Shutting down the original nodes

When shutting down and removing the original nodes you must send an AutoSupport message about the upgrade, destroy the mailboxes, power down the nodes, and remove the chassis.

Procedure

  1. Send an AutoSupport message from node1 and node2 to inform technical support of the upgrade: system node autosupport invoke -node node_name -type all -message "MAINT=2h Upgrading node_name from platform_original to platform_new"
  2. Disable high availability or storage failover on node1 and node2:
    If you have a... Enter...
    Two-node cluster
    1. cluster ha modify -configured false
    2. storage failover modify -node node_name -enabled false
    Cluster with more than two nodes storage failover modify -node node_name -enabled false
  3. Halt the node: system node halt -node node_name
    You can suppress the quorum check during the reboot process by using the -ignore-quorum-warnings option.
  4. Connect to the serial console, if you are not already connected. The node should be at the LOADER prompt. Use the boot_ontap maint command to boot into maintenance mode.
    A message might appear asking you to ensure that the partner node is down or takeover is manually disabled on the partner node. You can enter yes to continue.
  5. Record each original node's system ID, which is obtained through disk ownership information in Maintenance mode: disk show -v
    You need the system IDs when you assign disks from the original nodes to the new nodes.
    *> disk show -v
    Local System ID: 118049495
    DISK    OWNER               POOL    SERIAL NUMBER          HOME
    ----    -----               ----    -------------          ----
    0a.33   node1 (118049495)   Pool0   3KS6BN970000973655KL   node1 (118049495)
    0a.32   node1 (118049495)   Pool0   3KS6BCKD000097363ZHK   node1 (118049495)
    0a.36   node1 (118049495)   Pool0   3KS6BL9H000097364W74   node1 (118049495)
    ...
    
  6. If you have FC or CNA port configuration, display the configuration in Maintenance mode: ucadmin show
    You should record the command output for later reference.
    *> ucadmin show
    Current Current Pending   Pending
    Adapter Mode    Type      Mode    Type    Status
    ------- ------- --------- ------- ------- ------
    0e      fc      initiator -       -       online
    0f      fc      initiator -       -       online
    0g      cna     target    -       -       online
    0h      cna     target    -       -       online
    ...
    
  7. In Maintenance mode, destroy the node1 and node2 mailboxes: mailbox destroy local
    The console displays a message similar to the following:
    Destroying mailboxes forces a node to create new empty mailboxes, which 
    clears any takeover state, removes all knowledge of out-of-date plexes and 
    mirrored volumes, and will prevent management services from going online in
    2-node cluster HA configurations.
    Are you sure you want to destroy the local mailboxes?
  8. Destroy the mailboxes by entering y when you see a prompt similar to the following:
    .............Mailboxes destroyed
    Takeover On Reboot option will be set to ON after the node boots.
    This option is ON by default except on setups that have iSCSI or FCP license.
    Use "storage failover modify -node <nodename> -onreboot false" to turn it OFF.
    
    *>
  9. Exit Maintenance mode: halt
  10. Turn off the power to node1 and node2, and then unplug them from the power source.
  11. Label and remove all cables from node1 and node2.
  12. Remove the chassis containing node1 and node2.