Hi,

We are in the process of testing slony for replicating master-> single 
slave for now.
While testing it on static backup versions of our live database  we 
never ran into  problems.  Also while running it with postgresql-8.1 and 
slony1-bin 1.2.1-1 we experienced no strange behavior.
We are running debian etch and recently upgraded to postgresql-8.3 with 
slony1-bin 1.2.13-1 from debian sid.
By using the "by-hand" method of running slony we suddenly got this 
error in our frontend from postgresql:


[Mon Apr 07 13:00:59 2008] [error] [Hunchentoot] Database error XX000: 
cache lookup failed for type 76787


This was instantly fixed by stopping slony and using slonik to uninstall 
both nodes. To try to replicate the error i've set up a test again where 
the master db is static and not processing any queries. by inserting it 
with the query that caused the above error message nothing strange 
happend...
We also didn't see any messages in the slony logs that could be related.

I would like to test slony once more on our live db server but we can't 
really afford to run into this bug again. Looking at the replication 
solutions for pg we think slony would be the best solution for us, hope 
someone has a clue:)


Regards,

Gersom