I have a single server that is running ovirt as a host and a storage domain. I ran an update when the GUI told me that there was one. I shutdown all of the VM's, placed the host in maintenance mode and did the update. Upon rebooting the server, I was able to pull my VM's online without any issue, but then, all of a sudden, vm's stopped responding as well as the host. The host is showing as "Down" and I cannot "Activate" it. Also "Maintenance Mode" is greyed out. Storage pools are "down" but eventually come back online, but the host still will not come online. Every few minutes, the host goes to "UP" right before failing and going back "Down". To make it more confusing, like I mentioned before, this is all in one box, so it isn't like their is a network issue between the storage and the host boxes or anything. Currently trying to reinstall the host and see if that helps, but I would very much appreciate any sort of guidance, support or ideas! Fencing failed on Storage Pool Manager OVIRT-Host for Data Center OVIRT-Datacenter. Setting status to Non-Operational. Host OVIRT-Host failed to recover. _______________________________________________ Users mailing list Users(a)ovirt.org http://lists.ovirt.org/mailman/listinfo/users

Thanks for the quick response Charles! Only had a second so I have not taken a look at the logs yet... But after reinstalling the host, it looks like everything came back up at first. I have discovered what is causing the entire host to fail, crash and reboot is when I mess with my windows 10 vm through the graphics card. I have an old Radeon HD 6870 that I have been using in my windows 10 vm without issue (I even did some light gaming with no issues). It seems that the last update I did has caused not only my windows VM to crash, but also my entire OVIRT host. Really not sure why Ovirt has an issue with my graphics card on my windows host all of a sudden. Will check the logs soon. The main login page shows: Version 4.1.4.2-1.el7.centos Isn't 4.1.5 out? My centos7 box wont seem to find the 4.1.4 On Oct 7, 2017 5:20 PM, "Charles Kozler" <ckozleriii(a)gmail.com&gt; wrote: > Most answers reside in the agent.log under /var/log/hosted-engine-ha/. > Tail that for a bit and see what pops up. You can also review vdsm log > under /var/log > > I've found that looking at these two logs and just trying a few things > yields results > > On Sat, Oct 7, 2017 at 5:18 PM, Wesley Stewart <wstewart3(a)gmail.com&gt; > wrote: > >> I have a single server that is running ovirt as a host and a storage >> domain. I ran an update when the GUI told me that there was one. >> >> I shutdown all of the VM's, placed the host in maintenance mode and did >> the update. Upon rebooting the server, I was able to pull my VM's online >> without any issue, but then, all of a sudden, vm's stopped responding as >> well as the host. >> >> The host is showing as "Down" and I cannot "Activate" it. Also >> "Maintenance Mode" is greyed out. >> >> Storage pools are "down" but eventually come back online, but the host >> still will not come online. >> >> Every few minutes, the host goes to "UP" right before failing and going >> back "Down". To make it more confusing, like I mentioned before, this is >> all in one box, so it isn't like their is a network issue between the >> storage and the host boxes or anything. >> >> Currently trying to reinstall the host and see if that helps, but I >> would very much appreciate any sort of guidance, support or ideas! >> >> >> Fencing failed on Storage Pool Manager OVIRT-Host for Data Center >> OVIRT-Datacenter. Setting status to Non-Operational. >> Host OVIRT-Host failed to recover. >> >> >> _______________________________________________ >> Users mailing list >> Users(a)ovirt.org >> http://lists.ovirt.org/mailman/listinfo/users >> >> >

So it appears that things being passed through VIA USB passthrough are broken. My mouse does not seem to register in my Windows 10 VM, and the two keyboards I have tried have caused the VM and the host to crash. I am not trying to update my host to 4.1.5 by running "sudo engine-setup" on my host. It seems to detect the updates, but then fails saying "[ ERROR ] Failed to execute stage 'Setup validation': Cannot upgrade the Engine database schema due to wrong ownership of some database entities." Any ideas? I am tempted just to blow everything away and start over, but I dont know how difficult it is to use an old data domain in a new install. Would it be as simple as copying the files over into the new domain? On Sat, Oct 7, 2017 at 8:35 PM, Wesley Stewart <wstewart3(a)gmail.com&gt; wrote: > I checked under: > /var/log/hosted-engine-ha/ > > I do not have this directory in my /var/log file. > > Also I did some more experiments. I removed the graphics card and left > only a keyboard and mouse (USB inputs) added to the VM. Same as before, > everything works fine, but once I try to use the keyboard or mouse > > BOOM everything crashes (including the host, which reboots after). Next > step, I suppose it to try a different keyboard and mouse, but my guess is > that any USB added device will cause the VM and host to crash upon > interaction. Everything loads fine, and the VM even continues to continue > working normally, until one of these device is activated or used. > > Any ideas? > > On Sat, Oct 7, 2017 at 8:09 PM, Wesley Stewart <wstewart3(a)gmail.com&gt; > wrote: > >> Thanks for the quick response Charles! Only had a second so I have not >> taken a look at the logs yet... >> >> But after reinstalling the host, it looks like everything came back up >> at first. I have discovered what is causing the entire host to fail, crash >> and reboot is when I mess with my windows 10 vm through the graphics card. >> >> I have an old Radeon HD 6870 that I have been using in my windows 10 vm >> without issue (I even did some light gaming with no issues). It seems that >> the last update I did has caused not only my windows VM to crash, but also >> my entire OVIRT host. >> >> Really not sure why Ovirt has an issue with my graphics card on my >> windows host all of a sudden. Will check the logs soon. >> >> The main login page shows: >> Version 4.1.4.2-1.el7.centos >> >> Isn't 4.1.5 out? My centos7 box wont seem to find the 4.1.4 >> >> On Oct 7, 2017 5:20 PM, "Charles Kozler" <ckozleriii(a)gmail.com&gt; wrote: >> >>> Most answers reside in the agent.log under /var/log/hosted-engine-ha/. >>> Tail that for a bit and see what pops up. You can also review vdsm log >>> under /var/log >>> >>> I've found that looking at these two logs and just trying a few things >>> yields results >>> >>> On Sat, Oct 7, 2017 at 5:18 PM, Wesley Stewart <wstewart3(a)gmail.com&gt; >>> wrote: >>> >>>> I have a single server that is running ovirt as a host and a storage >>>> domain. I ran an update when the GUI told me that there was one. >>>> >>>> I shutdown all of the VM's, placed the host in maintenance mode and >>>> did the update. Upon rebooting the server, I was able to pull my VM's >>>> online without any issue, but then, all of a sudden, vm's stopped >>>> responding as well as the host. >>>> >>>> The host is showing as "Down" and I cannot "Activate" it. Also >>>> "Maintenance Mode" is greyed out. >>>> >>>> Storage pools are "down" but eventually come back online, but the host >>>> still will not come online. >>>> >>>> Every few minutes, the host goes to "UP" right before failing and >>>> going back "Down". To make it more confusing, like I mentioned before, >>>> this is all in one box, so it isn't like their is a network issue between >>>> the storage and the host boxes or anything. >>>> >>>> Currently trying to reinstall the host and see if that helps, but I >>>> would very much appreciate any sort of guidance, support or ideas! >>>> >>>> >>>> Fencing failed on Storage Pool Manager OVIRT-Host for Data Center >>>> OVIRT-Datacenter. Setting status to Non-Operational. >>>> Host OVIRT-Host failed to recover. >>>> >>>> >>>> _______________________________________________ >>>> Users mailing list >>>> Users(a)ovirt.org >>>> http://lists.ovirt.org/mailman/listinfo/users >>>> >>>> >>> >