One of the more unique and complicated areas of devops is dealing with database updates, especially those for databases with pre-defined schemas like relational databases. Databases generally:
- Need to stay online during updates
- Should not be destroyed and re-created
- Frequently have changes that can be long running and have impacts to the running application code
- Have implied or explicit dependencies between the database schema and application code or across shards in a sharded database
- Still have the same requirements as application code for being traceably updated
- And should be updated in an automated fashion as part of your devops strategy
This talk will cover tools and strategies the speaker has identified in over 15 years of working with automated database migrations for making sure your persistence stores are smoothly integrated into your devops workflow. We will cover:
- Handling database updates via database migration tools: how to get started and what in the speaker's experience is and is not important in a good migration tool
- Patterns, and tips and tricks, for addressing common issues with database migrations
- Workflow processes for developers, DBAs, and operations teams that enable a smooth, controlled database migration process as part of devops
- How automating your database migrations benefits database administrators