Horton

Re: [BUG]Update Toast data failure in logical replication

Dilip Kumar <dilipbalaut@gmail.com>

View thread
From: Dilip Kumar <dilipbalaut@gmail.com>
To: Amit Kapila <amit.kapila16@gmail.com>
Cc: Alvaro Herrera <alvherre@alvh.no-ip.org>, Petr Jelinek <petr.jelinek@enterprisedb.com>, Peter Eisentraut <peter.eisentraut@enterprisedb.com>, Tomas Vondra <tomas.vondra@enterprisedb.com>, Andres Freund <andres@anarazel.de>, Kuntal Ghosh <kuntalghosh.2007@gmail.com>, "tanghy.fnst@fujitsu.com" <tanghy.fnst@fujitsu.com>, pgsql-hackers <pgsql-hackers@postgresql.org>
Date: 2021-08-11T12:44:55Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. WAL log unchanged toasted replica identity key attributes.

  2. Add new wal_level, logical, sufficient for logical decoding.

  3. Add the notion of REPLICA IDENTITY for a table.

Attachments

On Wed, Aug 11, 2021 at 10:30 AM Amit Kapila <amit.kapila16@gmail.com> wrote:
>
> On Tue, Aug 10, 2021 at 8:08 PM Alvaro Herrera <alvherre@alvh.no-ip.org> wrote:
> >
> > On 2021-Jul-30, Amit Kapila wrote:
> >
> > Reading Dilip's last posted patch that day, I had some reservations
> > about the API of ExtractReplicaIdentity.  The new argument makes for a
> > very strange to explain behavior "return the key values if they are
> > unchanged, *or* if they are toasted" ... ???
> >
>
> I think we can say it as "Return the key values if they are changed
> *or* if they are toasted". Currently, we have an exception for Deletes
> where the caller always passed key_changed as true, so maybe we can
> have a similar exception when the tuple has toasted values. Can we
> think of changing the flag to "key_required" instead of "key_changed"
> and let the caller identify and set its value? For Deletes, it will
> work the same but for Updates, the caller needs to compute it by
> checking if any of the key columns are modified or has a toast value.
> We can try to see if the caller can identify it cheaply along with
> determining the modified_attrs as at that time we will anyway check
> replica key attrs.

Right

>
> Currently, in proposed patch first, we check that the tuple has any
> toast values and then it deforms and forms the new key tuple. After
> that, it checks if the key has any toast values and then only decides
> to return the tuple. If as described in the previous paragraph, we can
> cheaply identify whether the key has toasted values, then we can avoid
> deform/form cost in some cases. Also, I think we need to change the
> Replica Identity description in the docs[1].

Yeah we can avoid that by detecting any toasted replica identity key
in HeapDetermineModifiedColumns, check the attached patch.



--
Regards,
Dilip Kumar
EnterpriseDB: http://www.enterprisedb.com