You have two standbys and one primary HMaster. Your primary HMaster fails. Which of the
remaining HMasters becomes the new primary?
Whichever HMaster first responds to ZooKeeper
Whichever HMaster ZooKeeper randomly selects
Whichever HMaster creates the znode first
Whichever HMaster has the lower IP address
* The Hbase master server creates the zookeeper znode /hbase . This is then used
for hbase daemons to coordinate. Even the name of the active Hbase master is stored here. If the
hbase master dies, the backup hbase master overwrites the contents of the znode so clients and
region servers know about the new master. Apart from this, region info is maintained in zookeeper
znodes as well.
* Multi-master feature introduced in 0.20.0 does not add cooperating Masters; there is still just one
working Master while the other backups wait. For example, if you start 200 Masters only 1 will be
active while the others wait for it to die. The switch usually takes zookeeper.session.timeout plus a
couple of seconds to occur.