Horton

Re: pgsql: Store 2PC GID in commit/abort WAL recs for logical decoding

Nikhil Sontakke <nikhils@2ndquadrant.com>

View thread View on postgr.es
From: Nikhil Sontakke <nikhils@2ndquadrant.com>
To: Alvaro Herrera <alvherre@2ndquadrant.com>
Cc: Heikki Linnakangas <hlinnaka@iki.fi>, Michael Paquier <michael@paquier.xyz>, Simon Riggs <simon@2ndquadrant.com>, pgsql-hackers <pgsql-hackers@postgresql.org>
Date: 2018-06-17T08:34:26Z
Lists: pgsql-hackers 
Hi Alvaro,


>> There was a slight oversight in the twophase_gid length calculation in
>> the XactLogCommitRecord() code path in the cf5a1890592 commit. The
>> corresponding XactLogAbortRecord() code path was ok. PFA, a small
>> patch to fix the oversight.
>
> Forgot to add: maybe it would be useful to have tests in core where
> these omissions become evident.  Do you have some?
>
Thanks for the commit.

I do have some tests. They are part of the "logical decoding of 2PC"
patch which adds the needed infrastructure to *actually* use this code
present in the core as of now. I am going to submit it in the upcoming
commitfest.

Regards,
Nikhils
-- 
 Nikhil Sontakke                   http://www.2ndQuadrant.com/
 PostgreSQL Development, 24x7 Support, Training & Services

Commits

  1. Fix off-by-one bug in XactLogCommitRecord

  2. Fix confusion on the padding of GIDs in on commit and abort records.

  3. Store 2PC GID in commit/abort WAL recs for logical decoding