Dear Support Community,
I've read myself a little bit through the various ways on how to deploy a redundant VPLS PE setup and wanted to ask for your advice on the best way to do so.
Do you see the following idea as a possible (and good) way to implement such a setup? And/or what are your concerns and recommendations?
Setup:
- Redundant Cisco Catalyst 6500 and Cisco ME 3800X as current MPLS PE devices at a small numbers of locations (maybe around 5)
- At some locations the 6500 PE devices could be converted to a VSS setup (which simplifies the redundancy setup), but that's not possible with the ME 3800X
- STP is individual per location (RSTP, MST ..)
Idea/Sample config:
- My idea is based on the Cisco white paper: http://www.cisco.com/c/en/us/products/collateral/data-center-virtualization/data-center-interconnect/white_paper_c11_493718.html#wp9000295 and the Cisco Press book "Interconnecting Data Centers Using VPLS"
- MPLS, VPLS, STP and IGP are not part of the sample config
- STP would be local between the PE and the CEs, Root Bridge is on the CE side.
- All Loopbacks are redistributed over the underlying IGP
- Loopback80 is the Semaphore which will be monitored by the PEs
- Loopback90 is used for the termination of the VPLS VC
- The Primary PE polls every second the route of the Backup Semaphore Loopback (80.80.80.2/32) -> Under normal circumstances this is down
- The Backup PE polls every second the route of the Primary Semaphore Loopback (80.80.80.1/32) -> Under normal circumstances this is up
- If the Backup PE doesn't see the the route of the Primary Sempahore Loopback anymore, he changes his own Loopback interface (Loopback80) and the Loopback for the VPLS VC termination (Loopback90) to "no shut" -> The Backup PE is now active
- When the Primary PE returns, his Semaphore Interface (Loopback80) and the Interface for the VPLS VC termination (Loopback90) are active as well. As soon as the Primary PE receives the route of the Backup Semaphore (80.80.80.2/32) over the IGP, he sets his Interface for the VPLS VC termination (Loopback90) to "shut", but leaves the Semaphore Interface (Loopback80) enabled.
- As soon as the Backup PE receives the route of the Primary Semaphore (80.80.80.1/32) over the IGP, he waits for 100 seconds and then sets his Semaphore Interface (Loopback80) and his Loopback Interface for the VPLS VC termination (Loopback90) to "shut" -> The Backup PE is no longer active
- As soon as the Primary PE doesn't see the route of the Backup Semaphore (80.80.80.2/32) anymore, he will set his own Interface for the VPLS VC termination (Loopback90) to "no shut" again. -> The Primary PE is active again
Sample config:
Primary:
--------
interface Loopback80
description P-Semaphore
ip address 80.80.80.1 255.255.255.255
interface Loopback90
description P-PW
ip address 90.90.90.1 255.255.255.255
track timer ip route 1
track 10 ip route 80.80.80.2/32 reachability
event manager applet VPLS-B-Semaphore-Up
event track 10 state up
action 1.0 cli command “enable”
action 2.0 cli command “conf t”
action 3.0 cli command “int lo90”
action 3.1 cli command “shut”
action 9.0 syslog msg “Backup N-PE is Active, Primary PW goes Standby”
event manager applet VPLS-B-Semaphore-Down
event track 10 state down
action 1.0 cli command “enable”
action 2.0 cli command “conf t”
action 3.0 cli command “int lo90”
action 3.1 cli command “no shut”
action 9.0 syslog msg “Backup N-PE is Standby, Primary PW goes Active”
Backup:
-------
interface Loopback80
description B-Semaphore
ip address 80.80.80.2 255.255.255.255
shutdown
interface Loopback90
description B-PW
ip address 90.90.90.2 255.255.255.255
shutdown
track timer ip route 1
track 10 ip route 80.80.80.1/32 reachability
delay up 100
event manager applet VPLS-P-Semaphore-Up
event track 10 state up
action 1.0 cli command “enable”
action 2.0 cli command “conf t”
action 3.0 cli command “int lo90”
action 3.1 cli command “shut”
action 4.0 cli command “int lo80”
action 4.1 cli command “shut”
action 9.0 syslog msg “Primary N-PE is Active, Backup PW goes Standby”
event manager applet VPLS-P-Semaphore-Down
event track 10 state down
action 1.0 cli command “enable”
action 2.0 cli command “conf t”
action 3.0 cli command “int lo80”
action 3.1 cli command “no shut”
action 4.0 cli command “int lo90”
action 4.1 cli command “no shut”
action 9.0 syslog msg "Primary N-PE failed, Backup PW goes Active”
Thank you and best regards,
Lukas
