Re: vacuum, performance, and MVCC

Mark Woodward wrote:

> The update behavior of PostgreSQL is probably the *last* serious issue.
> Debate all you want, vacuum mitigates the problem to varying levels,
> fixing the problem will be a huge win. If the update behavior gets fixed,
> I can't think of a single issue with postgresql that would be a show
> stopper.

Nah, it's just *your* pet peeve.  Everyone has theirs.  Some people may
share yours, of course.  I agree it's a problem, but from there to
saying "it's _the last_ issue" there's a lot of distance.


Your idea of reusing a tuple's self pointer (t_ctid) does not work BTW,
because the self pointer must point to self.  The case where the pointer
does not point to exactly the same tuple, it must point to a newer
version.  If you change that invariant, a lot of things break; see for
example heap_get_lastest_tid.

-- 
Alvaro Herrera                                http://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.