Re: pg_upgrade Only the install user can be defined in the new cluster - Mailing list pgsql-general

From Dávid Suchan
Subject Re: pg_upgrade Only the install user can be defined in the new cluster
Date
Msg-id VE1P191MB111805B6886F6AD3E78E620BD6BC9@VE1P191MB1118.EURP191.PROD.OUTLOOK.COM
Whole thread Raw
In response to Re: pg_upgrade Only the install user can be defined in the new cluster  (Daniel Gustafsson <daniel@yesql.se>)
Responses Re: pg_upgrade Only the install user can be defined in the new cluster  (Tom Lane <tgl@sss.pgh.pa.us>)
Re: pg_upgrade Only the install user can be defined in the new cluster  (Daniel Gustafsson <daniel@yesql.se>)
List pgsql-general
So I tried upgrading into this brand new cluster I created(using initdb -D /somedatapathichose). Running the upgrade with --check worked, it returned message that the clusters are identical. Then I stopped the new cluster, and ran the pg_upgrade without --check, which resulted in another:
Checking database user is the install user
Only the install user can be defined in the new cluster.
Failure, exiting
I dont know what to check for anymore, the log files dont say anything other than "Only the install user can be defined in the new cluster" when postgres is the install user everywhere.
The ' SELECT COUNT(*) FROM pg_catalog.pg_roles WHERE rolname !~ '^pg_'; ' prints count 4 and ' SELECT rolname FROM pg_roles WHERE oid = 10; ' prints rolname postgres.

Od: Daniel Gustafsson <daniel@yesql.se>
Odoslané: štvrtok 16. marca 2023 10:28
Komu: Dávid Suchan <david.suchan@student.tuke.sk>
Kópia: pgsql-general@postgresql.org <pgsql-general@postgresql.org>
Predmet: Re: pg_upgrade Only the install user can be defined in the new cluster
 
> On 15 Mar 2023, at 16:39, Dávid Suchan <david.suchan@student.tuke.sk> wrote:
>
> It prints out:
>  count
> -------
>      1

I have a feeling the cluster you tried to upgrade to doesn't match this one, as
the check that failed will fail on values other than 1.  Did you create them
equally?  If you try to upgrade into this cluster, even just with the --check
option, does that yield more success?

--
Daniel Gustafsson

pgsql-general by date:

Previous
From: "Dolan, Sean"
Date:
Subject: RE: EXTERNAL: Re: "No Free extents", table using all allocated space but no rows!
Next
From: Adrian Klaver
Date:
Subject: Re: Is the PL/pgSQL refcursor useful in a modern three-tier app?