Why Xenon Is the Modern MySQL HA Alternative to MHA

MySQL High Availability Choices

During the era of traditional MySQL 5.5 replication, MHA (Master High Availability) was a mature solution, but it has not kept pace with newer MySQL versions that rely on GTID.

Why MHA Is Losing Relevance

MHA was developed by DeNA’s youshimaton. The last release was in 2018, and the author believes its value is limited in GTID environments. Modern deployments favor GTID + ROW + Semi‑Sync for better functionality, performance, and maintenance.

Introducing MySQL + Xenon

To replace MHA in GTID‑driven high‑availability scenarios, the open‑source tool Xenon (by RadonDB) offers a new approach.

What Is Xenon?

Xenon (https://github.com/radondb/xenon) is a Raft‑based, decentralized master election tool for MySQL clusters. It uses Semi‑Sync to guarantee no data loss and leverages MySQL 5.7+ parallel replication to reduce replica lag.

Xenon Architecture

Automatic leader election based on Raft (requires GTID) and enhanced Semi‑Sync.

Automatic failover using configurable leader‑start/stop commands or integration with Consul/ZooKeeper.

Xtrabackup backup scheduling integration .

How Xenon Works

In a three‑node cluster (one leader, two followers), each node maintains a GTID set. When the leader fails, followers compete; the follower with the highest GTID becomes the new leader, and the system automatically switches roles.

{Leader, [GTID:{1,2,3,4,5}]}
{Follower1, [GTID:{1,2,3,4,5}]}
{Follower2, [GTID:{1,2,3}]}

Followers detect leader loss and start election.

The follower with the highest GTID (Follower1) wins.

Election stops once a new leader is chosen.

Master‑slave switch completes automatically.

Enterprise‑Grade Features

One‑master‑multiple‑slave architecture ensuring financial‑grade strong consistency.

Second‑level master failover with automatic VIP handling for read/write traffic.

Automatic binlog expiration to free space.

Support for multiple MySQL versions (5.6, 5.7, 8.0) and deployment platforms (bare metal, VM/cloud, Kubernetes).

Improved stability, performance with GTID + MTS parallel replication, simpler architecture without a dedicated manager node.

Enhanced data safety via Semi‑Sync that never degrades to async.

Fully automated fault repair and fast node recovery using Xtrabackup.

Continuous updates by the RadonDB team.

References

https://github.com/radondb/xenon/tree/master/docs https://www.fatalerrors.org/a/separation-of-mha-atlas-for-mysql-high-availability.html https://github.com/yoshinorim https://code.google.com/archive/p/mysql-master-ha/ https://dev.mysql.com/doc/refman/5.6/en/replication-gtids-concepts.html