Virtual Switch Configuration

[soda0091]

I currently have 7 NICS allocated to each of my ESX servers. Below is my configuration

1 Nic for the service console.

1 Nic for V-motion.

2 Nics for my Virtual Machines.

1 Nic for my View clients (just started testing)

2 Nics for my ISCSI Equal logic storage.

What I would like know is if I should move any of the Nics around or is this a good setup? I am trying to follow the connection path that a user would take from connecting from their desktop and accessing a VM server. Please advise if this is correct.

Example

A client tries to RDP into a VM files server on ESX server 1. Where will most of the traffic take place? On the service console virtual Nic or on the 2 Virtual Nics allocated to the virtual machines?

Please let me know if my thought process is right. If so I am thiking about adding another physical Nic to the service console.

[marcelo_soares]

Your setup is good. I would only create a secondary service console (maybe together with the vmotion one?) in case you loose the first one...

About your question, traffic made on the VMs only use the nics for the Virtual Machine portgroups. Does not matter the service you are using. Service Console are only for management.

Marcelo Soares

VMWare Certified Professional 310/410

Virtualization Tech Master

Globant Argentina

Consider awarding points for "helpful" and/or "correct" answers.

[dragooner]

Hi, you guy:

The traffic will be on the vm port group,when you connected the VM which used RDP. You should read the introduction about the service console,vmkernel and the vitrual machine port group.

BTW:in the cluster environment the service console must be redundant.

[soda0091]

Thank you for the feedback. I have a pretty good understanding of what each of the virtual switches are used for. But I still think that there has to be traffic through the service console NIC to get to the VM. The reason I am suggesting this is look what happens when a ESX server goes offline. The virtual machine must be migrated over to another ESX server in order to bring it up in the new ESX servers resources. If the VM was using only the VM virtual switch it should be able to continue to run on the orginal server right?

Please provide your thoughts.

[dragooner]

Nice question.

First, you should know which traffic is the management information,and which is the data information.

Second, you described a HA happen time. And at this time you should know anther thing is that you had opened the VM detecting which can be configed at the cluster “edit setting”. Help with the vmtools, your VC and ESX hosts can always detect your VM running.

When you use RDP to connect with your VMs, the data traffic will use the VM port group,at the same time VC and your ESX hosts use the service console to detect you VM’s status.

So, you can see,the management traffic will always use the service console, and the community information with the VM will be always used the VM port group.

BTW:When a host offline, VMs on it are not migrated,but restart on the other hosts or you can say other hosts restart those VMs.

[dragooner]

Oh,by the way:When a ESX server goes offline, the virtual machine is not migrated,but restarted on the other hosts.Or you can say the other hosts restart the vm.

[dragooner]

This communities's servers are really response slowly. I refreshed twice, it done‘t print my edit message.