As Athos said on 24 of June last year - see 
https://bugs.launchpad.net/ubuntu/+source/qemu/+bug/2107529/comments/1 we'd 
need more information to make this actionable in any way. The problem is that 
those traces will create a lot of output, but we do not yet know what to look 
for. And then you can't predict when it happens which makes this even harder.
To be clear, strace/ltrace won#t hurt, but it also isn't the one solution.

Let me ask more specifically for what might help in this case, I think
even better would be

- set this up on a system alone (no other guests or activities to have quite 
clean logs)
- after it happened attach here
  - a timestamp when the problem occurred
  - `cat /var/log/libvirt/qemu/$guestname.log`
  - all from `journalctl -u libvirtd`

Furthermore, sadly, the steps as in "Go to edit -> Connections Details -> 
Storage" are ambiguous.
In which management stack, and then when you say "Add hardware -> passthrough 
(SCSI)" what kind of device do you select?

I can only assume this might be via virt-manager, but that is hard to recreate.
Could you, once you configured it, get the full guest config and then the 
related system info.

Like
- virsh dumpxml $guestname
And to then understand how this fits to scsi
- lsblk 
- multipath -ll

Furthermore around the time it failed please provide the 10 minutes of
general journalctl output before and after (use --since --until) as
something else in the system might relate to what happens in
libvirt/qemu.

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2107529

Title:
  Ubuntu24.04: KVM passthrough SCSI_disks IO fail

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/qemu/+bug/2107529/+subscriptions


-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to