Regarding FAQ entry "Active Directory login while in the registration VLAN": 

Make sure you do *not* have a wildcard A record for Active Directory's DNS 
subdomain. This would break _sites.DomainDnsZones and the like.

To support Mac/Linux krb5 clients and to avoid timeouts, allow ports 88,389,636 
for both UDP and TCP.

The FORWARD chain defaults to DROP. So you need to duplicate the custom -A 
input-internal-vlan-if rules, or maybe define a new named chain that gets 
appended to both input-internal-vlan-if and FORWARD. I don't think we want all 
of input-internal-vlan-if chain allowed for FORWARD.

Something like:

-A input-internal-vlan-if --protocol udp -m multiport -d (dc) --dports 
88,389,636,137,138 -j ACCEPT
-A input-internal-vlan-if --protocol tcp -m multiport -d (dc) --dports 
88,389,636,139,445 -j ACCEPT 

-A FORWARD -i eth1 -s (registration) -d (dc) -p udp -m multiport --dports 
88,389,636,137,138 -j ACCEPT 
-A FORWARD -i eth1 -s (registration) -d (dc) -p tcp -m multiport --dports 
88,389,636,139,445 -j ACCEPT 
-A FORWARD -i eth1 -s (isolation) -d (dc) -p udp -m multiport --dports 
88,389,636,137,138 -j ACCEPT 
-A FORWARD -i eth1 -s (isolation) -d (dc) -p tcp -m multiport --dports 
88,389,636,139,445 -j ACCEPT 
 
Next, not everyone has to worry about this, but I happen to have my domain 
controller, the PF management web interface, and the PF server interface 
closest to the domain controller on 3 different firewall interfaces. Loops, 
asymmetric routing, and leaks (firewall bypass) are possible if we're not 
careful. To ensure that packets in a flow go in and out the same interface, use 
route tables.

In this example:

10.1.1.0/24: Registration
192.168.1.0/24: Isolation
172.16.1.0/24: Production network including the DCs
172.16.1.1: PacketFence server internal router interface
172.16.1.254: Router facing PF internal router interface
172.17.1.1: PacketFence server management interface
172.17.1.254: Router facing PF management interface

Tell your router to use PF as gateway to the reg/isolation networks. In 
Cisco-speak:

ip route 10.1.1.0 255.255.255.0 172.16.1.1
ip route 192.168.1.0 255.255.255.0 172.16.1.1

# edit /etc/iproute2/rt_tables to include:
200 ROUTE10
201 ROUTE172

# There's probably a more "correct" way to do this, but I put
# it in /etc/rc.local.
# Make sure that traffic is routed out the correct interfaces,
# to avoid asymmetric routing.
ip rule add from 10.1.1.1/24 to 172.16.1.0/24 table ROUTE10
ip rule add from 192.168.1.0/24 to 172.16.1.0/24 table ROUTE10
ip rule add from 172.16.1.1 to 172.16.1.0/24 table ROUTE10
ip route list table ROUTE10
ip rule add from 172.17.1.1 to 172.16.1.0/24 table ROUTE172
ip route list table ROUTE172
ip route add default via 172.16.1.254 table ROUTE10
ip route add default via 172.17.1.254 table ROUTE172
ip rule ls
ip route flush cache

Wouldn't it be much easier just to NAT instead of route? Yes. But then your 
logs are wrong.

------------------------------------------------------------------------------
Virtualization & Cloud Management Using Capacity Planning
Cloud computing makes use of virtualization - but cloud computing 
also focuses on allowing computing to be delivered as a service.
http://www.accelacomm.com/jaw/sfnl/114/51521223/
_______________________________________________
Packetfence-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/packetfence-users

Reply via email to