It turned out that I had to re-install the Nvidia drivers so in the end quite simple. I downloaded the drivers direct from Nvidia and re-installed them and now working again without issue.
I always had issue using the Nvidia drivers from Pacman, though reckon in this instance custom installing them is what caused my problems to begin with.
Hope this helps someone else.
An update, after some searching I was able to remove the libvirt errors, I had to use
virsh --pool-undefine <image_name>
and now those errors have stopped showing up, I still get a error from libvirtd: Internal error:Failed to find path for dmidecode binary but I don’t think this is causing things to fail.
Looking further down through the log the next red section pertains to cinnamon-sessio process doing a core dump, I don’t really understand the stack trace but there is the line #0 0x00007fd53a61404f raise (libc.so.6) and that libc.so is where I thought I saw the error when I did the update.
Looking through the pacman logs I can see that one of the packages updated was linux itself, from version 4.6.4-1 to 4.7-1 and I reckon that is where the issue lies.
I will post again once I have solved this
Last night I had 14 updates available for my system, I did them as I normally did (didn’t really pay any attention to what was being updated) and it went through, a prompt came up that certain files couldn’t be found,the only one I recall was libc.so and that might even be wrong
I closed the prompt and actually noticed things were misbehaving a bit so restart the computer, GRUB comes up without issue but when the system tries to boot I get some network errors that always happen and then just a plain black screen.
Not 100% certain what is going on, I do have VGA passthrough set-up for a gaming Windows VM and am using the proprietarey Nvidia diriver,though I don’t think that is what is causing the issue. I have been able to load up my system with just the command line, can’t remember the specific kernal parameter now and using journalctl --boot=-1 (I let the system boot then restarted into the command line) it looks like the error before everything goes belly up is the following lines:
libvirtd: unsupported configuration: cannot fine logical volume group name ‘windows_gaming’
libvirtd: internal error: Failed to autostart storage pool ‘windows_vm’: unsupported configuration: cannot find logical volume group name ‘windows_gaming’
libvirtd: cannot open directory ‘/home/richard/OldCompStuff/Stuff 3’: no such file or directory
libvirtd: internal error: Failed to autostart storage pool ‘Stuff 3’
Also a few lines after that systemd-coredump reports: Process 570 (cinnamon-sessio) of user 1000 dumped core.
Not sure if those lvm’s are causing the issue but the odd thing is I no longer use them anymore, when I was getting things set-up I believe I set-up a lvm called windows_gaming but I abandoned that idea and just have a partition with a qcow file on it. I was fairly certain that I removed the lvm I created. Also that directory used to contain my old HDD’s from my old Windows set-up which is no longer present.
/etc/fstab is not trying to mount any of those locations so I wander if libvirt updated and this new version doesn’t like something I did before that has managed to persist, toruble is I have no idea where to start looking, lvdisplay has no entries in it. I also tried pvdisplay and vgdisplay but they all returned nothing.
Thanks for any help
Ok so I have managed to solve my own problem by creating my own package that provides ‘libgl’, then the various qt packages updated without issue.
Hi I have an issue where I can’t update qt5 (or qt4) because it asks me to choose a provider for libgl and when I choose nvidia-libgl (or any option actually) it tells me there are file conflicts and the files already exist then bombs out and fails.
I believe the problem is that I am using VGA Passthrough (which is freaking amazing btw) and I had to install the nvidia driver direct from the Nvidia website. If I install it through Pacman then I don’t get a graphical environment when I reboot. Not sure exactly why this is but believe it is related to the vfio kernal I am using.
So I believe I already have nvidia-libgl installed but Pacman doesn’t realise that I do and wants to install it but then fails when the files are already on my system. Is there a way around this or to get Pacman to ignore looking for libgl?
Thought I would just post how it went, I downloaded the nvidia package (361.18-2) and had to download a couple of extra packages (nvidia-utils, and linux 4.4) downloaded those packages and installed all fine but on restart got a black screen. Not sure what it was and had to power off my machine multiple times. I tried disabling the graphics to boot to a command line (can’t remember the option I tried) but that didn’t work and after some faffing with a live cd discovered that the powering off I did killed my SSD :(
Luckily for me it was practically brand new and I could return it without any quibbles and I had the data on a seperate HDD so was down only for a couple of days. I am not sure if upgrading the linux kernal was a good idea and is probably what caused it, I have later found out I could have downgraded the kernal version which might have saved the repeated power offs but I am not going to try this again, for a while anyway.
If I use
pacman -U downloaded.package.tar.gzto install a package will it all still work properly when a new version of said package becomes available through the repos and I do an update through the update manager?
Specifically I am thinking about updating my Nvidia driver manually as the new (hopefully bug free) version I need is currently in testing and whilst looking at the package online I realised I could download the latest version of it rather than continue to wait for it to come out of testing.