At times, you simply outgrow your old database system, like you can outgrow a pair of old jeans – it has served you well for many years, but now it is simply holding you back. Our client was using an 8 yr old MySQL table with 120 million records as the foundation of their application. They needed to migrate to PostgreSQL for scalable growth and were not allowed to experience downtime since their customers relied on that table 24 hours a day. How do you pull off a migration of this scale and not have anyone miss a beat? Spoiler alert...you do not simply turn off the lights and cross your fingers!
Knackforge set up an Amazon Aurora PostgreSQL cluster with a partitioned table approach based on the year of the record as the partition key.
To facilitate a faster migration and not completely saturate the application’s network bandwidth, we used eight AWS Data Migration Service (DMS) tasks to filter records by year.
With Knackforge Cloud Services in place, the customer experienced: