Categories
00 - Hybrid Cloud Infrastructure

How to install SAP Data Intelligence on Red Hat OpenShift

In case you want to test-drive SAP Data Intelligence (SDI), here is what you can do to get a quick setup on an existing OpenShift Cluster.

If you do not know what SAP data intelligence is, I recommend reading this article: What is SAP Data Intelligence

Prerequisites

To install SAP Data Intelligence (SDI) you need a running OpenShift Cluster with at least 3 Worker nodes and access to Block and object storage.

The worker nodes need to have the following minimum requirements:

  • 8 CPUs
  • 32 GB memory
  • 100 GB local ephemeral storage

In addition, 250GB persistent volumes and 90GB for the container registry is needed
See also Minimum Sizing for SAP Data Intelligence

For the checkpoint storage feature and the data lake feature you also need S3 storage.

We have tested SDI with Red Hat OpenShift Container Storage (OCS) and this article will explain how to use this for providing the S3 storage. More Information on OpenShift Container Storage can be found here.

Please note that you either need additional nodes for OCS or you need additional resources (10 vCPUs and 24GB of memory) on each of the 3 worker nodes.

In case you do not have an OpenShift cluster available the following table describes a feasible setup in AWS with separate OCS and SDI nodes:

TypeCountOperating SystemvCPURAM (GB)Storage (GB)AWS Instance Type
Bootstrap

 

1

 

RHCOS

2

16

120

i3.large

Master

3+

RHCOS

4

16

120

m4.xlarge

Compute

3+

RHEL 7.6 or RHCOS

4

32

120

m4.2xlarge

Storage

3+

RHCOS

10

24

120 + 2048

m5.4xlarge

For a POC it is feasible to have only three Worker Nodes and you can aggregate the requirements. In production environments it is recommended to have SDI and OCS on separate worker nodes, which enables you to easier scaling of one or the other.

High Level Installation Flow

For deploying SAP Data Intelligence the cluster the following steps need to be performed:

  1. Label the worker nodes which should get SDI contents
  2. Change the configuration of the SDI Worker Nodes to suite SAP’s requirements
  3. Deploy sdi-observer monitoring and installation helper tool from Red Hat
  4. Prepare required S3 storage pools
  5. Apply required lifted permissions to the SDI project
  6. Deploy the SAP LifeCycle Manager Container Bridge (SLCB) for installing SDI
  7. Launch the installation of SDI in SLCB web frontend

During this process you need the following data:

  • Your SAP S-User name and password for downloading the software
  • login credentials to Red Hat Portal
  • login credentials to access the OpenShift Cluster with admin permissions

Verify / prepare the management workstation

You need to prepare a management workstation from where you can access your OpenShift cluster, run ansible playbooks and a web browser session.
To make it easy, we assume a linux workstation, e.g. based on RHEL or Fedora. If you are on Windows, MacOS or other Linux distribution you need to adapt the settings accordingly

  1. Login to your management workstation
  2. Ensure the following software is installed
    • ansible for automating the setup together with the python modules for managing OpenShift:
      • python3-pyyaml
      • python3-urllib3.noarch
      • python3-requests
      • python3-requests-oauthlib
      • python3-openshift (from EPEL)
    • yum-utils for managing repositories
    • git for loading data from github

      On RHEL 8 issue the following commands as root:

      dnf -y install https://dl.fedoraproject.org/pub/epel/epel-release-latest-8.noarch.rpm
      dnf -y ansible python3-pyyaml python3-urllib3 python3-requests python3-requests-oauthlib python3-openshift yum-utils
  3. make sure jq version 1.6 is installed for parsing JSON
    curl -L -o /usr/local/bin/jq https://github.com/stedolan/jq/releases/download/jq-1.6/jq-linux64
    chmod a+x /usr/local/bin/jq
  4. Install an OpenShift client according to your OCP version
    OCP_VERSION=4.7.2 wget https://mirror.openshift.com/pub/openshift-v4/clients/ocp/${OCP_VERSION}/openshift-client-linux-${OCP_VERSION}.tar.gz
    sudo tar zxvf openshift-client-linux-${OCP_VERSION}.tar.gz -C /usr/bin
    sudo rm -f openshift-client-linux-${OCP_VERSION}.tar.gz /usr/bin/README.md
    sudo chmod +x /usr/bin/oc /usr/bin/kubectl
  5. setup bash completion (optional)
    oc completion bash | sudo tee /etc/bash_completion.d/openshift > /dev/null
  6. Clone the github repository with the ansible playbooks and scripts for configuring the cluster. You will find the playbooks and scripts used in this blog in the demo subdirectory:
git clone https://github.com/redhat-sap/sap-data-intelligence.git

