Connecting two subnets together through procurve switches

[chrisjako]

We have had a Virgin Media 1GB Ethernet extension fibre link installed today between both our upper and lower schools which I’m running into our core HP 5406 switches at both ends.

 

They are on separate subnets upper school is 10.5.224.0/21 and lower school is 10.11.224.0/21. The domain and active directory are the same and also setup correctly in sites and services (they are currently connected through a 40mb VLAN)

 

Fibre end runs to the core switch 10.5.225.102 at the upper school is on port B22

The other end runs to the core switch 10.11.226.151 at the lower school on port B22

 

only thing I did was add an IP of the other subnet in the VLAN 1 of both switches and I also enabled ip rip in the config of both switches

 

Upper School Switch 10.5.225.102

(config) >ip rip

(config) >vlan 1

(vlan 1) >ip address 10.11.231.253/21

(vlan 1) >write mem

 

Josephs Switch 10.11.226.151

(config) >ip rip

(config) >vlan 1

(vlan 1) >ip address 10.5.231.253/21

(vlan 1) >write mem

 

From the above CLI commands I can get the switches to communicate with each other, and the switches themselves can communicate with any other device on both subnets however other devices and other switches still cannot communicate cross subnets

 

Hope this make sense can anyway help me to let me know what I’m missing or doing wrong please?

[SchoolsBroadband]

What does the routing table on both switches show?

[SYNACK]

They are both in vlan 1 so are not separated in layer 2 leaving all of your broadcasts traversing the link. Your client default gateways need to be the switch on each side so that all traffic checks there for a route and you should have a separate vlan inbetween them (vlan 2 ) for the link, this way you can have each core switch have vlan 1 with ip address 10.11.231.253/21 and vlan2 with 192.168.100.100

have a static route for the other network

ip route 10.5.231.253/21 192.168.100.101 (address of the other switch on the link) and a default route on each switch to their local internet connection.

 

This should mean that when a packet is received that is addressed to something non-local it will hit the switch, the switch will check and see the route for the other network (if that is where it is headed) and forward it to the vlan2 link, if not it will go out to the net (assuming no silly proxy rubbish in which case you would have no default route)

 

It needs to be setup on both ends as IP connectivity is end to end, ie it needs to know where to go at every step of the way to get there and then back.

[chrisjako]

They are both in vlan 1 so are not separated in layer 2 leaving all of your broadcasts traversing the link. Your client default gateways need to be the switch on each side so that all traffic checks there for a route and you should have a separate vlan inbetween them (vlan 2 ) for the link, this way you can have each core switch have vlan 1 with ip address 10.11.231.253/21 and vlan2 with 192.168.100.100

have a static route for the other network

ip route 10.5.231.253/21 192.168.100.101 (address of the other switch on the link) and a default route on each switch to their local internet connection.

 

This should mean that when a packet is received that is addressed to something non-local it will hit the switch, the switch will check and see the route for the other network (if that is where it is headed) and forward it to the vlan2 link, if not it will go out to the net (assuming no silly proxy rubbish in which case you would have no default route)

 

It needs to be setup on both ends as IP connectivity is end to end, ie it needs to know where to go at every step of the way to get there and then back.

 

how should the vlan 2 be tagged?

 

vlan1 untagged on port b22 and vlan2 tagged on port b22?

[SYNACK]

All vlans untagged, your link ports on each switch are untagged on vlan 2, the rest are all Vlan1, the router engine in the switches takes any packets that actually need to travel across the link and puts them across vlan 2 to be dumped out on the other network by the other switch. This way only packets that need to travel across the link actually do so you don't end up flooding it with rubbish and leaking stuff from one site to the other via broadcasts

[chrisjako]

o.k still not getting it to work

 

ive tagged port b22 on vlan 1 (forbidding cuts me off completely), untagged on vlan2 on both switches

 

my client machine is on the 10.5.224.0/21 subnet - I've set its default gateway IP address to 10.11.226.151 (which is the VLAN 1 ip address of the switch on the other network)

 

on the switches ive set the following:

 

switch on the 10.5.224.0/21 network

VLAN 1 IP = 10.11.226.151

VLAN 2 IP = 192.168.100.101

went into config cli and typed in : ip route 10.11.224.0/21 192.168.100.100

 

switch on the 10.11.224.0/21 network

VLAN 1 IP = 10.11.225.102

VLAN 2 IP = 192.168.100.100

went into config cli and typed in : ip route 10.5.224.0/21 192.168.100.101

 

with all the above in place its still not working, where am i going wrong!!

 

Thanks for your help so far btw!

[Trapper]

How far apart are your two sites?

[chrisjako]

How far apart are your two sites?

 

they are a mile apart

[SYNACK]

Default gateway should be the local switch, try ping from a client to the vlan2 address of the local switch, if this works routing is working on that one, next ping vlan 2 remote switch, then remote vlan 1 ip

 

you can also do:

tracert remoteVlan1IP

 

and see where it stops

Edited July 24, 2014 by SYNACK

[chrisjako]

SYNACK YOU ARE A LEGEND!!!!

 

I could ping the switch perfectly at the other site however couldn't ping the exchange server connected to it, change the default gateway of the exchange server to the local switch and everything started working fine

 

BIG BIG thank you for helping me out