Bug#787473: systemd: Segfault while adding a new job

[Tino]

Hi Michael, nice to read you! :)

Yes I indeed do. And I was almost about to build a new selfmade NAS some
days ago just to get rid of the regular annoyance of rebooting.

Back in August 2015 I was using the formerly up to date systemd of
Stretch (221-1) on my Jessie system. With this version, systemd only
crashed when i "spoke" to it directly via systemctl or journalctl and
upon a monthly cronjob.
Because i did not get automatic updates i changed my apt policies and
tried to update systemd to the current up to date version of Stretch
(228-3 or 228-4, not sure). That completely bricked the system and i had
to boot to a former snapshot and revert that update.
Then i thought, well, there have been some updates of systemd within
Jessie as well. Maybe there is already a fix for this in the current
Jessie-Version. So i changed to the current up to date systemd of Jessie
(215-17+deb8u3) and i am back to daily crashes again. These seem to
happen around the daily cronjob at ~6:30. I guess the crash is triggered
when one of the cronjobs tries to talk to systemd, just like with the
monthly cronjob on 221-1. The services keep running for 1-2 days after
systemd crashed, but then i have to flip the power switch, because i
can't even login via ssh anymore.

Am 28.01.2016 um 18:26 schrieb Michael Biebl:
> Hi Tino
>
> On Sat, 22 Aug 2015 13:13:20 +0200 Tino <zottsoft at gmx.net> wrote:
>> Hi everyone,
>>
>> it has been one month since the last crash. I wanted to check if
>> everything is ok today and right in to moment i tried to access systemd,
>> it crashed. Unfortunately, it received an ABRT signal right after the
>> SEGV, which had overwritten the core dump. So i couldn't figure out
>> where the SEGV was comming from. I compared the ps faux of the crashed
>> system with the rebooted one and there was no difference except PID 1
>> (which is usually /sbin/init). Since systemd was crashing right in the
>> moment of the "systemctl" command, the origin of the SEGV should be
>> systemd, right?
>>
>> I don't know if the problem of these SEGV is still the same with version
>> 221-1. At least the system is somewhat stable if i don't touch anything.
> It's been a while.
> Are you still running into this issue?
>


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: OpenPGP digital signature
URL: <http://alioth-lists.debian.net/pipermail/pkg-systemd-maintainers/attachments/20160129/79d19d9c/attachment-0002.sig>