[cluster] SplitBrainMergeValidationOperation Ignoring join check from [NODEB]:5701, because this node is not master...

[wellprado]

Hi guys, How are you ?

Something strange happened about 2 times in 2 months of hazelcast usage.
Today was the second day.

I stopped/started the tomcat server nodeA, nodeB Become the "master" node. When NodeA was online again and join the hazelcast cluster, some messages was writen in catalina.out:

...
[[2m2017-05-15 14:48:59.754[[0;39m [[32m INFO[[0;39m [[35m2567[[0;39m [[2m---[[0;39m [[2m[thread-Acceptor][[0;39m [[36mc.h.nio.tcp.SocketAcceptorThread        [[0;39m [[2m:[[0;39m [NODEA]:5710 [hazelepp] [3.8] Accepting socket connection from /NODEB:46454
[[2m2017-05-15 14:48:59.754[[0;39m [[32m INFO[[0;39m [[35m2567[[0;39m [[2m---[[0;39m [[2m[cached.thread-6][[0;39m [[36mc.h.nio.tcp.TcpIpConnectionManager      [[0;39m [[2m:[[0;39m [NODEA]:5710 [hazelepp] [3.8] Established socket connection between /NODEA:5710 and /NODEB:46454
[[2m2017-05-15 14:48:59.763[[0;39m [[32m INFO[[0;39m [[35m2567[[0;39m [[2m---[[0;39m [[2m[ration.thread-0][[0;39m [[36mc.i.o.SplitBrainMergeValidationOperation[[0;39m [[2m:[[0;39m [NODEA]:5710 [hazelepp] [3.8] Ignoring join check from [NODEB]:5701, because this node is not master...
2017-05-15 14:56:59.754  INFO 2567 --- [ration.thread-0] c.i.o.SplitBrainMergeValidationOperation : [NODEA]:5710 [hazelepp] [3.8] Ignoring join check from [NODEB]:5701, because this node is not master...
[[2m2017-05-15 14:50:59.754[[0;39m [[32m INFO[[0;39m [[35m2567[[0;39m [[2m---[[0;39m [[2m[ration.thread-0][[0;39m [[36mc.i.o.SplitBrainMergeValidationOperation[[0;39m [[2m:[[0;39m [NODEA]:5710 [hazelepp] [3.8] Ignoring join check from [NODEB]:5701, because this node is not master...
[[2m2017-05-15 14:52:59.754[[0;39m [[32m INFO[[0;39m [[35m2567[[0;39m [[2m---[[0;39m [[2m[ration.thread-0][[0;39m [[36mc.i.o.SplitBrainMergeValidationOperation[[0;39m [[2m:[[0;39m [NODEA]:5710 [hazelepp] [3.8] Ignoring join check from [NODEB]:5701, because this node is not master...
[[2m2017-05-15 14:54:59.754[[0;39m [[32m INFO[[0;39m [[35m2567[[0;39m [[2m---[[0;39m [[2m[ration.thread-0][[0;39m [[36mc.i.o.SplitBrainMergeValidationOperation[[0;39m [[2m:[[0;39m [NODEA]:5710 [hazelepp] [3.8] Ignoring join check from [NODEB]:5701, because this node is not master...
...
2017-05-15 15:16:59.754  INFO 11893 --- [ration.thread-0] c.i.o.SplitBrainMergeValidationOperation : [NODEA]:5710 [hazelepp] [3.8] Ignoring join check from [NODEB]:5701, because this node is not master...

Even with this messages above, every thing seemed to work normal, but for this messages stop, I stopped NODEB for NODEA to become "Master" and started NODEB again, but I did this just this time today. After that I stopped/started many times each node in differente sequences and the error not appear any more. It's something intermittent that occurs rare times.

Do you have some guess about that ?

My env 64 bits: Linux RedHat 7.3, Tomcat 8.5, hazelcast-tomcat85-sessionmanager-1.1.jar, hazelcast-all-3.8.jar

Thank you very much!!!

[metanet]

Hi @wellprado

This message should be safe and should not create any problem. Did you see the log message printed repeatedly without terminating?

Regards,

[wellprado]

Hi @metanet

Yes, the message repeated but I didn't realize any termination. So, are we good ?

Thank you very much

Regards,

[metanet]

Ok, as I understand, when NodeA is restarted, it successfully joins to the master NodeB but still you see that message repeatedly printed. Can you provide any log files so that we investigate the issue more efficiently. Especially the logs of NodeA's restart that lead to these log messages.

Regards,

[mmedenjak]

@wellprado are you able to provide some more logs :

• from nodeB when nodeA has shut down
• from nodeA and node B when nodeA is started

[mdogan]

@wellprado: Without more logs, we can't identify the exact problem. Reason might be a short network interruption, or some long pause in the system/application.

For now closing this issue, we can reopen when you have more logs.