Home / server / Questions / 790313
Accepted
Oleksandr
Asked: 2016-07-18 06:46:05 +0800 CST

Why does DFSZKFailoverController kills Namenode process in hadoop?

  • 772

I try to configure hadoop high availability cluster by following this tutorial:
http://www.edureka.co/blog/how-to-set-up-hadoop-cluster-with-hdfs-high-availability/

When I follow that article I faces with two main problems:
1. hdfs namenode -bootstrapStandby (I can't use this command because Namenode on the Standby node isn't started.) To solve this problem I manually run namenode on the Standby node before use this command.
2. When I run second ZKFC (on the Standby node) it kills Namenode process and I can't even manually start it. That is why Namenode started only on the Active node. If we kill the Active node the Standby node continue doing nothing (it doesn't starts Namenode).

Does somebody know what is the problem with that article?

hadoop hdfs high-availability zookeeper

1 Answers

  1. Best Answer

    format active namenode and then run hdfs namenode -bootstrapStandby on standby namenode command to copy matadata from active namenode

    there is no need to start seceoundry namenode to copy matadata from active namenode.

    • 1