--- Michael Plump <plumpy@skylab.org> wrote:
> On Wed, 6 Oct 1999, Mike Mascari wrote:
>
> > You show yourself attempting to dump a database
> > called "darwin", while connecting to and
> displaying
> > tables in the "template1" database. If the data
> you
> > wish to dump is in template1, you should issue the
> > command:
>
> But, as his message indicated, the error says:
> > > pg_dump: couldn't find the template1 database.
> You
> > > are really hosed.
>
> Hence, he was showing you that template1 really does
> exist.
>
>
I see. I guess I wasn't convinced darwin existed.
Knowing that they both do, the source to pg_dump
shows the error after the following select:
SELECT oid from pg_database where datname =
'template1'
in the routine which attempts to determine the
maximum OID. Nikos should see the template1 database
when performing the select in "darwin". If he
does not, something is seriously wrong. However,
he may still be able to dump the darwin database
if he does not use the -o switch to dump OIDs,
or he may wish to manually INSERT the missing
records into pg_database:
INSERT INTO pg_database VALUES ('template1', uid, 0,
'template1');
where uid is the usesysid from pg_user of the owner
of the database (usually, 'postgres').
It would be interesting if he could list the details
on the file:
/usr/local/pgsql/data/pg_database
Of course, if he's dumping OIDs for a reason
(say he's using them as pk/fk) then he may
be forced, if the above fails, to attempt to dump the
schema, and then dump the data by using COPY :-(
Hope that helps,
Mike Mascari
(mascarim@yahoo.com)
=====
__________________________________________________
Do You Yahoo!?
Bid and sell for free at http://auctions.yahoo.com