Re: pg_upgrade libraries check

Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> writes:
> Well, the scheme I had in mind would require pg_upgrade to verify that
> the new cluster contains an extension control file for each extension in
> the old cluster (which is something it really oughta do anyway, if it
> doesn't already). After that, though, it ought not be looking at the
> individual functions defined by an extension --- it is the extension's
> business which libraries those are in.

I have some plans that we will be discussing later in the new dev cycle
and that would impact such a method if we're to follow them. To better
solve both the per-system (not even cluster) and per-database extension
versions and the inline/os-packaged extension discrepancy, I'm thinking
that we should move the extension support files from their shared OS
location to a per-database location at CREATE EXTENSION time.

That would mean managing several copies of those so that each database
can actually implement a different version and upgrade cycle of an
extension, including of its shared object libraries (some more session
state) in simplest cases (no extra shared object dependencies).

I don't intend to be already discussing the details of that proposal
here, that's just a hint so that you know that I intend to rework the
current control file strategy that we have, so any work on that in
pg_upgrade in next cycle will possibly be impacted.

> The only reason for pg_upgrade to still look at probin at all would be
> to cover C functions that weren't within extensions. In the long run it
> might be possible to consider those unsupported, or at least not common
> enough to justify a safety check in pg_upgrade.

I'm thinking about optionally forbidding creating functions written in C
or installed into pg_catalog when not done via an extension script, and
maybe later down the road changing the default to be the forbidding.

The pg_catalog case makes sense as going via an extension's script is
the only way I know about to dump/restore the function.

Regards,
--
Dimitri Fontaine
http://2ndQuadrant.fr PostgreSQL : Expertise, Formation et Support