vCenter

I am using 4.5.2.2 to monitor a number of servers.

These include a number of Ubuntu 10.04 LTS running Tomcat 6.0.24.

The hq-agent is running as the tomcat6 user on these boxes.

It seems like every other reporting cycle the tomcat instance on several nodes is reporting as unavailable, then coming back up and the back down.

I have seen a few errors in the logs, but never anything constistent.

The most recent one I am seeing is:

2012-10-02 12:32:05,662 DEBUG [Thread-4] [MxServerDetector] Cannot configure local jmx.url: State.Name.re=java|jsvc,State.Name.Pne=jsvc,Args.*.sw=-Dcatalina.base=/var/lib/tomcat6 java.io.IOException: Could not map vmid to user Name
java.io.IOException: State.Name.re=java|jsvc,State.Name.Pne=jsvc,Args.*.sw=-Dcatalina.base=/var/lib/tomcat6 java.io.IOException: Could not map vmid to user Name
        at org.hyperic.hq.product.jmx.MxUtil.getUrlFromPid(MxUtil.java:397)
        at org.hyperic.hq.product.jmx.MxServerDetector.configureLocalMxURL(MxServerDetector.java:126)
        at org.hyperic.hq.product.jmx.MxServerDetector.getServerResource(MxServerDetector.java:351)
        at org.hyperic.hq.plugin.tomcat.TomcatServerDetector.getServerResource(TomcatServerDetector.java:255)
        at org.hyperic.hq.product.jmx.MxServerDetector.getServerResources(MxServerDetector.java:379)
        at org.hyperic.hq.plugin.tomcat.TomcatServerDetector.getServerResources(TomcatServerDetector.java:226)
        at org.hyperic.hq.autoinventory.scanimpl.NullScan.scan(NullScan.java:83)
        at org.hyperic.hq.autoinventory.Scanner.start(Scanner.java:198)
        at org.hyperic.hq.autoinventory.ScanManager.mainRunLoop(ScanManager.java:141)
        at org.hyperic.hq.autoinventory.ScanManager.access$000(ScanManager.java:41)
        at org.hyperic.hq.autoinventory.ScanManager$1.run(ScanManager.java:107)

I've checked the jmx settings, and most of them seem to be working and reporting.

Any ideas or suggestions on resolving this wsould be greatly appreciated.

I usually recommend checking for time sync issues. See my post last month on the subject: Troubleshooting False Alerts in Hyperic with Time Sync

That is interesting, and certainly looks like it could account for some of the false alerts.

It doesn't really address the error I posted related to the jmx issue, however.

I will do some testing to see what sort of impact those tips have.