Yellow Bricks

by Duncan Epping

RE: VMFS 3 versions – maybe you should upgrade your vmfs?

· ·

I was just answering some questions on the VMTN forum when someone asked the following question:

Should I upgrade our VMFS luns from 3.21 (some in 3.31) to 3.46 ? What benefits will we get?

This person was referred to an article by Frank Brix Pedersen who states the following:

Ever since ESX3.0 we have used the VMFS3 filesystem and we are still using it on vSphere. What most people don’t know is that there actually is sub versions of the VMFS.

  • ESX 3.0 VMFS 3.21
  • ESX 3.5 VMFS 3.31 key new feature: optimistic locking
  • ESX 4.0 VMFS 3.33 key new feature: optimistic IO

The good thing about it is that you can use all features on all versions. In ESX4 thin provisioning was introduced but it does need the VMFS to be 3.33. It will still work on 3.21. The changes in the VMFS is primarily regarding the handling of SCSI reservations. SCSI reservations happens a lot of times. Creation of new vm, growing a snapshot delta file or growing thin provisioned disk etc.

I want to make sure everyone realizes that this is actually not true. All the enhancements made in 3.5, 4.0 and even 4.1 are not implemented on a filesystem level but rather on a VMFS Driver level or through the addition of specific filters or even a new datamover.

Just to give an extreme example: You can leverage VAAI capabilities on a VMFS volume with VMFS filesystem version 3.21, however in order to invoke VAAI you will need the VMFS 3.46 driver. In other words, a migration to a new datastore is not required to leverage new features!

Reader Interactions

Comments

  1. Doug says

    Great information — I have never been able to find this anywhere else.

    I do have a question about one of the statements here:
    “In ESX4 thin provisioning was introduced but it does need the VMFS to be 3.33. It will still work on 3.21”

    Should the first sentence read, “In ESX4 thin provisioning was introduced but it does NOT need the VMFS to be 3.33?”

    • Doug says

      Nevermind… I haven’t had enough coffee yet this morning and didn’t follow the quoting properly.

  2. Rick Vanover says

    I still think that if we can afford the time and I/O to migrate off and reformat; it should be considered.

    Plus, that may be a good time to wipe out an 1, 2 or 4 MB block size datastores!

  3. Brandon says

    Are there any differences in the VMFS 3.x filesystem itself, or is it just a flag in the metadata that identifies this verion of the driver formatted this volume? From your post it looks like there are no changes at the filesystem level at all.

    I think people get curious because when you look at the properties of a datastore, you see a “version” number there. I can see the purpose of that being in there, perhaps so you can see the version differences when there are major releases, ie. 2 -> 3.

  4. Ronald says

    Can you elaborate a little on the differences with regard to the scsi reservations? Does it make a difference in production environments. Also, do you know the changes in 3.46? Further, great article, very useful to know that the driver is key.

    • Duncan Epping says

      Difference for SCSI Reservations from 3.0 to 3.5 were huge as optimistic locking was introduced. It basically means that we combine several locks on a single host in a single combined batch to reduce overhead and decrease the “time”. This was even more improved when VAAI was introduced.

      • Ronald says

        Thanks for the reply. So basically upgrading your VMFS from 3.21 to 3.31 would have benefit, when it comes to SCSI-reservations? And upgrading to 3.46 would be even more beneficial, if you can use VAAI?

        • Ronald says

          Nevermind, just read the article (thoroughly this time ;)), and see that it is not.

  5. djlaube says

    Thanks Duncan for this post. However you don’t go into changing the VMFS driver? For example, I was onsite at a customer and they wanted to use thin-provisioning but their VMFS “version” was below the version that supported it. How do we fix that without migration to a new datastore? (which is what I did for them, using Storage vMotion)..
    Many thanks! Cheers!

    • Duncan says

      No, changing the VMFS driver is as simple as upgrading your ESXi host to the next release. There’s no need to do anything else.

  7. cgrvy says

    What are the features that we can expect to see in the next version of VMFS?

    Thoughts/wishes anyone?

  8. Florian says

    statement still correct using ESX5 but VMFS3?
    “…a migration to a new datastore is not required to leverage new features!”

  10. ela2014 says

    hello
    in my network i have vsphere server 4.1 and i want upgrade to 5.what happen if i want vmfs 3 upgrade to vmfs 5 for my VMs in vsphere?

  11. ela2014 says

    already we have used vsphere 4 with cracked license and we have plane to upgrade to vsphere 5,meanwhile we have vsphere 5 cracked license .what problem

    will be occur for license during upgrade from vsphere 4 to 5 ?

    tnx all

    • Duncan says

      Are you seriously asking me if I can help while you have a “cracked license” for vSphere? You do realize I am a VMware employee?

  12. Troy says

    @ela2014 ha ha ha what an awesome question.
    Duncan maybe you can give him the answer but he will have to find a crack to be able to read it.

    • Mo says

      @ela2014 – Why use a cracked license, when you can generate 60 days licenses via email address every two months. you will still be legal (for evaluations) and will be able to ask questions freely!

  13. Nick says

    Wondering if your assertion that: ‘All the enhancements made in 3.5, 4.0 and even 4.1 are not implemented on a filesystem level but rather on a VMFS Driver level or through the addition of specific filters or even a new datamover”, applies to ESx5i aswell?
    For example after upgrading to ESX5i, do I need to upgrade my datastores to VMFS5 to get the benefits of >2tb datastores etc?