Re: Stefan's bug (was: max_standby_delay considered harmful) - Mailing list pgsql-hackers

From Fujii Masao
Subject Re: Stefan's bug (was: max_standby_delay considered harmful)
Date
Msg-id AANLkTikWt21uZelewDV4CMFUiE-a-BnsNdKPYEsFjhIZ@mail.gmail.com
Whole thread Raw
In response to Re: Stefan's bug (was: max_standby_delay considered harmful)  (Robert Haas <robertmhaas@gmail.com>)
Responses Re: Stefan's bug (was: max_standby_delay considered harmful)
Re: Stefan's bug (was: max_standby_delay considered harmful)
List pgsql-hackers
On Mon, May 17, 2010 at 10:20 PM, Robert Haas <robertmhaas@gmail.com> wrote:
> OK, I think I understand now.  But, the SIGTERM sent by the postmaster
> doesn't kill the recovery process unconditionally.  It will invoke
> StartupProcShutdownHandler(), which will set set shutdown_requested =
> true.  That gets checked by RestoreArchivedFile() and
> HandleStartupProcInterrupts(), and I think that neither of those can
> get invoked until after the control file has been updated.  Do you see
> a way it can happen?

Yeah, the way is:
StartupXLOG() --> ReadCheckpointRecord() --> ReadRecord() -->
XLogPageRead() --> XLogFileReadAnyTLI() --> XLogFileRead() -->
RestoreArchivedFile()

ReadCheckpointRecord() is called before pg_control is updated.


ISTM that walreceiver might be invoked even after shutdown is requested.
We should prevent the postmaster from starting up walreceiver if
Shutdown > NoShutdown?

Regards,

--
Fujii Masao
NIPPON TELEGRAPH AND TELEPHONE CORPORATION
NTT Open Source Software Center


pgsql-hackers by date:

Previous
From: Robert Haas
Date:
Subject: Re: Performance problem in textanycat/anytextcat
Next
From: Fujii Masao
Date:
Subject: Re: Stefan's bug (was: max_standby_delay considered harmful)