Dev1 Galaxy Forum / Encrypted volume shutdown issue going from Jessie to ascii

Re: Encrypted volume shutdown issue going from Jessie to ascii

dummy@example.com (Geoff 42) — Sat, 14 Apr 2018 10:36:14 +0000

A quick look around shows that there are rules in /lib/udev/rules.d/69-lvm-metad.rules so that it would appear that eudev is supposed to work with lvmetad.

Geoff

Re: Encrypted volume shutdown issue going from Jessie to ascii

dummy@example.com (Geoff 42) — Fri, 13 Apr 2018 13:33:02 +0000

If I have read the docs correctly, lvmetad is used to remember your lvm set-up, so that it can boot faster. If it can't talk to lvmetad, then lvm has to search the hardware itself, which is slower. lvmetad uses udev to get its information. As Ascii uses eudev instead of udev, the question must be, "has eudev been tested against lvmetad?"

Geoff

Re: Encrypted volume shutdown issue going from Jessie to ascii

dummy@example.com (fsmithred) — Tue, 10 Apr 2018 16:58:34 +0000

Similar issue, possible fix for delayed shutdown of encrypted system:
https://dev1galaxy.org/viewtopic.php?pid=8289#p8289

Re: Encrypted volume shutdown issue going from Jessie to ascii

dummy@example.com (Geoff 42) — Tue, 10 Apr 2018 16:04:11 +0000

I am not using encrytion, but I am using LVM2 (for occasional use of Xen). Since upgrading from Jessie to Ascii, I am also seeing the WARNING: Failed to connect to lvmetad message. I can also see a message in daemon.log lvmetad[2207]: Failed to accept connection errno 11.. lvmetad does appear in a ps listing. messages report random: lvm: uninitialized urandom read (4 bytes read). There do appear to be some pauses in booting. I vaguely recall reading something about random being slower in order to get some good data.

Geoff

Re: Encrypted volume shutdown issue going from Jessie to ascii

dummy@example.com (zenkaon) — Sat, 22 Jul 2017 00:40:14 +0000

This *feature* is discussed extensively at https://bugs.debian.org/cgi-bin/bugrepo … bug=791944, which seems to point the blame at udev.

The udev in strecth/ascci is 232-25, whereas in Buster/Ceres it's 234-1, with the changelog http://metadata.ftp-master.debian.org/c … _changelog suggesting that bug #791944 was fixed in udev 233.

After setting some pin priorities, apt-get -t ceres install udev, updated me to 234-1. I still get the annoying long shutdown, so this has not resolved the issue.

I've just put stretch back on my laptop with the same encrypted LVM partitioning as shown in the first image above, with udev 232-25, and I have no problems shutting down at all.

Incidentally, the "WARNING: Failed to connect to lvmetad" message is also present when stretch boots, so I think that's a red herring here.

Encrypted volume shutdown issue going from Jessie to ascii

dummy@example.com (zenkaon) — Wed, 19 Jul 2017 01:27:32 +0000

Hi all,

I like to setup an encrypted volume with an LVM containing swap and /, as can be seen at:
https://ibb.co/iALmzk

When I boot in Jessie, I see this:
https://ibb.co/cb2MX5

This gives no problems on shutdown. However, once upgraded to ascii I see on boot:
https://ibb.co/c8SoC5
Note the additional line "WARNING: Failed to connect to lvmetad"

This results in a long shutdown time (2 minutes or so) with the screen looking like:
https://ibb.co/hYNAQQ

Any ideas?

This is really annoying on shutdown - it takes ages. The encrypted volume and LVM appear to work fine, indeed I'm typing this on the encrypted ascii system experiencing the shutdown problem.