Enclose latest version of Hot Standby. This is the "basic" patch, with
no must-fix issues and no known bugs. Further additions will follow
after commit, primarily around usability, which will include additional
control functions for use in testing. Various thoughts discussed here
http://wiki.postgresql.org/wiki/Hot_Standby_TODO
Patch now includes a set of regression tests that can be run against a
standby server using "make standbycheck" (requires setup, see
src/test/regress/standby_schedule). This helps make explicit in code
which commands and variants are allowed and disallowed in hot standby.
Complete with full docs and comments.
Barring resolving a few points and subject to even more testing, this is
the version I expect to commit to CVS on Wednesday. I would appreciate
further objective testing before commit, if possible.
Last remaining points
* NonTransactionalInvalidation logging has been removed following
review, but AFAICS that means VACUUM FULL doesn't work correctly on
catalog tables, which regrettably will be the only ones still standing
even after we apply VFI patch. Did I misunderstand the original intent?
Was it just buggy somehow? Or is this hoping VF goes completely, which
seems unlikely in this release. Just noticed this, decided better to get
stuff out there now.
* Are we OK with using hash indexes in standby plans, even when we know
the indexes are stale and could return incorrect results?
Software also available via git on the repo used by Heikki and myself.
ssh://git@git.postgresql.org/users/heikki/postgres.git
branch = hs-riggs
Patch prepared using from my private repo using
git diff pgsql/master..hs-simon | filterdiff --format=context
so please look out for any weirdness that might introduce.
--
Simon Riggs www.2ndQuadrant.com
