December 11, 201510 yr As part of my ongoing upgrades to my server, I have bought a small SSD for my Plex database and config files. I do not want to replace my cache drive with an SSD as this is a seed box as well, and the cache drive gets a lot of use. I installed the new SSD, formatted it XFS and mounted it with the Unassigned Devices plugin. I then copied my plex config folder on to the SSD and updated the volume mapping in my Plex docker to point at the new location. When I start up plex it doesn't see the config files at all and it starts up as a new install - its like the mapping totally fails. When I start a bash session in the container I can see files in the /config directory, but they are not the ones on my disk - and when I make a text file in that folder, it doesn't show up anywhere in the host file system - basically the symptoms of a failed mount. Any idea what would cause this? I tried copying the files with cp -p to maintain all the ownership and permissions - I don't know what else it could be! There's no help in the docker log file - it shows the same stuff no mater which directory I use. Thoughts?
December 11, 201510 yr Docker can't access mounts that were created after the Docker service was started. If you manually restart the Docker service (the whole service, not the individual containers) it should work.
December 11, 201510 yr Author OK! now we are on to something.. I'll try this when I get home. Is there a way to delay the docker start until after the disk has mounted?
December 11, 201510 yr I use my cache drive for Plex, so I don't have experience with this issue myself. But see the OP of the Unassigned Devices plugin: https://lime-technology.com/forum/index.php?topic=38635.msg359360#msg359360
December 12, 201510 yr I struggled with this issue for two weeks. Finally got it to work but unfortunately don't know how I did it! Symptoms were the same as you. After the SSD was mounted and a volume mapped to it, the folder appeared empty inside the docker container. Issuing: "docker exec plex mount" in terminal showed the mount point wasn't the drive (sdx1) but rather it was being mounted as /rootfs. I finally gave up after devoting ungodly amounts of time trying to make it work. Later I had need to temporarily mount an old spinner. Just for kicks I tried mapping the volume to it instead of the SSD and it worked flawlessly. That lead me to think something was up with the SSD. Here's where I'm not sure what got it working. I can only describe what I did. From the command line I manually re-partitioned and formatted the SSD to EXT4 then again into XFS. Updated all my unRaid plugins. Updated unRaid from 6.1.3 to 6.1.4. After the next boot the SSD was showing a generic label in unAssigned Devices. I clicked 'rescan disks' which made it show the expected drive name. After that the drive was able to work properly. There might have been some other things but that's all I remember doing specifically. I also remember poking around the unAssigned Devices config file but doubt anything I did in there helped matters. But if you need to manually adjust settings or clear things out that's where you can look.
December 21, 201510 yr Author I just restarted Docker after mounting the disk and the problem is solved.
Archived
This topic is now archived and is closed to further replies.