Re: update to 16.2 - Mailing list pgsql-general

From Adrian Klaver
Subject Re: update to 16.2
Date
Msg-id 72847831-73d2-4bd6-8a92-43c64e29ec3b@aklaver.com
Whole thread Raw
In response to update to 16.2  (Matthias Apitz <guru@unixarea.de>)
List pgsql-general
On 3/8/24 00:53, Matthias Apitz wrote:
> 
> Hello,
> 
> We plan to update our customers on SuSE Linux from 11.4, 13.1, 14.1 and
> 15.1 to 16.2. Do I understand the release notes correct that for this
> the way is only dump/restore? The release notes say:
> 
> https://www.postgresql.org/docs/release/16.2/
> ...
> .
> ...
> It does not say definitely that for all other versions a dump/restore is
> required.

In the current versioning scheme(10+) the version is X.x where X is 
major version and x is minor version. Any time you move from a major to 
a major version the binaries are not guaranteed to be compatible and you 
will need to do a dump/restore, pg_upgrade or logical step to move to 
the new version. A minor to minor upgrade within a major version in 
general does not require those steps. The 'A dump/restore is not 
required for those running 16.X' note is there in case there is an 
instance when a dump/restore is needed in which case it would be changed 
to '... is required ...' . This is the case in the development releases 
of a new version(before X.0) where you usually do have to dump/restore, 
etc between minor version changes. The note is there to remove any 
ambiguity.

> 
> We could even unload in the databases the ~400 tables to COPY format,
> re-create the tables in a new 16.2 server and load the ~400 files (all this
> relatively easy per scripts we created to migrate from Sybase and Oracle
> to PostgreSQL). But I think, producing the dump with the old version,
> setup new cluster and load the dump with the 16.2 sql command will work.
> 
> Any comments?
> 
>     matthias
> 
> 

-- 
Adrian Klaver
adrian.klaver@aklaver.com




pgsql-general by date:

Previous
From: David Gauthier
Date:
Subject: Re: v11.5- v15.3 upgrade (linux)
Next
From: Ron Johnson
Date:
Subject: Re: creating a subset DB efficiently ?