Horton

Re: straightening out backend process startup

Andres Freund <andres@anarazel.de>

View thread View on postgr.es
From: Andres Freund <andres@anarazel.de>
To: Robert Haas <robertmhaas@gmail.com>, Kyotaro HORIGUCHI <horiguchi.kyotaro@lab.ntt.co.jp>
Cc: "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
Date: 2021-08-05T19:39:17Z
Lists: pgsql-hackers 
Hi,

Thanks Robert, Horiguchi-san for looking.

On 2021-08-04 16:34:52 -0400, Robert Haas wrote:
> On Mon, Aug 2, 2021 at 12:41 PM Andres Freund <andres@anarazel.de> wrote:
> > - AuxiliaryProcessMain() is used for two independent tasks: Start bootstrap /
> >   checker mode and starting auxiliary processes. In HEAD there's maybe 5 lines
> >   out 250 that are actually common to both uses.
> >
> >   A related oddity is that we reserve shared memory resources for bootstrap &
> >   checker aux processes, despite those never existing.
> >
> >   This is addressed in patches 1-7
>
> This all looks pretty mechanical and, I would guess, not very controversial.

Pushed these patches.


> > - The order of invocation of InitProcess()/InitAuxiliaryProcess() and
> >   BaseInit() depends on EXEC_BACKEND. Due to that there often is no single
> >   place initialization code can be put if it needs any locks.
> >
> >   This is addressed in patches 8-9
> >
> > - PostgresMain() has code for single user and multi user interleaved, making
> >   it unnecessarily hard to understand what's going on.
> >
> >   This is addressed in patches 10
>
> This stuff I'd need to study more in order to have an intelligent opinion.

Unless somebody protests soon I plan to push at least the
"process startup: Always call Init[Auxiliary]Process() before BaseInit()."
portion, as the inconsistent order between EXEC_BACKEND/!EB is making life
hard for me in other patches.

I don't have a dependency on
"process startup: Split single user code out of PostgresMain()."
but it does make the code a good bit more readable imo...

Greetings,

Andres Freund

Commits

  1. process startup: Split single user code out of PostgresMain().

  2. process startup: Do InitProcess() at the same time regardless of EXEC_BACKEND.

  3. process startup: Initialize PgStartTime earlier in single user mode.

  4. pgstat: Schedule per-backend pgstat shutdown via before_shmem_exit().

  5. pgstat: Bring up pgstat in BaseInit() to fix uninitialized use of pgstat by AV.

  6. process startup: Always call Init[Auxiliary]Process() before BaseInit().

  7. process startup: Centralize pgwin32_signal_initialize() calls.

  8. process startup: Remove bootstrap / checker modes from AuxProcType.

  9. process startup: Move AuxiliaryProcessMain into its own file.

  10. process startup: auxprocess: reindent block

  11. process startup: Separate out BootstrapModeMain from AuxiliaryProcessMain.

  12. process startup: Rename postmaster's --forkboot to --forkaux.