Re: BUG #6706: pg_upgrade fails when plpgsql dropped/re-created

On 09/12/2014 5:26 PM, Tom Lane-2 [via PostgreSQL] wrote:
> cowwoc <[hidden email]
> </user/SendEmail.jtp?type=node&node=5829826&i=0>> writes:
> > Tom Lane-2 wrote
> >> That's bizarre.  It sounds like there's something corrupt about your
> >> pg_extension system catalog.  Does "select * from pg_extension" show
> >> reasonable-looking data?  If so, try "REINDEX TABLE pg_extension".
>
> > This error is quite random, so I'll have to try again the next time
> I run
> > into it. I get the following output when the database is in a "good"
> state:
>
> If it was non-reproducible, you really should have mentioned that to
> start
> with.  That fact changes the likely causes significantly.  Now I'm
> wondering about flaky hardware, for example.

Highly unlikely. This is running off my local machine which has been
stable for many years now. If flaky hardware was at play, I'd expect to
see other intermittent failures elsewhere.

> > extname, extowner, extnamespace, extrelocatable, extversion, extconfig,
> > extcondition
> > "plpgsql";10;11;f;"1.0";"";""
> > "hstore";13288111;2200;t;"1.2";"";""
>
> > Is that "reasonable-looking data"?
>
> Yeah, although that OID for the owning role of hstore seems pretty large.

I haven't ever run AUTOVACUUM on this machine (the pg installation is
about 3 months old). Could it cause some of these problems. This is a
very low-volume developer machine that only runs small jobs I trigger
manually.

Gili




--
View this message in context:
http://postgresql.nabble.com/BUG-6706-pg-upgrade-fails-when-plpgsql-dropped-re-created-tp5714220p5829834.html
Sent from the PostgreSQL - bugs mailing list archive at Nabble.com.