vCSA restore failure in vSphere 6.7 environment

[Asahi]

Hello,

Does anyone have any information about vCSA restore failure on Veeam B&R?

Our customers use Veeam B&R to get a backup of their vSphere environment.
ESXi 6.7 Update2 Build:13981272
vCSA 6.7 Update 2c Build:14070457
Veeam Backup & Replication 9.5 Update 4b (9.5.4.2866)

The environment has three ESXi hosts and is connected to shared storage.
The original vCSA exists in ESXi1.

As a restore test, the original vCSA was stopped from ESXi1 and deleted.
After that, Veeam B & R performed Entire VM Restore on the ESXi1 host.
However, the restore fails with the following error:

Code: Select all

==============
[17.10.2019 18:48:10] <01> Info [Soap] CreateVirtualDisk, vmRef '58', diskLabel 'Hard disk 5', controllerKey '1000', diskUnitNumber '4', diskKey '2004', diskPath '[datastore_name] vcsa/vcsa_4.vmdk', capacity '10737418240', thinProvisioned 'True', zeroEaged 'False', profileId ''
[17.10.2019 18:48:10] <01> Info [VimApi] ReconfigVM, type "VirtualMachine", ref "58"
[17.10.2019 18:48:27] <01> Error Unable to write VMX file: /vmfs/volumes/5d4412be-742bb8b0-88af-20677cf09644/vcsa/vcsa.vmx.
[17.10.2019 18:48:27] <01> Error Failed to lock the file
[17.10.2019 18:48:27] <01> Info Host [xx.xx.xx.xx] is not a virtual center, skipping vddk access routines
[17.10.2019 18:48:27] <01> Info [Soap] Connection 'xx.xx.xx.xx:443:root:False::0:1' is disposing.
[17.10.2019 18:48:27] <01> Info [AgentsSession] Disposing...
[17.10.2019 18:48:27] <01> Info [AgentsSession] Removing agents connection
==============

This restore succeeds without problems if you restore it with a different name instead of the original.
Attempting to restore to the original location with vCSA removed will fail.

Kind Regards,
Asahi,
Climb Inc.

[HannesK]

Hello,
interesting... what is the case number for that issue? If you have none, please open a case and post the case number here for reference.

Best regards,
Hannes

[Asahi]

Hi Hannes!

Thank you for reply.
Case number is below number.
03826838

Kind Regards,
Asahi,
Climb Inc.

[Asahi]

Hi Hannes!

In this case, there is an update.
vSphere HA may be the cause of the problem.

To isolate the problem, after removing vCSA, we restored the information to the original datastore using VM Files Restore.
After that, VM inventory registration was performed on the ESXi host.

Then vSphere HA worked, moved from the original ESXi1 to ESXi3 and started.

After that, vSphere HA was disabled before removing vCSA.
When vSphere HA is disabled, Entire VM Restore was successful even after vCSA was deleted.
Perhaps there is a possibility that the file is locked on the VMware side by executing the restore when vSphere HA is enabled.

Detailed information was uploaded to 03826838 along with the logs.

Kind Regards,
Asahi,
Climb Inc.

[Asahi]

Hello,

This issue was caused by vSphere HA.
A workaround from Veeam has also been published in KB.

https://www.veeam.com/kb3141

Kind Regards,
Asahi,
Climb Inc.

[HannesK]

Hello Asahi,
thanks for posting the solution.

Best regards,
Hannes