Verify OpenShift Cluster

make sure you have OpenShift Cluster admin rights

oc whoami system:admin
Code language: CSS (css)

In case you do not have cluster-admin permissions, login with a user which has. In this article we assume this user is named admin:

oc login -u admin

check, that min requirements are met (example with 3 worker nodes):

oc get nodes ip-10-0-133-218.ec2.internal Ready master 47m v1.20.0+5fbfd19 ip-10-0-141-94.ec2.internal Ready worker 37m v1.20.0+5fbfd19 ip-10-0-154-232.ec2.internal Ready master 47m v1.20.0+5fbfd19 ip-10-0-159-127.ec2.internal Ready worker 40m v1.20.0+5fbfd19 ip-10-0-167-89.ec2.internal Ready master 48m v1.20.0+5fbfd19 ip-10-0-175-27.ec2.internal Ready worker 43m v1.20.0+5fbfd19
Code language: CSS (css)

You should see 3 worker nodes and 3 master nodes

NOTE: If you see something like this after your systems shut down:

# oc get nodes NAME STATUS ROLES AGE VERSION ip-10-0-137-27.ec2.internal NotReady worker 2d23h v1.20.0+5fbfd19 ip-10-0-141-89.ec2.internal NotReady master 3d v1.20.0+5fbfd19 ip-10-0-154-182.ec2.internal NotReady master 3d v1.20.0+5fbfd19 ip-10-0-159-71.ec2.internal NotReady worker 2d23h v1.20.0+5fbfd19 ip-10-0-165-90.ec2.internal NotReady worker 2d23h v1.20.0+5fbfd19 ip-10-0-168-27.ec2.internal NotReady master 3d v1.20.0+5fbfd19
Code language: CSS (css)

This can happen if the systems are shutdown for more then 24hrs and the certificates are set invalid. 

The following command will manually approve the new certificates:

# oc adm certificate approve $(oc get csr | grep Pending | awk {'print $1'})
Code language: PHP (php)

Check the hardware resources of the cluster Nodes

Store the worker names in a variable for later use:

# WORKER=$(oc get nodes | awk ' ( $3 ~ "worker" ) {print $1 }')
Code language: PHP (php)

Run the following command to get information on the capacity of the worker nodes in the OpenShift cluster and make sure they satisfy the minimum requirements for SDI and OCS:

# oc describe node $WORKER | grep -A 6 Capacity Capacity: attachable-volumes-aws-ebs: 25 cpu: 16 ephemeral-storage: 125293548Ki hugepages-1Gi: 0 hugepages-2Mi: 0 memory: 64792280Ki -- Capacity: attachable-volumes-aws-ebs: 25 cpu: 16 ephemeral-storage: 125293548Ki hugepages-1Gi: 0 hugepages-2Mi: 0 memory: 64792280Ki -- Capacity: attachable-volumes-aws-ebs: 25 cpu: 16 ephemeral-storage: 125293548Ki hugepages-1Gi: 0 hugepages-2Mi: 0 memory: 64792280Ki
Code language: PHP (php)

So the minimum requirements are met with 64 GiB memory on 3 worker nodes and 120 GiB local storage.

prepare OCP cluster for SDI

Switch to the directory with the ansible playbooks.
The playbook ocp_prep_nodes.yml will label all worker nodes in your cluster for use with SDI.
Please change the variable sdi_configure_ocp_worker_nodelist if you want something different.
you can also change the when statement, so that additional properties can be used for selecting nodes for SDI.

