Creating a VMFS datastore

Use the IBM Storage® Virtualize Plugin for vSphere to automate the creation of vSphere Virtual Machine File System (VMFS) datastores.

Before you begin

Identify the vSphere cluster for which you want to create the new VMFS datastore(s).

Note: In multi-IO-group configurations, ensure that hosts are configured and zoned to all IO-groups on the storage system - volumes created in IO-groups that are not accessible to the host will not be accessible.

About this task

This workflow will initially evaluate storage connectivity to ensure all hosts in the selected vSphere cluster have sufficient access to the storage system. If any hosts in the selected vSphere cluster are deemed not to be correctly defined on the storage system, or if conditions exist that prevent the host status being reported as online by the underlying storage system, an alert will be displayed. If hosts are not defined in a host cluster on the storage system, then an advisory message will be displayed. Where possible, consider using host clusters to simplify management and storage provisioning for clustered host environments.

Volume to be created:
  • The name specified in the workflow will be used both when creating the volumes on the storage system and for the VMFS datastore(s).
  • The specified capacity will define the size of the volume created in the storage system. Note that a small area of the volume is allocated to the VMFS partition and so the available capacity for the VMFS datastore may differ from the underlying volume's capacity.

To ensure that each ESXi host in the selected vSphere cluster has been correctly configured on the selected storage system and that all members of the selected vSphere cluster can demonstrate uniform storage connectivity, the host storage adapter WWPN details are evaluated and compared with host definitions on the storage system. This is to support environments where the name of the host as defined on the storage system does not match the name as displayed within vSphere client.

In configurations where Object-Based Access Control (OBAC) is in use, the owner of each host is also compared to the user account for the chosen storage system. If there are discrepancies between the owner names of the two, an ownership error is raised and a link to the storage system is shown to resolve.

Assuming all hosts are reporting online, there is an optional step to assign the datastore to a volume group. Volume groups and any assigned snapshot or replication policies must be predefined on the storage system. This workflow does not allow for the creation of volume groups or creation/assignment of policies to existing ones.

The workflow will create the requested number of volumes of the required size on a selected storage system, in a selected pool.

The appropriate host/hostcluster volume mappings will be created for all members of the vSphere cluster, and if a volume group is selected, the newly created volume will be placed into the volume group.

By default, all volumes will be created as Thin Provisioned/Space Efficient volumes on the storage system, if this is not the preferred behavior consider assigning a Provisioning Policy to the storage pool on the storage system. This Provisioning Policy will define how volumes are created by default and will override the default behavior.

Finally, a VMFS datastore will be created from each volume, and the storage adapters for each vSphere cluster member will be rescanned to correctly mount the new datastore(s).

Note: Version VMFS6 will automatically be used for all datastores created by the plugin. In accordance with best practices, a single volume will be created for each VMFS datastore requested.

Procedure

  1. From the vSphere Client, select the vSphere Cluster that requires the new VMFS datastore(s) and from the Actions menu (or by right click menu), select IBM Storage.
  2. From the IBM® Storage plug-in submenu, select Create VMFS Datastore.
  3. Define a name for the new VMFS datastore. The same name will be used when creating the volume on the storage system.
  4. Specify the required capacity (TB or GB).
  5. Specify the required number of VMFS datastores to be created. If more than 1 datastore is being created, specify the starting index.
    The starting index value can be used to continue existing naming convention for datastores that are already in use.
  6. Click Next.
  7. In the Storage System Selection step, select the storage system for which the datastore will be created on.
    Note: It can take a few moments while all ESXi hosts in the VMware cluster are cross-referenced with host objects and assigned ports. Depending on the number of hosts in the vSphere cluster, the cross-referencing can take up to 1 minute. Any discrepancies that might exist between the host objects or port assignments are indicated. A hyperlink to the management interface for the selected storage system is available for easy access to the element manager.
  8. Select the pool to create the volume in and click Next.
    Note: You cannot select pools that are not in an "online" state, or ones that are in different ownership groups to the registered user.
    Note: If there is only 1 pool registered, the pool selection page is skipped and the single registered pool is automatically chosen.
  9. Optionally, select a volume group to place the volume in once created. On select of the volume group, the snapshot and replication policy details are displayed above the datagrid. If skipping the selection click Skip, otherwise click Next.
    Note: If using replication policies, you can only select production location volume groups.
    Note: If there are no volume groups defined on the storage system, this step is automatically skipped.
  10. Review the summary and click Create.
  11. A progress bar will appear within the view, and vSphere tasks will be generated for each step in the workflow.
  12. Monitor the Recent Tasks view in vSphere to confirm the task is completed successfully.

    After all tasks are successfully completed, the VMFS datastore(s) will be available for use.