Re: Make mesage at end-of-recovery less scary.

From: Kyotaro Horiguchi <horikyota(dot)ntt(at)gmail(dot)com>
To: vignesh21(at)gmail(dot)com
Cc: aleksander(at)timescale(dot)com, pgsql-hackers(at)lists(dot)postgresql(dot)org, stark(dot)cfm(at)gmail(dot)com, alvherre(at)alvh(dot)no-ip(dot)org, pryzby(at)telsasoft(dot)com, jchampion(at)timescale(dot)com, andres(at)anarazel(dot)de, ashu(dot)coek88(at)gmail(dot)com, pashkin(dot)elfe(at)gmail(dot)com, michael(at)paquier(dot)xyz, bossartn(at)amazon(dot)com, david(at)pgmasters(dot)net, peter(dot)eisentraut(at)2ndquadrant(dot)com, jtc331(at)gmail(dot)com, robertmhaas(at)gmail(dot)com
Subject: Re: Make mesage at end-of-recovery less scary.
Date: 2024-01-11 07:18:16
Message-ID: 20240111.161816.1259268425616323411.horikyota.ntt@gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

At Fri, 5 Jan 2024 16:02:24 +0530, vignesh C <vignesh21(at)gmail(dot)com> wrote in
> On Wed, 22 Nov 2023 at 13:01, Kyotaro Horiguchi <horikyota(dot)ntt(at)gmail(dot)com> wrote:
> >
> > Anyway, this requires rebsaing, and done.
>
> Few tests are failing at [1], kindly post an updated patch:

Thanks!

The errors occurred in a part of the tests for end-of-WAL detection
added in the master branch. These failures were primarily due to
changes in the message contents introduced by this patch. During the
revision, I discovered an issue with the handling of empty pages that
appear in the middle of reading continuation records. In the previous
version, such empty pages were mistakenly identified as indicating a
clean end-of-WAL (that is a LOG). However, they should actually be
handled as a WARNING, since the record curently being read is broken
at the empty pages. The following changes have been made in this
version:

1. Adjusting the test to align with the error message changes
introduced by this patch.

2. Adding tests for the newly added messages.

3. Correcting the handling of empty pages encountered during the
reading of continuation records. (XLogReaderValidatePageHeader)

4. Revising code comments.

5. Changing the term "log segment" to "WAL
segment". (XLogReaderValidatePageHeader)

regards.

--
Kyotaro Horiguchi
NTT Open Source Software Center

Attachment Content-Type Size
v28-0001-Make-End-Of-Recovery-error-less-scary.patch text/x-patch 23.7 KB

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Bertrand Drouvot 2024-01-11 07:49:47 Re: Synchronizing slots from primary to standby
Previous Message Umair Shahid 2024-01-11 06:54:21 Re: Update docs for default value of fdw_tuple_cost