Immutable SQL Server Clusters

Swapping nodes in and out of our SQL clusters is exactly what we did; running in prod with 0 downtime and through an automation framework.

While we were able to automate the building of a cluster to an hour in AWS it was still static and managing it became just like it was in the old hosting providing.

Often databases get forgotten about when it comes to automation and the app servers have all the fun but databases don’t always need to be pets. They can become long living cattle.

This talk will go over the steps we took to enable this automation and what lessons we learnt while doing it during a pressure filled migration from our previous hosting partner to AWS.



John Bowker


I’ve had a varied career in IT spanning the development and infrastructure sides of the industry.

7 years were spent in the online gaming industry where I developed and later lead a team developing online poker rooms. After this I changed industries and countries and went to work for Gameloft in Montreal as their Release Manager. I now have a love/hate relationship with Canadian snow.

After surviving various degrees of coldness I relocated to New Zealand and started working for Xero and lead their DevOps enablement team and later moved into the Product Owner for Data where my teams manage the data infrastructure that runs Beautiful Accounting Software.

Oh and a big fan (not expert just competent) of python after spending many years in the C++, Java and Actionscript world; each day I use Flask is a delight.

John is the Product Owner for Data Infrastructure at Xero, a cloud-based accounting software company with over a million subscribers.

John’s team builds and maintains the SQL Server infrastructure and tooling that Xero’s applications depend on. John has over 15 years experience in the management of software products and teams in the gaming as well as SaaS product industries.