MammothJerk

after just 2 hours of uptime i first stopped all dockers and then tried to stop the array

The problem occurs with share_size script preventing the array from stopping successfully.

i SIGTERM'ed the PIDs in htop to be able to successfully stop the array.

i then disabled docker in settings (VMs was already off), i ticked the box for 'reboot in safe mode' and rebooted.

2 hours later i was able to stop the array with no issue.

i then rebooted again (without enabling docker or VMs)

waited 2 hours again stopped the array without issue.

i re-enabled the docker service and then started the array.

i instantly got those du -sb... commands in htop.

on a hunch i stopped my hass-unraid docker and about 5-10 minutes later the du commands disappeared from htop.

it would appear the hass-unraid docker is polling the shares for their sizes and with a share with multiple small files like my photos share it is taking a long time for the scan to finish thus not letting the array stop without finishing those commands. ill bring it up to the author.
https://github.com/IDmedia/hass-unraid/issues/48

Another new 14900k and the issue seems to be fixed.

It would appear (at least in my case) that the cache SSD (kingston a400) was not of sufficient quality for my usecase.
i switched to a pair of kingston kc3000 nvme SSDs and the issue is completely gone.
 
My dockers feel more responsive than ever.
 
Although they are hitting pretty high temps (60-70c) so i might have to pick up some heatsinks for them 😬

the culprit was plex trying to analyze a old .mkv file that links to 2 other .mkv files.
 
the 2 mkvs were originally released as anime episodes but then got bundled into a movie, so to save space i used segment linking to create a file that acts as the movie without using double the space.