MIME-Version: 1.0
References: 
 <CAHGQGwFzNYroAxSoyJhqTU-pH=t4Ej6RyvhVmBZ91Exj_TPMMQ@mail.gmail.com>
 <CAA4eK1+CrQNqiPDKv1wYfdkbX0FARJoi1=0ioaAqkLzbq2vG1w@mail.gmail.com>
 <CAHGQGwHABvuCoyM24HUiFZ5oJq_CoFomjt_cqD-0cJLMjFXJjQ@mail.gmail.com>
 <CAA4eK1LzZGfRANPAnv6NpKCH2ENuZO6HswgY14A=xsOXmucPhw@mail.gmail.com>
 <CABdArM6nepct0uxizCnZqy-kAjjTOndvu7bWtaNcmknuxx82Hg@mail.gmail.com>
 <CAJpy0uBhJB4HqouLXegD=miSkZfZp86tTC2K0K6nf=bHCcsPLQ@mail.gmail.com>
 <CABdArM6Mfhk2+9TVR_D3cgfPWPfHDuZEg7MOc5KqULwt0OcQUg@mail.gmail.com>
 <CAA4eK1+d2vN80-Yvy_Hr=ATF3XL5db+_W-sXF=2Vxm+OFBO82w@mail.gmail.com>
 <CAA4eK1KxU4b53GUor41A55x+Bx-DdOaQ9g1DqURyRY2Cg-hhPQ@mail.gmail.com>
 <CABdArM6pmn5yFqiU33KTYBXYM=Vny2ULnJY_gqFbsMEdt+1dPA@mail.gmail.com>
 <CAHGQGwEsEjXbCthXb02=HScBW7C=BHmygv9SK1VGCcc-9bmTsw@mail.gmail.com>
 <CABdArM5rrhSmFvVL4C5LL0iea-R0HRtB=ZvD=ereoTDa1Tm=NA@mail.gmail.com>
 <CAJpy0uD5brcAOpM8+UGfpzrm4pu-vsSfQUT2zgmcKsXEb8++4Q@mail.gmail.com>
 <CABdArM6E2aBj0FUjr2870sZUQqc6mpV1xSM=sqUFd2tKbfUQHA@mail.gmail.com>
 <TY4PR01MB16907318494FF276465E97DF79453A@TY4PR01MB16907.jpnprd01.prod.outlook.com>
 <CABdArM5RqLna+Y3CF=0BKew-WyxoAKXgcaOnXG1DZpMEiogCqw@mail.gmail.com>
In-Reply-To: 
 <CABdArM5RqLna+Y3CF=0BKew-WyxoAKXgcaOnXG1DZpMEiogCqw@mail.gmail.com>
From: Fujii Masao <masao.fujii@gmail.com>
Date: Wed, 1 Apr 2026 18:48:59 +0900
Message-ID: 
 <CAHGQGwF1=zu478YX1kNaX+qaNp9DH_p+5ZdSfFdTa=JuynJufQ@mail.gmail.com>
Subject: Re: Use SIGTERM instead of SIGUSR1 for slotsync worker to exit during
 promotion?
To: Nisha Moond <nisha.moond412@gmail.com>
Cc: "Zhijie Hou (Fujitsu)" <houzj.fnst@fujitsu.com>,
 shveta malik <shveta.malik@gmail.com>,
	Amit Kapila <amit.kapila16@gmail.com>,
	PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Archived-At: 
 <https://www.postgresql.org/message-id/CAHGQGwF1%3Dzu478YX1kNaX%2BqaNp9DH_p%2B5ZdSfFdTa%3DJuynJufQ%40mail.gmail.com>
Precedence: bulk

On Wed, Apr 1, 2026 at 2:34=E2=80=AFPM Nisha Moond <nisha.moond412@gmail.co=
m> wrote:
>
> On Tue, Mar 31, 2026 at 3:56=E2=80=AFPM Zhijie Hou (Fujitsu)
> <houzj.fnst@fujitsu.com> wrote:
> >
> > On Tuesday, March 31, 2026 2:02 PM Nisha Moond <nisha.moond412@gmail.co=
m> wrote:
> > >
> > >
> > > Please find the updated patch (v6) attached.
> >
> > Thanks for updating the patch. One minor comment:
> >
> > I think we could avoid interrupting and reporting an ERROR when
> > IsSyncingReplicationSlots() returns false to avoid reporting ERROR unne=
cessarily
> > when the slotsync has already finished.
> >
>
> Thanks for the review. Fixed above in v7.

Thanks for updating the patch! It looks good to me, with just a few minor p=
oints
. If those are addressed, I'd like to push it.

+ * a new worker (or a new API call) that starts after the old worker was

"API" feels a bit vague. It might be clearer to explicitly say
"pg_sync_replication_slots()".

+ PROCSIG_SLOTSYNC_MESSAGE, /* ask slotsync worker/API to stop */

"API" here also feels a bit vague. So I'd like to use "ask slot synchroniza=
tion
to stop" as the comment, instead.

+ * We cannot rely solely on 'stopSignaled' here because:
+ * 1) It resides in shared memory and is visible to all processes, so chec=
king
+ *    it directly in ProcessInterrupts() would require additional checks t=
o
+ *    ensure only the synchronizing process acts on it.
+ * 2) It has different lifetime semantics and cannot be reset after handli=
ng,
+ *    as it also guards against postmaster and promotion race conditions.
+ * 3) Accessing it requires acquiring a spinlock, which can be too expensi=
ve
+ *    or undesirable for every ProcessInterrupts() call.

Now that PROCSIG_SLOTSYNC_MESSAGE is in place, using SlotSyncShutdownPendin=
g
is intuitive. So it seems more useful to explain why stopSignaled is still
needed rather than why SlotSyncShutdownPending is used (i.e., why stopSigna=
led
is not sufficient). Since that rationale is already covered in the SlotSync=
Ctx
comments, I'd suggest removing this comment block.


As for backpatching, this looks like it should go back to v17, where slotsy=
nc
was introduced. Thought?

Regards,

--=20
Fujii Masao