In this tutorial, we'll walk through over several scenarios where things can get complex with your Percona XtraDB Cluster (PXC)/Galera installation. We will break things in these scenarios and then you'll have a chance to diagnose and fix them. Learn some of the most complex troubleshooting techniques.
This tutorial assume that you are somehow familar with the Percona XtraDB Cluster.
When talking about Galera, the main key points are the following:
- Node sync: Boostraping, State Transfer, Node joining/leaving. Data replication.
- Network layer (Maintain the cluster as a whole): Quorum, Primary Component, Cluster Stability.
- Data consistency (Guarantee data is the same): Certification, Flow Control.
Besides the omnipresent topic on databases, and our reason to live: Performance.
All the scenarios will take place on a basic 3-node PXC Cluster
Creation of the VMs and the cluster are available at the Environment link: here
The only sign of a node failure is the loss of connection to the node as seen by other nodes. That's why understanding network in a galera context is key to a happy PXC journeys
Getting the pieces together. You only have a cluster when you have the nodes connected and synced.