[syslog-ng]syslog-ng dies on reload

[Florian Heissenhuber]

Hi,

I'm using syslog-ng on our central syslog server and it works really 
great. However I've found the following problem as I added a new filter 
rule:

I've added a filter rule that filters specific error messages and sends 
these error messages to a shell script that generates a SNMP traps. The 
  relavant configuration lines look as follows:

destination snmpsend { program("/data/scripts/pix_snmp_send.ksh "); };
filter f_SNMPTRAP1 { match('PIX-1-10100[1-5]|PIX-1-102001'); };
log { source(net); filter(f_SNMPTRAP1); destination(snmpsend); };

If I send a reload signal to syslog-ng (kill -HUP) the first time, 
everything is OK. Syslog-ng reloads its configuration and the following 
messages is found in /var/log/messages:

Mar 17 11:15:47 pixsysl01 syslog-ng[11020]: new configuration initialized
Mar 17 11:15:46 pixsysl01 syslog-ng[11020]: SIGHUP received, restarting 
syslog-ng

If I send a reload signal to syslog-ng (kill -HUP) the next time, 
syslog-ng dies with the following message:

Mar 17 11:17:36 pixsysl01 syslog-ng[11020]: new configuration initialized
Mar 17 11:17:35 pixsysl01 syslog-ng[11020]: SIGHUP received, restarting 
syslog-ng
Mar 17 11:17:36 pixsysl01 syslog-ng[11020]: syslog-ng version 1.5.26 
going down


Since I'm using logrotate for automatic log rotation, syslog-ng is 
reloaded more than once a day (and dies with the second reload).

Any idea what's wrong? I've already tried version 1.6rc1, but it has the 
same effect!

Regards,
Florian


-- 
Florian Heissenhuber