My unraid server used to run on a 3600X/B450 combo about a month ago and I experienced an error where both graceful AND forced shutdown failed and the server would just infinitely hang on terminal. I can type into terminal and interact with Unraid but it just can't shutdown. My only option is to power off the system by holding the power button.

 

I sold off my parts and moved to a 12700/Z690 combo and I am still experiencing the same problem. When I connect to my KVM i see the following

 

Waiting up to 90 seconds for graceful shutdown...

[...]

Forcing shutdown...

Starting diagnostics collection... sh: -c: line 1: unexpected EOF while looking for matching `"

 

Diagnostics attached. I created them immediately after my server booted up.devoraid-diagnostics-20230306-2008.zip

 

Can someone point me toward what is causing shutdown to hang?

 

Thanks.

 

Edited March 6, 20233 yr by LumpyCustard

Do you have the same issue if you first manually shutdown the VM(s)?

What are the names of the VMs you're running.  Any special characters or oddities on the naming?

Since moving to an Intel based server and utilising QSV/NVENC i opted to retire my Server 2016 VM. I haven't done anything funky with IOMMU groups either.

 

When i WAS using a VM it was simply named "Windows Server 2016"

 

I currently have no VM's in Unraid and have disabled the VM service.

Edited March 8, 20233 yr by LumpyCustard

When running the diagnostics from the GUI, does the popup hang?  What does it show when it hangs?

No, when i boot my server after failing to force shutdown it SOMETIMES says improper shutdown, other times it doesn't.

 

When i generate diagnostics it completes in about 10 seconds.

 

Once the array has started though the diagnostics take a long while to complete as it seems to be going through thousands of files in my Urbackup directory.

 

Edited March 8, 20233 yr by LumpyCustard

7 minutes ago, LumpyCustard said:

as it seems to be going through thousands of files in my Urbackup directory.

Why do you think that?

I posted a screenshot above. I have generated diagnostics dozens of times and i've never seen it do this before. Urbackup is a docker container i use to handle backups for my personal PC and a friend of mine who backs up over the internet. I've never seen it go through tens of thousands of files to generate a diagnostic before.

 

It's been generating a diagnostic for the last 5 minutes now and it's still going.

 

edit: 15 minutes now and it's still going lol.

Edited March 8, 20233 yr by LumpyCustard

OK i had to stop the diagnostic, after 20+ minutes my browser had reached 32GB of memory usage and things were getting weird. Why is the diagnostic tool killing itself by going through my nvme cache / urbackup folder?

You have mover logging enabled (Settings, Schedules, Mover Settings) so everything that gets moved gets logged, and obviously a ton of lines that urbackup is logging and then mover logs it, and it all needs to get anonymized

I did enable mover logging some weeks back for testing purposes, it has now been disabled. This wouldn't be the cause of the shutdown issue would it?

Doubtful

Back to square one then

And do the diagnostics work if you get them on the console?