Now verify and run the playbook:

# ansible-playbook -i myhosts -vv ocp_prep_nodes.yml
Code language: CSS (css)

The playbook will perform the following steps on all nodes in sdi_configure_ocp_worker_nodelist:

  1. Label SDI compute Nodes with node-role.kubernetes.io/sdi=""
  2. Enables net-raw capability for containers on schedulable nodes
  3. Pre-load additional needed kernel modules for SDI (e.g. NFS and iptables)
  4. Increasing PID limits to 16384
  5. Associate MachineConfigs defined in step in 2-3 to the nodes with the label sdi

     

    It may take a while until all nodes are updated. The following
    command can be used to wait until the change gets applied to all the
    worker nodes:

       oc wait mcp/sdi --all --for=condition=updated

    The following command lists the status of the nodes:

       oc get mcp

    NoteIf the update is not working, check the machineconfig operator

  6. Now double check that settings are made

     you can use the following script to double check that all required
    changes are made on the OpenShift Worker nodes

    #!/usr/bin/bash
     
    # CHECK OCP (Note Files may change after update)
     
    for worker in `oc get nodes|awk '/worker/{print $1}'`; do
      echo "Checking node $worker ------------------------------------------------------------------------------"
      # Check for additional kernelmodules
      oc debug node/$worker -- chroot /host cat /etc/crio/crio.conf.d/90-default-capabilities  2> /dev/null
      # Check for additional kernelmodules
      oc debug node/$worker -- chroot /host cat /etc/modules-load.d/sdi-dependencies.conf 2> /dev/null
      # check for module load service
      oc debug node/$worker -- chroot /host systemctl status sdi-modules-load.service 2> /dev/null
      # check for pidsLimit:
      oc debug node/$worker -- chroot /host cat /etc/crio/crio.conf.d/01-ctrcfg-pidsLimit
      echo "--------------------------------------------------------------------------------------------------------"
    done 

Configure Storage

Run the playbook ocs-create-S3buckets.yml to create a new project sdi-infra that keeps the S3 buckets for SDI:

ansible-playbook -i myhosts -vv ocs-create-S3buckets.yml
Code language: CSS (css)

By default, two buckets will be created. You can list them this way:

# oc project sdi-infra # bash <(curl -s https://raw.githubusercontent.com/redhat-sap/sap-data-intelligence/master/utils/mksdibuckets) list Bucket claim namespace/name: sdi/sdi-checkpoint-store (Status: Bound, Age: 7m33s) Cluster internal URL: http://s3.openshift-storage.svc.cluster.local Bucket name: sdi-checkpoint-store-ef4999e0-2d89-4900-9352-b1e1e7b361d9 AWS_ACCESS_KEY_ID: LQ7YciYTw8UlDLPi83MO AWS_SECRET_ACCESS_KEY: 8QY8j1U4Ts3RO4rERXCHGWGIhjzr0SxtlXc2xbtE Bucket claim namespace/name: sdi/sdi-data-lake (Status: Bound, Age: 7m33s) Cluster internal URL: http://s3.openshift-storage.svc.cluster.local Bucket name: sdi-data-lake-f86a7e6e-27fb-4656-98cf-298a572f74f3 AWS_ACCESS_KEY_ID: cOxfi4hQhGFW54WFqP3R AWS_SECRET_ACCESS_KEY: rIlvpcZXnonJvjn6aAhBOT/Yr+F7wdJNeLDBh231

Deploy SDI OBSERVER

SDI Observer is a redhat monitoring software that controls the behaviour and the proper installation of SDI.

