I have set up a replication cluster consisting of 3 nodes. Node 1 is the 
master, Nodes 2 and 3 are the slaves. Node 2 listens to events generated 
on node 3 via node 1 and vice versa. (Both 2 and 3 listen to events 
generated on node 1 on node 1, of course - I have 6 listen paths 
altogether). When I am trying to failover to node 3, here's what happens: 
 
The failover command is issued on node 3 and succeeds. 
Tables on node 2 are updated fine, i.e. according to sl_set on node 2, set 
1 now originates from node 3, but.. 
... according to sl_set on node 3, set 1 still originates on node 1!  As a 
result, I cannot drop node 1, not to mention the fact that the replication 
process doesn't work as it should.... 
 
Does anyone have any idea what goes wrong here? I have been trying to 
solve this problem for quite some time now, but to no avail.... 
Thanks in advance, 
 
Kris