Import large Virtual Machines into the ITS Private Cloud

Based on our experience importing virtual machines via OVA or OVF to the ITS Private Cloud, we have defined as “large” any OVA or VMDK larger than 300GiB.

OVA and OVF with VMDK files smaller than 200GiB, can easily be import with the process defined https://eis-vss.atlassian.net/wiki/spaces/VSSPublic/pages/1356955721 .

Anything larger than 300GiB may cause delay or even fail due to the overhead to the environment. In this case, we recommend following the process described below:

 Instructions

  1. If you have an OVA, you’d need to unzip/untar.

    1. For nix users:

      tar -xvf vm-1.ova
    2. For windows users:

      1. Update the extension from ova to zip.

      2. Extract the contents and you will have an ovf file and vmdk file(s).

  2. Upload the VMDK files via either the vss-cli or https://vskey-stor.eis.utoronto.ca

    # with the vss-cli vss-cli stor ul vm-disk-1.vmdk Upload vm-disk-1.vmdk to /vm-disk-1.vmdk in progress ⏫ Upload complete to vm-disk-1.vmdk, etag=e1a011438b19a919cea1bb5b632bfe7f-110, version-id=None ✅ # with scp and your vss credentials scp vm-disk-1.vmdk {username}@vskey-stor.eis.utoronto.ca:/ut-vss
  3. Publish (sync) the vmdk files to the ITS Private Cloud either via the Portal or the vss-cli:

    1. Follow steps listed

    2. Run the following command:

      vss-cli compute vmdk personal sync id : 66 status : SUBMITTED task_id : c8ac46c3-47b7-4c1b-9ed4-5069260f3242 message : Request has been accepted for processing
  4. List available files:

  5. Deploy a shell (empty) vm:

  6. Delete the empty 1GiB disk:

  7. Create a new disk backed by the OVF vmdk aka, import disk from vskey-stor:

 

Disk transfers with this methodology are faster but the conversion process, may take a couple of hours based on the size.

University of Toronto - Since 1827