MySQL High Availability at GitHub and Code Refactor vs Rewrite


Details
18:00 - 18:30 - Gathering Food & Drinks
18:30 - 18:50 - Refactor or Rewrite - Properly Disposing of Legacy Code by Roman Labunsky, Software Architect @ datree.io
In this session, we will talk about the challenges we face when working with legacy code, consider our options for dealing with legacy systems while minimising risk and producing value from the get go. We’ll also have a practical guide on how to actually take a legacy component and replace it with a new one.
19:00 - 19:45 - MySQL High Availability at GitHub by Shlomi Noach, Senior Infrastructure Engineer @ GitHub
This session describes the automated MySQL failover setup at GitHub, based on orchestrator, Consul and GLB/HAProxy, and which delivers hands-free high availability of our MySQL clusters. Our setup is tolerant to master, intermediate master and complete DC failures, and recovery makes the best of a situation based on state rather than configuration. We will illustrate our multi-DC orchestrator/raft, Consul & load balancers setups, as well as our production testing setup. We will discuss the choices we've made and the path we took while reducing MySQL outage without introducing false positives.
Want to submit a talk/propose a location? Hit us at meetup@datree.io, let's talk!

MySQL High Availability at GitHub and Code Refactor vs Rewrite