Re: Support logical replication of DDLs - Mailing list pgsql-hackers

From vignesh C
Subject Re: Support logical replication of DDLs
Date
Msg-id CALDaNm3ELSfqTi-GHuNkOqcuN12rmkR_JAquMvL+R266FZW9Ow@mail.gmail.com
Whole thread Raw
In response to Re: Support logical replication of DDLs  (vignesh C <vignesh21@gmail.com>)
List pgsql-hackers
On Wed, 24 May 2023 at 11:08, vignesh C <vignesh21@gmail.com> wrote:
>
> On Mon, 22 May 2023 at 11:27, shveta malik <shveta.malik@gmail.com> wrote:
> >
> > On Wed, May 17, 2023 at 4:45 PM vignesh C <vignesh21@gmail.com> wrote:
> > >
> > > On Wed, 17 May 2023 at 15:41, shveta malik <shveta.malik@gmail.com> wrote:
> > > >
> > > > On Fri, May 12, 2023 at 12:03 PM shveta malik <shveta.malik@gmail.com> wrote:
> > > > >
> > > > > On Tue, May 9, 2023 at 4:23 PM shveta malik <shveta.malik@gmail.com> wrote:
> > > > > >
> > > > > > On Mon, May 8, 2023 at 4:31 PM shveta malik <shveta.malik@gmail.com> wrote:
> > > > > > >
> > > > > > > On Mon, May 8, 2023 at 3:58 PM shveta malik <shveta.malik@gmail.com> wrote:
> > > > > > > >
> > > > > > > > On Tue, May 2, 2023 at 8:30 AM shveta malik <shveta.malik@gmail.com> wrote:
> > > > > > > > >
> > > > > > > > > On Fri, Apr 28, 2023 at 5:11 PM Amit Kapila <amit.kapila16@gmail.com> wrote:
> > > > > > > > > >
> > > > > > > > > > Now, I think we can try to eliminate this entire ObjTree machinery and
> > > > > > > > > > directly from the JSON blob during deparsing. We have previously also
> > > > > > > > > > discussed this in an email chain at [1]. I think now the functionality
> > > > > > > > > > of JSONB has also been improved and we should investigate whether it
> > > > > > > > > > is feasible to directly use JSONB APIs to form the required blob.
> > > > > > > > >
> > > > > > > > > +1.
> > > > > > > > > I will investigate this and will share my findings.
> > > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > > Please find the PoC patch for create-table after object-tree removal.
> > > > > > >
> > > > > >
> > > > >
> > > > > Please find the new set of patches attached for object-tree removal.
> > > >
> > > > Please find the new set of patches for object-tree Removal.  The new
> > > > changes are in patch 0008 only. The new changes incorporate the
> > > > object-tree removal for 'alter table' command.
> > >
> >
> > Please find the new set of patches for object-tree Removal.  The new
> > changes are in patch 0008 only. The new changes address object tree
> > removal for below commands.
> >
> > create sequence
> > alter sequence
> > alter object owner to
> > alter object set schema
> > alter object rename
> >
> > In this patch 0008, ddldeparse.c is now object-tree free for all the
> > table related commands. Index related commands are yet to be done.
>
> I found few comments while making some changes to the patch:
> 1) Now that objtree is removed, these comments should be modified:
>  * Deparse object tree is created by using:
>  * a) new_objtree("know contents") where the complete tree content is known or
>  *     the initial tree content is known.
>  * b) new_objtree("") for the syntax where the object tree will be derived
>  *     based on some conditional checks.
>  * c) new_objtree_VA where the complete tree can be derived using some fixed
>  *     content or by using the initial tree content along with some variable
>  *     arguments.
>  *

Modified

>  2) isgrant can be removed as it is not used anymore:
> +/*
> + * Return the given object type as a string.
> + *
> + * If isgrant is true, then this function is called while deparsing GRANT
> + * statement and some object names are replaced.
> + */
> +const char *
> +stringify_objtype(ObjectType objtype, bool isgrant)
> +{
> +       switch (objtype)
> +       {
> +               case OBJECT_TABLE:
> +                       return "TABLE";
> +               default:
> +                       elog(ERROR, "unsupported object type %d", objtype);
> +       }
> +
> +       return "???";                           /* keep compiler quiet */
> +}

Modified

> 3) This statement is not being handled currently, it should be implemented:
> "alter table all in tablespace tbs1 set tablespace"

Modified

> 4) This pub_ddl is selected as the 7th column, it should be 7 instead of 9 here:
> @@ -6405,6 +6418,8 @@ describePublications(const char *pattern)
>                 printTableAddCell(&cont, PQgetvalue(res, i, 4), false, false);
>                 printTableAddCell(&cont, PQgetvalue(res, i, 5), false, false);
>                 printTableAddCell(&cont, PQgetvalue(res, i, 6), false, false);
> +               if (has_pubddl)
> +                       printTableAddCell(&cont, PQgetvalue(res, i,
> 9), false, false);
>                 if (has_pubtruncate)
>                         printTableAddCell(&cont, PQgetvalue(res, i,
> 7), false, false);
>                 if (has_pubviaroot)

Modified

The following patches were changed to fix the above issues: 1)
0001-Deparser-for-Table-DDL-commands-and-extending-event-2023_05_29.patch
2) 0005-DDL-replication-for-Table-DDL-commands-2023_05_29.patch and 3)
0008-ObjTree-Removal-for-multiple-commands-2023_05_29.patch

The attached patch has the changes for the same.

Regards,
Vignesh

Attachment

pgsql-hackers by date:

Previous
From: "Drouvot, Bertrand"
Date:
Subject: Re: BF animal dikkop reported a failure in 035_standby_logical_decoding
Next
From: Giuseppe Broccolo
Date:
Subject: Re: vector search support