On 8/3/2005 7:23 PM, Andrew Gold wrote:
> Hi everyone,
> 
> I've used the sample scripts to set up my first attempt at replication. 
> I screwed up on one of the lines and forgot to comment out the line 
> where it adds an index to the public.history table. The script quits as 
> it should but cannot be rerun as it exists because the new namespace 
> (_slony1, in my example) has been created.
> 
> I would like to rerun the script, but I need to clean up the partial 
> initialization (init cluster). Can I safely delete the entry in 
> pg_namespace, or are there side effects? Is there another way to clean 
> up a failed initialization? I am adding a pre-existing table from a live 
> database, so I need to be sure that removing the entry in pg_namespace 
> doesn't cause problems, although I suppose I could just create the same 
> scenario on a test machine ...
> 
> The table is a junk table which I created for testing, but the parent 
> database is valid.
> set add table (set id=1, origin=1, id=1, fully qualified name = 
> 'public.spam', comment='test table')
> 
> I'd like to know for sure.

There's a special slonik command for that.

http://gborg.postgresql.org/project/slony1/genpage.php?slonik_commands#stmt_uninstall_node


Jan

> 
> Thanks,
> 
> Andrew Gold
> _______________________________________________
> Slony1-general mailing list
> Slony1-general at gborg.postgresql.org
> http://gborg.postgresql.org/mailman/listinfo/slony1-general