On Mon, 4 Dec 2006, Heikki Linnakangas wrote: > > o Test WAL replay more thoroughly. > > I've had that problem too with a lot of things I've hacked. I've used a > shell script that does the operation under test, runs a select, kills > and restarts postmaster, and reruns the select. If the select after > crash returns the same result as before, presumably WAL code works. But > you need to watch out for full page writes that might mask bugs in the > redo code. > > Anyone have a more sophisticated method?
Well, I've done a combination of what you did and replaying a bunch of operations using PITR. Thanks, Gavin ---------------------------(end of broadcast)--------------------------- TIP 9: In versions below 8.0, the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match