[Pacemaker] Logging in syslog daemon can't be stopped

[Andrew Beekhof]

On Wed, Sep 26, 2012 at 2:45 AM, Andrew Daugherity <adaugherity at tamu.edu> wrote:
>> Message: 2
>> Date: Tue, 25 Sep 2012 11:54:44 +1000
>> From: Andrew Beekhof <andrew at beekhof.net>
>> To: The Pacemaker cluster resource manager
>>       <pacemaker at oss.clusterlabs.org>
>> Subject: Re: [Pacemaker] Logging in syslog daemon can't be stopped
>> Message-ID:
>>       <CAEDLWG0rRdYK=C3M1V4jxezO16BQf-9yi+M8xPF30mtFoTpF+g at mail.gmail.com>
>> Content-Type: text/plain; charset=ISO-8859-1
>>
>> On Tue, Sep 25, 2012 at 9:21 AM, Larry Brigman <larry.brigman at gmail.com> wrote:
>>> On Mon, Sep 24, 2012 at 4:08 PM, Andrew Beekhof <andrew at beekhof.net> wrote:
>>> There is at least one if not more pacemaker processes that don't
>>> honor the facility.  It always goes to daemon.
>>> crmd I know doesn't honor the config settings and always goes to daemon.
>>> pengine also logs
>>>
>>> I'm using version 1.1.5. on  Linux.
>>
>> I have a hard time believing that, all the daemons expect the lrmd use
>> the same logging initialisation function.
>> In any case, I just verified that with
>> https://github.com/beekhof/pacemaker/commit/da8809c, all this works as
>> expected.
>
> I realize this may have changed, and if so, great (or possibly these programs are not really part of pacemaker and I'm barking up the wrong tree), but as of the versions that shipped with SLES 11 SP1 (looks like it was 1.1.2, now running 1.1.5 after updates), I had to filter out several programs that didn't respect the 'syslog_facility: local1' setting in corosync.conf or 'logfacility local1' in logd.cf, specifically cluster-dlm, stonith, and ocfs2_controld/ocfs2_hb_ctl.

None of those are from pacemaker, so my changes wont affect them.
'stonith' is the cli from cluster-glue I think.
Btw. logging is MUCH improved over 1.1.2 (and even 1.1.5), so perhaps
its less necessary to disable syslog.

>
> This is what I have in my rsyslog.conf:
> ====
> # keep cluster stuff that doesn't respect the log facility setting in
> # corosync.conf or logd.cf out of /var/log/messages
> #
> # NB: '$syslogseverity <= 5' is equivalent to a selector of '*.notice'
> if      ( ($programname == 'cluster-dlm') or ($programname == 'stonith') or \
>         ($programname startswith 'ocfs2') ) and \
>         ($syslogseverity <= 5) then \
>         -/var/log/cluster/cluster.log
> if      ( ($programname == 'cluster-dlm') or ($programname == 'stonith') or \
>         ($programname startswith 'ocfs2') ) then \
>         -/var/log/cluster/cluster-debug
> if      ( ($programname == 'cluster-dlm') or ($programname == 'stonith') or \
>         ($programname startswith 'ocfs2') ) then \
>         ~
>
>
> # Cluster logs
> local1.debug                            -/var/log/cluster/cluster-debug
> local1.notice                           -/var/log/cluster/cluster.log
> ====
>
> Again, I'm not saying you're wrong, just that when I set up my cluster I had to work around programs that ignored the syslog facility setting.  If they have since been fixed, all the better!
>
>
> Andrew Daugherity
> Systems Analyst
> Division of Research, Texas A&M University
>
> _______________________________________________
> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org