Removing a control enclosure and its expansion enclosures

You can use the management GUI and the command-line interface (CLI) to remove the control enclosure and all its attached expansion enclosures.

Important: Before you remove a control enclosure and its expansion enclosures, note that all FlashCopy mappings on the control enclosure are lost if their dependencies are not resolved before the controller is removed.
  1. Make sure that your volume data is backed up.
  2. Determine the volumes, hosts, and arrays that are affected by the removal:
    1. All the volumes that belong to the I/O group and the hosts that require access to them.
    2. Which arrays have member drives in the enclosures.
    3. The list of volumes that depend on these arrays.

    You can use the management GUI and command-line interface (CLI) filtering capabilities to obtain this information.

  3. Decide which volumes are still necessary, and plan for their migration. Consider both available storage pool capacity and the system configuration limits.
  4. If FlashCopy copies with their associated bitmaps are on the control enclosure that you plan to remove, you must first resolve the dependencies. Otherwise, the bitmaps are lost, which causes target volumes to go offline. For help, refer to the rmnodecanister command before you remove the second node canister.

The procedure provides the necessary steps for removing volumes.

  1. Delete any volumes that are no longer needed and are dependent on the enclosure that you plan to remove.
  2. Move any remaining volumes that are accessed through the I/O group to a different I/O group.
    Any hosts that use these volumes must be reconfigured so they retain access.
  3. Migrate any remaining volume data from the arrays in the I/O group:
    • Migrate the entire volume copy to another storage pool. This method is preferable when most of the arrays in the current storage pool are from the I/O group.
    • Migrate the volume data extents to other arrays or managed disks (MDisks) within the current storage pool. Deleting an array forces migration of all volume extents that are allocated on it to other arrays in the pool. You can also migrate extents for each volume individually. This method is preferable when most of the storage capacity in the pool is provided by other I/O groups or external controllers.
    Important: Wait for the migrations to complete. You can use the management GUI to monitor the migration progress.
  4. Delete any remaining arrays that are formed from drives in the I/O group.
    If any arrays still show dependent volumes, repeat the previous step and migrate the data for these volumes.
  5. Mark all the drives (including any configured as spare drives) in the enclosures to be removed as unused.
  6. Ensure no I/O is running to any volumes that are from the I/O group of which the enclosure is part.
  7. Use the chiogrp command to put the I/O group into maintenance mode.
    For more information about the chiogrp command, see chiogrp.
  8. Unmanage and remove enclosures by using the management GUI. Remove the expansion enclosures first and the control enclosure last. Select Monitoring > System. On the System - Overview page, select the directional arrow next to the enclosure that you are removing to open the Enclosure Details page. Select Enclosure Actions > Remove.
  9. Remove the node canisters by using the rmnodecanister command or the management GUI.
  10. The control enclosure and attached expansions are now removed from the system. Remove power from the enclosures to shut them down.
  11. Use the chiogrp command to take the I/O group out of maintenance mode.
Parent topic: Managing nodes