Josh Berkus <josh@agliodbs.com> writes:
> Hmmm. As someone who often deploys pg.conf changes as part of a
> production code rollout, I actually like the "atomic" nature of updating
> postgresql.conf -- that is, all your changes succeed, or they all fail.
If we actually *had* that, I'd agree with you. The problem is that it
appears that we have such a behavior, but it fails to work that way in
corner cases. My proposal is aimed at making the corner cases less
corner-y, by adopting a uniform rule that each backend adopts all the
changes it can.
regards, tom lane