One PPtP tunnel will be fine. You will create a unique EoIP tunnel (unique tunnel-id) for each “zone”. You will create a unique bridge per “zone” as well. Something like:
/interface bridge
add name=zoneAbridge
add name=zoneBbridge
(and so forth)
Then, you will add (under /interface bridge ports) the appropriate EoIP tunnels and ethernet ports for each zone.
Thanks!
Now I’m configuring first router (left router).
DHCP servers are only in private and service ethernet, the same that have a bridge (guest netwok isn’t bridged and eoip).
When connect cable to private I get correct ip, if I connect to service eth I get the same subnetwork ip.
This is relevant configuration code. A note:
First internal network: name=private 192.168.10.x/24 - bridged
Second internal network: name=service 192.168.11.x/24 - bridged
Third internal network: name=guest 192.168.20.x/24 - NOT bridged
This has moved beyond what I can offer free support for. Perhaps some others can offer input. FWIW, the configuration you posted looks fine at first glance, other than IP assignments, which belong on the bridge (not the physical interface)
The information I offered works here in my lab and in the 4 or 5 other networks I’ve configured with this scenario. Verify the bridge configurations on both sides to ensure that you do not have a bridge loop. Also, the IP addresses and DHCP server should be configured on the actual bridge interface (not the physical ports). I am assuming that you are not using a vlan configuration on top of the bridges, as this can cause some heartache under certain circumstances.
The interface facing the other router will need an IP address. This IP is used for the PPtP tunnel. The PPtP tunnel has an IP on both ends (of course) and this IP is used to terminate the EoIP tunnel. The Bridge that is used to bridge the EoIP tunnel and some other interface has an IP for the purpose of providing IP services to the end users. The interface that is in the bridge does NOT need an IP, as the bridge has one already.