One of the partner I am working with do not want to use RAID on master nodes.
They are looking for some answers to below questions.
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
1) Do you think we can rely on HA for handling service
interruptions due to disk failures
2) Will HA failover kick if the drive that has the mount for
service configuration files (XML files) fail
3) Will HA failover kick if the drive that has the mount for log
files fails
4) If a service uses local storage for metastore, will HA failover
kick if the drive that has the mount for metastore fails.
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
What is your opinion on these?
@Chris Nauroth , @sdutta@hortonworks.com, @vnair@hortonworks.com