AnsweredAssumed Answered

How to handle EC2 instance and volume failure?

Question asked by communityadmin on Jun 22, 2011
Latest reply on Jun 23, 2011 by Ted Dunning
I am trying to figure out how to handle instance and volume failures of my NameNode in an EC2 environment.  DRBD seems like a nice way to do this, but it doesn't work in EC2.

Some options include:

 - glusterfs to replicate namenode data (didn't seem to work in my first test)
 - raid1 ebs volumes with manual failover to standby machine.

Neither of these address the problem of how to get the datanodes to recognize the new namenode without restarting everything in the entire cluster.

What are others doing on EC2?