My wife went to use our kitchen pc which is a windows vm running on unraid.  The vm which was running did not wake up when she moved the mouse.  I opened a VPN connection and went to the VM tab and it said Libvirt Services failed to start.   Then while on the phone with my wife she said nevermind it came back up.  I am home now writing this post from the VM and it appears to be working fine, however the VM tab still says Libvirt Sevices failed to start.  

 

I have attached a copy of diagnostics.  Any suggestions.  Something like this happed a couple of weeks ago (same message) and I rebooted the unraid server and everything came back fine.  Since its happened twice...should I delete the libvirt image and recreate my vm's?  I know I have backups of the xmls.   

 

Thanks,

Dan

tower-diagnostics-20210520-1338.zip

13 minutes ago, goinsnoopin said:

I am home now writing this post from the VM and it appears to be working fine, however the VM tab still says Libvirt Sevices failed to start.  

This is very strange, and I don't see nothing in the logs about libvirt issues, the only thing logged about that are the scheduled trims.

@JorgeB   I just found the logs for the last crash, it was only 5 days ago...not the couple of weeks I stated in my original posting.  Attached are those logs.   

 

All I did after this crash on 5/15/2021 was reboot the server and all came back fine. 

 

Thanks for looking!

tower-diagnostics-20210515-0746.zip

I can see this:

May 15 07:39:08 Tower emhttpd: shcmd (6380905): /etc/rc.d/rc.libvirt stop
May 15 07:39:08 Tower root: error: failed to connect to the hypervisor
May 15 07:39:08 Tower root: error: Failed to connect socket to '/var/run/libvirt/libvirt-sock': Connection refused
May 15 07:39:08 Tower root: Waiting on VMs to shutdownerror: failed to connect to the hypervisor
May 15 07:39:08 Tower root: error: Failed to connect socket to '/var/run/libvirt/libvirt-sock': Connection refused
May 15 07:39:08 Tower root:
May 15 07:39:08 Tower root: error: failed to connect to the hypervisor
May 15 07:39:08 Tower root: error: Failed to connect socket to '/var/run/libvirt/libvirt-sock': Connection refused
May 15 07:39:08 Tower root: Stopping libvirtd...
May 15 07:39:09 Tower root: error: failed to connect to the hypervisor
May 15 07:39:09 Tower root: error: Failed to connect socket to '/var/run/libvirt/libvirt-sock': Connection refused
May 15 07:39:09 Tower root: /etc/rc.d/rc.libvirt: line 157: kill: (23854) - No such process
May 15 07:39:12 Tower root: Stopping virtlogd...
May 15 07:39:13 Tower root: Stopping virtlockd...
May 15 07:39:14 Tower emhttpd: shcmd (6380906): umount /etc/libvirt
May 15 07:39:14 Tower root: umount: /etc/libvirt: target is busy.

 

VM service failed to stop and unmount libvirt, so it's normal for a reboot to fix it, can't say why it failed to stop though.

Since I have backup of the VMs xml, any reason not to delete libvirt.img and start fresh?

Likely won't make any difference, and should have a backup of the image itself, not just the XMLs.

Sometimes when I have this problem on my test box I found stopping docker then starting VMs then starting Docker again sometimes works.

Also does the same thing happen if you uninstall the libvirtwol plugin?