The mailing list for listmasters using Sympa

["Eric Sandquist" <address@concealed>]

Well this is interesting....

I did #2, and it has been running for over 6hrs now...  with no errors...
but, it  is attached to my telnet connection... so I obviously cannot run
this way for ever....  I will try number #1 on Thursday and see if something
shows up there....  But past logs, at level 4, showed nothing... the logging
just stopped...

Eric Sandquist

-----Original Message-----
From: Olivier Salaun - CRU [mailto:address@concealed]
Sent: Tuesday, December 24, 2002 1:23 AM
To: Eric Sandquist
Cc: address@concealed
Subject: Re: [sympa-users] Sympa is dieing for no appearant reason...


Hi Eric,

I can think about too investigating ways :

        1/ run processes in normal (or log_level 1) and look for common points
        in logs when processes die : same list, same action, same client
address,...

        2/ run sympa.pl attached to the TTY : ~sympa/bin/sympa.pl -d
        There must be an error message just before it dies, but it is written 
to
STDERR
        which is lost when running as a daemon.

You did not tell it (which is bad ;-) but I guess you are running a 3.4.2 ?
Any recent changes in your Sympa code ? changes on the server ?

Good luck.

Eric Sandquist wrote:
> I don't know if this has happened to anyone else, but for some reason
> sympa.pl, task_manager.pl, bounced.pl, and archived.pl all die on a fairly
> regular basis.
>
> I have searched though the logs and have found nothing.  It seems to
happen
> about an hour+ after restarting sympa, but has happened in as little as 10
> minutes.  The only thing that shows in the log is that sympa.pl and the
> others nolonger appear in the report.  I have logging set for 4(debug)...
> wwsympa.fcgi is working fine, and it's logging continues.
>
> I have almost no hair left :/
>
> I have found no evidence of anything "killing" them in any of the logs.
>
> Does anyone have any idea what could be causing these processes to die and
> leave no record that anything had gone wrong?  I keep looking for a
pattern
> or clue but have found none.  I have been experiementing with restarting
> sympa with a cron job every 10-20 minutes, but this solution creates its
own
> sets of problems.  I would rather figure out what is causing the 4 of them
> to die and correct the issue.

--
Olivier Salaun
Comite Reseau des Universites