Installing Global Data Platform

From the Services page, enable the IBM Fusion Global data platform service.

1. Go to the Services page in the IBM Fusion user interface.
2. In the Available section, click Global data platform tile to install and configure storage service.
   Note: You cannot install other dependent services until you install and configure a storage service.
3. In the Global Data Platform page, go through the details and click Install.
   If you have less than six storage nodes, then the Install is disabled. Go to the Nodes page > Discovered tab and add nodes to the cluster so that it has enough storage for Global Data Platform. After you add a minimum of six nodes, the Install is enabled and you continue with the installation.
   After you enable the Global Data Platform, you can view the service version and health status. In the Services page, you can see Global data platform tile under Installed section.

   Table 1. Health states Global Data Platform service

   State	Description
   Healthy	The Running and Ready states for the following items indicate that the service is healthy: All the pods in ibm-spectrum-scale namespace are running. Scale operator pod in ibm-spectrum-scale-operator namespace. ScaleManager CR in the Fusion namespace. All the pods in ibm-spectrum-scale-csi namespace are running. GPFS Scale service is active on all the pods. The filesystem is mounted on all the nodes.
   Degraded	If any of the following conditions are not met, then the service goes to a degraded state. Any one of the scale core pod statuses is not running per Recovery Group (RG) GPFS service is not active on any one of the pods per RG. If one of the PM collector pods is not running or all containers are not in the ready state. If one of the GUI pods is not running or all containers are not in the ready state.
   Critical	2 scale core pods are not running per RG with erasure code parity 4+2P 2 or 3 scale core pods are not running per RG with erasure code parity 4+3P 2 or 3 scale core pods are not running per RG with erasure code parity 8+3P GPFS service is not active on 2 scale core pods per RG (4+2P per RG) GPFS service is not active on 2 or 3 scale core pods per RG (4+3P per RG) GPFS service is not active on 2 or 3 scale core pods per RG (8+3P per RG) If both the PM collector pods are not running or all containers are not in the ready state. If any of the CSI pods are not running or all containers are not in the ready state. If both the GUI pods are not running or all containers are not in the ready state. Add Filesystem is not mounted on more than the tolerable number of nodes.
   Down	More than 2 scale core pods are not running (4+2P per RG) More than 3 scale core pods are not running (4+3P Per RG) More than 3 scale core pods are not running (8+3P Per RG) GPFS service not Active on more than 2 scale core pods per RG (4+2P per RG) GPFS service not Active on more than 3 scale core pods per RG (4+3P Per RG) GPFS service not Active on more than 3 scale core pods per RG (8+3P Per RG)

   From the ellipsis menu, you can download logs and view documentation. After you successfully collect the logs, a success notification gets displayed. The notification disappears automatically after some time.

   For more information about Global data platform service issues during IBM Fusion HCI installation, see Installation issues.

4. Verify Global Data Platform installation.

   Validate Global Data Platform operators, deployments, and pods from the OpenShift management console

   1. Log in to the OpenShift web console.
   2. From the menu navigation, click Operators > Installed Operators.
   3. From the Installed Operators list, click IBM Storage Fusion operator. The Details tab is open by default.
   4. Open Scale tab, and check whether the storagemanager is in the Scales list and is healthy.
      Note: For Global Data Platform Remote Mount version 6.0.0.1, the storage-manager is not created.
   5. From the menu navigation, click Workloads > Deployments.
   6. Check whether the following namespaces and their deployments are available:

      ibm-spectrum-fusion-ns project

      • isf-storage-service-dep
      • isf-storage-operator-controller-manager

      ibm-spectrum-scale-operator project

      ibm-spectrum-scale-controller-manager

      ibm-spectrum-scale-csi project

      ibm-spectrum-scale-csi-operator

      Note: For Global Data Platform Remote Mount version 6.0.0.1, the ibm-spectrum-scale-csi-operator pod is no longer available.

   7. From the menu navigation, click Workloads > Pods.
   8. Check whether you can see the following pods:

      ibm-spectrum-fusion-ns project

      isf-storage-operator-controller-manager

      isf-storage-service-dep

      ibm-spectrum-scale-csi project

      ibm-spectrum-scale-csi-operator

      ibm-spectrum-scale-csi-provisioner

      ibm-spectrum-scale-csi-attacher

      ibm-spectrum-scale-csi

      Note: For Global Data Platform Remote Mount version 6.0.0.1, the ibm-spectrum-scale-csi-operator pod is no longer available.

      ibm-spectrum-scale project

      compute-1-ruX

      ibm-spectrum-scale-gui-0

      ibm-spectrum-scale-gui-1

      ibm-spectrum-scale-pmcollector-0

      ibm-spectrum-scale-pmcollector-1

      ibm-spectrum-scale-operator

      ibm-spectrum-scale-controller-manager

   Validate other pods from the CLI

   1. To list all DNS pods, run the OC command:

      ./oc get pods -n ibm-spectrum-scale-dns

      For example, see the following result:

       
      oc get po -n ibm-spectrum-scale-dns
      NAME            READY   STATUS    RESTARTS   AGE
      coredns-4gmrg   1/1     Running   3          7d16h
      coredns-bjpv7   1/1     Running   5          7d16h
      coredns-hnrph   1/1     Running   3          7d16h
      coredns-j4dxq   1/1     Running   3          7d16h
      coredns-ln6gj   1/1     Running   3          7d16h
      coredns-rtfdg   1/1     Running   3          7d16h

      Make sure all the pods are up and running.

   Verify the storage class.

   Note: Ensure that the following storage classes are created.

   oc get sc