Bah, probably bug.

[Johan De Meersman]

I've already found bug #37454 <http://bugs.mysql.com/bug.php?id=37454> but
apparently it's not considered a bug as described there. To me it damn well
seems like one.

Here's how to replicate:

   1. CREATE USER on master -- user gets created on both sides
   2. apply same create user statement on master again -- master reports
   error; slave replication fails on the statement.

Now, I might be mistaken here, but aren't failing statements supposed to not
end up in the binlog ? This is on 5.0.51a-24+lenny4-log. Can anyone confirm
this behaviour, or tell me if they know in which release it's been fixed ?

I usually don't bother with create user - I trust the first GRANT to manage
things. A colleague killed replication several times in the last few days
doing this, however.

Full step-by-step on http://pastebin.com/K43rmgLH


-- 
Bier met grenadyn
Is als mosterd by den wyn
Sy die't drinkt, is eene kwezel
Hy die't drinkt, is ras een ezel