Create a service account to pull images from Red Hat registry

Before running SDI Observer you need to create a service account for registry.redhat.io at
https://access.redhat.com/terms-based-registry/, download and save
it to rht-registry-secret.yaml

deploy SDI Observer

Now run the following playbook to deploy SDI observer

ansible-playbook -i myhosts -vv deploy-sdi-observer.yml
Code language: CSS (css)

The playbook does the following;

  1. It create Namespaces for SDI Observer, SDI and the Install bridge:
    • sdi-observer
    • sdi
    • sap-slcbridge
  2. It creates a pull secret for redhat registry within sdi-observer
    namespace.
  3. It defines a couple of variables for SDI Observer & deploys it.
    You can influence the behaviour of SDI observer by changing/adding certain variables in the playbook. See https://access.redhat.com/articles/5100521 section 4.1 for the
    variable lists.
    In particular we want sdi-observer to deploy a SDI compliant registry and to make sure that an OpenShift route is created after the installation of SDI.
    Note: You can change or at variables at a later time with the following command:
    oc set env -n sdi-observer dc/sdi-observer <variable name>=<value>
    Note: you can list the variables with
    oc set env -n sdi-observer --list dc/sdi-observer
    Note: If you change a variable afterwards a rebuild of sdi-observer might be required you can trigger this with the following command: oc start-build -n sdi-observer -F bc/sdi-observer
  4. wait until sdi-observer and registry pods are running:
     $ oc get pods
     NAME                                READY   STATUS      RESTARTS   AGE
     container-image-registry-1-build    0/1     Completed   0          3m20s
     container-image-registry-1-deploy   0/1     Completed   0          82s
     container-image-registry-1-jkrx8    1/1     Running     0          79s
     deploy-registry-4gccn               0/1     Completed   0          3m26s
     sdi-observer-1-build                0/1     Completed   0          5m48s
     sdi-observer-1-deploy               0/1     Completed   0          3m53s
     sdi-observer-1-xphzw                1/1     Running     0          3m49s
    

    You can follow the deployment processes with e.g.
    oc logs sdi-observer-1-build -f 
    oc logs container-image-registry-1-build -f 

get the registry credentials and distribute the certificate in openshift

The following script will store the registry credentials in the file registry-credentials.txt in the local directory. If the registry connection is ok, press enter to distribute the certificate.

#!/bin/bash ## Change Namespace to sdi-observer NAMESPACE="${NAMESPACE:-sdi-observer}" oc project sdi-observer ## Obtain registry credentials reg_credentials=$(oc get -n "${NAMESPACE:-sdi-observer}" secret/container-image-registry-htpasswd -o jsonpath='{.data.\.htpasswd\.raw }' | base64 -d ) reg_user=$(echo $reg_credentials| cut -d: -f1) reg_pw=$(echo $reg_credentials| cut -d: -f2) ## Obtain registry hostname reg_hostname="$(oc get route -n "${NAMESPACE:-sdi-observer}" container-image-registry -o jsonpath='{.spec.host}')" echo "=================================================" | tee registry-credentials.txt echo "Using registry: $reg_hostname" | tee -a registry-credentials.txt echo "USER: $reg_user" | tee -a registry-credentials.txt echo "PW : $reg_pw" | tee -a registry-credentials.txt echo "=================================================" | tee -a registry-credentials.txt if [ -z "$reg_user" -o -z "$reg_pw" ]; then echo "Something went wrong. Check if the pods are running" exit 1 fi #set -x ### Obtain Ingress Router's default self-signed CA certificate mkdir -p "/etc/containers/certs.d/${reg_hostname}" router_ca_crt="/etc/containers/certs.d/${reg_hostname}/router-ca.crt" oc get secret -n openshift-ingress-operator -o json router-ca | \ jq -r '.data as $d | $d | keys[] | select(test("\\.crt$")) | $d[.] ' | base64 -d > ${router_ca_crt} ### test via curl curl -k -I --user ${reg_credentials} --cacert ${router_ca_crt} "https://${reg_hostname}/v2/" ### test via podman echo $reg_pw | podman login -u $reg_user --password-stdin ${reg_hostname} reg_login_ok=$? if [ $reg_login_ok ]; then # Configure Openshift to trust container registry (8.2) echo "Configure Openshift to trust container registry" echo "CTRL-C to stop, ENTER to continue" read zz caBundle="$(oc get -n openshift-ingress-operator -o json secret/router-ca | \ jq -r '.data as $d | $d | keys[] | select(test("\\.(?:crt|pem)$")) | $d[.]' | base64 -d)" # determine the name of the CA configmap if it exists already cmName="$(oc get images.config.openshift.io/cluster -o json | \ jq -r '.spec.additionalTrustedCA.name // "trusted-registry-cabundles"')" if oc get -n openshift-config "cm/$cmName" 2>/dev/null; then # configmap already exists -> just update it oc get -o json -n openshift-config "cm/$cmName" | \ jq '.data["'"${reg_hostname//:/..}"'"] |= "'"$caBundle"'"' | \ oc replace -f - --force else # creating the configmap for the first time oc create configmap -n openshift-config "$cmName" \ --from-literal="${reg_hostname//:/..}=$caBundle" oc patch images.config.openshift.io cluster --type=merge \ -p '{"spec":{"additionalTrustedCA":{"name":"'"$cmName"'"}}}' fi # Check that the certifcate is deployed sleep 20 # wait for distribution of certificates echo "======== Configured Registries ==========" oc rsh -n openshift-image-registry "$(oc get pods -n openshift-image-registry -l docker-registry=default | \ awk '/Running/ {print $1; exit}')" ls -1 /etc/pki/ca-trust/source/anchors else echo "Registry setup failed, please repair before you continue" fi

