Virtual Connect Fail over

chuckk281 — Fri, 03 Dec 2010 14:42:46 GMT

Erkan was needing some advice on implementing Virtual Connect failover links.

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

Hello All,

How can we configure the HA with Virtual connect , it isn’t working as an HA

The configuration is below ,

if Second or first Virtual connect module fail ( We powerd off the module ) All link fail 90 seconds , and then connection comes after minute

We are using the cookbook senaryo 1.3 (Scenario 1:3 – Multiple Simple Networks Providing Redundancy and Link Aggregation 802.3ad (LACP) with VLAN Tunneling – VMwareESX ,)

What is problem ? or which configuration can we use for Tunnel mode ( ideal configuration )

Thank you for your help

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

Lots of comments but I will post the one from Steve:

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

Erkan,

Some things have changed since that Scenario was written. Page 50 of the Cookbook states that BEACON PROBING should be used. This is no longer the case. SmartLink was not supported back then, and is now. You should enable SmartLink, configure for Link status only and route based on virtual port ID, as below. Also, ensure PortFast is enable on the upstream switch ports VC is connected to. Also, optionally enable BPDU Guard on the switch ports too.

If that does not work, please advise the following;

- What model of server do you have?

- What version of VC FW?

- What version of ESX, NIC fw and driver?

- Assuming Standard vSwitch?

Also, try your first tests by simply unplugging the uplink cables from the VC module, assuming that works fine, then try powering the module off or shutting down the switch ports.

I just when through this with a customer this morning and he now has it down to a single ping lost during fail-over.

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

Any other suggestions or questions for Erkan?

topic Virtual Connect Fail over in BladeSystem - General

Virtual Connect Fail over