From: | Peter Geoghegan <pg(at)heroku(dot)com> |
---|---|
To: | Pg Hackers <pgsql-hackers(at)postgresql(dot)org> |
Cc: | Noah Misch <noah(at)leadboat(dot)com> |
Subject: | Re: amcheck (B-Tree integrity checking tool) |
Date: | 2016-10-14 23:56:39 |
Message-ID: | CAM3SWZSa+BD=Pg1HpKXmRMUB-egVfO1GQPt8sP58YOeW0ewPwQ@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Mon, Feb 29, 2016 at 4:09 PM, Peter Geoghegan <pg(at)heroku(dot)com> wrote:
> To recap, the extension adds some SQL-callable functions that verify
> certain invariant conditions hold within some particular B-Tree index.
> These are the conditions that index scans rely on always being true.
> The tool's scope may eventually cover other AMs, including heapam, but
> nbtree seems like the best place to start.
Noah and I discussed possible future directions for amcheck in person
recently. I would like to get Noah's thoughts again here on how a tool
like amcheck might reasonably target other access methods for
verification. In particular, the heapam. MultiXacts were mentioned as
a structure that could receive verification in a future iteration of
this tool, but I lack expertise there.
I've placed a lot of emphasis on the importance of having a
low-overhead verification process, particularly in terms of the
strength of heavyweight lock that the verification process requires.
Ideally, it would be possible to run any new verification process in a
fairly indiscriminate way with only limited impact on live production
systems.
--
Peter Geoghegan
From | Date | Subject | |
---|---|---|---|
Next Message | Amit Kapila | 2016-10-15 07:43:03 | Re: Speed up Clog Access by increasing CLOG buffers |
Previous Message | Magnus Hagander | 2016-10-14 23:51:58 | Re: pg_basebackup stream xlog to tar |