Answer to q1, q2, and q3:
A DHCP address request broadcast heard on that interface will be forwarded to each of the servers, x.x.x.x and x.x.x.y. The requests will be sent sequentially, but it will be sent to both DHCP servers.
Each DHCP server that receives the forwarded request, will look for a scope that matches the IP subnet that the original broadcast was heard on. (The router or multilayer switch puts information in the request that it forwards, which allows the DHCP server to identify the correct scope.)
If the DHCP server has a scope for that subnet and some available addresses in that range, it will offer one and reply back to the router/switch that forwarded the original request.
The router/switch will then send the reply back out on the interface where the request originated.
The DHCP client will receive one or more offers. Generally, the client completes the address lease/negotiations with the first server that it receives an offer from. ***Note: This isn't always the first server listed in the configuration.***
Even though the helper-addresses are listed sequentially in the configuration, that does not mean that the first one listed always responds to the client first. If that server is busy and the second one is not, then the second server's reply will probably get there first. Or, if the second one's CPU is faster, it may process the request sooner.
In practice, it's a good idea to divide the address pool in half, and assign one half to one DHCP server, and the other half to the other DHCP server. Because you can't predict with 100%certainty how the addresses will be handed out. If the two servers are comparable in speed and utilization, though, it should work out pretty close to a 50%/50% distribution.
Hope this helps.
