Horton

Re: BUG #17103: WAL segments are not removed after exceeding max_slot_wal_keep_size

Kyotaro Horiguchi <horikyota.ntt@gmail.com>

View thread View on postgr.es
From: Kyotaro Horiguchi <horikyota.ntt@gmail.com>
To: alvherre@alvh.no-ip.org
Cc: tgl@sss.pgh.pa.us, jeff.janes@gmail.com, mk@071.ovh, pgsql-bugs@lists.postgresql.org
Date: 2021-08-02T06:00:39Z
Lists: pgsql-bugs 
At Thu, 29 Jul 2021 17:19:24 -0400, Alvaro Herrera <alvherre@alvh.no-ip.org> wrote in 
> On 2021-Jul-29, Kyotaro Horiguchi wrote:
> 
> > At Thu, 29 Jul 2021 16:20:38 +0900 (JST), Kyotaro Horiguchi <horikyota.ntt@gmail.com> wrote in 
> > > So it seems to me we need to explicitly prevent unexpected checkpoints
> > > from happening maybe by enlarging max_wal_size temporily.
> > 
> > I ended up with the attached. It causes a checkpoint reliably exactly
> > at the aimed timing without.
> 
> Thanks for looking into it.  The explanation makes sense, so I pushed
> your patch and also fixed one outdated comment I noticed while reading
> the related code.  I hope this commit will solve the problem ... it is
> quite low probability, so we'll have to wait at least two weeks to make
> any conclusion.

Agreed. I hope this is the last failure..

Thanks.

-- 
Kyotaro Horiguchi
NTT Open Source Software Center

Commits

  1. Close yet another race condition in replication slot test code

  2. Make new replication slot test code even less racy

  3. Make new replication slot test code less racy

  4. Advance old-segment horizon properly after slot invalidation