is anybody else facing with error when starting up a VM on the vVol datastore? I have an error failing to create .vSphere-HA folder for HA protected machine list.
From VASA provider log the message as follows:
Caused by: com.netapp.offtap3.ontap.ApiFailedException: Invalid VM name: _vSphere_HA. VM names must be between 1 and 254 characters in length. The first letter must be one of A-Z or a-z. The remaining characters must be A-Z, a-z, 0-9, space or valid ASCII punctuation. (errno=13001)
Does anybody knows how to solve that, or is there a "secret" patch for this issue?
The .vSphere-HA folder is created automatically by vSphere and it is not the VM!
Any news about this issue? I have found, that there is KB article from June with linking to Bug 1212933, but I have been creating SR in March and till that time NetApp was not able to solve this issue?? How long it will take?
We have identified changes in ONTAP and the virtual appliance (VASA Provider) to address this. The ONTAP changes are targeted to 9.6P1 and 9.5P8. The virtual appliance changes are in the 9.6 release, targeted for August release. Both are undergoing QA testing, so these may change. I hope this helps.
I tried 9.6P3 and as you guessed, it has not been fixed there yet. I then downgraded my test environment to 9.5P8.
It does look like 9.5P8 improves the .vSphere-HA behavior when you enable HA. I've created both NFS and iSCSI-based vVols and I'm now cloning some VMs to them with HA enabled. For the iSCSI vVols I was able to see .vmdk files appear in OnCommand's LUNs pane as expected. I am getting an "Object type storage alarm" in the Datastores navigation pane, but that could be due to Protocol Endpoint issues with my test environment and/or the downgrade.
Thank you, it was just one of my two test hosts that was throwing the errors, so your vvold solution may have helped. What I ended up doing was deleting all vVols, disabling VASA in the VSC, re-enabling it, then recreating the vVols. I'm not getting any errors now, and I would say that 9.5P8 has been solid as far as vVols go. I won't upgrade my test environment to 9.6Px until the release notes say that vVols have been fixed in that version.