Mailinglist Archive: opensuse-bugs (4247 mails)

< Previous Next >
[Bug 1019886] Upgrade from 42.1 to 42.2 fails to initialize LVM VG with use_lvmetad = 1
  • From: bugzilla_noreply@xxxxxxxxxx
  • Date: Wed, 18 Jan 2017 07:30:03 +0000
  • Message-id: <>

--- Comment #14 from zhen ren <zren@xxxxxxxx> ---
(In reply to Youri Matthys from comment #13)
(In reply to zhen ren from comment #12)
Thanks for your confirmation. Following your steps, I failed to reproduce
this issue on my side. Everything works well after upgraded to Leap42.2. The
difference is my setup doesn't use md raid.

I will continue to try lvm2 on top of md-raid setup.

My pleasure.

As noted the parameter "use_lvmetad = 1" seems to require (in my specific
case) LVM modules included in the initrd. Setting this parameter to 0
"solves" the issue on my system.

See bug#1017034 comment#8,#43.
LVM activates LVs in different way depending on the use_lvmetad value in
lvm.conf. If it's set to 0, then systemd will call system-generators/lvm2-
activation-generator to generate several service file to activate LVs during
boot. While it's set to 1, then LVs will be activated by udev rules in 69-dm-
lvm-metad.rules. So we need to fix the problem separately.

Could I propose adding the dracut rule for LVM by default to avoid issues in
the future whenever any VG is detected? Maybe we could even have the default
lvm.conf set to use_lvmetad = 0 by default?

I'd say no. This should be a real bug.

It appears this issue is related to some upstream reports:

This post is about "md device (PV) is not found", not your case?

This post is very helpful! I guess it's because the "lvmetad" is not running
when running "udevadm trigger".

Thanks for your feedback. Let me know if I can assist further.

Could you please try the lvm2 rpm in my repo?

You are receiving this mail because:
You are on the CC list for the bug.
< Previous Next >