O'Reilly logo

Apache Accumulo for Developers by Guðmundur Jón Halldórsson

Stay ahead with the world's most comprehensive technology and business learning platform.

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, tutorials, and more.

Start Free Trial

No credit card required

Failover

By design, Hadoop's single point of failure is when the NameNode goes down, and there isn't a secondary NameNode running. This limitation requires Hadoop to be configured for NameNode failover. For more information about NameNode failover, visit http://wiki.apache.org/hadoop/NameNodeFailover.

When nodes in the ZooKeeper cluster are started, the first task the nodes do is to find the leader. After the leader has been chosen, the rest of the ZooKeeper nodes will follow that leader. The concept of "follow the leader" means that there is always a leader machine and (in a multi-node environment), some followers. The follower can be behind the leaders, but that will not affect the clients because they will always connect to the same ZooKeeper ...

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, interactive tutorials, and more.

Start Free Trial

No credit card required