Ever wanted to migrate your DaaS to a new cluster including storage? Actually this is not possible according to VMware DaaS.
We all know that with some tricks you can accomplish a lot ;-).
So how did I migrate my vdi workload to a new cluster and new storage?
Don’t forget all my clusters are in the same vCenter!
Lets start!
– Put your floating pool to 0.
– Migrate your Imported VM’s en Images to the new cluster.
– And perhaps your dedicated VM’s needs to migrated to the new cluster.
After the migration is done, you need to re-assign the tenant resources.
This trick can also be used when a tenant has to much resources and you need to decrease them!
– Un-assign the network from the desktop manager.
– Un-assign the compute resource from the desktop manager
– Reassign the compute and choose the new compute values
– Assign network
When you are done with these actions you have to check you images and Imported VM’s. It could be that the images and VM’s have a status Error. When you check the desktone log you will see:
2019-02-21 09:39:33,914 INFO [com.desktone.provision.service.impl.ProvisionServiceImpl]-[https-openssl-apr-4443-exec-9] Refreshed the datastores for ComputePool ‘d816489a-b039-462e-bf2d-9634a862b9ca’
2019-02-21 09:39:33,914 WARN [com.desktone.provision.service.impl.ProvisionServiceImpl]-[https-openssl-apr-4443-exec-9] Unable to determine datastore validity for gold pattern VM ‘DemoImage’, datastore ‘NFS01’ does not exist on ComputePool ‘d816489a-b039-462e-bf2d-9634a862b9ca’
solution:
edb=# select vm_path from t_general_machine where name = ‘DemoImage’ ;
vm_path
————————————————-
[NFS01] DemoImage/DemoImage.vmx
(1 row)
For each image and dedicated vm you need to update the vm_path column!
edb=# update t_general_machine set vm_path='[NewStore] DemoImage/DemoImage.vmx’ where name = ‘DemoImage’;
UPDATE 1
Important:
If you have instant clones you need to reseal that image(s), otherwise you cannot expand the assignment.
This will recreate the template, replica and parent objects.