I have what appears to be a busted VM capability. I am unable to start any VM’s: every attempt leads to the following execution error:

 

“Hook script execution failed: internal error: Child process (LC_ALL=C PATH=/bin:/sbin:/usr/bin:/usr/sbin Home=/ etc/libvirt/hooks/qemu < Name of VM in single quotes> prepare begin -) unexpected exit status 255”

 

Virt-Mgr (docker) also throws an error during startup, the beginning of which reads: “The libvirtd service does not appear be installed.” As best I can tell, however, the service does appear to be running; it is possible that I have mis-configured the docker container & that this issue is not relevant to the larger problem.

 

I also have an Undead VM which, every time I remove it with prejudice...comes back when I delete (and rebuild) the libvirt image. The dadgum zombie even survives virsh undefine –nvram (&tc.), still returning when the image is rebuilt. It does not (however) start: it simply generates the same execution error as any other VM I create.

 

Interestingly (if, perhaps, irrelevant) the Undead VM seems to have no associated XML; only the GUI is undead. (The Idiot Within suspects this to imply that the Undead VM has no soul.)

 

The whole problem persists through restart of both the array or the server, including cold start. It appears to be similar to https://forums.unraid.net/topic/50194-hook-script-execution-failed-internal-error/ which, apparently,  has never been resolved in public.

 

A diagnostics file is attached.

 

Thanks in advance for your very kind attention!

cuailnge-diagnostics-20210516-1555.zip

9 hours ago, not_fred999 said:

I have what appears to be a busted VM capability. I am unable to start any VM’s: every attempt leads to the following execution error:

 

“Hook script execution failed: internal error: Child process (LC_ALL=C PATH=/bin:/sbin:/usr/bin:/usr/sbin Home=/ etc/libvirt/hooks/qemu < Name of VM in single quotes> prepare begin -) unexpected exit status 255”

 

Virt-Mgr (docker) also throws an error during startup, the beginning of which reads: “The libvirtd service does not appear be installed.” As best I can tell, however, the service does appear to be running; it is possible that I have mis-configured the docker container & that this issue is not relevant to the larger problem.

 

I also have an Undead VM which, every time I remove it with prejudice...comes back when I delete (and rebuild) the libvirt image. The dadgum zombie even survives virsh undefine –nvram (&tc.), still returning when the image is rebuilt. It does not (however) start: it simply generates the same execution error as any other VM I create.

 

Interestingly (if, perhaps, irrelevant) the Undead VM seems to have no associated XML; only the GUI is undead. (The Idiot Within suspects this to imply that the Undead VM has no soul.)

 

The whole problem persists through restart of both the array or the server, including cold start. It appears to be similar to https://forums.unraid.net/topic/50194-hook-script-execution-failed-internal-error/ which, apparently,  has never been resolved in public.

 

A diagnostics file is attached.

 

Thanks in advance for your very kind attention!

cuailnge-diagnostics-20210516-1555.zip 157.36 kB · 0 downloads

 

Did you ever modify or added hooks to libvirt?

Especially file qemu in /etc/libvirt/hooks/

It seems they are failing because of that hook.

3 hours ago, ghost82 said:

 

Did you ever modify or added hooks to libvirt?

Especially file qemu in /etc/libvirt/hooks/

It seems they are failing because of that hook.

Nope.

It looks like I've stumbled over a resolution.  As it happens, I'm still working with my original geek stick, upgraded many times since early 2015.  There was still a "domain.img" file under config/plugins/dynamix.kvm.manager.  A bit of spelunking showed that the Undead VM was mentioned several times in that image.  I'm guessing that the Unraid was still processing that image when rebuilding libvirt.img.  Since the timestamp on the file was quite recent (just yesterday), it was probably still updating it, too.  Mere speculation, of course.

 

Renaming of the file & deletion of the libvirt image, followed by a server reboot, seems to have eliminated the resurrection of that Undead VM.  The file in question appears not to have been replaced by a new one, either.

 

Also, VM's now start.  No error banners are displayed.  I'll pop back in if oddities re-occur.

 

On a related note, it looks to me like I need to figure out which of those old plugin directories are no longer of any use.  Not to mention figuring out which ones are causing problems.