On Fri, Oct 4, 2019 at 09:43:31AM -0700, Peter Geoghegan wrote:
> On Fri, Oct 4, 2019 at 9:09 AM Tom Lane <tgl@sss.pgh.pa.us> wrote:
> > > You can't REINDEX safely regarding that note.
> >
> > Actually running into that problem is quite unlikely; and if you did
> > hit it, it'd just mean that the REINDEX fails, not that you have any
> > urgent problem to fix. I'd encourage you to just go ahead and REINDEX,
> > if you have indexes that could benefit from the other changes.
>
> Right. It is hard to imagine an application that evolved to fully rely
> on the previous slightly higher limit, and cannot tolerate a reduction
> in the limit by only 8 bytes. The limit applies to a tuple *after*
> TOAST compression has been applied.
Right. Pg_upgrade is fast, but we don't want it limiting file format
changes that can improve Postgres. Allowing REINDEX to fix things is
the best of both worlds --- fast upgrades, and after some REINDEX-ing,
faster Postgres.
--
Bruce Momjian <bruce@momjian.us> http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so once was I. As I am, so you will be. +
+ Ancient Roman grave inscription +