From: | Andres Freund <andres(at)anarazel(dot)de> |
---|---|
To: | Georgy Buranov <gburanov(at)gmail(dot)com> |
Cc: | pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: PostgreSQL logical decoder output plugin - unchanged toast data |
Date: | 2018-08-31 15:39:12 |
Message-ID: | 20180831153912.zl5zpln2lpw7r7ji@alap3.anarazel.de |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On 2018-08-31 17:34:02 +0200, Georgy Buranov wrote:
> > Again, you can set REPLICA IDENTITY to FULL and it'll be there.
> >
>
> So, why I think this is complicated
>
> * We use primary keys for all tables, so we do not need REPLICA
> IDENTITY full actually. As far as I understand, it will make
> master/slave replication ineffective as well
What do you mean with "ineffective"? That there's more data in the WAL?
Sure. Otherwise I don't know what you could mean.
There's no free lunch :/
> * I need the information about this primary key in rd_replidindex of
> relation (to send it later to kafka). As far as I understand, with
> REPLICA IDENTITY FULL the rd_replidindex will not be the primary key
If you want the pkey, that's in rd_pkindex. rd_replidindex will only
differ if the identity is manually set to another candidate key
(possibly because there's no pkey).
Greetings,
Andres Freund
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2018-08-31 16:02:04 | Re: pg_verify_checksums and -fno-strict-aliasing |
Previous Message | Georgy Buranov | 2018-08-31 15:34:02 | Re: PostgreSQL logical decoder output plugin - unchanged toast data |