Connecting with on-premises probes

You can connect IBM® Netcool® Operations Insight® probes directly to IBM Cloud Pak for AIOps using an ObjectServer pair running within the Cloud Pak. By default, the ObjectServers are not started up. If you plan to connect IBM® Netcool® Operations Insight® probes, you must update the installation resource. This will change the number of replicas of the aiops-ir-core-ncoprimary and aiops-ir-core-ncobackup stateful sets and the aiops-ir-core-ncodl-if deployment to one.

The update that you need to make is to the Installation custom resource. After installing IBM Cloud Pak for AIOps, use the following steps:

  1. Find your installation name and set an environment variable for it.

    1. Find the name of your IBM Cloud Pak for AIOps installation.

      oc get installation -n <project>
      

      Where <project> is the project that your IBM Cloud Pak for AIOps installation is deployed in.

      Example output, where ibm-cp-aiops is the name of the IBM Cloud Pak for AIOps installation:

      NAME          PHASE    LICENSE   STORAGECLASS STORAGECLASSLARGEBLOCK AGE
      ibm-cp-aiops  Running  Accepted  rook-cephfs  rook-ceph-rbd          34h
      

    2. Export environment variables.

      export INSTALLATION_NAME=<name>
      export PROJECT_CP4AIOPS=<project>
      

      Where:

      • <name> is the name returned by the previous command, for example ibm-cp-aiops.
      • <project> is the project that your IBM Cloud Pak for AIOps installation is deployed in.
  2. Run the following command to edit the installation.

    oc edit installation ${INSTALLATION_NAME} -n ${PROJECT_CP4AIOPS}
    
  3. Set the values as follows to scale up nocprimary and ncobackup:

    spec:
      pakModules:
        - enabled: true
          name: aiManager
        - enabled: true
          name: aiopsFoundation
        - enabled: true
          name: applicationManager
          config:
           - name: ir-core-operator
             spec:
               issueresolutioncore:
                 customSizing:
                   deployments:
                   - name: ncodl-if
                     replicas: 1
                   statefulSets:
                   - name: ncoprimary
                     replicas: 1
                   - name: ncobackup
                     replicas: 1
    

These pods will start aiops-ir-core-ncoprimary-0 and aiops-ir-core-ncobackup-0

Shutting down nocprimary and ncobackup

  1. Find your installation name and set an environment variable for it.

    1. Find the name of your IBM Cloud Pak for AIOps installation.

      oc get installation -n <project>
      

      Where <project> is the project that your IBM Cloud Pak for AIOps installation is deployed in.

      Example output, where ibm-cp-aiops is the name of the IBM Cloud Pak for AIOps installation:

      NAME          PHASE    LICENSE   STORAGECLASS STORAGECLASSLARGEBLOCK AGE
      ibm-cp-aiops  Running  Accepted  rook-cephfs  rook-ceph-rbd          34h
      

    2. Export environment variables.

      export INSTALLATION_NAME=<name>
      export PROJECT_CP4AIOPS=<project>
      

      Where:

      • <name> is the name returned by the previous command, for example ibm-cp-aiops.
      • <project> is the project that your IBM Cloud Pak for AIOps installation is deployed in.
  2. Run the following command to edit the installation.

    oc edit installation ${INSTALLATION_NAME} -n ${PROJECT_CP4AIOPS}
    
  3. To shut down nocprimary and ncobackup, scale them down to 0.

    spec:
      pakModules:
        - enabled: true
          name: aiManager
        - enabled: true
          name: aiopsFoundation
        - enabled: true
          name: applicationManager
          config:
           - name: ir-core-operator
             spec:
               issueresolutioncore:
                 customSizing:
                   deployments:
                   - name: ncodl-if
                     replicas: 0
                   statefulSets:
                   - name: ncoprimary
                     replicas: 0
                   - name: ncobackup
                     replicas: 0
    

Monitoring the processes

ncoprimary and ncobackup are not managed by deployment. They are managed by statefulsets (sts).

If you run the following command to grep the pods for ncoprimary and ncobackup, you will not see anything:

oc get pod | grep aiops-ir-core-ncoprimary-0

If you then check the statefulset using the following command:

oc get sts | grep nco

You can see you can see that the pods are listed under the statefulset, but they are not running:

aiops-ir-core-ncobackup                             0/0     49d
aiops-ir-core-ncoprimary                            0/0     49d

To get these pods up and running, complete the following steps:

  1. Scale up the statefulsets.

    Scale up aiops-ir-core-ncobackup:

    oc scale --replicas=1 sts/aiops-ir-core-ncobackup
    

    Expected output:

    statefulset.apps/aiops-ir-core-ncobackup scaled
    

    Scale up aiops-ir-core-ncoprimary:

    oc scale --replicas=1 sts/aiops-ir-core-ncoprimary
    

    Expected output:

    statefulset.apps/aiops-ir-core-ncoprimary scaled
    
  2. Check the pods for ncoprimary and ncobackup.

    Check for ncoprimary:

    oc get pod | grep aiops-ir-core-ncoprimary-0
    

    Expected output:

    aiops-ir-core-ncoprimary-0                                        1/1     Running     0               5m24s
    

    Check for ncobackup:

    oc get pod | grep aiops-ir-core-ncobackup-0
    

    Expected output:

    aiops-ir-core-ncobackup-0                                         2/2     Running     0               5m54s
    
  3. Grep pvc for nco.

    oc get pvc | grep nco
    

    Expected output:

    database-aiops-ir-core-ncobackup-0                                Bound    pvc-534c0182-0ce9-47d1-a134-20168781965c   5Gi        RWO            ocs-storagecluster-cephfs     18m
    database-aiops-ir-core-ncoprimary-0                               Bound    pvc-4507add6-9d6f-432b-9a53-644c742c61cd   5Gi        RWO            ocs-storagecluster-cephfs     17m
    

Integrating stand-alone on-premises probes to IBM Cloud Pak for AIOps

If you want to connect stand-alone on-premises probes to IBM Cloud Pak for AIOps, you can use NodePorts to securely connect to the in-cluster ObjectServer.

These instructions can be used to connect an on-premises deployment of IBM® Netcool® Operations Insight® to IBM Cloud Pak for AIOps.

Complete the steps in both of the following topics.

Note: IBM Cloud Pak for AIOps does not support connecting with a uni-directional gateway.