Deactivating and Activating Brokers all the time in Conv Cluster normal?

Question

hi *,
i just wanted to know if this is a normal behaviour.
we do have a conventional cluster with 8 + 1 node
8 workes and 1 master.
in our cluster jmq logfiles we are very often getting this messages:
[26/Sep/2008:13:54:16 CEST] [B1179]: Activated broker
        Address = mq://example.com:36105/?instName=jmqDivAccA2&brokerSessionUID=6667329240416509952
        StartTime = 1222429874806
        ProtocolVersion = 410
[26/Sep/2008:13:54:16 CEST] [B1072]: Closed cluster connection to broker mq://example.com:36105/?instName=jmqDivAccA2&brokerSessionUID=6667329240416509952
[26/Sep/2008:13:54:16 CEST] [B1180]: Deactivated broker
        Address = mq://example.com:36105/?instName=jmqDivAccA2&brokerSessionUID=6667329240416509952
        StartTime = 1222429874806
        ProtocolVersion = 410
[26/Sep/2008:13:54:16 CEST] [B1179]: Activated broker
        Address = mq://example.com:39105/?instName=jmqCzeAccA2&brokerSessionUID=8358993850445347840
        StartTime = 1222429866204
        ProtocolVersion = 410
[26/Sep/2008:13:54:16 CEST] [B1072]: Closed cluster connection to broker mq://example.com:39105/?instName=jmqCzeAccA2&brokerSessionUID=8358993850445347840
[26/Sep/2008:13:54:16 CEST] [B1180]: Deactivated broker
        Address = mq://example.com:39105/?instName=jmqCzeAccA2&brokerSessionUID=8358993850445347840
        StartTime = 1222429866204
        ProtocolVersion = 410
[26/Sep/2008:13:54:16 CEST] [B1179]: Activated broker
        Address = mq://example.com:37100/?instName=jmqSsfAccA1&brokerSessionUID=1687473902759115264
        StartTime = 1222235763760
        ProtocolVersion = 410
[26/Sep/2008:13:54:16 CEST] [B1072]: Closed cluster connection to broker mq://example.com:37100/?instName=jmqSsfAccA1&brokerSessionUID=1687473902759115264
[26/Sep/2008:13:54:16 CEST] [B1180]: Deactivated broker
        Address = mq://example.com:37100/?instName=jmqSsfAccA1&brokerSessionUID=1687473902759115264
        StartTime = 1222235763760
        ProtocolVersion = 410 any idea why this happens so often?
right now all the nodes reside on the same hardware (very low delay) and use almost non CPU usage.
regards chris

Answer 1

hi linda,
this cluster is for test cases on the same machine and using the very same binaries.
therefore there is for shure no version conflict or something similar.
can you send me the line how to turn on debug for cluster communication?
every broker is started like this:
nohup imqbrokerd -name jmqXXXXXXXXA1 -port 39100 -bgnd -Dimq.cluster.url=file:/home/caps/XXXXXXXXXX/broker.config -varhome /home/caps/XXXXXXXX/jmqXXXXA1 > jmqXXXXA1.out 2>&1 &and broker config is like this:
imq.cluster.brokerlist=servername:35100,servername:35105,servername:36100,servername:36105,servername:37100,servername:37105,servername:38100,servername:38105,servername:39100,servername:39105
imq.cluster.masterbroker=servername:34100after our system was patched yesterday i can not reproduce this behaviour anymore.
BUT it would be great if you can send me the lines for debuging....
regards chris