kern.maxfiles limit exceeded error

[Steven Gould]

I keep getting "kern.maxfiles limit exceeded by uid 80, please see tuning (7)" displayed on my freenas screen over and over.... i cant access the graphic interface thru http, or login thru putty. I also cant shut it down directly from the screen on the freenas unit under item 11. It accepts the command says its getting ready to shutdown but never does. I can open a shell but the message keeps appearing every 1/2 second I tried the shutdown command but same result as item 11. I didnt want to just turn it off since I have a failed harddrive in my ZFS. I did have plex running and i think that is the root of my problem but how can I kill that message from shell?

Running Freenas 9.2.1?
24 gb Ram
6x 3TB Hard Drives.

Don't make me pull the plug :(

Thanks,

 

[dlavigne]

You have run out of file descriptors. What is the current value shown when you type sysctl kern.maxfiles ?

 

[anodos]

I keep getting "kern.maxfiles limit exceeded by uid 80, please see tuning (7)" displayed on my freenas screen over and over.... i cant access the graphic interface thru http, or login thru putty. I also cant shut it down directly from the screen on the freenas unit under item 11. It accepts the command says its getting ready to shutdown but never does. I can open a shell but the message keeps appearing every 1/2 second I tried the shutdown command but same result as item 11. I didnt want to just turn it off since I have a failed harddrive in my ZFS. I did have plex running and i think that is the root of my problem but how can I kill that message from shell?

Running Freenas 9.2.1?
24 gb Ram
6x 3TB Hard Drives.

Don't make me pull the plug :(

Thanks,

In early versions of 9.2.1 (I think 9.2.1.3) there was a bug related to inotify and CIFS that caused it to launch smbd processes running at 100% like hamsters having babies. The solution was to add the auxiliary parameter to the CIFS config "kernel change notify = no" (which has been the default ever since).

I was sometimes able to properly shutdown by connecting a keyboard and mouse and accessing the CLI directly. Other times I had to smother the server with a pillow.

 

[Steven Gould]

dlavigne,
Everytime I type something in the cli it disappears before I can read it from the error message repeating..... I'm thinking its smother with pillow time :(.

 

[Ericloewe]

Other times I had to smother the server with a pillow.

You can say I'm no fun, but holding the ATX power button for a few seconds seems like a better option than starving the server of airflow until the system shuts down due to a CPU overheat.

 

[anodos]

dlavigne,
Everytime I type something in the cli it disappears before I can read it from the error message repeating..... I'm thinking its smother with pillow time :(.

Sometimes I was able to get around the problem by opening a new virtual terminal i.e. "ctrl alt f2", logging in, and immediately "poweroff". Alternatively, if you're running CIFS you can try 'pkill smbd', which might allow you to then issue a poweroff command.

 

[danb35]

holding the ATX power button for a few seconds

It's even possible that pressing it momentarily will cause the server to shut down properly. Seems worth a try, anyway.

 

[Steven Gould]

Well had to do a hard shutdown..... thanks for the help I now have a new problem only 3 of the 6 harddrives show up in the pool...... but I'm going to research abit and make a new post if I have to.