> The workaround seems good, but there still tends to be issues bringing up
> the multipath devices on boot with QEMU in my testing; just the workaround
> isn't always sufficient: break=pre-multipath then resuming helps out there.
May you describe the issues you're seeing?
I guess I have hit more than a few so far.
Specifically, check bug 1429327 ('multipath' runs *before* all disks show up), and bug 1432062 (udev rules fail w/ disk-names w/ spaces when user_friendly_names isn't set).
Hi Mathieu,
> The workaround seems good, but there still tends to be issues bringing up
> the multipath devices on boot with QEMU in my testing; just the workaround
> isn't always sufficient: break=pre-multipath then resuming helps out there.
May you describe the issues you're seeing?
I guess I have hit more than a few so far.
Specifically, check bug 1429327 ('multipath' runs *before* all disks show up), and bug 1432062 (udev rules fail w/ disk-names w/ spaces when user_friendly_names isn't set).