Nagios Support Forum

HI TEAM ,

This is in continuation to the previously mentioned issue below , that we have been still receiving lots of JMX alerts daily that return null value .

http://support.nagios.com/forum/viewtop ... 20&t=22263

The thing is that we are configuring Nagios Xi on a new server B and are making use of the existing configuration files of server A.

While Server A is a Solaris machine, Server B is a Linux One.

And also when we receive these alerts from Server B , we do not receive similar alerts from Server A at the same time.

The wrapper version of JMX is the same on both Solaris and Linux servers

Hosta A :

(PROD)emarkets@ap-em5p-a:/me01/apps/emarkets/nagiosjmxdaemon-1.0.2/bin$ ./wrapper -v
Wrapper (Version 3.2.1) http://wrapper.tanukisoftware.org


INFO | jvm 1 | 2014/02/14 10:36:26 | java version "1.5.0_22"


Host B :

(PROD XI)root@dkd1mr7954:/me01/apps/emarkets/nagiosjmxdaemon-1.0.2/bin# ./wrapper -v
Wrapper (Version 3.2.1) http://wrapper.tanukisoftware.org



INFO | jvm 1 | 2014/02/14 10:36:26 | java version "1.5.0_22"


Please let us know if you need any details on configuration or JMX application logs to debug further
Thanks
Clive

Are the issues intermittent, or are the failing checks consistently failing? Can you post one of the commands that is failing?

Well this happens consistently once every week ?
We were wondering if it is impcated because of some issues related to load and if some Java app could be attacting more load ?

Are you recording performance checks on this remote host? Can you check if there was a load spike around the time that the jmx checks fail? Additionally, do the jmx services restart at all during the week?

Looks like some of the JMX based commands that are supposed to provide passive results did not have the "check_command" option defined when we were revieing the services concerned.

Now we have select the appropritae command and "active" option is being checked .

Hope that works

ok. Let us know if you have further issues.

For now we have been restarted JMX demon every 5 minutes as we have seen that restarting JMX after we have received these bulk messages with null values returned has helped.

hmmm, that does not sound like a good solution. Any hints from the jmx server logs?

Fixing the "check_command" has not helped entirely as it seems.

Can you suggest anything else that I could be missing ?

Do you see anything having to do with the queries you are running against JMX from it's logs?