You will find the script named configure-sdi-observer-registry.sh in the github repository. Make sure that your registry will be shown at the end below “Configured Registries”. You can rerun the script at any time if that is not the case.

Ensure that the project service accounts have correct priviledges

SDI uses a couple of serviceaccounts that need additional priviledges

oc login -u admin oc project sdi oc adm policy add-scc-to-group anyuid "system:serviceaccounts:$(oc project -q)" oc adm policy add-scc-to-user privileged -z "$(oc project -q)-elasticsearch" oc adm policy add-scc-to-user privileged -z "$(oc project -q)-fluentd" oc adm policy add-scc-to-user privileged -z default oc adm policy add-scc-to-user privileged -z mlf-deployment-api oc adm policy add-scc-to-user privileged -z vora-vflow-server oc adm policy add-scc-to-user privileged -z "vora-vsystem-$(oc project -q)" oc adm policy add-scc-to-user privileged -z "vora-vsystem-$(oc project -q)-vrep"
Code language: JavaScript (javascript)

Install SDI Install Bridge

Now that the SDI observer is running, we can install the SDI Install
Bridge container that is used to install SDI on the cluster. The
following steps should run as the user admin.

# oc login -u admin # oc whoami admin
Code language: PHP (php)

Note

This document assumes your cluster has direct internet access. If you
require proxy settings follow the steps in
https://access.redhat.com/articles/5100521 section 5.1

Download the SAP install bridge from SAP (you need your S-User now)

Go to the maintenance planner (MP) at https://apps.support.sap.com/sap/support/mp and click “Plan a New System:
image

Select “CONTAINER based”
image

Select “SAP DATA INTELLIGENCE”, DI – Platform full, latest version (currently 3.1) and click Next
image

Click Next
image

Select Linux on x86_64 and Confirm Selection
image

Select “SL CONTAINER BRIDE” and Click Next
image

Select SLCB01_*.EXE, Click “Push to Download Basket” and Next
image

Click “Execute Plan”
image

Note: Keep the browser tab open. You need to return here after the installation of SLCB.

Use SAP Software Downloader to download the previously selected “SLCB01*.EXE” from the Downloadbasket.

Alternatively you can download SLCB01_<Version>.EXE for Linux directly from https://support.sap.com/, rename it to slcb and make it executable (click Software Downloads and enter Software Lifecycle Container Bridge in the search field)

mv SLCB01_*.EXE /usr/bin/slcb chmod +x /usr/bin/slcb

Install SDI Install Bridge

Note: This tutorial is tested with version 1.1.63. You can always install this version with `slcb init --bridgebaseVersion 1.1.63, but any later version should do, too.

