Re: pg_ctl stop -m immediate on the primary server inflates sequences

From: Martijn van Oosterhout <kleptog(at)svana(dot)org>
To: Boszormenyi Zoltan <zb(at)cybertec(dot)at>
Cc: Bernd Helmle <mailings(at)oopsware(dot)de>, pgsql-hackers(at)postgresql(dot)org, Hans-Juergen Schoenig <hs(at)cybertec(dot)at>
Subject: Re: pg_ctl stop -m immediate on the primary server inflates sequences
Date: 2010-04-10 20:02:15
Message-ID: 20100410200215.GC6944@svana.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Sat, Apr 10, 2010 at 02:36:41PM +0200, Boszormenyi Zoltan wrote:
> >> The above is quite reproducable, "pg_ctl stop -m immediate"
> >> "usually" inflated my serial sequence, but I had two occasions
> >> when not. The 69 -> 70 was one. The inflated increase is always 33:
> >
> > AFAIKS sequences are pre-logged with 32 values to WAL to avoid
> > overhead. I suspect this is why you are seeing those gaps.
>
> Then it should happen all the time, even with "-m fast" or "-m smart", no?

Nope, because on a normal shutdown it writes out the actual value. When
you say "immediate" you mean "right now, don't bother with anything not
important", like for example gaps in sequences. You're essentially
crashing the DB.

Have a ncie day,
--
Martijn van Oosterhout <kleptog(at)svana(dot)org> http://svana.org/kleptog/
> Patriotism is when love of your own people comes first; nationalism,
> when hate for people other than your own comes first.
> - Charles de Gaulle

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Robert Haas 2010-04-11 01:43:28 Re: Virtual Private Database
Previous Message Robert Haas 2010-04-10 19:04:27 Re: extended operator classes vs. type interfaces