MySQL with Glusterfs Storage vs MySQL replication with hostPath storage

MySQLreplication

I am trying to search right road in a project.

We have currently three nodes openshift cluster(okd 3.11) and deployed MySQL with Glusterfs storage. its working fine but look like over engineered to me.

I am thinking to replace above with MySQL Replication with hostpath storage so 1 node be master and other node will act as slave. this way no need for extra steps to deploy glusterfs storage.

We may not extend this cluster it will be three and if we may have to deploy two node OKD cluster MySql replications will be ideal choice as Glusterfs dont recommend to use less then 3 nodes.

Main goal is HA so graceful failover can be achieved. Any thoughts?

Best Answer

The minimum for an HA setup is 3 nodes, geographically distributed.

If you put all the nodes in a single data center, you are vulnerable to a datacenter crash. (They do happen.)
If you use Master + Slave(s), failover is not well automated. See MHA as a possible alternative.
With Galera cluster, you can write to all nodes, thereby helping minimize downtime. (I don't know about Glusterfs.)

Related Solutions

Mysql – Replication in thesql with many servers

If your Servers get out of sync they play back the events from the binlog, so as long you have all statements writen into your binlog, the master/slave is able to catch up. I had to catch up 12days without any problems, but in a cluster this is not what you want to happen. And be careful when using "unsafe" statements, this are statements which the server can't reproduce on your master/slave setup.

Mysql Article on safe/unsafe statements - http://dev.mysql.com/doc/refman/5.1/en/replication-rbr-safe-unsafe.html

MySQL Cluster – Management Node Redundancy

If you have split exactly in half brain situation MySQL Cluster resolves it designating one node as master. If you look into ndb_mgm output below you will notice asterisk at the end of node 10 line. It marks master.

It means that if your cluster will be split in two equal parts the half which contains master will continue running. Nodes from the other half will be shut down.

ndb_mgm> show
Connected to Management Server at: 10.1.1.10:1186
Cluster Configuration
---------------------
[ndbd(NDB)]     2 node(s)
id=10   @10.1.1.20  (mysql-5.6.24 ndb-7.3.9, Nodegroup: 0, *)
id=11   @10.1.1.21  (mysql-5.6.24 ndb-7.3.9, Nodegroup: 0)

Best Answer

Related Solutions

Mysql – Replication in thesql with many servers

MySQL Cluster – Management Node Redundancy

Related Question