Virtual Connect (VC) Active/Passive failover based on bandwidth - It works!

chuckk281 — Thu, 02 May 2013 15:21:11 GMT

Some test results from Dan:

*******************

So I had heard about this before but up until recently did not have a chance to test it.

Let say you have 2 uplinks from VC Bay 1 and 2 more from VC Bay 2, and you are in an active/passive design with all 4 uplinks in a single SUS.

If Bay 1’s 2 uplinks are active, and a single cable is pulled or the SFP fails or something, your bandwidth is effectively cut in half.

Virtual connect in this in instance will almost immediately upon detecting the failure will fail over to Bay 2 because it has more aggregate bandwidth still available.

Tested this at a customer site last week and it worked very well.

All failure testing of VC modules resulted in 1 ping loss or less for every failure test we threw at it.

If you are like me, you have to see it to believe it, but hopefully this anecdotal evidence will increase your confidence level when talking about this feature.

**********

Dave also relpied:

******************

Dan, We simulate this failover with the ‘Disable’ option inside of the SUS.

Select an Active port pair and drop the Speed from ‘Auto’ to change it to ‘Disabled’.

No need to be onsite to pull cables…

****************

Comments or questions?

topic Virtual Connect (VC) Active/Passive failover based on bandwidth - It works! in BladeSystem - General

Virtual Connect (VC) Active/Passive failover based on bandwidth - It works!