Qs about UCS

minja.miletic · ‎04-25-2010

Hi,

I have a couple of questions for all of you. I have not seen something similar here, so that is why I am posting...

1. When I am creating vNIC, I have a option to chose should traffic go through Fabric A or B. I also have a pin group as a option. So I am wondering what is going on, if I chose Fabric A, and ping group where I am pointing to Fabric B some uplink port? Is this collision, or it is going to work somehow?

2. I have also to select in vNIC wizard, if I want to enable failover. In Network Control Policy I have a option to select vNIC link down, if uplink goes down. So my question here is if I do not select failover, but I have a Network Control Policy, and if Fabric A goes down, according my first rule here failover is not going to occur, but my vNIC is going to go down, according to the Network Control Policy, isn't by default(if vNICs are in teaming) second vNIC going to forward traffic to the Fabric B?

3. What is exactly the new feature in UCSM, vNIC/vHUB placement Policy? And vCon1 and vCon2 in the wizard?

Thanks

mm

minja.miletic · ‎04-29-2010

Hi to all again

I've got the answers so I am sharing it with u (I think it is very interesting):

1.

Selecting Fabric A or B would basically set the primary path for the vNIC.If you enable failover then if the primary path fails it will use secondary path. Esle it will not failover.

Eg. If you select A for vNIC1, then if FI-A fails the vNIC1 will fail.

If you had enabled failover then VNIC1 would go through FI-B.

Best practice - Do not enable failover if you are using ESX, as we have seen issues with failover. Better to use the vswitches take care of the failovers.

Pin group is used for specifying which uplink from the FI (A or/and B) the vNIC should use. If not defined, it will still get automatically pinned to one of the uplinks. If you select VNIC1 to Fabric A and pin group on Fabric B, it basically means that VNIC1 primary path will use FI-A, while if failover is enabled and VNIC1 has to go through FI-B it will use the pin group you have defined. If no pin group is defined it will just get automatically pinned.

Best practice- unless you want to particularly pin the VNIC traffic, its better to leave it automatic.

2.

For each VNIC there is a vif created on the respective FI. If you enable failover there will be a vifs created on each FI (A and B) for 1 VNIC.

In that case when one of the Uplinks on the FI fails, the particular vif on that FI will be marked as down and other vif will be up.

3.

This is basically for venture with 2 Mezzanine cards on it. Each VCON is associated to a particular physical card. That way we can control which HBA vnic goes to which physical card. Each VCON goes to both Fabrics. Again conside each VCON as a physical Mezzazine card.

Jeremy Waldrop · ‎04-29-2010

Very useful information, thank you for sharing.