Thank you for your suggestion. My initial example was different to the environment I am working on, and so to progress this I built up the example network. Both switches are now N2048p units.
Note I only added LACP to be as close to my development network as possible. Only when I plug a device into Gi 1/0/2 on Switch A do I get a route showing:
Default Gateway is 172.16.1.1
S *0.0.0.0/0 [253/0] via 172.16.1.1, Vl2
C *172.16.1.0/24 [0/1] directly connected, Vl2
C *172.16.2.0/24 [0/1] directly connected, Vl3
If nothing is plugged in I get:
Default Gateway is 172.16.1.1
S *0.0.0.0/0 [253/0] via 172.16.1.1, Vl2
C *172.16.1.0/24 [0/1] directly connected, Vl2
As I want to make all ports setup for 802.1x, there is no guarantee a device will be present 100% of the time to keep the route up. As a minimum I would like the route there permanently for testing. Is there a way of achieving this? That is, the switch keep the route alive as it is the gateway for the subnet.
I looked into the loopback option and it works for management, but as it is apart from the switching backbone it cannot be applied to other ports.
The reason these are not on the trunk now is to avoid sending unneeded packets across a link, optimising both the link and core by reducing load. It does provide one solution which is plan B at this time and not the solution I am trying to achieve. There are a lot of cases, and I have seen it myself, where a device continuously broadcasting takes down a vlan/switch. This solution is in part to reduce the issue to a building rather than potentially impact on the core, as this is also a BYOD environment.
Looking at other vendors for ideas it seems Procurve can do this, but Cisco and Juniper have similar issues. Other than sacrificing a port it does not seem possible, but as I am new to Dell it is here I am hoping to discover the best way,
On reflection, I do have enough to achieve the results required however any future testing/fault finding will require extra steps. I can live with this. Thank you for confirming the characteristics of the N2000 series switches.
acoyle
4 Posts
0
December 20th, 2015 15:00
Thank you for your suggestion. My initial example was different to the environment I am working on, and so to progress this I built up the example network. Both switches are now N2048p units.
Config A:
!
configure
vlan 2-3
exit
slot 1/0 9 ! Dell Networking N2048P
stack
member 1 9 ! N2048P
exit
ip routing
interface vlan 1
ip address dhcp
exit
interface vlan 2
ip address 172.16.1.2 255.255.255.0
exit
interface vlan 3
ip address 172.16.2.1 255.255.255.0
exit
ip default-gateway 172.16.1.1
username "Admin" password d41e98d1eafa6d6011d3a70f1a5b92f0 privilege 15 encrypted
!
interface Gi1/0/1
channel-group 1 mode active
exit
!
interface Gi1/0/2
switchport access vlan 3
exit
!
interface port-channel 1
description "Example LACP trunk"
switchport mode trunk
switchport trunk allowed vlan 1-2
exit
snmp-server engineid local 800002a203f8b156940cae
exit
Config B:
!
configure
vlan 2
exit
slot 1/0 9 ! Dell Networking N2048P
stack
member 1 9 ! N2048P
exit
ip routing
interface vlan 1
ip address dhcp
exit
interface vlan 2
ip address 172.16.1.1 255.255.255.0
exit
ip route 172.16.2.0 255.255.255.0 172.16.1.2
username "admin" password Passw0rd privilege 15 encrypted
!
interface Gi1/0/1
channel-group 1 mode active
exit
!
interface port-channel 1
description "Example LACP trunk"
switchport mode trunk
switchport trunk allowed vlan 1-2
exit
snmp-server engineid local 800002a203f8b156940ca6
exit
Note I only added LACP to be as close to my development network as possible. Only when I plug a device into Gi 1/0/2 on Switch A do I get a route showing:
Default Gateway is 172.16.1.1
S *0.0.0.0/0 [253/0] via 172.16.1.1, Vl2
C *172.16.1.0/24 [0/1] directly connected, Vl2
C *172.16.2.0/24 [0/1] directly connected, Vl3
If nothing is plugged in I get:
Default Gateway is 172.16.1.1
S *0.0.0.0/0 [253/0] via 172.16.1.1, Vl2
C *172.16.1.0/24 [0/1] directly connected, Vl2
As I want to make all ports setup for 802.1x, there is no guarantee a device will be present 100% of the time to keep the route up. As a minimum I would like the route there permanently for testing. Is there a way of achieving this? That is, the switch keep the route alive as it is the gateway for the subnet.
acoyle
4 Posts
0
December 21st, 2015 21:00
Thanks Daniel
I looked into the loopback option and it works for management, but as it is apart from the switching backbone it cannot be applied to other ports.
The reason these are not on the trunk now is to avoid sending unneeded packets across a link, optimising both the link and core by reducing load. It does provide one solution which is plan B at this time and not the solution I am trying to achieve. There are a lot of cases, and I have seen it myself, where a device continuously broadcasting takes down a vlan/switch. This solution is in part to reduce the issue to a building rather than potentially impact on the core, as this is also a BYOD environment.
Looking at other vendors for ideas it seems Procurve can do this, but Cisco and Juniper have similar issues. Other than sacrificing a port it does not seem possible, but as I am new to Dell it is here I am hoping to discover the best way,
acoyle
4 Posts
0
December 22nd, 2015 14:00
On reflection, I do have enough to achieve the results required however any future testing/fault finding will require extra steps. I can live with this. Thank you for confirming the characteristics of the N2000 series switches.