Re: Minimal logical decoding on standbys

From: Amit Kapila <amit(dot)kapila16(at)gmail(dot)com>
To: "Drouvot, Bertrand" <bertranddrouvot(dot)pg(at)gmail(dot)com>
Cc: Andres Freund <andres(at)anarazel(dot)de>, Robert Haas <robertmhaas(at)gmail(dot)com>, Jeff Davis <pgsql(at)j-davis(dot)com>, Thomas Munro <thomas(dot)munro(at)gmail(dot)com>, Alvaro Herrera <alvherre(at)2ndquadrant(dot)com>, Ibrar Ahmed <ibrar(dot)ahmad(at)gmail(dot)com>, Amit Khandekar <amitdkhan(dot)pg(at)gmail(dot)com>, fabriziomello(at)gmail(dot)com, tushar <tushar(dot)ahuja(at)enterprisedb(dot)com>, Rahila Syed <rahila(dot)syed(at)2ndquadrant(dot)com>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>, Melanie Plageman <melanieplageman(at)gmail(dot)com>
Subject: Re: Minimal logical decoding on standbys
Date: 2023-04-04 05:57:02
Message-ID: CAA4eK1LzWv2p+97vsHekhftSuMR0m-qVEdoCpASUApN6pXtZpQ@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Mon, Apr 3, 2023 at 8:51 PM Drouvot, Bertrand
<bertranddrouvot(dot)pg(at)gmail(dot)com> wrote:
>
> On 4/2/23 10:10 PM, Andres Freund wrote:
> > Hi,
> >> restart_lsn = s->data.restart_lsn;
> >> -
> >> - /*
> >> - * If the slot is already invalid or is fresh enough, we don't need to
> >> - * do anything.
> >> - */
> >> - if (XLogRecPtrIsInvalid(restart_lsn) || restart_lsn >= oldestLSN)
> >> + slot_xmin = s->data.xmin;
> >> + slot_catalog_xmin = s->data.catalog_xmin;
> >> +
> >> + /* the slot has been invalidated (logical decoding conflict case) */
> >> + if ((xid && ((LogicalReplicationSlotIsInvalid(s)) ||
> >> + /* or the xid is valid and this is a non conflicting slot */
> >> + (TransactionIdIsValid(*xid) && !(LogicalReplicationSlotXidsConflict(slot_xmin, slot_catalog_xmin, *xid))))) ||
> >> + /* or the slot has been invalidated (obsolete LSN case) */
> >> + (!xid && (XLogRecPtrIsInvalid(restart_lsn) || restart_lsn >= oldestLSN)))
> >> {
> >
> > This still looks nearly unreadable. I suggest moving comments outside of the
> > if (), remove redundant parentheses, use a function to detect if the slot has
> > been invalidated.
> >
>
> I made it as simple as:
>
> /*
> * If the slot is already invalid or is a non conflicting slot, we don't
> * need to do anything.
> */
> islogical = xid ? true : false;
>
> if (SlotIsInvalid(s, islogical) || SlotIsNotConflicting(s, islogical, xid, &oldestLSN))
>
> in V56 attached.
>

Here the variable 'islogical' doesn't seem to convey its actual
meaning because one can imagine that it indicates whether the slot is
logical which I don't think is the actual intent. One idea to simplify
this is to introduce a single function CanInvalidateSlot() or
something like that and move the logic from both the functions
SlotIsInvalid() and SlotIsNotConflicting() into the new function.

--
With Regards,
Amit Kapila.

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Richard Guo 2023-04-04 06:47:19 Re: same query but different result on pg16devel and pg15.2
Previous Message tender wang 2023-04-04 05:53:52 Re: same query but different result on pg16devel and pg15.2