Re: amcheck prototype

From: Peter Geoghegan <pg(at)heroku(dot)com>
To: Pg Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: amcheck prototype
Date: 2015-06-08 06:09:48
Message-ID: CAM3SWZRB=ba1jz408ZdSt32LE==S9GUjYyw3+BpSNr818Mvubg@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Wed, Nov 19, 2014 at 2:09 PM, Peter Geoghegan <pg(at)heroku(dot)com> wrote:
> Attached is a revision of what I previously called btreecheck, which
> is now renamed to amcheck.

This never really went anywhere, because as a project I don't think
that it has very crisp goals. My sense is that it could be developed
in a new direction, with the goal of finding bugs in the master
branch. This seems like something that could be possible without a
large additional effort; committing the tool itself can come later.

Right now, the code that is actually tested by the tool isn't
particularly likely to have bugs. I used a slightly revised version to
constantly verify B-Trees as the regression tests are run. That didn't
catch anything, but since the tool doesn't consult the heap at all I'm
not surprised. Also, I didn't incorporate any testing of recovery with
that stress test.

I wrote amcheck with the assumption that it is useful to have a tool
that verifies several nbtree invariants, a couple of which are fairly
elaborate. amcheck *is* probably useful for detecting corruption due
to hardware failure and so on today, but that is another problem
entirely.

--
Peter Geoghegan

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Michael Paquier 2015-06-08 06:48:51 Memory leak fixes for pg_dump, pg_dumpall, initdb and pg_upgrade
Previous Message Amit Kapila 2015-06-08 05:39:25 Re: Reducing tuple overhead