Cluster does not work after a while

Question

Hi :
          I have 2 clustered EJB server, the IP address is 192.168.0.226 and
          192.168.0.227.
          A servlet server is calling these two EJB servers with
          t3://192.168.0.226,192.168.0.227:7001
          All three machine use weblogic 5.1 sp9, Win2K Advanced Server.
          If I restart weblogic service, then the cluster works fine. but after while
          the cluster does not work, all the client request goto only one machine, in
          most case it's 227.
          It has no use even if I restart the service in the losted server. I have to
          restart all the server.
          The weblogic.log in the losted server is like this :
          Thu Apr 26 00:04:16 GMT 2001:<I> <RJVM> Signaling
          peer -6817319611378695685S192.168.0.204:[7001,7001,7002,7002,7001,-1] gone:
          weblogic.rjvm.PeerGoneException:
          - with nested exception:
          [java.io.EOFException]
          Thu Apr 26 00:04:47 GMT 2001:<I> <RJVM> Signaling
          peer -2123734719233546013S192.168.0.227:[7001,7001,7002,7002,7001,-1] gone:
          weblogic.rjvm.PeerGoneException:
          - with nested exception:
          [java.io.EOFException]
          Thu Apr 26 00:14:16 GMT 2001:<E> <ConMan> Attempt to sendMsg using a closed
          connection
          Thu Apr 26 00:14:16 GMT 2001:<E> <RJVM> Exception on send :
          weblogic.rmi.ConnectException: Attempt to sendMsg using a closed connection
          Please help me
          thanks
          andrew
          

Answer 1

Are you running with any proxy? Is the servlet server also the static httpd
          server?
          What type of network equipment are you using? If its high-end (Cisco) make sure
          all ports are set to 100/Full.
          Also, try setting the SendDelay slightly higher (ie. 25, 30, 35). Experiment
          with different values.
          weblogic.cluster.multicastSendDelay=25
          andrew wrote:
          > Mike, thanks for reply
          > The 192.168.0.204 is the servlet server.
          > I changed the NICs to use 100Mbps/Full duplex. it looks better: the error
          > msg is less than before, the it still happened.
          > Any suggestion?
          > thanks
          > andrew
          >
          > "Mike Kincer" <[email protected]> wrote in message
          > news:[email protected]...
          > > Which box is 192.168.0.204 ??
          > > It is communicating via multicast, which I would say is suspect.
          > > Otherwise, I'd say you have some network issues.
          > > Make sure all switch ports and all NICS are NOT set to "auto" config.
          > Select
          > > 100Mbps/Full duplex on all ports.
          > >
          > > andrew wrote:
          > >
          > > > Hi :
          > > > I have 2 clustered EJB server, the IP address is 192.168.0.226 and
          > > > 192.168.0.227.
          > > > A servlet server is calling these two EJB servers with
          > > > t3://192.168.0.226,192.168.0.227:7001
          > > >
          > > > All three machine use weblogic 5.1 sp9, Win2K Advanced Server.
          > > >
          > > > If I restart weblogic service, then the cluster works fine. but after
          > while
          > > > the cluster does not work, all the client request goto only one machine,
          > in
          > > > most case it's 227.
          > > >
          > > > It has no use even if I restart the service in the losted server. I have
          > to
          > > > restart all the server.
          > > >
          > > > The weblogic.log in the losted server is like this :
          > > >
          > > > Thu Apr 26 00:04:16 GMT 2001:<I> <RJVM> Signaling
          > > > peer -6817319611378695685S192.168.0.204:[7001,7001,7002,7002,7001,-1]
          > gone:
          > > > weblogic.rjvm.PeerGoneException:
          > > > - with nested exception:
          > > > [java.io.EOFException]
          > > > Thu Apr 26 00:04:47 GMT 2001:<I> <RJVM> Signaling
          > > > peer -2123734719233546013S192.168.0.227:[7001,7001,7002,7002,7001,-1]
          > gone:
          > > > weblogic.rjvm.PeerGoneException:
          > > > - with nested exception:
          > > > [java.io.EOFException]
          > > > Thu Apr 26 00:14:16 GMT 2001:<E> <ConMan> Attempt to sendMsg using a
          > closed
          > > > connection
          > > > Thu Apr 26 00:14:16 GMT 2001:<E> <RJVM> Exception on send :
          > > > weblogic.rmi.ConnectException: Attempt to sendMsg using a closed
          > connection
          > > >
          > > > Please help me
          > > > thanks
          > > > andrew
          > >
          > > --
          > > /\/\i|<e
          > >
          > > Mike Kincer
          > > Solutions Developer/Engineer
          > > Atlas Commerce "ebusiness evolved"
          > > Office phone: [phone]
          > > mailto:[email protected] [http://www.atlascommerce.com]
          > >
          > >
          /\/\i|<e
          Mike Kincer
          Solutions Developer/Engineer
          Atlas Commerce "ebusiness evolved"
          Office phone: [phone]
          mailto:[email protected] [http://www.atlascommerce.com]