AnsweredAssumed Answered

Starting resourcemanager and nodemanager fails..

Question asked by pukkimonkey on Aug 5, 2015
Latest reply on Aug 6, 2015 by pukkimonkey
I have a small 5 node cluster, with one client node, control node and 3 data nodes.

After installing mapR, I login through the web. I noticed:

 1. In control node tab (hostname - NameNode)  - ResourceManager has failed to start
 2. In data node tab (hostname - DataNode2) - NodeManager has failed to start.

Strangely, the nodemanager on the other two data nodes ( DataNode1, DataNode3) are running fine.

Regarding 1, I look into the log file of resourcemanager and I find this

    2015-08-05 14:18:41,231 FATAL org.apache.hadoop.yarn.server.resourcemanager.ResourceManager: Error starting ResourceManager
    org.apache.hadoop.yarn.exceptions.YarnRuntimeException: Problem binding to [] Address already in use; For more details see:

When I look at :8031, It looks fine to me  

    java    10184 mapr  207u  IPv6  911101      0t0  TCP *:8031 (LISTEN)  
    java    10184 mapr  251u  IPv6 1419902      0t0  TCP namenode:8031->datanode1:40190 (ESTABLISHED)  
    java    10184 mapr  253u  IPv6 1421374      0t0  TCP namenode:8031->datanode3:54147 (ESTABLISHED)  

Regarding 2: I looked into nodemanager log file on datanode2, I find this.

    2015-08-05 14:44:12,940 FATAL org.apache.hadoop.yarn.server.nodemanager.NodeManager: Error starting NodeManager
    org.apache.hadoop.yarn.exceptions.YarnRuntimeException: java.lang.RuntimeException: Unable to create proxy to the ResourceManager null