Admin Node Hanging when Monitoring Cluster in Console

Question

          I'm running into a very unusual situation with my cluster. The administrative node
          is failing. I'm running all of my nodes as background services. The admin node
          usually fails after about an hour. That's usually after all of the nodes are completely
          up.
          The only application deployed to the admin is the console.war. I typically will
          log into the console and monitor the progress of all of my managed nodes starting.
          I typically go to the server tab a watch all of my nodes change from Red to Green.
          I do this by clicking the refresh icon. It refreshes every ten seconds. Once I
          close my browser and reattempt to log back into the console, it ends up hanging.
          I log onto the server and notice that there are no errors in my log files, and
          that the beasvc.exe is still running for the admin. All indications are that it
          is running. The only way to resolve the issue is to kill the service and restart
          it. I've tried to replicate this in the foreground, but I get a similar error.
          The admin node fails and in the dos command window, I get the following error
          at the bottom of my thread dump:
          "# An unknown exception code exception has been detected in native code outside
          the VM. # Program counter=0x77fb16cc"
          

Answer 1

JDK 1.3.1_02 is certified for WLS 6.0 SP2 on Solaris only.
          Kumar Allamraju wrote:
          > what version of JDK?
          > I presume you are using the JDK s hipped with WLS.
          > If yes, we have seen JVM crashing at other customer sites.
          >
          > I would suggest you to try a new JDK (preferably 131_02) and see
          > if that makes any difference. Until we see thread dumps we cannot
          > really figure out the problem in WLS, so pls make sure that the JVM
          > is not crashing.
          >
          > --
          > Kumar
          >
          > Steve Feldman wrote:
          >
          > > I'm running into a very unusual situation with my cluster. The administrative node
          > > is failing. I'm running all of my nodes as background services. The admin node
          > > usually fails after about an hour. That's usually after all of the nodes are completely
          > > up.
          > >
          > > The only application deployed to the admin is the console.war. I typically will
          > > log into the console and monitor the progress of all of my managed nodes starting.
          > > I typically go to the server tab a watch all of my nodes change from Red to Green.
          > > I do this by clicking the refresh icon. It refreshes every ten seconds. Once I
          > > close my browser and reattempt to log back into the console, it ends up hanging.
          > >
          > >
          > > I log onto the server and notice that there are no errors in my log files, and
          > > that the beasvc.exe is still running for the admin. All indications are that it
          > > is running. The only way to resolve the issue is to kill the service and restart
          > > it. I've tried to replicate this in the foreground, but I get a similar error.
          > > The admin node fails and in the dos command window, I get the following error
          > > at the bottom of my thread dump:
          > >
          > > "# An unknown exception code exception has been detected in native code outside
          > > the VM. # Program counter=0x77fb16cc"
          > >
          Rajesh Mirchandani
          Developer Relations Engineer
          BEA Support