Re: pg_upgrade and publication/subscription problem - Mailing list pgsql-hackers

From Marcos Pegoraro
Subject Re: pg_upgrade and publication/subscription problem
Date
Msg-id CAB-JLwYkPjD=C+RNtxd8RTSQXLJHFTO3HefhxKB46LsrvmPV_Q@mail.gmail.com
Whole thread Raw
In response to Re: pg_upgrade and publication/subscription problem  (Amit Kapila <amit.kapila16@gmail.com>)
Responses Re: pg_upgrade and publication/subscription problem
List pgsql-hackers
The reason is after an upgrade, there won't be any data in
pg_subscription_rel, and only when you tried to refresh it is trying
to sync again which leads to the "duplicate key value ..." problem you
are seeing.

So, is pg_upgrade populating pg_subscription and not pg_subscription_rel ? It is doing 50% of his job ?

Don't you want to eventually upgrade the publisher node as well? You
can refer to blog [1] for the detailed steps.

It is possible but I don´t think changing publisher will solve anything, will ?

Yes, the way you are doing I think it is bound to happen. There is
some discussion about why this is happening in email [2]. AFAIK, it is
not documented and if so, I think it will be a good idea to document

And my problem remains the same, how to solve it ? All records on pg_subscription_rel are initialize with srsubstate null. How can I replay only updates since yesterday. This replication is a auditing database, so I cannot loose all things happened since that pg_upgrade. [1] points me how to upgrade but if I did the wrong way, how to solve that ?

pgsql-hackers by date:

Previous
From: Virender Singla
Date:
Subject: Re: TOAST - why separate visibility map
Next
From: Zhihong Yu
Date:
Subject: Re: Non-decimal integer literals