It’s another T-SQL Tuesday. Thanks to Wayne Sheffield (b|t) for hosting this month. His T-SQL Tuesday challenge is to write about Brick Walls we have faced. (If you are unfamiliar with the T-SQL Tuesday party, check out the website for the full backstory.)
It took me a bit to figure out what to write about for this topic. But that’s not the type of brick wall I want to post about today.
The Brick Wall:
I was working with a highly transactional system. The table with the most writes was also the table with the most reads. To add to the problem, the data being updated the most was the same data we need to read the most. The next set of popular tables, i.e. the next most read and updated tables, had triggers that used the most popular table for some of the data needed. As you can guess, we had a lot of deadlocks in our system. Oh, did I mention that this was running SQL Server 2000?