Shutting down the system

Learn how to power off IBM® Cloud Pak for Data System version 2.x.

About this task

The procedure applies to all 2.x systems with or without NPS installed. Some steps are only required for NPS systems, as noted.
Note: It might take up to one hour to power off the nodes that are shut down using the core user.


  1. Log in to the control plane node and verify the state of the system:
    [root@e1n1 ~]# ap state -d
    System state is 'Ready'
    Application state is 'Ready'
    Platform management state is 'Active'
  2. NPS systems only: Run the following steps:
    Note: In the following examples the NPS namespace is named nps-1. In case of multitenant deployments, the commands must be run for each NPS namespace. The following command can be used to identify all NPS namespaces:
    oc get sts --all-namespaces | grep ipshost  | awk '{print $1}'
    1. Run nzstop from the host container:
      oc exec -it ipshost-0 -n nps-1 -- bash -c "su - nz -c 'nzstop'"
    2. Run the following command to shut down SPU nodes:
      oc exec -it ipshost-0 -n nps-1 -- bash -c "su - nz -c 'nzpush -a power off'"
    3. Scale down the NPS host pod:
      oc scale statefulset.apps/ipshost --replicas=0 -n nps-1
  3. Run apstop.
  4. Run apstop --service.
  5. Run the following command to stop the LDAP VM running on the control nodes:
    for node in e1n{1..3}; do ssh $node "virsh shutdown ${node}-ldap"; done 
  6. Shut down GPFS file system from the first control node. From e1n1 run:
    mmumount all -a
    mmshutdown -a
  7. Verify GPFS file systems are unmounted:
    [root@e1n1 ~]# mmlsmount all -L
    Example output:
    [root@e1n1 ~]# mmlsmount all -Lmmcommon: mmremote command cannot be executed. Either none of the
    nodes in the cluster are reachable, or GPFS is down on all of the
    nodes. mmlsmount: Command failed. Examine previous error messages to
    determine cause.
  8. Run the following commands. Replace n in {n..1} with the number of enclosures in your system:
    for ip in e1n4 e{n..2}n{4..1}; do ssh core@$ip sudo shutdown -h; done
    for ip in e{n..1}n{4..1}bmc; do echo [$ip]; ipmitool -I lanplus -H $ip -U USERID -P PASSW0RD power status;done 
    for ip in e1n{3..1}; do ssh $ip sudo shutdown -h; done
    Example for a base + 4 system without NPS installed:
    for ip in e1n4 e{6..2}n{4..1}; do ssh core@$ip sudo shutdown -h; done
  9. Verify that all nodes are powered off. Wait 5 - 10 minutes.

    Check that the power LEDs are blinking on all the nodes.

  10. Physically remove power from the D2 enclosures, the fabric switch, and management switch.

What to do next

The power on procedure is described in Powering on the system.