Since upgrading to 7.1.2 I am having a Docker issue. Docker starts normally on boot and all the containers come up properly, but I've had 5-7 times in the last 2 weeks where the Unraid GUI says the Docker Service Failed to Start (hours or days later) and the containers have issue seeing each other (Radarr/Sonarr will not see Sab, etc). Rebooting solves this, but the message crops up again after a day or two and I need to reboot again. Docker has never actually failed to start on a reboot.. it's just after a period of time this appears.

I do have another issue with one of my cache drives (the Data pool with one device). It's a nvme drive that went unmountable about a week ago. I've tried reseating it without success, but eventually tried reformatting and using it again (it's Adata with no advanced RMA and a 18-35 day turnaround so I want to make sure I need to replace before going through that). After reformatting the drive worked again for a day or two, but has died again and will need to be replaced.

Appdata is on the main cache drive as is the docker container. the data pool was just a 4tb drive to use as cache for downloads, so I wouldn't expect it's death to cause the Docker issue, but the timing is suspicious so wanted to mention. I have been refreshing some of my movie content so Radarr and Sab have been using that drive, and Plex would be looking at it as well for the media updates, so don't know if that's part of the issue.

I am looking to remove the data cache completely from the system and redirect the couple of shares that reference it to the main cache drive. I realize this may fix the docker issue as well, but wanted to get more experienced eyes on the logs to help solve if the 2 issues are unrelated.

cydstorage-diagnostics-20250528-1333.zip

Temporarily uninstall the Connect plugin and retest, if it still happens, post the output from (after the issue starts):

ps -aux | grep dockerd

cat /var/run/dockerd.pid

ls /proc/<proc from above>

So, it took a while to get to a point where I'd solved my other issues (cache pool, nic), and did see this Docker issue pop up again just over a week ago. I removed the Connect plugin and haven't seen the issue since... though it's somewhat intermittent so while I am not 100% sure it's gone, I've been updating a bunch of media content, which seemed to be a trigger when this issue occured previously, and I don't seem to have the issue now while doing similar work, so am trusting the issue may be addressed by removing the plugin (for whatever reason).

Should I be re-adding the plugin and see if it occurs again?

As an aside... I did try the commands you posted above when Docker crashed (before removing Connect). I could find the dockerd.pid, but the ls /proc/proc# didn't return any result.

There are now other confirmed reports of the issue being caused by Connect, so it's most likely that.

Okay, so the recommendation is to not use Connect until a new version is released that should address this issue?

If you are affected, yes.