Horton

Re: Don't choke on files that are removed while pg_rewind runs.

Michael Paquier <michael@paquier.xyz>

View thread View on postgr.es
From: Michael Paquier <michael@paquier.xyz>
To: Masahiko Sawada <masahiko.sawada@2ndquadrant.com>
Cc: Daniel Gustafsson <daniel@yesql.se>, Justin Pryzby <pryzby@telsasoft.com>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>, Heikki Linnakangas <heikki.linnakangas@iki.fi>, Fujii Masao <masao.fujii@oss.nttdata.com>
Date: 2020-07-13T07:56:29Z
Lists: pgsql-hackers

Attachments

On Mon, Jul 13, 2020 at 03:59:56PM +0900, Masahiko Sawada wrote:
> On Mon, 13 Jul 2020 at 15:34, Daniel Gustafsson <daniel@yesql.se> wrote:
>> Yeah, I agree with that, seems like the call should've been PQgetisnull(res, i, 1);
>> to match the loop.
> 
> +1

Good catch, Justin.  There is a second thing here.  The second column
matches with the file size, so if its value is NULL then atol() would
just crash first.  I think that it would be more simple to first check
if the file size is NULL (isdir and link_target would be also NULL,
but just checking for the file size is fine), and then assign the
result values, like in the attached.  Any thoughts?
--
Michael

Commits

  1. Fix handling of missing files when using pg_rewind with online source

  2. Don't choke on files that are removed while pg_rewind runs.