Horton

Re: Two pg_rewind patches (auto generate recovery conf and ensure clean shutdown)

Alvaro Herrera <alvherre@2ndquadrant.com>

View thread View on postgr.es
From: Alvaro Herrera <alvherre@2ndquadrant.com>
To: Paul Guo <pguo@pivotal.io>
Cc: Michael Paquier <michael@paquier.xyz>, Thomas Munro <thomas.munro@gmail.com>, pgsql-hackers <pgsql-hackers@postgresql.org>, Jimmy Yih <jyih@pivotal.io>, Ashwin Agrawal <aagrawal@pivotal.io>, Alexey Kondratov <a.kondratov@postgrespro.ru>
Date: 2019-09-25T17:48:12Z
Lists: pgsql-hackers 
CC Alexey for reasons that become clear below.

On 2019-Sep-25, Paul Guo wrote:

> > Question about 0003.  If we specify --skip-clean-shutdown and the cluter
> > was not cleanly shut down, shouldn't we error out instead of trying to
> > press on?
> 
> pg_rewind would error out in this case, see sanityChecks().
> Users are expected to clean up themselves if they use this argument.

Ah, good.  We should have a comment about that below the relevant
stanza, I suggest.  (Or maybe in the same comment that ends in line
272).

I pushed 0001 with a few tweaks.  Nothing really substantial, just my
OCD that doesn't leave me alone ... but this means your subsequent
patches need to be adjusted.  One thing is that that patch touched
pg_rewind for no reason (those changes should have been in 0002) --
dropped those.

Another thing in 0002 is that you're adding a "-R" switch to pg_rewind,
but we have another patch in the commitfest using the same switch for a
different purpose.  Maybe you guys need to get to an agreement over who
uses the letter :-)  Also, it would be super helpful if you review
Alexey's patch: https://commitfest.postgresql.org/24/1849/


This line is far too long:

+   printf(_("  -s, --skip-clean-shutdown      skip running single-mode postgres if needed to make sure target is clean shutdown\n"));

Can we make the description more concise?

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

Commits

  1. Improve test coverage of pg_rewind

  2. Improve handling and coverage of --no-ensure-shutdown in pg_rewind

  3. Fix issues in pg_rewind with --no-ensure-shutdown/--write-recovery-conf

  4. Fix --dry-run mode of pg_rewind

  5. pg_rewind: test new --write-recovery-conf functionality

  6. pg_rewind: Allow writing recovery configuration

  7. Have pg_rewind run crash recovery before rewinding

  8. Split out recovery confing-writing code from pg_basebackup

  9. Switch TAP tests of pg_rewind to use non-superuser role, take two