Problem basically solved. See Edit 2 below. Thanks to anyone that takes the time to read this entire post. I'm just hoping to learn something.

 

I was following SpaceInvader's video to install Virt-Manager and at the step in the video HERE I think I must have screwed something up. I opened the file using nano, made the required changes (listen_addr = "0.0.0.0") and wrote out the file. I even went back into the file to ensure the changes stuck and all looked well. I then stopped the array and restarted it. Unfortunately after that I get "Libvirt Service failed to start" when clicking the VMS tab. A review of the system logs is throwing the following error:

 

Jan 26 23:41:39  xxxxxxx root: 2019-01-27 06:41:39.113+0000: 9765: error : main:1165 : Can't load config file: configuration file syntax error: /etc/libvirt/libvirtd.conf:1: expecting a name: /etc/libvirt/libvirtd.conf

 

I have SSH'd into the server and the libvertd.conf file is not in the /etc/libvert folder (it is empty). I did find a master config file in /etc/libvertd-/ but copying that to the /etc/libvert/ folder and rebooting does not work. In fact a reboot deletes the file I copied leaving /etc/libvert/ empty again.

 

As a disclaimer I am a weekend warrior and while I have enough knowledge to get around, I do not understand what is happening and am hoping not to loose the week of time I have spent setting up what had turned into a great server. I had several VM's running flawlessly along with about 6 dockers, plugins, etc. My only goal was to install virt-manager to keep xml changes to the config file persistent when using a gui.

 

I have not included diagnostics because my hope is that somebody more knowledgeable than myself can tell what is wrong from the error above. I assume I will need to find and/or download the correct file and put it somewhere but thats where my ideas end. If this is incorrect then I'm up for whatever it takes.

 

The next lines of the log file state:

 

Jan 26 23:41:39 xxxxxxx emhttpd: shcmd (109): exit status: 1

Jan 26 23:41:39 xxxxxxx emhttpd: shcmd (111): umount /etc/libvirt

Jan 26 23:41:39 xxxxxxx emhttpd: nothing to sync

 

I'm thinking that the /etc/libvert-/ folder is where the file goes and then that folder gets mounted as /etc/libvert but as I said I have no idea. I'm hoping to be able to correct this in some fashion without needing to re-install. I'd also like to figure out what went wrong to pick up some knowledge along the way.

 

TIA to anybody that has any suggestions! 

 

EDIT: Well I was able to successfully get the service restarted by deleting the libvert image, turning off the service, rebooting and re-enabling the service. Of course all my VM's were gone but I thought I was smart by making a copy of the old image. I repeated these steps and rebooted with the service left off. I then copied the old image file back to the image directory, left the service off and re-booted. When I turned the service on this time I got the same error. This leads me to believe that the problem is within the image file but I have no way of knowing if (or how) to edit that file. I really don't want to loose my VM configs so if anybody knows how to edit that I would appreciate it. Barring none I guess I will just need to re-configure all the prior machines.

 

Edit 2: I am back where I was before encountering this issue. I also have virt-manger running within a Fedora VM as outlined in Spaceinvaders video. All VM's have been re-created and re-configured properly. Other than loosing 12 hours I'm thrilled but still miffed and would appreciate any of you that are way more qualified than me chiming in. I believe when using nano the first time I must have fat-fingered something which was then written to the libvirt.img. Assuming that is in fact the case, I'd really like to know if there is a way to edit that image file or at least look inside of it. While I will certainly be more careful with future edits, I'd still like to pick-up some knowledge.

 

Thanks.

Edited January 27, 20197 yr by CraigGivant

Hi Craig, 

 

Ran into this same issue, you're write up helped me get my VM's back, thanks! 

If you lose the files in /etc/libvirt for example libvirtd.conf qemu.conf etc. then the best approach will be to delete the libvirt.img via gui delete function. Restart the unraid server => then you will get a new configuration with all necessary files in /etc/libvirt. But the old configuration (VMs) will be lost. Just replace the libvirt.img from your backup,  then restart the server again. Everything will be like before all VMs will appear again. If needed you can replace the whole domains ans system folder, as long as you have regularly backups. Therefore it is important to create backups, in my case I use the (daily) snapshot fuction via ZFS Master triggered by a scirpt (spaceinvaderone). Furthermore these snapshots are being rsynced to two different servers (qnap), whereas one of the servers is acting only as a backup server with no internet access, in order to prevent a possible randsonware attack.  Otherwise both of my qnap servers could be compromised.