Skip to content

Comment on A race condition in Aurora RDSparent

Comments

Yeah I agree, this seems like a pretty critical feature to the Aurora product itself. We saw a similar behavior recently where we had a connection pooler in between which indicates something wrong with how they propagate DNS changes during the failover. wtf aws

Whenever we have to do any type of AWS Aurora or RDS cluster modification in prod we always have the entire emergency response crew standing by right outside the door.

Their docs are not good and things frequently don't behave how you expect them to.

Oh, well, it’s always DNS!

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.