Horton

Re: Bug in Physical Replication Slots (at least 9.5)?

Greg Stark <stark@mit.edu>

View thread View on postgr.es
From: Greg Stark <stark@mit.edu>
To: Kyotaro HORIGUCHI <horiguchi.kyotaro@lab.ntt.co.jp>
Cc: Michael Paquier <michael.paquier@gmail.com>, Jonathon Nelson <jdnelson@dyn.com>, PostgreSQL-development <pgsql-hackers@postgresql.org>, PostgreSQL Bugs <pgsql-bugs@postgresql.org>
Date: 2018-01-19T11:28:58Z
Lists: pgsql-bugs, pgsql-hackers 
On 19 January 2017 at 09:37, Kyotaro HORIGUCHI
<horiguchi.kyotaro@lab.ntt.co.jp> wrote:
>
> Though I haven't look closer to how a modification is splitted
> into WAL records. A tuple cannot be so long. As a simple test, I
> observed rechder->xl_tot_len at the end of XLogRecordAssemble
> inserting an about 400KB not-so-compressable string into a text
> column, but I saw a series of many records with shorter than
> several thousand bytes.

I think the case to check is a commit record with many thousands of
subtransactions. I'm not sure you can fill a whole segment though.


-- 
greg

Commits

  1. Fix scenario where streaming standby gets stuck at a continuation record.