If an attemptis made to stop the array and there is disk in use then the Array will not be stopped and a looping message (every 5 seconds or so) will appear indicating the problem.

 

1) If the disk in question is not part of the Array why does it matter at all? Can't it be disregarded?

 

2) Why not state the mount point that is causing the issue; ideally allowing the user to unmount the offending drive(s) and/or kill any process associated with that disk that is keeping it mounted.

 

I ran into the first case because I have a scratch drive on a USB stick that I place the log files onto. It is not part of the array.

If an attemptis made to stop the array and there is disk in use then the Array will not be stopped and a looping message (every 5 seconds or so) will appear indicating the problem.

 

1) If the disk in question is not part of the Array why does it matter at all? Can't it be disregarded?

 

2) Why not state the mount point that is causing the issue; ideally allowing the user to unmount the offending drive(s) and/or kill any process associated with that disk that is keeping it mounted.

 

I ran into the first case because I have a scratch drive on a USB stick that I place the log files onto. It is not part of the array.

unRAID only attempts to un-mount the disk drives affiliated with the array.

A disk cannot be un-mounted if it is busy.

A disk is busy if ANY file on it is open.

A disk is busy if any process has it as the current working directory.

A disk will be busy if a process that is invoked resides on that disk.

A disk will be busy if there is another file-system mounted on it, or on any directory on it, or if a loop-back file-system exists mounted on it, or if a swapfile is active on it.

 

In other words, just logging in, and then changing directory to a disk, will make that disk busy.

  log in as root then type:

    cd /mnt/disk1

At that point, disk 1 is busy until you change directory off of it.

 

If instead, you start a program you added on disk1 as a background process..  that disk will be busy until that process is terminated, even if you log off with your login shell.

 

If you see unRAID attempting to un-mount a specific disk you can type

 

lsof | grep mnt

to see a list of open files.

 

You can type:

losetup -a

to see any loop back devices.

 

You can type

swapon -s

to see if there is a swap file active (typically, unRAID has no swap file unless you added one)

 

unRAID has never not shut down if other disks not involved in the array are mounted.

 

it would be better if all the add-ons were coded to stop prior to an array shutdown.  In the past, there was no easy way to know when the array was being stopped.  In the past year or two, there has, but not many people used it.  Recently, in the past few betas, there is actually a stopping array event your add-ons can use... alas, not many do.

 

I agree, it might make it a tiny bit easier to print the mount-point being un-mounted, but I seem to remember it is being printed in the syslog, at the beginning of the series of umount messages.

 

 

Thanks Joe.  That helps to explain it.

 

Now I need to understand why my scratch drive which is not part of the Array is appearing to cause the Array not to be able to stop, or was there some cockpit error where I had the 'cwd' part of the array; probably the latter.  Your post will help me isolate the issue tomorrow.

BubbaTools has a component that adds an alert to the GUI if any files are open, and will show them to you, and let you kill the process that has each one open.