Qlik Replicate - Galera Clustering

[MichaelMockus]

Qlik Replicate Release notes for 6.6 mentionsupport for Galera Clustering and page 377 of the Replication Setup Guide mentions setting one setting: log_slave_updates. Could you please share were I can find architecture diagrams of how it is support and the how Replicate behaves in the various states of a failure and failover in a Galera cluster and any of the best practices on how to achieve HA and DR with the tool when using a Galera MySQL cluster as a source.

[Ola_Mayer · Accepted solution]

Hello Michael,

The behavior of Galera cluster is competently transparent to Replicate. We see all the nodes in the cluster as one connection point, as if it is a single node. In our lab we tested shutting down nodes in the cluster, but as long as at least one of the nodes stayed operational, Replicate wasn't impacted at all.

For instructions on how Galera handles nodes failure and fail over I recommend to review Galera documentation.

[MichaelMockus]

When connected as the instructions state

"To be able to replicate clustered (NDB) tables (i.e. by connecting Qlik Replicate to any of the cluster nodes), the following parameters must be configured in MySQL's my.ini (Windows) or my.cnf (UNIX) files"
Replicate does NOT fail over at all and simply stops to replicate when a node fails.
More verbose information on configuration and requirements would e greatly appreciated or best practices for leveraging Replicate on a Galera cluster.