VMware Active/Active configuration to failover-Amount of time required for Virtual Connect

[chuckk281]

Gautham was have some issues at a customer:

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

We were performing some acceptance testing for a VMware environment built on Blades + Virtual Connect in an Active/Active configuration(i.e. SUS_A & SUS_B). One of the tests involved removing all the uplinks from 1 VC module.

We noted that there were 6 dropped ping packets before the other NIC became active.

The customer is asking if this is normal?(The time/amount of packets before failover occurred). Is there any way to make the failover quicker?

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

Question from Kevin:

No.. You should sustain 1 packet drop.. Is the infrastructure switch configured for Portfast ?

Reply from Mahmoud and Gautham:

We found that the “admin-edge” feature on the ProCurve switch wasn’t enabled.

After we enabled it on all the ports, we received 1 timeout only if we remove all the links from one interconnect.

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

Other input or questions?