IP address config for multiple switches with multiple VLANS

[KTCBOT]

All I am trying to figure out is if I have say 6 HP Procurve switches. Say my core switch is setup with Default_VLAN 192.168.1.1 , VLAN 2 192.168.2.1, VLAN 3 192.168.3.1. Now I want to configure my other Procurve switches to work with the existing core swtich VLANS. I know the ip address for the Default_VLAN must be unique (192.168.1.2). Now for VLAN 2 is the IP address 192.168.2.1 which is exactly the same as my core switch or do I have to have a unique IP (192.168.2.2)? I would say normally that each switch per VLAN should have a unique IP address, but how would that work with DHCP? Handing out DHCP I give all the computers the same default gateway, so it would not know how to route properly unless I had each IP address for each VLAN on each swtich exactly the same?

Any and all help will be greatly appreciated...

 

[VinceWhirlwind]

Generally, only one of your switches would be running as a layer-3 switch.
So the "Core" switch has all the default GWs on it, while each of the other switches just has a single IP address (in any one VLAN) which is used for no other purpose than managing that switch.

Alternatively, if every switch is running at Layer3 (which would not normally be sensible), then all the subnets(VLANs) don't exist on all the switches and each switch is the default GW for its subnet.

 

[KTCBOT]

Currently it is running with a Cisco 2801 as the main routing source (router-on-a-stick) with a HP Procurve 6600 acting as a core switch. From there is goes to 2800, 2500 and 2500 series switches. All of these switches have basic routing capability. So if I have my router and the 6600 switch configured with my VLANs and their routes. How do the rest of the swtiches need to be configured.If my router has IPs for VLANs that are say 192.168.1.3, 192.168.2.3, 192.168.3.3, core swtich has VLANS as 192.168.1.2, 192.168.2.2, 192.168.3.2, then what should my 2800, 2600, and 2500 series routers have defined for the VLANS. On the VLAN IP settings, I have disabled, manual and DHCP. Do I choose disabled and specify a manual IP. If manual do I start it with 192.168.1.4, 192.168.2.4 and 192.168.3.4 or do I match the IP settings of the core swtich?

Thank you

 

[VinceWhirlwind]

Your switches should each have necessary VLANs created on them, with no IP address assigned to each VLAN, EXCEPT the "management" VLAN.
So the switches should have one IP address only. This address is not used by any of the hosts on the network, except for accessing the switch for management purposes.

Make sure your "Core" switch is not misconfigured - if you have a router doing the routing, you shouldn't have a switch doing routing as well. What are all those extra addresses on the "Core" switch actually doing?

 

[KTCBOT]

Thanks for the reply VinceWhirlwind. From what I read my network is badly misconfigured and it seems as though the extra IPs on the core switch maybe causing problems or at the very least using up IPs.

I appreciate your help.

 

[VinceWhirlwind]

It's worth reviewing a sample traffic flow across your network at layer2 & layer-3 to understand what the switches do and what the router does:
- PC1 to PC2 on the same subnet:
PC1 has PC2s's IP address but not its MAC address, so PC1 broadcasts an ARP request which is propagated by all the switches on PC1's VLAN to every switchport on PC1's VLAN.
PC2 answers the broadcast, every other host ignores it.
PC1 addresses packets to PC2's IP address. It puts these on the network by encapsulating the packets in frames addressed to PC2's MAC address.
PC1 gives the frames to its switch. The switch reads the MAC address and uses its MAC address table to forward the frame out the correct interface.

If PC2 is on a different subnet, you get the same process, except:
- packets addressed to PC1
- frames addressed to default GW

So as you see, the IP address is totally irrelevant to your switches.

 

[KTCBOT]

IT seems to be working better after your recommendations. I appreciate all your help...