go-sakila-remora

A sidecar process written in Go to monitor MySQL slaves / masters for replication errors or lag.

Welcome!

Hi! and thanks for stopping by this project in it's early days.

Inspired by GitHub's interesting article: https://githubengineering.com/context-aware-mysql-pools-via-haproxy/

It's similar to methods I've used previously to manage traffic being sent to Asterisk VoIP servers based on various metrics.

Project goals

Ultimately, I'm pretty new to Go and thought this would be an interesting challenge and an awesome product if everything turns out well.

Phase 1:

The initial iteration will handle reading YAML configuration, executing some basic checks against MySQL and presenting a JSON based endpoint for health checking against. After this iteration I should have a decent BVP - Barely Viable Product.

Phase 2:

Let's not get ahead of ourselves.

Percona master / slave for testing

I needed a really quick way to be able to:

• Spin up a master and slave with replication configured
• Create lag on the slave
• Stop lag on the slave

To do this I've thrown together a really dirty docker-compose file and a Makefile with the following targets:

Target	Description
make up	Brings up the master, slave and sets up replication
make down	Runs docker-compose down cleaning everything up
make start-lag	Uses MASTER_DELAY=300 to start the slave lagging
make stop-lag	Stops slave lag by setting MASTER_DELAY=0
make slave-lag	Shows the current value of Seconds_Behind_Master
make stuff	Does some SQL stuff - a CREATE and DROP of a database

MySQL Ports

When running the containers you can access MySQL on host: 127.0.0.1 with:

• User: root
• Pass: secret
• Master Port: 3306
• Slave Port: 3307

Obviously: Please! Please! Please! Don't ever use these settings or config In production (or any other application environment). This is for quick and dirty testing and development of go-sakila-remora.

Simulating slave lag

To simulate slave lag for testing is super easy:

1. make up to bring up a master / slave and configure replication
2. make start-lag to start the slave lagging (up to 300s)
3. make stuff to do some SQL stuff and things
4. make slave-lag a few times to see the lag increasing
5. make stop-lag stops the slave lagging