Weblogic Console hangs when connecting to one of two servers in a cluster

Question

          Hello,
          Please email me as well as posting reply if possible as resolution is required
          immediately.
          We have two Weblogic 5.1 servers (named server1 & server2) running in a cluster
          using multicast over a VLAN and a DNS entry of "mycluster" to point to the list
          of two IP addresses.
          Both servers say they have joined the cluster upon startup, however, only one
          of them (server2) reports when the other one (server1) has joined, but not vice
          versa.
          Also, the Weblogic Console only displays server1 in the cluster, and hangs when
          it tries to attach to server2, which generates exceptions on server2. The error
          is as follows:
          <Kernel> ExecuteRequest failed. NullPointerException at weblogic.cluster.MemberStash.getManagedSubObjects(MemberStash.java:259).
          I expect this would also prevent replication from working properly.
          One other thing that might be related to the above is that when the URL "mycluster:7001"
          is used directly in the web browser, it sometimes (not always, not sure if it
          matters on which machine) gives the error: "Inconsistent cookie: primary !=
          me".
          Two IIS web servers are being proxied to the cluster URL, and round-robin load
          balancing of the app servers appears to work fine from this perspective.
          Both servers appear to be working properly otherwise.
          Many thanks for any clues,
          Ben
          

Answer 1

          primary != me occurs when you restart WLS but do not exit the browser, then hit
          WLS (the browser has a stale cookie).
          Mike
          "Ben Yenko-Martinka" <[email protected]> wrote:
          >
          >Hello,
          >
          >Please email me as well as posting reply if possible as resolution is
          >required
          >immediately.
          >
          >We have two Weblogic 5.1 servers (named server1 & server2) running in
          >a cluster
          >using multicast over a VLAN and a DNS entry of "mycluster" to point to
          >the list
          >of two IP addresses.
          >Both servers say they have joined the cluster upon startup, however,
          >only one
          >of them (server2) reports when the other one (server1) has joined, but
          >not vice
          >versa.
          >Also, the Weblogic Console only displays server1 in the cluster, and
          >hangs when
          >it tries to attach to server2, which generates exceptions on server2.
          > The error
          >is as follows:
          ><Kernel> ExecuteRequest failed. NullPointerException at weblogic.cluster.MemberStash.getManagedSubObjects(MemberStash.java:259).
          >I expect this would also prevent replication from working properly.
          >
          >One other thing that might be related to the above is that when the URL
          >"mycluster:7001"
          >is used directly in the web browser, it sometimes (not always, not sure
          >if it
          >matters on which machine) gives the error: "Inconsistent cookie: primary
          >!=
          >me".
          >
          >Two IIS web servers are being proxied to the cluster URL, and round-robin
          >load
          >balancing of the app servers appears to work fine from this perspective.
          >Both servers appear to be working properly otherwise.
          >
          >Many thanks for any clues,
          >Ben