3 Replies Latest reply: May 18, 2012 4:15 AM by Galder Zamarreño RSS

    can't generate topology of clustering modes

    wang yong Newbie

      i want to configure a distributed cluster.(5.1.4.FINAL)

      now i have two machines : A(192.168.1.114) and B(192.168.1.115).

      node1: 192.168.1.114 port:6800

      node2: 192.168.1.114 port:6801

      node3: 192.168.1.115 port:6800

       

      i use these command:

      startServer.bat -l 192.168.1.114 -p 6800 -m 20 -t 20 -c infinispan.xml -r hotrod

      startServer.bat -l 192.168.1.114 -p 6801 -m 20 -t 20 -c infinispan.xml -r hotrod

      startServer.bat -l 192.168.1.115 -p 6800 -m 20 -t 20 -c infinispan.xml -r hotrod

       

       

      The problem is that :the generated cluster can't detect node3,

      and i'm sure that there is no firewall in each machine

       

      Logs:

      2012-05-17 11:54:56,765 DEBUG [TcpTransportFactory] (main) Statically configured servers: [/127.0.0.1:6801]

      2012-05-17 11:54:56,765 DEBUG [TcpTransportFactory] (main) Load balancer class: org.infinispan.client.hotrod.impl.transport.tcp.RoundRobinBalancingStrategy

      2012-05-17 11:54:56,765 DEBUG [TcpTransportFactory] (main) Tcp no delay = true; client socket timeout = 60000 ms; connect timeout = 60000 ms

      2012-05-17 11:54:56,890 INFO  [Codec11] (main) ISPN004006: New topology: [/127.0.0.1:6801, /127.0.0.1:6800]

      2012-05-17 11:54:56,890 INFO  [TcpTransportFactory] (main) ISPN004014: New server added(/127.0.0.1:6800), adding to the pool.

      2012-05-17 11:54:56,890 INFO  [TcpTransportFactory] (main) ISPN004016: Server not in cluster anymore(/192.168.1.115:6800), removing from the pool.

        • 1. Re: can't generate topology of clustering modes
          Sanne Grinovero Master

          Hi,

          from the logs it looks like your servers are bound to 127.0.0.1; When bound to localhost they won't listen to the external IP.

          • 2. Re: can't generate topology of clustering modes
            wang yong Newbie

            I changed all localhost to specific ip address(192.168.1.114),the result is the same.

             

            the hotrod client logs:

            2012-05-18 10:29:35,406 DEBUG [TcpTransportFactory] (main) Statically configured servers: [/192.168.1.114:6800, /192.168.1.114:6800, /192.168.1.115:6801]

            2012-05-18 10:29:35,406 DEBUG [TcpTransportFactory] (main) Load balancer class: org.infinispan.client.hotrod.impl.transport.tcp.RoundRobinBalancingStrategy

            2012-05-18 10:29:35,406 DEBUG [TcpTransportFactory] (main) Tcp no delay = true; client socket timeout = 60000 ms; connect timeout = 60000 ms

            2012-05-18 10:29:35,625 INFO  [Codec11] (main) ISPN004006: New topology: [/192.168.1.114:6800, /192.168.1.114:6801]

            2012-05-18 10:29:35,625 INFO  [TcpTransportFactory] (main) ISPN004016: Server not in cluster anymore(/192.168.1.115:6800), removing from the pool.

             

            netstat ( 192.168.1.114)

            TCP    127.0.0.1:7800         0.0.0.0:0              LISTENING       5796

            TCP    127.0.0.1:7801         0.0.0.0:0              LISTENING       5132

            TCP    127.0.0.1:7801         127.0.0.1:4298         TIME_WAIT       0

            TCP    127.0.0.1:7801         127.0.0.1:4309         ESTABLISHED     5132

             

            TCP    192.168.1.114:6800      0.0.0.0:0              LISTENING       5796

            TCP    192.168.1.114:6801      0.0.0.0:0              LISTENING       5132

             

            netstat (192.168.1.115)

            TCP    127.0.0.1:7800         0.0.0.0:0              LISTENING       3184

            TCP    192.168.1.115:139       0.0.0.0:0              LISTENING       4

            TCP    192.168.1.115:6800      0.0.0.0:0              LISTENING       3184

             

            pls check my jgroups-tcp.xml  configration.

            • 3. Re: can't generate topology of clustering modes
              Galder Zamarreño Master

              Looks like you .114 and .115 machines don't cluster. Take your jgroups config and run it through the tests in http://www.jgroups.org/manual/html/ch02.html from .114 and .115.

               

              Btw, your JGroups configuration looks misconfigured. You cannot have both MPING and TCPPING configured. You use either one or the other. I'd suggest you go back to the jgroups-tcp.xml that we ship and start from scratch.