Database schema migrationin highly available services
Abstract: Since the rise of DevOps, companies have been trying to provide rapid and agile cycles on the development of new products, reducing time and costs to go from definition to production. However, whenever a MySQL system with a zero-downtime requirement is found, the database deployment cycle is still manual, leaving it less evolved than other parts of the development cycle, especially when applying schema updates without affecting the availability of a MySQL system (Database Online Schema migrations) on Continuous Integration and Continuous Delivery pipelines. As a consequence, it delays the process. There is a strong belief that the different affected technologies are ready to automatize Online Schema migrations in production. In this thesis, we verify the viability and study the reliability of automatizing Online Schemamigrations with a zero-downtime requirement. We achieved this by testing the most common errors in a setup without error detection mechanisms and then testing our proposed design. We conclude that is viable and reliable to apply these techniques and strategies for providing Continuous Delivery. Our design prevents most common errors when updating SQL schemas online with some inherent implications and flaws in MySQL systems.
AT THIS PAGE YOU CAN DOWNLOAD THE WHOLE ESSAY. (follow the link to the next page)