Not sure where to post this to please move if need be.

 

This has happend twice to be and finally figured out what is causing it.

 

I have my Dockers/VMs on a unassigned ZFS pool of 2 SSDs - via the unRAID-ZFS plugin - works great..... but

 

If something goes wrong and the ZFS does not import and mount, unRAID sees the docker.img missing and does not attempt to start the dockers - which is good - but....

 

after stopping the array, and getting the mistake fixed that made the ZFS spool not import, and restarting the array - unRAID sees the docker.img and is happy - but....

 

in the previous attempt to start the docker service and seeing the docker.img missing, it is changing something in the docker system because upon restart and docker.img found - unRAID thinks there are no containers installed - so they have to all be re added via the templates that are saved on the flash drive.

 

fix:  when unRAID sees docker.img is missing on start and does not attempt to bring them up - it needs to stop there and not change the installed containers config - so when docker.img is back - your containers are back and will restart.

 

Hope I explained that so it makes sense....

 

Myk

 

 

 

I moved it here, to the Defects board, but if you can, would you conform it more like the format they have requested for defect reports (see Guidelines for Defect Reports)?

 

I'm sure they may want to add the obligatory disclaimer that they don't officially support ZFS, but the issue does not appear to be directly related to ZFS.  The issue appears to be that on array start, if docker.img is missing, then the Docker configuration is reset, and you would like to request an option that it not be reset.  That seems possibly problematic as there are probably cases where it SHOULD be reset if no docker.img (but I'm no expert here).