storage vmotion to convert disks to thin not working

[kwg66]

I've been migrating and converting disk formats from thick to thin and vice versa for over 15 years, now all of a sudden a bunch of migrations from VMFS to NFS datastores some of the VMDKs did not convert to thin and are still showing up as eager zero thick.

Does anyone have an explanation for this issue?   These disks need to be thin, why aren't they converting to thin during the migration process?   Even when to manually edit each disk to be thin format at the destination they are still showing as eager zero thick.  

Using VMware ESXi, 6.7.0, 19997733

Thanks

Ken

[maksym007]

So you are trying to convert disks during storage migration from VMFS to NFS, right? 

[kwg66]

@ makssym007 

yes, the existing storage is block VMFS6, new storage is NFSv3 hosted on NetApp.   All we are doing is migrating.  I even selected the option to manually edit the disk type during the migration making sure all were set to thin.   Past migrations to NFS datastores hosted in a Nutanix cluster we never had to do this, it automatically just converted to thin.   That said, we have VIBs from NetApp installed on the host that would allow for an eager zero thick disk if desired, but at this time this isn't desired, thin was specifically selected.   This was brought to me attention by some of our admins and I tested it out myself to make sure they were correct that this happening, and it is.   We have already migrated 100s of VMs into this new array, all thin. no problems, so I'm stumped as to why a handful of VMs this is not working as it should.   

[maksym007]

Now I am risking writing that - but what if - it will be exactly the issue. That you are migrating from VMFS - to NFS. 

And not like from VMFS to VMFS. 

have you tested in that way?

[Rubeck]

Hi.. 

Even though I'm sure you've already seen this I'm linking it anyway as it might be related:

https://kb.netapp.com/Advice_and_Troubleshooting/Data_Storage_Software/vStorage_APIs_-_Array_Integra... 

/Rubeck

[kwg66]

Hi Folks - great thoughts and contributions, but no go,

Here are the things we've tried in our efforts to remediate this issue:

1) uninstalled the Netapp VAAI plugin so the disk creation operations are handled by the host instead of off loaded to array, and went through the NetApp KBs on this to verify our settings and configurations 

2) tried migration both powered on and off

3) removed VM from inventory and reregistered as new machine 

4) tested migrations between VMFS and NFS in every direction and also from NFS to NFS and VMFS to VMFS datastores 

5) verified driver version for compatibility and known bugs 

Note, I saw something on spice works where  someone states that they ran sDelete in the guest OS, not sure why this would have any affect on the operations at the host level but any thoughts on this?    

Thanks

[kwg66 · Accepted solution]

Hey Folks - VMware support finally replied to me regarding this issue, and you aren't going to believe it.

They pointed me to this KB:

https://kb.vmware.com/s/article/68107

Note the language within it:

“Disk provisioning type should be seen as a creation hint.”

Talk about bull**bleep**, VMware has gone to **bleep** and I'm saying it right here and now.   Starting w 6.7 and newer yoru VMDK creation type is a hint, not guaranteed.   

Who ever came up with this idea instead of a fix is a complete moron.   MORONS!!!!@E$^@#$^@#$^!    

Please forward my  remarks to the manager of the product team, or whoever made this douchbag decision....

VMWare really does want everyone to migrate to AWS don't they...