Please read this before reporting a bug:
https://wiki.archlinux.org/title/Bug_reporting_guidelines
Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.
REPEAT: Do NOT report bugs for outdated packages!
https://wiki.archlinux.org/title/Bug_reporting_guidelines
Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.
REPEAT: Do NOT report bugs for outdated packages!
FS#56759 - libvirt 3.9.0-2: attempting to start a VM with virsh results in hard freeze
Attached to Project:
Arch Linux
Opened by Jimi Bove (Jimi-James) - Sunday, 17 December 2017, 22:10 GMT
Last edited by Doug Newgard (Scimmia) - Wednesday, 20 December 2017, 15:38 GMT
Opened by Jimi Bove (Jimi-James) - Sunday, 17 December 2017, 22:10 GMT
Last edited by Doug Newgard (Scimmia) - Wednesday, 20 December 2017, 15:38 GMT
|
DetailsDescription: Trying to start a virtual machine after I upgraded resulted in a freeze so hard that even trying `systemctl restart libvirtd` resulted in systemctl freezing indefinitely. It and all virsh-related processes could be killed with SIGTERM just fine, but they'd freeze again if I tried to run them again. For some reason, with none of those processes running (after I had killed them), my computer froze trying to shut down and I had to hold the power button down. Downgrading back to 3.8.0-1 seems to be a working workaround.
Additional info: libvirt 3.9.0-2 -> 3.8.0-1 My VM is a QEMU VM running Windows 10 with OVMF for PCI passthru, and I'm passing through a GPU and a USB card. I haven't yet had the chance to troubleshoot and figure out whether passing that hardware through has anything to do with this issue. Steps to reproduce: -Upgrade libvirt to 3.9.0-2 -Try to start a VM, possibly specifically the kind of VM I have (Windows 10, PCI passthru), with `virsh start <domain>` |
This task depends upon
-libvirt's version has nothing to do with it (meaning it's most likely a kernel bug, because qemu was not part of the upgrade that started this issue). Downgrading libvirt didn't solve the problem--restarting my computer did. And doesn't always.
-The exact bug isn't that virsh freezes. The exact bug is that libvirtd freezes when it first runs in my system. And a valid workaround so far seems to be having libvirtd.service be not enabled, and instead starting it manually after my computer has finished booting up into a GUI.
I'm going to have to do more troubleshooting and submit a completely different bug, maybe not even to the Arch Linux tracker. This bug is now invalid.