Andrew Sullivan wrote:
> On Tue, Feb 19, 2008 at 04:07:52PM -0500, Geoffrey wrote:
>> We have come to the conclusion that our replication data is useless, so 
>> we are going to drop the replication databases and revisit this from 
>> scratch.
> 
> The big puzzle is how it got that way.  If you have anything like a history
> of things done on these systems, and are willing to share the schema and its
> history for debugging purposes, I'd be very interested in seeing it (and I
> bet others would be too).

Yes, the big puzzle is terribly bothersome.  Particularly if once we 
re-replicate, we don't see the problem again.  I really want to know 
what caused the problem and therefore, how to avoid it in the future.

The production environment is a logistics tracking system.  There are 12 
separate databases running on 12 different ports, and therefore, 12 
different postmasters.  There is one 'master' database where data is 
archived to from the other 11 databases once a load is complete.

I am certainly willing to share more info with anyone who might be 
interested in this issue.

I don't know that I can share the schema, as it's likely considered 
proprietary.  I'll check though.

We did have some bizarre errors last Friday following initial 
replication of our databases.  The errors were generated by a process 
that communicates with all the databases and shuffles some data around. 
    Researching the error, I found references to problems with temp 
tables, although I'm told that this process does not use temp tables.

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

-- 
Until later, Geoffrey

Those who would give up essential Liberty, to purchase a little
temporary Safety, deserve neither Liberty nor Safety.
  - Benjamin Franklin