Hey everyone! I'm trying to set up a new RDS instance from a snapshot and I want to make sure the EBS volume is prepped to avoid any speed bumps during the first read. Right now, I've got a script that pulls all data from every table, but running that takes over 24 hours because our database is huge—over 15TB! Is this how people usually handle the hydration process, or are there better methods out there? Thanks for any advice!
2 Answers
Hydrating a database is tricky. A `SELECT *` runs a full table scan, which does hydrate your table blocks, but you might miss out on fully hydrating the indexes. They only get hydrated if they’re close enough to those data blocks, so if you end up hitting an un-hydrated block while scanning the index, you might notice a performance drop. I don’t have a specific solution, but just doing a `SELECT *` might not give you the 'like new' performance you're hoping for. If you really want good index performance, rebuilding the indexes could help, but you'll need to assess how feasible that is without affecting your DB's operation.
Your timing seems pretty normal. I manage a couple of 1TB databases and during our DR tests, it usually takes about 90 minutes for the select all to finish up.
Thanks for chiming in! I guess I just expected there would be a more efficient solution for this common issue.
That’s a solid point! It’s frustrating that there doesn’t seem to be a straightforward fix for this.