Re: Logical replication from PG v13 and below to PG v14 (devel version) is not working.

On Tue, Sep 22, 2020 at 12:02 PM Amit Kapila <amit(dot)kapila16(at)gmail(dot)com> wrote:
>
> On Tue, Sep 22, 2020 at 8:34 AM Ashutosh Sharma <ashu(dot)coek88(at)gmail(dot)com> wrote:
> >
> > On Mon, Sep 21, 2020 at 6:58 PM Amit Kapila <amit(dot)kapila16(at)gmail(dot)com> wrote:
> > >
> > > On Mon, Sep 21, 2020 at 6:27 PM Ashutosh Sharma <ashu(dot)coek88(at)gmail(dot)com> wrote:
> > > >
> > > > Thanks Dilip for the patch. AFAIU, the fix looks good. One small comment:
> > > >
> > >
> > > Thanks Ashutosh and Dilip for working on this. I'll look into it in a
> > > day or two.
> > >
> >
> > Just a thought:
> >
> > Should we change the sequence of these #define in
> > src/include/replication/logicalproto.h?
> >
> > #define LOGICALREP_PROTO_MIN_VERSION_NUM 1
> > #define LOGICALREP_PROTO_STREAM_VERSION_NUM 2
> > #define LOGICALREP_PROTO_VERSION_NUM 1
> > #define LOGICALREP_PROTO_MAX_VERSION_NUM LOGICALREP_PROTO_STREAM_VERSION_NUM
> >
> > I would have changed above to something like this:
> >
> > #define LOGICALREP_PROTO_VERSION_NUM 1
> > #define LOGICALREP_PROTO_STREAM_VERSION_NUM 2
> >
> > #define LOGICALREP_PROTO_MIN_VERSION_NUM LOGICALREP_PROTO_VERSION_NUM
> > #define LOGICALREP_PROTO_MAX_VERSION_NUM LOGICALREP_PROTO_STREAM_VERSION_NUM
> >
>
> I am not sure if this suggestion makes it better than what is purposed
> by Dilip but I think we can declare them in define number order like
> below:
> #define LOGICALREP_PROTO_MIN_VERSION_NUM 1
> #define LOGICALREP_PROTO_VERSION_NUM 1
> #define LOGICALREP_PROTO_STREAM_VERSION_NUM 2
> #define LOGICALREP_PROTO_MAX_VERSION_NUM LOGICALREP_PROTO_STREAM_VERSION_NUM
>

The only reason I proposed that was because for the *_MAX_VERSION_NUM
we are using the latest PROTOCOL version name in its definition so why
not to do the same for defining *_MIN_VERSION_NUM as well. Other than
that, I also wanted to correct the sequence so that they are defined
in the increasing order which you have already done here.

> Another thing is can we also test by having a publisher of v14 and
> subscriber of v13 or prior version, just reverse of what Ashutosh has
> tested?
>

I've tested LR from PGv12 (Publisher) to PGv14 (Subscriber) and it works fine.

--
With Regards,
Ashutosh Sharma
EnterpriseDB:http://www.enterprisedb.com