Support Questions

Find answers, ask questions, and share your expertise

Replacing NameNode in Hadoop cluster running on Azure VM's


I am running a 6 Node Hadoop Cluster(1 Edge Node, 1 NameNode, 1 Secondary namenode, 2 Datanodes) hosted on Azure VM's. I want to do some performance tests so i want to upgrade Namenode to higher capacity VM. I am running a Non HA Cluster. But in Production we would be hosting HA cluster.

What should be the ideal approach for moving/upgrading Namenode to VM of higher capacity? I donot want to loose any data and want to avoid any failure.




Of course test this first so you have a hardened playbook to go by

Enable HA and get ride off secondary namenode. Then add new nodes and use move namenode function in the HDFS section of Ambari. You can also add new nodes first, then enable HA pointing to new nodes, including journal nodes, that's probably a better approach.


@Artem Ervits

I can afford some downtime since it is a dev cluster. I dont think i would need HA. Anyways it looks a cleaner approach. Do you have any reference links/documents which i can follow to achieve this?