Service Console redandancy

[jerry2096]

For service console redandancy, the kb article 1002641 saying that "Having an additional service console on a separate vSwitch bound to a different vmnic which is attached to a different physical switch ensures that VMware HA does not power off virtual machines if primary service console network goes down. If primary service console network fails, it continues functioning by using the 2^nd service console.

But how about assign 2 VMNICS to a pair of redandant network switches which setup same VLAN ID for the 2 ports, and then join them to the same Service Console port group?

Is that NIC teaming also possible?

If possible which is better for service console redandancy in an HA environemnt?

[avlieshout]

I use a dedicated vSwitch for my COS and have 2 physical nics teamed up, which connect to different physical switches.

For the HA settings. In ESX 3.0.x the default setting is to powerdown the vms if the host gets isolated. I have changed this default to leave the vm's powered on, because I don't want HA to check my network.

I only want HA to kick in when my ESX host crashes.

In ESX 3.5 the default setting is to leave th vms powerd on.

The knowledgebase article wants you to configure a second COS on another network using another ip address. This is to reduce the HA false positive if only the COS network will be unreachable.

If you just do what I do and the COS network is unreachable on the complete network, than both adapters "fail" and HA kicks in. But as I said before I changed the default and I don't want HA to check my network.

Arnim van Lieshout

Visit my blog at: http://www.van-lieshout.com

[AntonVZhbankov]

My vSwitch with SC has 2 physical NICs. I just created 2 port groups on it - Service Console and VMkernel. For Service Console I set NIC0 active and NIC1 standby, for VMkernel NIC1 active and NIC0 standby.

---

VMware vExpert '2009

http://blog.vadmin.ru

[depping]

A secondary SC is more redundant. But if you increase the isolation response time to 60 seconds and just add an additional nic to your Service Console portgroup you should be fine. That's what I usually do to keep the config simple. Of course you would want to connect this secondary nic to a different physical switch, and you might want to add an extra isolation address to prevent false positives from happening.

Duncan

VMware Communities User Moderator

-

---

Blogging:

Twitter:

If you find this information useful, please award points for "correct" or "helpful".

[jerry2096]

Thank you everyone for the answers!

Yes, I think I want to try this simple configuration - teamed 2 NICs for both SC and VMkernel.

But I have questions about Active/Standby according to Cisico's recomandation showing in the attached picture.

1. Why use Active/Standby instead of Active/Active? Just for seperating the traffics for SC and VMkernel? Or any other concerns?

2. The Active/Standby configuration should be set at "Port Properties" or at "vSwitch Properties" or both?

thanks

[avlieshout · Accepted solution]

See it as a simple loadbalancing option.

You should configure it at the portgroup level. Select the optionbox to override vSwitch defaults.

If configured at the vSwitch level you are setting the options for all portgroups and that is not what you want at this point.

-Arnim van Lieshout

-

---

Blogging: http://www.van-lieshout.com

Twitter:

If you find this information useful, please award points for "correct" or "helpful".