that's consuming the resource (and there are other things beyond inotify including but not limited to tail) may still error out until after a reboot.