trurl

Things eventually break. Memtest is easy enough to do even if just to eliminate that as a possibility.

Have you done memtest?

no

The array is not started in those diagnostics. Can you start the array and post new diagnostics?

Diagnostics would tell whether cache is being used to make dockers more efficient.

Click the arrow in the corner of that graphic link I posted.

Post your docker run command as explained at this very first link in the Docker FAQ:

So did you get appdata, domains, system moved back to cache?

Go to Tools - Diagnostics and attach the complete Diagnostics ZIP file to your NEXT post in this thread.

Do you know how to examine the contents of cache? Could be some of the files that are still on cache are files you want to keep.

Sounds like a plan, with the Unassigned Devices plugin.

Do you actually have a separate controller board? I assumed you had a motherboard with 2 different controller chips. Check your motherboard documentation for which ports are which.

OK, I mostly wanted to take a look at how your user shares are currently set to use cache, especially appdata, domains, system. Those shares should be all on pool and configured to stay on pool, but that is not what you currently have. If you have appdata, domains, system on the array your docker and VM performance will be impacted by slower parity updates, and they will keep array disks spinning since files will be open. We can work on getting those all on pool after you get your new pools configured. Mover will only move cache-yes shares from cache to the array, so to get everything moved from cache to the array you will have to set all your shares to cache-yes. Also, mover can't move open files. You will have to disable docker and VM services before appdata, domains, system can be moved. Not the individual dockers/VMs, but the services. Go to Settings - Docker and disable, same for Settings - VM Manager. And finally mover won't move duplicates, so there may be some additional cleanup to do after you get most of it moved to the array. Do you actually have any VMs? The multiple pools feature is great. I have appdata, domains, system on NVMe fast pool and another SSD pool for caching.

Do you have a link?

Why are you having unclean shutdowns? Do you have cache?

Looks like you rebooted so diagnostics aren't showing any of the problems in the log snippet you posted. All disks mounted and SMART for all OK. Possibly a connection issue with cache, which is likely the cause of the loop2 errors also since docker is on cache. Which controller is cache plugged in to?

You should always just Go to Tools - Diagnostics and attach the complete Diagnostics ZIP file to your NEXT post in this thread.

I have some questions about some of your assumptions, and the easiest way to get answers would be for you to post your diagnostics. Go to Tools - Diagnostics and attach the complete Diagnostics ZIP file to your NEXT post in this thread.

Even so ethtool says Link detected: no

The screenshot you gave here is not part of stock Unraid

Sounds like shrinking the array is what you want to do. Wiki already linked. Copy or move the data to other disks then follow the wiki. I personally recommend the rebuild parity method in that wiki, I think some have had problems with the clear script. Also would mention that I don't think there is any good reason to do what you want to do.

Unassigned Devices are not part of user shares. You could make the plugin backup a folder on an Unassigned Device, but then it wouldn't be set to backup appdata for other dockers that were in the appdata user share.

I assume you didn't bother to read the first post in this support thread then. Here I quote the relevant section:

Do you have any working dockers?