Re: pg_upgrade does not upgrade pg_stat_statements properly - Mailing list pgsql-hackers

From Bruce Momjian
Subject Re: pg_upgrade does not upgrade pg_stat_statements properly
Date
Msg-id 20210729161251.GM9600@momjian.us
Whole thread Raw
In response to Re: pg_upgrade does not upgrade pg_stat_statements properly  ("David G. Johnston" <david.g.johnston@gmail.com>)
List pgsql-hackers
On Thu, Jul 29, 2021 at 08:39:32AM -0700, David G. Johnston wrote:
> On Thu, Jul 29, 2021 at 8:36 AM Dave Cramer <davecramer@gmail.com> wrote:
> 
> 
> 
> 
>         I have an issue with the fragment "whether they are from contrib" - my
>         understanding at this point is that because of the way we package and
>         version contrib it should not be necessary to copy those shared object
>         files from the old to the new server (maybe, just maybe, with a
>         qualification that you are upgrading between two versions that were in
>         support during the same time period).
> 
> 
>     Just to clarify. In no case are binaries copied from the old server to the
>     new server. Whether from contrib or otherwise.
>
> I had used "binaries" when I should have written "shared object files".  I just
> imagine a DLL as being a binary file so it seems accurate but we use the term
> differently I suppose?

Uh, technically, the _executable_ binary should only use shared object /
loadable libraries that were compiled against that binary's exported
API.  Sometimes mismatches work (if the API used by the shared object
has not changed in the binary) so people get used to it working, and
then one day it doesn't, but it is never a safe process.

If two people here are confused about this, obviously others will be as
well.  I think we were trying to do too much in that first sentence, so
I split it into two in the attached patch.

-- 
  Bruce Momjian  <bruce@momjian.us>        https://momjian.us
  EDB                                      https://enterprisedb.com

  If only the physical world exists, free will is an illusion.


Attachment

pgsql-hackers by date:

Previous
From: Robert Haas
Date:
Subject: Re: Out-of-memory error reports in libpq
Next
From: Bruce Momjian
Date:
Subject: Re: pg_upgrade does not upgrade pg_stat_statements properly