LMS 3.2 has died

nawas · ‎11-30-2009

LMS3.2 on Solaris 10. It was working all until this weekend, no changes were made and I have noticed log files (var/adm/CSCOpx) started growing which I took care of by using logrot.pl. This morning my CiscoWorks was hosed, I managed to start the services and was able to login but now almost 14 processes are down. JRM is down and my daemonsbackup.log is about 95GB. config archive doesn't work and error out that JRM is down. I'm not able to start the UT acquisition either. Please advise what should I be looking at.

Joe Clarke · ‎11-30-2009

Post the output of the pdshow command.

nawas · ‎11-30-2009

Joe

After several times restarting the daemon and killing the porcess manually it has come back up and seem to be running fine for past one hour.

I have some questions though

1. I have noticed that it all started when var/adm/CSCOpx/log started growing and daemonsbackup.log was upto 95GB. Can the growing file system crash the database? I do run logration on a normal bases but appears that somehow it was filling file systen in minutes.

2. I'm seeing tons of invalid messages (from wireless controllers) coming into var/adm/messages and var/log/syslog_info. I do not manage wireless controller in RME, how do I stop receiveing these invalid messages?

Thank you for your response.

Joe Clarke · ‎11-30-2009

1. Not unless /var/adm/CSCOpx and /opt/CSCOpx share a file system. The two are typically independent. However, contents of the daemonsbackup.log may point to the real underlying cause of the problem.

2. Filter the messages at the source. Reconfigure the WLCs not to send syslogs to this Solaris box.