This talk will cover what a few Oracle experts (but MySQL newbies) learned in a recent migration at NPR.org. It discusses how we separated our previously monolithic database into smaller, more functionally cohesive databases and some of the criteria we used to decide what tables ended up where. This architecture allowed for better load balancing and fault tolerance but it was outside our budget before we move to MySQL. There will be a discussion of some specific differences in SQL to be aware of between Oracle and MySQL. My team put together a cheat sheet of changes in function names or signatures that I will share. We also found equivalents for sequences and Oracle Text indexes and other Oracle specific functionality in MySQL. A few free tools we found helpful when writing code and debugging issues will be demoed and Ill explain some differences between how these tools work verses some of the Oracle tools you might be used to. Finally Ill describe a few of the test cases we used that revealed issues with our MySQL server configuration and our own code. This will touch on some of the challenges we encountered with concurrency and encoding.
