[Xen-users] Missing udev env variables & broken pvscan in Stretch & Xenial

[Mark Pryor <tlviewer@xxxxxxxxx>]

Users,

I've encountered a bug in lvm2/udev/systemd since December on Debian Stretch & Ubuntu Xenial (lately). I wrote up my experience with a workaround here:

http://107.185.144.55/xen/debian/stretch/46x/lvm2_bug/

Many people in #xen told me to file a bug on Debian BTS. Well, I've since found an existing bug report from July 2015! How can it be such an old bug and still exist? I think a solution was proposed within 2 weeks. See here

https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=791869#209

When you boot your Stretch OS & Xen, if you have VM's on LVM, if you use

use_lvmetad=1 

try `sudo lvm dumpconfig global/use_lvmetad`

then a pvscan should be run on your pv device. If my pv device is sdc1 (major:8, minor:33), then

systemctl status lvm2-pvscan@8:33.service

should show success and tell me it found so many volumes. With our subject bug, this pvscan is never run.

The bug makes proper boot sequence difficult and VM's on LVM are detected too late for xendomains service.

PryMar56

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxx
http://lists.xen.org/xen-users