Hi

We have a relatively simple replication setup - a single cluster with a 
single replication set, from a master node (1) to a single slave node 
(2). This has been working fine for a number of months.

I recently tried adding a second slave node (3) to this cluster, as a 
cascaded slave from node 2. While the replication *worked*, we were 
seeing some unexpected behaviour on the other two nodes - I wanted to 
ask whether this is usual, and just a misunderstanding on our part as to 
how cascaded replication works.

The intention was to set up node 3 with no direct access to the master 
node (1) - however, this appears to not be possible. Having added node 3 
(with a provider of node 2) to the cluster, the initial full data sync 
appeared to be made from node 1, not node 2 as I was expecting. After 
that point, while the replication was working correctly, the log-table 
truncates on node 1 weren't occurring until they had happened on node 2 
(and hence, if node 3 was lagging, the log tables would bloat on the 
master *and* on the first slave).

My assumption originally was that node 3 would be isolated from 1, and 
so node 1 wouldn't see any additional load caused by node 3 lagging. Is 
this purely a misunderstanding on my part, or does it sound like there 
has been a misconfiguration in our replication setup?

Any info would be extremely useful!

Many thanks
Ben