Horton

Re: Possible bug in logical replication.

Michael Paquier <michael@paquier.xyz>

View thread View on postgr.es
From: Michael Paquier <michael@paquier.xyz>
To: Kyotaro HORIGUCHI <horiguchi.kyotaro@lab.ntt.co.jp>
Cc: alvherre@2ndquadrant.com, simon@2ndquadrant.com, a.sher@postgrespro.ru, sawada.mshk@gmail.com, k.knizhnik@postgrespro.ru, pgsql-hackers@lists.postgresql.org
Date: 2018-06-22T09:39:54Z
Lists: pgsql-hackers 
On Fri, Jun 22, 2018 at 04:33:12PM +0900, Kyotaro HORIGUCHI wrote:
> pg_advance_replication_slots can advance uninitialized physical
> slots and that might not be good. (Logical slots always have
> initial invalid values in thw two lsn columns.)

The current logic is careful that users willing to move to a new
position cannot move in the future, but the logic is visibly wanted to
accept past values.  Petr, what do you think?  KeepLogSeg() won't return
negative values so some applications may take advantage of that.  Or
should advancing be simply disabled for non-initialized slots? 
--
Michael

Commits

  1. Rewrite comments in replication slot advance implementation

  2. Fix a couple of bugs with replication slot advancing feature

  3. Ability to advance replication slots