Known issue and its workaround

Before you proceed further configuring additional namenodes, you must be aware of the following known issue.

Issue 1

In case the JournalNodes were restarted on a cluster recently, the Add new NameNode wizard for HDFS service might fail to bootstrap the new NameNode, if there was not a new fsImage created since the restart of the JournalNodes. However, while restarting the JournalNodes the edit logs were rolled in the system.: If the bootstrap fails during the Add new NameNode wizard, you must perform the following steps:

Delete the newly added NameNode and FailoverController

Move the active HDFS NameNode to safe mode

Perform the Save Namespace operation on the active HDFS NameNode

Leave safe mode on the active HDFS NameNode

Try to add the new NameNode again

note
Entering safe mode disables write operation to HDFS and causes a service disruption. If you can not afford to enter safe mode, just delete the newly added NameNode and FailoverController in the HDFS service. You must wait until HDFS automatically creates a new fsImage and try adding the new NameNode again with the wizard.

After adding three NameNodes, performance impact like slowness in the cluster is detected while performing heavy read and write operations.: Tune the following parameters:

dfs.client.failover.sleep.base.millis=1

dfs.client.failover.max.attempts=25