Re: Issue upgrading from 9.5 to 13 with pg_upgrade: "connection to database failed: FATAL: database "template1" does not exist"

On 11/13/20 6:32 AM, Jeremy Wilson wrote:
> I’m running CentOS 8 on an EC2 instance and attempting to upgrade a 9.5 database to 13 using pg_upgrade. Both are running on the same box and pass initial tests but it fails during the later part of the process.
>
> ---
>
> bash-4.4$ /usr/pgsql-13/bin/pg_upgrade --old-bindir /usr/pgsql-9.5/bin --new-bindir /usr/pgsql-13/bin --old-datadir /var/lib/pgsql/9.5/data --new-datadir /var/lib/pgsql/13/data --old-port=54320 --new-port=5432 --socketdir=/var/run/postgresql/
> Performing Consistency Checks
> -----------------------------
> Checking cluster versions ok
> Checking database user is the install user ok
> Checking database connection settings ok
> Checking for prepared transactions ok
> Checking for reg* data types in user tables ok
> Checking for contrib/isn with bigint-passing mismatch ok
> Checking for tables WITH OIDS ok
> Checking for invalid "sql_identifier" user columns ok
> Checking for invalid "unknown" user columns ok
> Checking for roles starting with "pg_" ok
> Creating dump of global objects ok
> Creating dump of database schemas
> ok
>
> connection to database failed: FATAL: database "template1" does not exist
>
> could not connect to target postmaster started with the command:
> "/usr/pgsql-13/bin/pg_ctl" -w -l "pg_upgrade_server.log" -D "/var/lib/pgsql/13/data" -o "-p 5432 -b -c synchronous_commit=off -c fsync=off -c full_page_writes=off -c vacuum_defer_cleanup_age=0 -c listen_addresses='' -c unix_socket_permissions=0700 -c unix_socket_directories='/var/run/postgresql'" start
> Failure, exiting
>
> ---
>
> When I manually run that command it works, although when I attempt to connect to the ‘postgres’ database on it, it complains it doesn’t exist.

To me it seems the initdb for the 13 instance did not complete
successfully. Have you tried clearing /var/lib/pgsql/13/data and doing
the init over again? If you do try it monitor the output carefully.

>
> I can’t use the dump/restore method to upgrade this server as the data in question is fairly massive and in my testing it took 45 hours to complete. Any help appreciated.
>
>
>
>

--
Adrian Klaver
adrian(dot)klaver(at)aklaver(dot)com