This post is apart of the original VCAP6 3.2 Objective Post: Link
First, I’ll migrate my LAN/Management network. Keep in mind I’m using VM Workstation for this so the setting may be different in a different environment. I.E. vlan IDs, security policies etc…
ESXi-01
ESXi-02 (With running VM)
Single ping was dropped which is expected. Successful migration of the management/LAN vswitch. Now we can go into the vSS on each host and remove it since it’s not needed anymore.
Confirm the standard switch can be removed. (Nothing is configured on it).
Now let’s continue to do the same for the rest of the switch migrations.
vMotion, NFS and iSCSI.
vMotion Switch Migration (Template Mode)
Add hosts
Adding both hosts (Only using 2 until I get to the vSAN objectives).
Select our template host (ESXi-01 in this case)
Migrating physical adapters and vmkernel’s.
Assign Uplinks to template host
Once complete, apply to all to copy the configuration to the rest of the hosts.
Migrate vmkernel adapters
No impact. Good.
All good!
Let’s test to make sure we can still vMotion.
Success!
Let’s do the remaining pieces (NFS and iSCSI)
NFS Switch
No downtime. Performed the same steps as above.
Now for iSCSI. We make sure we get our port bindings setup correctly. Everything else is the same.
On the vDS Switch.
Bind Uplink 1 to iSCSI1 port group. Do the opposite to the other port group. See below.
Bind Uplink 2 to iSCSI2 port group
It should look like this when complete.
Now let’s proceed but we’re going to skip a step to show the importance of what we’re missing.
Let’s proceed with the migration of the iSCSI vSS to vDS like we did with the other switches.
One thing to note is that we will want to make sure our port groups a bound to the vmkernel’s correctly because we want to use multi-pathing like we are going on our standard switches. See below.
Click Next.
Now we have an issue. We can’t migrate our iSCSI networking because it’s in use. So how do we accomplish this without any downtime? That’s easy!
Scroll down.
In order for us to migrate our iSCSI networking without incurring downtime we need to remove 1 iSCSI binding from our vSS so we can migrate it over to our vDS. Now in doing this, we are losing our MPIO and redundancy but we are able to migrate without downtime and this operation shouldn’t take very long to accomplish and allow for us to bring back MPIO and redundancy on our vDS. So let’s get started!
We need to remove 1 iSCSI binding from the standard switch.
Add host
No impact! Good!
Once finished, we need to go into the network bindings for the iSCSI adapter and re-bind the new port on the vDS.
Rescan, make sure you have 2 paths once again.
Proceed with migrating the other iSCSI vmkernel and port group.
Once complete, go back into the iscsi adapter and verify the bindings and paths are still alive.
Review the standard switch. If it looks like the one below we can remove it since we don’t need it any longer.
Successful migration of all vSS to vDS switches and didn’t miss a single ping!