VMFS 3 versions and upgrade paths (1005325)
All VMFS 3 versions are fully-supported from VMware ESX 3.0 and onward. For example, VMware ESX 3.0 can store and access files on a VMFS v3.33 datastore created by VMware ESX 4.1, and the same VMware ESX 4.1 host can also store and access files on VMFS 3.21 volumes created by a VMware ESX 3.0 host.
VMFS 3, which was released initially with ESX 3.0.0 as version 3.21, has evolved with new minor versions:
- ESX 3.0.0 is provided with VMFS 3.21 (initial release)
- ESX 3.5.0 is provided with VMFS 3.31
- vSphere (ESX 4.0) is provided with VMFS 3.33
- vSphere (ESX 4.1) is provided with VMFS 3.46
- vSphere (ESXi 5.0) is provided with VMFS 5.54
It is possible to upgrade your VMFS minor version.
Warning: This removes the formatting of the LUN and all the data on the datastore. Relocate your virtual machines and files prior to removing the datastore.
To upgrade your VMFS minor version:
- Open and log into the VMware vSphere or Infrastructure Client.
- Migrate all the data off of the VMFS datastore that you are upgrading. For more information, see Migrating virtual machines (1017769).
- Select the Configuration tab and the Storage view.
- Browse the datastore in question and confirm that there is no remaining content. If there are templates, ISOs, or virtual machines still residing on the datastore, migrate them or remove them.
- When ready, right-click on the datastore and click Remove. Alternatively, select the datastore in the Storage view, and select Remove.
Note: If there are any ongoing locks or any objects still referencing the datastore, the removal attempt will fail, preventing removal of the datastore while it's potentially in-use. Verify your inventory items and confirm that there are no outstanding references to the datastore if this occurs.
- Recreate the datastore from a current VMware ESX or ESXi host. This ensures that the LUN is formatted with your intended or desired version of VMFS. Click Storage > Add Datastore.
- Migrate the virtual machines back to the respective datastore using the same procedures in Step 2.