On Wed, 2006-12-20 at 16:13, Norman Yamada wrote:
> Initial replication on the medium sized table on this slave node took  
> about 16 hours(!), and the replication of the entire database is  
> still running some 20 hours after I started.

I had this kind of symptoms when the indexes on the slave were not
dropped during the sync due to some issues related to failing truncate
due to foreign key references. This issue is supposed to be fixed in
1.2.2, so maybe there's another scenario when the indexes are not
dropped ? Is that the problem at all ? (easy to check, just see if you
have active indexes on the currently copied table).

Cheers,
Csaba.