(Another) Dockers dissapear after array restart

[Doowle]

I've done some research on this but I can't see anything to help me.

 

Each time my array restarts I loose all my dockers, if I add them back they work fine with persistant data / config.

 

I'm sure this is me, something I've done, I just can't see what.

 

Thanks you in advance.

 

Doowle

unhump-diagnostics-20230911-1837.zip

[JorgeB]

Sep  8 12:56:15 unhump root: Creating new image file: '/mnt/cache/system/docker/docker.img' size: 20G

 

A new docker image was created, suggesting one didn't exist before, share is configured to cache=yes, meaning the image would be move to the array, but only if it didn't already exist and the docker service was stopped, do you have any custom scrips running on array stop?

[Doowle]

9 minutes ago, JorgeB said:

Sep  8 12:56:15 unhump root: Creating new image file: '/mnt/cache/system/docker/docker.img' size: 20G

 

A new docker image was created, suggesting one didn't exist before, share is configured to cache=yes, meaning the image would be move to the array, but only if it didn't already exist and the docker service was stopped, do you have any custom scrips running on array stop?

No custom scripts.

 

If one exists in the array, would that stop it moving to the array?

[JorgeB]

Yes, post the output of:

 

find /mnt -name docker.img

 

[Doowle]

I did a stupid / bad thing.

 

I just removed the system and domain folders in /mnt/user, thinking they didn't need to be there (there was a docker.img) there.

 

Now it won't shut the array down!

Sep 11 19:39:47 unhump root: rmdir: failed to remove '/mnt/user': Device or resource busy 
Sep 11 19:39:47 unhump emhttpd: shcmd (3085): exit status: 1 
Sep 11 19:39:47 unhump emhttpd: shcmd (3087): /usr/local/sbin/update_cron 
Sep 11 19:39:47 unhump emhttpd: Retry unmounting user share(s)... 
Sep 11 19:39:52 unhump emhttpd: shcmd (3088): /usr/sbin/zfs unmount -a 
Sep 11 19:39:52 unhump emhttpd: shcmd (3089): umount /mnt/user 
Sep 11 19:39:52 unhump root: umount: /mnt/user: target is busy. 
Sep 11 19:39:52 unhump emhttpd: shcmd (3089): exit status: 32 
Sep 11 19:39:52 unhump emhttpd: shcmd (3090): rmdir /mnt/user

Been doing that for a while, I'm nipping out.  If it's not stopped by the time I am back, what to do..

 

D

[itimpi]

Doing that would have removed files docker was trying to use such as docker.img which is in the ‘system’ share.   I think you are going to have to force a shutdown by holding down the power button.

[Doowle]

Not my best decision...

 

Hard Powering now.

[Doowle]

FWIW, my mad and dumb approach seemed to work.  thank you for putting me on the right track.

 

J