Horton

Re: [COMMITTERS] pgsql: Fix freezing of a dead HOT-updated tuple

Alvaro Herrera <alvherre@alvh.no-ip.org>

View thread View on postgr.es
From: Alvaro Herrera <alvherre@alvh.no-ip.org>
To: Peter Geoghegan <pg@bowt.ie>
Cc: "Wood, Dan" <hexpert@amazon.com>, Michael Paquier <michael.paquier@gmail.com>, "Wong, Yi Wen" <yiwong@amazon.com>, PostgreSQL Hackers <pgsql-hackers@postgresql.org>
Date: 2017-10-06T17:49:56Z
Lists: pgsql-hackers 
Peter Geoghegan wrote:
> On Fri, Oct 6, 2017 at 6:18 AM, Alvaro Herrera <alvherre@alvh.no-ip.org> wrote:
> > Wood, Dan wrote:
> >> Yes, I’ve been testing 9.6.  I’ll try Alvaro’s patch today.
> >>
> >> I would prefer to focus on either latest 9X or 11dev.
> >
> > I tested my patch on 9.4 and 9.5 today and it seems to close the problem
> > (with the patch, I waited 10x as many iterations as it took for the
> > problem to occur ~10 times without the patch), but I can reproduce a
> > problem in 9.6 with my patch installed.  There must be something new in
> > 9.6 that is causing the problem to reappear.
> 
> What problem persists? The original one (or, at least, the original
> symptom of pruning HOT chains incorrectly)? If that's what you mean, I
> wouldn't be so quick to assume that it's the freeze map.

I can tell that, in 9.6, REINDEX still reports the error we saw in
earlier releases, after some of the runs of my reproducer scripts.  I'm
unable to reproduce it anymore in 9.3 to 9.5.  I can't see the one Dan
originally reported anywhere, either.

I don't know if it's really the freeze map at fault or something else.

-- 
Álvaro Herrera                https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

Commits

  1. Backport addition of rs_old_rel to rewriteheap's state.

  2. Perform a lot more sanity checks when freezing tuples.

  3. Revert bogus fixes of HOT-freezing bug

  4. Fix traversal of half-frozen update chains

  5. Fix freezing of a dead HOT-updated tuple

  6. During index build, check and elog (not just Assert) for broken HOT chain.

  7. Fix WAL replay of locking an updated tuple

  8. Change the way we mark tuples as frozen.