Dear Maintainer,

After upgrading to latest Sid my QNAP TS-419 failed to boot.

By breaking into the initramfs shell I figured out that local-top/lvm2
was hanging in lvchange_activate(). Running the same command manually
from the initramfs shell prompt also resulted in a hang:

lvm lvchange -ddddddd -aay -y --sysinit --ignoreskippedcluster /dev/qnap-vg
  lvmetad is not active yet, using direct activation during sysinit
[  132.878220] device-mapper: uevent: version 1.0.3
[  132.887704] device-mapper: ioctl: 4.33.0-ioctl (2015-8-18)
initialised: dm-devel@redhat.com
<hang>

There aren't many tools in the initramfs to help debug this, but Sysrq-T
showed:

[  186.603029] lvm             S c036e5a8     0   134    128 0x00000000
[  186.609431] [<c036e5a8>] (__schedule) from [<c036e7a0>] (schedule+0xa8/0xb8)
[  186.616511] [<c036e7a0>] (schedule) from [<c01646ec>] (SyS_semtimedop+0x514/0x76c)
[  186.624113] [<c01646ec>] (SyS_semtimedop) from [<c000a2e0>] (ret_fast_syscall+0x0/0x38)

Which seems to suggest that it is waiting for some event which
presumably never occurs (I've left it for several hours).

I have narrowed this down to upgrading libudev0 and udev to the Stretch
version. If I downgrade only those to the Jessie version then the
problem no longer occurs. I've filed this against lvm2 firse since it
supplies the script, but maybe this belongs to udev.

Ian.

Forgot to say, the good versions of {libudev1,udev} are 215-17+deb8u2
and known bad are 227-2.

I'll do a little bit of bisection over
http://snapshot.debian.org/package/systemd/ and see if I can narrow it
down a bit.

Ian.

226-4 is the last good and 227-1 is the first bad revision. which is
probably not all that useful...

Ian