HSRP and Ethernet Internet Handoffs

Sep 25, 2006 5 Replies
HSRP and Ethernet Internet Handoffs open original image

In the past - on Internet gateway routers I would track the serial interface (say HSSI for DS3) with HSRP running on the inside fast ethernet interface. Generally I'd have two Internet gateway routers sharing an HSRP VIP on the inside. Today it is common for Internet providers to give you an ethernet hand-off for say a 100Mb or 1000Mb handoff to the Internet. The ISP facing FE or GE interface is unlikely to go down and is not a reliable guage of whether connectivity to a peer is up or down. What would really be ideal would be if HSRP could track BGP states. e.g. if a connection to ISP A off inetgateA goes down - the inside HSRP VIP becomes active on router inetgateB. As far as I know that's just wishful thinking. But clearly other engineers have to deal with this. What approaches are you taking.



For reference here's Cisco DOC on HSRP. Note how tracking only refers to interface upness which isn't much help with an ethernet handoff..



formatting link
Thanks.



Michael



Hi amigan:

HSRP/VRRP are designed to be router/gateway redundancy protocols. They are used to backup the active router/gateway with the backup one in the case of link-down or device failure. They are not designed to monitor any L2 links (although "track" does exist, because when the link goes down, it makes the active router/gateway unusable, just like the router/gateway was down) or L3 sessions. So to monitor BGP sessions, you have to look for other solutions.

Cisco have altered how "track" works to give a more general mechanism.

rather than implement track for HSRP / VRRP / GLBP and various other clients that want to react to topology changes, they have built a way for routes, interfaces, pings etc to affect an "object" in the router, and then for that to be used to affect HSRP.

if you want more do a search on object tracking on their site, or see

formatting link

1st time i used this we had to have enterprise plus code on the router (in 12.2?) - you need to check what is required for the features you want.

Here's a thought.. If the ISPs would permit it - you could create a GRE tunnel from each inetgate to a device at the respective ISP. Use two a static route to the ISP GRE tunnel end point - the preferred to the neighbor and a higher cost backup route to dev null. That's to make sure the tunnel goes down when BGP goes down - else the traffic could flow to inetgate2 and to the ISP via the other ISP. Now you have an interface (say tunnel1) which goes down on in the event that BGP goes down. If HSRP tracks tunnel1 - now the VIP will move as required.

What do you think?

Hi amigan ,

Your idea is feasible. But before you decide, you should have a look at what stephen said, and the link he provided.

Cool - somehow I had missed Stephen's post. I will get that into the lab asap. Thank-you and thanks Stephan!

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required