Portal Home > Knowledgebase > VMware Knowledge Base > Best practices for VMware ESX or ESXi when scheduling SAN downtime

Best practices for VMware ESX or ESXi when scheduling SAN downtime


This article provides best practice procedures for VMware ESX or ESXi when scheduling downtime for connected SAN storage.


When physically moving an array with running virtual machines, follow this best practice procedure:
  1. Shut down all virtual machines utilizing the targeted SAN storage, or migrate the virtual machines to standby/alternate storage.
  2. Unpresent the targeted SAN's Logical Units (LUs) from the ESX hosts, according to the applicable steps below:

    VMware ESX 3.x

    1. Unmask the SAN LUs from your VMware ESX Servers (remove visibility or access). For additional steps and information, refer to your SAN vendor documentation.
    2. Select an ESX Server in your host inventory, its Configuration tab, then Storage Adapters.
    3. Rescan the ESX Server's Host Bus Adapters (HBAs).
    4. Verify that the LUs no longer appear in the lower pane when selecting each of your HBAs.

    VMware ESX 4.x

    For instructions on how to safely unpresent LUs from a ESX 4.x host, see Removing a LUN containing a datastore from VMware ESXi/ESX 4.x (1029786).

    After completing these steps, verify that the array's LUs are no longer visible to the VMware ESX 4.x hosts, and no dead paths persist.

    : Virtual machines will display as Inaccessible if their datastores are not accessible or mounted. Through the detailed procedure in this article, this behavior is expected. They will return to a normal state after the original storage devices are re-presented to the VMware ESX servers.

  3. Complete your scheduled downtime on the SAN. Proceed to the remaining steps only after the SAN components are confirmed available or online.
  4. Re-present the LUs to the ESX hosts, in accordance to your SAN's documentation and best-practices.
  5. Rescan the VMware ESX Servers.
  6. Confirm that the virtual machines return to a normal state, as opposed to Inaccessible.
  7. Power on the virtual machines.

Alternate method

  1. Shut down the virtual machines.
  2. Shut down the VMware ESX hosts for the duration of the downtime.
  3. Start the VMware ESX hosts and confirm that the LUs are discovered and accessible.
  4. Start the virtual machines.
Based on VMware KB 100277

Also Read