For one of my projects we have to perform a lifecycle on the hosts within a Management Cluster. To make this work we have to migrate a Hybrid vSAN Cluster to an All-Flash vSAN Cluster. The current hosts in the Management Cluster have magnetic spinning disks and vSAN was therefore running in a vSAN Hybrid configuration. The new hosts are equipped with nice NVMe drives and vSAN should therefore be running in an All-Flash (AF) configuration to have all the nice features available like Erasure Coding, Deduplication and Compression.
I had never done this before and the vSAN documentation (link) describes only replacing the disks of the hosts, and does not describe the procedure for replacing the Hybrid vSAN hosts by All-Flash vSAN hosts.
Well this is a good opportunity to get my hands dirty again with some labbing. To the Lab!
Lab Setup
- Software used
- vCenter 7.0 u1
- vSphere 7.0 u1
- vSAN 7.0 u1
- Hosts – Hybrid – 1x 8GB Cache SSD 1x 64GB Capacity HDD
- dcesxi71.vmbaggum.local
- dcesxi72.vmbaggum.local
- dcesxi73.vmbaggum.local
- Hosts – All-Flash – 1x 8GB Cache SSD 1x 64GB Capacity SSD
- dcesxi74.vmbaggum.local
- dcesxi75.vmbaggum.local
- dcesxi76.vmbaggum.local
Let’s get started
I’ve prepared vSphere Cluster “CL-03” with the 3 Hybrid hosts and made sure the vSphere and vSAN Cluster status is healthy.
vSAN is enabled, and as you can see everything is pretty default.
Under Configure -> vSAN -> Disk Management the vSAN Disk Groups are shown. Also in the column “Type” you can see that the vSAN Disk Groups are of the type Hybrid.
After adding the 3 All-Flash ESXi hosts to the vSphere Cluster “CL-03” they also show up in the vSAN Disk Management. But as you can see no disks are in use for the newly added hosts.
The next step is toch select Claim Unused Disks, review the settings and make sure the correct disks are claimed for the Cache and Capacity Tier. Select Create to claim the disks for vSAN.
After the disks are claimed and added to the vSAN Disk Groups they now show up as “All Flash” in the column “Type”.
The next step is to exit the hosts from Maintenance Mode and let them participate in the (vSAN) Cluster.
Select “Yes” to exit the hosts from Maintenance Mode.
Next up is the removal of the “old” Hybrid hosts, enter the hosts into Maintenance Mode.
Select “Yes” to enter the hosts into Maintenance Mode.
In the next step, make sure you select “Full data migration”! And select OK.
After all VMs and data is migrated off the Hybrid hosts, they are put into Maintenance Mode.
Now you can check if your VM objects are placed on the new All-Flash hosts. You can do this under Monitor -> Virtual Objects -> Select a Hard Disk -> View Placement Details
Now all VM objects are running on the new All-Flash hosts, we can remove the Hybrid hosts from vSAN. Under Configure -> Disk Management -> Select a Disk Group (from a Hybrid host) -> … -> Remove
Read the warning messages and select Remove if you accept the warning messages.
After the previous step has been applied to all hosts that need to be removed from the vSAN Cluster, the vSAN Disk Management overview will display no available Disk Groups on the Hybrid hosts.
Now it is time to remove the Hybrid hosts from the vSphere Cluster. Select the Hybrid hosts and select Remove from Inventory.
Select Yes to remove the Hybrid hosts from the inventory.
Make sure the correct Hybrid hosts were selected and select Yes.
After the removal task has been completed, you’re left with a nice and clean All-Flash vSAN Cluster!
Now you can even make use of a feature only available under All-Flash vSAN like Space Efficiency and enable Deduplication and Compression.
As you can see super simple and straight forward process, some test VMs that I ran on the cluster while performing the actions above weren’t impacted. I hope this article helped you, leave a comment below if you have any further questions.
