Geoffrey <lists at serioustechnology.com> writes:
> Andrew Sullivan wrote:
>> On Tue, Feb 19, 2008 at 06:57:20PM -0500, Geoffrey wrote:
>>> Apparently there were 100's of these errors generated, yet, when
>>> researching the production data, it does not appear that there was
>>> any corruption.  Here is the text of one of the errors:
>>>
>>> relation with OID 394006 does not exist
>> Aha!  I'll bet lunch someone was doing DDL in their "shuffling data
>> around". That is an error message telling you that a
>> previously-planned query is
>> looking for a now-dropped relation.  One way to get this is indeed with temp
>> tables, but you can get it by dropping regular tables &c.  I'd have a very
>> good look at how that procedure from last Friday worked.
>> If you do DDL under Slony without passing it through EXECUTE SCRIPT,
>> you
>> _will_ get very strange results.
>
> I only wish that was the problem.  Our application does not remove any
> tables.

That's not the exact implication, though.  I know we ran into this
sort of thing on an occasion where we dropped replication from an
"origin" node; the trouble was that query plans were expecting the
Slony-I-defined tables to still be there.

The missing relation might not be one of your tables; it might be
something else about the schema.
-- 
"cbbrowne","@","cbbrowne.com"
http://cbbrowne.com/info/lisp.html
Rules of the Evil Overlord #229. "If I have several diabolical schemes
to destroy the hero,  I will set all of them in  motion at once rather
than wait for them to fail and launch them successively."
<http://www.eviloverlord.com/>