Craig,
* Craig Ringer (craig@2ndquadrant.com) wrote:
> There are questions about whether this is a good idea, and there's still
> discussion ongoing. It doesn't look like it's in a state where we can
> confidently say "let's include this for 9.3" to me, but I'd like other
> viewpoints.
For my part, I think the straight-up 'lock_timeout' piece, which is in
the latest patch, is in pretty good shape. It's much less invasive and
provides a capability which other RDBMS's have and is reasonably
straight forward.
> Should we bump this to the next CF? It's clearly still a viable idea,
> just possibly not ready yet.
Unless a committer steps up to take on the statement-level lock-wait
timeout, it's not going to get into 9.3, imv. Waiting to add that
(whatever it ends up being) until post-9.3 makes sense to me, but I'd
hate to miss getting the simpler lock_timeout capability into 9.3.
Thanks,
Stephen