HBase HA Solution

HBase HA

HMaster in HBase allocates Regions. When one RegionServer service is stopped, HMaster migrates the corresponding Region to another RegionServer. The HMaster HA feature is brought in to prevent HBase functions from being affected by the HMaster single point of failure (SPOF).

Figure 1 HMaster HA implementation architecture

The HMaster HA architecture is built based on Ephemeral nodes (temporary nodes) created in the ZooKeeper cluster.

Upon startup, HMaster nodes try to create a master znode in the ZooKeeper cluster. The HMaster node that creates the master znode first becomes the active HMaster, and the other is the standby HMaster.

It will add watch events to the master node. If the service on the active HMaster is stopped, the active HMaster disconnects from the ZooKeeper cluster. After the session expires, the active HMaster disappears. The standby HMaster detects the disappearance of the active HMaster through watch events and creates a master node to make itself be the active one. Then, the active/standby switchover completes. If the failed node detects existence of the master node after being restarted, it enters the standby state and adds watch events to the master node.

When the client accesses the HBase, it first obtains the HMaster's address based on the master node information on the ZooKeeper and then establishes a connection to the active HMaster.