We have a locally developed syslog-ng pkg that works perfectly on 200+ systems of various architectures and kernel revs. The SMF service created with the package has been great so far on all of these systems, however there is one system that just doesn't work with SMF. Whenever it tries to start it goes into maintenance mode:
svc:/system/system-log-ng:default (system log) State: maintenance since Wed Oct 22 19:35:32 2008 Reason: Restarting too quickly. See: http://sun.com/msg/SMF-8000-L5 See: syslogd(1M) See: /var/svc/log/system-system-log-ng:default.log Impact: This service is not running. [ Oct 22 19:35:32 Executing start method ("/lib/svc/method/system-log-ng") ] syslog-ng service starting. [ Oct 22 19:35:32 Method "start" exited with status 0 ] [ Oct 22 19:35:32 Stopping because all processes in service exited. ] [ Oct 22 19:35:32 Executing stop method (:kill) ] [ Oct 22 19:35:32 Restarting too quickly, changing state to maintenance ] Ok, so it seems like it may be a problem with the start method - but I can run that manually just fine: # ps -ef|grep syslog-ng root 1614 29276 0 22:07:07 pts/9 0:00 grep syslog-ng # /lib/svc/method/system-log-ng syslog-ng service starting. # ps -ef|grep syslog-ng root 1622 29276 0 22:07:18 pts/9 0:00 grep syslog-ng root 1620 1 0 22:07:16 ? 0:00 /usr/local/sbin/syslog-ng and like I said before - this same package runs fine on many other systems. How can I further debug this service start? -- This message posted from opensolaris.org
