I'm on Unraid 6.9.2.  A failing Cache SSD had me worried (a couple of unrecoverable errors showing up in SMART), but I didn't feel like I *needed* it at the moment, so I just wanted to switch back to running my array without cache for now. 

 

I was following SpaceInvader One's instructions for replacing a cache drive (https://youtu.be/ij8AOEF1pTU), with the caveat that after migrating everything off the cache drive, I was just going to get it running again without cache for now and worry about replacing that later.  Flipped all the shares with data on the cache to Yes, run Mover, everything relocated back to the array just fine.  Deassigned the cache SSD, then downed Unraid gracefully and uninstalled the suspect drive.

 

Powering back up, I wanted to get Docker up and running again.  Deleted the Docker vDisk as recommended, rebooted again, then went to re-enable Docker recreating the vDisk, but after spinning awhile, nothing happened.  Flipped over to the Docker tab to see the "Docker Service failed to start." error message.   Checking the log, the one notable error I'm seeing is this one:  

 

May  4 15:25:21 TacoUnraid root: starting dockerd ...
May  4 15:25:42 TacoUnraid nginx: 2021/05/04 15:25:42 [error] 3144#3144: *590 upstream timed out (110: Connection timed out) while reading upstream, client: 10.10.10.99, server: , request: "POST /update.php HTTP/1.1", upstream: "fastcgi://unix:/var/run/php5-fpm.sock:", host: "10.10.10.11", referrer: "http://10.10.10.11/Settings/DockerSettings"
May  4 15:25:50 TacoUnraid emhttpd: shcmd (95): umount /var/lib/docker

 

Any idea what I'm doing wrong?  Diagnostics attached here:  

 

Thanks.

 

 

tacounraid-diagnostics-20210504-1745.zip

No obvious reason why this happened.  What happens if you switch the image to being a xfs image instead of a btrfs image in settings - docker?

Unfortunately both image types fail in exactly the same way.   I can't see why Docker won't start.  It's not giving me very comprehensive log output.  

 

Is there some way that I can enable some better log output? 

Also, I actually hadn't gotten very far with my Docker and VM setup yet...  so wasn't that attached to my now-uncooperative Unraid config.   There's a part of me that was just wanting to try resetting Unraid to a clean image and starting over...  however there's a lot of data (~10TB) that I already migrated over to my array disks.

 

If I were to try starting clean instead, is it possible to boot a fresh deploy of Unraid, but then re-import the array's data and parity disks intact without needing to copy the data off and back on somehow?   

I'm not an expert at this, but it's my understanding that every drive designation is remembered by unraid (I have mine written down). Regardless of what you change (understanding you might have parity, disk1, etc), you can (un)rebuild if you know every drive that belongs in its place.

 

1 hour ago, Yonder said:

is it possible to boot a fresh deploy of Unraid, but then re-import the array's data and parity disks intact

 

Again, no expert here, but there's no reason to "re-import" as long as you know where/what each disk was before the "(re)deploy". In short, there's no way to "re-import" to different drives, but/and/or, the original drives are still very much ready to be deployed the same way they were before.

 

I hope this helps.

 

6.

 

 

If you want to restart Unraid back to defaults, you just delete the contents of /config on the flash drive (but don't delete the .key file or the super.dat file)

 

 

Thanks so much for the information @Squid.   I'm very grateful for your help.  

 

I'm going to give that a try, rather than spending a while banging my head against the wall trying to fix the mysterious problem.