CSS Interface Redundancy

jmcglashan · ‎03-02-2004

Is it possible to configure interface redundancy so that if 1 interface on the master CSS fails, it fails over all interfaces configured on the master css, effectively failing over the whole box if there is an interface failure.

d.parks · ‎03-02-2004

If you are running box-to-box redundancy, you can apply the "redundancy-phy" command to the interfaces that you'd like to trigger a failover.

If you are running vip/interface redundancy, the redundancy-phy command does not apply.

jmcglashan · ‎03-02-2004

We are not running box to box redundancy, We are using virtual routers, we would like to failover all interfaces if one should fail.

stevehall · ‎03-02-2004

This concept of failing over all virtual routers when a single fails is what is referred to as "fate sharing". Typically it is done by configuring identical critical services to all virtual routers on a CSS. This is sufficient for the vast majority of installations.

There are some rare situations where this will not be the best solution. In software version 7.3 a new feature called "VRID Peering" was introduced. You can tie multiple routers together so they are bound. This would be similiar to HSRP tracking on a Cisco router.

You can read about fate sharing at the following link:

http://www.cisco.com/univercd/cc/td/doc/product/webscale/css/css_730/redundgd/vipredun.htm#1063408

You can read about VRID Peering at the following link:

http://www.cisco.com/univercd/cc/td/doc/product/webscale/css/css_730/redundgd/vipredun.htm#1109382

Hope this helps...

-Steve

d.parks · ‎03-02-2004

Another way to account for this if your traffic volume allows, would be to trunk both/all of your circuits over a single physical interface. I do this on one of my gigabit interfaces and it works well with VIP/Int redundancy.