Proper method to determine what is preventing array from unmounting?

[kaiguy]

Hello!

 

Over the last few RC's (and now to 6.1.0), I have noticed that something is keeping my array from unmounting.  Specifically, according to the log, something is keeping the cache drive active:

 

Sep  1 08:34:12 titan logger: rmdir: failed to remove '/mnt/disk12': No such file or directory
Sep  1 08:34:12 titan emhttp: shcmd (1703): umount /mnt/disk13 |& logger
Sep  1 08:34:12 titan logger: umount: /mnt/disk13: not found
Sep  1 08:34:12 titan emhttp: shcmd (1704): rmdir /mnt/disk13 |& logger
Sep  1 08:34:12 titan logger: rmdir: failed to remove '/mnt/disk13': No such file or directory
Sep  1 08:34:12 titan emhttp: shcmd (1705): umount /mnt/cache |& logger
Sep  1 08:34:12 titan logger: umount: /mnt/cache: device is busy.
Sep  1 08:34:12 titan logger:         (In some cases useful info about processes that use
Sep  1 08:34:12 titan logger:          the device is found by lsof( or fuser(1))
Sep  1 08:34:12 titan emhttp: Retry unmounting disk share(s)...

I am not sure if I am doing this correctly, but I ran commands such as the following:

lsof /mnt/cache/*

lsof /mnt/cache

fuser /mnt/cache/*

fuser /mnt/cache

 

Which gave me nothing.  I further ran ps -A and reviewed the results line-by-line.  Nothing stood out to me.

 

Every time I can never figure out whats hanging my array, so I just end up issuing a powerdown command. 

 

I'm running 4 Dockers from linuxserver (which I wouldn't think would cause a problem), and the following plugins: Powerdown 2.18, Open-VM-Tools (yeah, I know it's not bare metal, but maybe someone can be so kind as to help me), cache_dirs (which I did not see in ps -A, tried to killall cache_dirs but nada), Dynamix active streams, and Community Applications.  Nothing in my 'go' file of merit.

 

If someone can give me some correct commands or areas to check when this happens in the future, that would be awesome. 

 

Thanks for any insight!

 

[kaiguy]

Even though I figured out my issue, I'd still love to know what are the best practices for identifying what is keeping a device busy and ultimately keeping the array from unmounting.

 

In my case, I stupidly created an SMB mountpoint under the cache drive.  I only figured it out after I specifically ran 'mount' to see what mountpoints were still active.  Once I unmounted it, the array was able to unmount.

 

I believe (and correct me if I'm wrong) the correct command to use lsof would be for other types of array unmounting issues would be:

 

lsof +D /mnt/

 

to see what open files are still within a directory under /mnt (again, not helpful in my case)... Any other commands that would be of use?

[ctviggen]

I know this is an old thread, but I had the same problem today.  A new installation, but still I could not shutdown through the web interface.  I had to put on my monitor and keyboard and use "powerdown".  I had stopped anything I thought might be accessing the server prior to this.

 

Is there anyway to determine why I can't unmount?

[ctviggen]

By the way, by "new installation", I mean really new.  New hardware, new everything.  Running basically nothing, the only apps are binhex-minidlna (used for my HD Homerun system, which records TV), and community applications.  That's it.

 

I took down all computers in the house, the HD homerun, and my HT computer.  By "took down", I mean unplugged if I had to or shutdown if it was a computer.  As far as I could tell, nothing was connected to the unraid server.

 

I did have one computer running to run the web interface, but that's all that was running. 

 

The new server is in its new location, and it's running a parity check, which will take 15 or so hours I think.  So, I can't attempt to shut it down until tomorrow.