New cache SSD, docker's fried.

[ehcorn]

I don't really know where to start with this, so I've got a screenshot of what I see, the errors I get from unraid are literally "undefined error" showing in red from time to time.

 

this is right after a cache drive replacement, I tried to follow spaceinvader one's instructions online of set the shares to "no", the I used appdatabackup to take a full docker backup, and then shutdown once everthing seemed fine, then swap the drives. That's when I noticed a few of my dockers were missing.

 

I tried nuking all my docker containers, deleting my docker image and re-installing everything from the CA store and initially it looked fine but I've got the same error now. oddly enough if I go to the docker page, it says my array must be started for docker to work... even though containers are clearly running and I can interact with them in between weird errors.

 

I've now tried to go back into settings to disable docker, but I can't for some reason. I don't get any errors, the buttons appear normal just when I click apply nothing happens.

 

my gut feeling is my boot USB is fried as there are errors across my whole install and settings aren't saving on anything I try.

 

edit: It seems I can't download diagnostics either, before anyone else asks.

Edited January 6, 2022 by ehcorn
update

[trurl]

attach diagnostics to your NEXT post in this thread.

[ehcorn]

I can't. when I try to download the logs, it just sits at download... /boot/logs

[trurl]

Do you have a flash backup?

 

Shutdown, put flash in PC and checkdisk. While there make a backup. Reboot from a USB2 port then post diagnostics

[ehcorn]

tower-diagnostics-20220105-2145.zip

 

pulled the drive and made a backup of all the files at least, checkdisk didn't show anything bad and on a new boot I was able to pull the logs. Never posted these before so I hope this is right.

 

update: I deleted all docker containers again and started from scratch this time, my problem appears to be resplved but if someone with more knowledge than me feels like looking at the logs for anything abnormal that would be appreciated.

 

update2: I can reproduce the error by firing up Machinaris an setting it to plot, when I do the web UI goes completely bonkers and the system more or less stops responding until reboot, even if I shut down the Machinaris docker. I'm now getting a specific error:

Warning: file_pib_contents():Only -1 of XXX (the number of bytes changes) written, possible out of free disk space in /user/local/emhttp/plugins/unassigned.devices/include/lib.php on line 75

 

Final update and solution (so far):

Machinaris is mapped to plot using a ramdisk, this script was interrupted when I switched my cache for some reason so /mnt/ram didn't exist, and the docker container dumped the files to who knows where, causing something that wasn't big enough for the files to fill up nearly instantly. So this is all related to some core directory of the unraid install running out of space and that caused the issues.

Edited January 7, 2022 by ehcorn
update