Re: Speeding up pg_upgrade

Tom, David,

* Tom Lane (tgl(at)sss(dot)pgh(dot)pa(dot)us) wrote:
> Stephen Frost <sfrost(at)snowman(dot)net> writes:
> > Well, if it's dropped, I think we need to make sure that users are aware
> > of that going in and that's why I was suggesting a switch. If you've
> > got a better idea for that, great, but having certain pg_upgrade
> > migrations require running ANALYZE and some migrations not require it is
> > something we need to make users *very* clear about. No, I don't think a
> > note in the release notes is really enough..
>
> Seems like we could make this reasonably transparent if pg_upgrade
> continues to emit an analyze script that you're supposed to run
> afterwards. It just has to vary how much that script does.

That's post-upgrade though, meaning it's entirely likely to catch
someone by surprise. That's the issue that I have with it.

That said, I'm remembering that there's a '--check' option to
pg_upgrade. Let's make sure to include in that --check output what
additional post-upgrade work will have to happen, that'll work well
enough to deal with this issue and I have little sympathy for people who
run pg_upgrade blindly and without running --check first, given that
there's an option for it.

* David G. Johnston (david(dot)g(dot)johnston(at)gmail(dot)com) wrote:
> On Thu, Dec 7, 2017 at 12:04 PM, Stephen Frost <sfrost(at)snowman(dot)net> wrote:
> > If you've
> > got a better idea for that, great, but having certain pg_upgrade
> > migrations require running ANALYZE and some migrations not require it is
> > something we need to make users *very* clear about. No, I don't think a
> > note in the release notes is really enough..
>
> Ideally this would be covered by:
>
> """
> 13. Post-Upgrade processing
>
> If any post-upgrade processing is required, pg_upgrade will issue warnings
> as it completes. It will also generate script files that must be run by the
> administrator. The script files will connect to each database that needs
> post-upgrade processing. Each script should be run using: ...
> """
>
> https://www.postgresql.org/docs/10/static/pgupgrade.html
>
> Then whether is instructed as per #14 would be conditional.
>
> Its arguable whether a warning/notice that "ANALYZE doesn't need to be run"
> message would be useful.

I'm fine with updating the docs to make it conditional but our users
need an easy way to figure out if their particular from/to PG versions
require ANALYZE to be run or not, so they're able to prepare for it.
Having that be included in the --check output would work to handle that
nicely.

Also, really, we should update those steps to have --check run earlier,
imv, and before saying "stop the clusters".

Thanks!

Stephen