You are not logged in.

#1 2014-12-19 07:28:27

Registered: 2014-02-20
Posts: 7

syslog-ng 3.6.1 segfaults in

After the latest syslog-ng upgrade:

[ALPM] warning: /etc/syslog-ng/syslog-ng.conf installed as /etc/syslog-ng/syslog-ng.conf.pacnew
[PACMAN] upgraded syslog-ng (3.5.6-1 -> 3.6.1-1)

and using the default packaged syslog-ng.conf, syslog-ng.service fails due to syslog-ng segfaulting in
From 'journalctl -b':

kernel: syslog-ng[330]: segfault at 0 ip 00007f6e920d6b0e sp 00007f6e78d78870 error 4 in[7f6e920d1000+c000]
systemd[1]: syslog-ng.service: main process exited, code=killed, status=11/SEGV
systemd[1]: Unit syslog-ng.service entered failed state.
systemd[1]: syslog-ng.service failed.
systemd-coredump[335]: Process 268 (syslog-ng) of user 0 dumped core.

... after which systemd attempts to restart the syslog-ng a few times before giving up. The journald.conf is the default one, i.e. Storage=auto and ForwardToSyslog=no.

Attempting to run syslog-ng manually causes the same crash.
All this is happening with the default source section in the /etc/syslog-ng/syslog-ng.conf :

source src {

Is this a bug ? Anybody else getting this ? Thanks in advance for any help.

P.S. As hinted in this post , when changing the above to:

source src {

the syslog-ng will run without crashing. However, it seems it will still need 'ForwardToSyslog=yes' to get anything from journald.


#2 2014-12-19 14:18:36

Registered: 2008-11-14
Posts: 2,127

Re: syslog-ng 3.6.1 segfaults in

I've got the same crash with an i686 machine, no crash with a 64bit one.
Also, the workaround posted is harmless and just reverts the syslog-ng to the old (better and lighter imho) behaviour.

Last edited by kokoko3k (2014-12-19 14:18:53)

Help me to improve ssh-rdp !


Board footer

Powered by FluxBB