6500 VSS fast-hello

[DuJin0509]

Hello folks,

I understand that the fast-hello is for dual-active detection, upon detecting the dual-active condition, the original active chassis enters into recovery mode and brings down all of its interfaces expect the VSL.

I am eager to know that what will happen if the VLS is up but the fast-hello link is down ? If active chassis will proceed the normal behaviour of  fast-hello mentioned above ?

In VPC enviroment, I believe that the if keepalive linkd down only that won't incur any service impact .

Your prompt reply would be highly appreciated.

[paul driver]

Hello

I am eager to know that what will happen if the VLS is up but the fast-hello link is down

This will be as if you haven't enabled fast hello at all, And as such if no outage is incurred between the active and standby switches then nothing will happen however is you do incurr such outage to the vsl link a situation will arise that both switches will think they should be active( dual active or split brain) scenario.

If this occurs then inter-vlan routing will be disrupted and multiple errors will be reported such as duplicate ip addressing, stp issues. etc.

res
Paul

[DuJin0509]

Hello Paul,

Thanks a lot for your prompt reply.

So may I understant it as in this scenario (fast-hello link down and VSL is up) there will be no service impact at all.  Am I correct ?

In this case,  why Cisco recommands to setup redundant fast-hello link on 6500? I recall that on Nexus plantform we ultilized mgt port to setup vPC keepalive link to save resource.

[paul driver]

Hello
Apologies I may have mislead you a little bit here-
I was stating the affects of not having the Dual active fast hello in a VSS environment.

Unfortunately I have not had any real exposure to DC VPCs but I am on the understanding the VPC keep-alive is a prerequisite to establishing a VPC peer link so if this wasn't configured the VPC peer link between each switch wouldn't establish.

Now if the VPC peer keep-alive and VPC peers were established then the keep-alive peer failed I am assuming the VPC peer link wont then have anyway to monitor each others active state thus the VPC peer link will also fail and you'll then incur a dual active state as mentioned earlier.

res
Paul

[Rolf Fischer]

You're correct, there is no service impact when the fast-hello link goes down while the VSL is up. However, you loose an important mechanism for avoiding very ugly problems in the case of a dual-active scenario.

To be on the safe side, it is recommended to use more than one dual-active detection mechanism for redundancy reasons. You can combine fast-hello (advantage: it is very fast) and ePAgP (advantage: no extra link needed), or use more than one fast-hello link or just use more than one ePAgP port-channel if you need to safe the link(s). Anything is better than having no dual-active detection.

HTH

Rolf

[paul driver]

Hello Rolf 

Regarding the OP query about vpc keep alive -You saying the vpc keep-alive in a vpc environment wouldn't bring down the vpc peer link even it's required for  a vpc peer link to establish 

res

Paul