When answering the installer questions, the following selections are important:

  • Installation Type: Expert Mode
  • Service Type
    • on AWS choose Loadbalancer. You do not need to give annotations
    • on all other environments choose NodePort
  • Have a look at https://access.redhat.com/articles/5100521, if you need to configure proxies – This article assumes direct connection to the internet
  • You need to provide the following information:
    • Address of the Container Image Repository
    • Image registry user name
    • Image registry password
    • Your S-User + password
    • admin user password

Now you need the different credentials and information you noted earlier. Execute slcb init. Here is an example log:

$ slcb init 'slcb' executable information Executable: slcb Build date: 2021-03-26 03:45:45 UTC Git branch: fa/rel-1.1 Git revision: 4f99471a2f764f65da2d72ef74c5259e8639697e Platform: linux Architecture: amd64 Version: 1.1.62 SL Core version: 1.0.0 SLUI version: 2.6.67 Arguments: init Working dir: /home/generic_emea_mkoch Schemata: 0.0.62, 1.13.62 Explanation of supported shortcuts: <F1>: Display help for input value. <ENTER> or <Ctrl-N>: Confirm and continue to next input value. <F12> or <Ctrl-B>: Go back to previous input value. <r>: Retry current step. <e>: Edit a multi-line input value. <Ctrl-C>: Abort current processing and return to the Welcome dialog of the SLC Bridge Base. Ctrl-C is not explicitly shown as an option in the command line prompt but you can always use it. <Tab>: Completion of input values. In dialogs that accept only a restricted set of values (like files, directories etc) use the <Tab> key to cycle through the values or for completion of incomplete input. Execute step Download Bridge Images *********************************** * Product Bridge Image Repository * *********************************** Enter the address of your private container image repository used to store the bridge images. You require read and write permissions for this repository. Choose action <F12> for Back/<F1> for help Address of the Container Image Repository: container-image-registry-sdi-observer.apps.cluster-bf86.bf86.example.opentlc.com ************************ * Image Registry User * ************************ The user name used to logon to "container-image-registry-sdi-observer.apps.cluster-bf86.bf86.example.opentlc.com". Choose action <F12> for Back/<F1> for help Image registry user name: user-q5j0lq Choose action <F12> for Back/<F1> for help Image registry password: *************************** * Enter Logon Information * *************************** You require S-User credentials to log on to the SAP Registry ("rhapi.repositories.cloud.sap") for product version "SL TOOLSET 1.0" (01200615320900005323) Choose action <F12> for Back/<F1> for help S-User Name: S0001234567 Choose action <F12> for Back/<F1> for help Password: Copying image slcb://01200615320900005323.dockersrv.repositories.sapcdn.io/com.sap.sl.cbpod/slcbridgebase:1.1.62 to "container-image-registry-sdi-observer.apps.cluster-bf86.bf86.example.opentlc.com" Copying image slcb://01200615320900005323.dockersrv.repositories.sapcdn.io/com.sap.sl.cbpod/nginx-sidecar:1.1.62 to "container-image-registry-sdi-observer.apps.cluster-bf86.bf86.example.opentlc.com" Checking prerequisite Execute step Check Prerequisites I0331 13:01:04.372152 6354 request.go:621] Throttling request took 1.153431509s, request: GET:https://api.cluster-bf86.bf86.example.opentlc.com:6443/apis/flows.knative.dev/v1beta1?timeout=32s Checking prerequisite Kubernetes Server Version ************************ * Prerequiste Check * ************************ Checking the prerequisites for "SL Container Bridge" succeeded. Kubernetes Cluster Context: Cluster name: api-cluster-bf86-bf86-example-opentlc-com:6443 API server URL: https://api.cluster-bf86.bf86.example.opentlc.com:6443 Editable Prerequisites Enter the path to the "kubectl" configuration file. The configuration information contained in this file will specify the cluster on which you are about to perform the deployment. Choose action <Tab> for completion/<F1> for help Path to the "kubeconfig" file: ESC[1G Path to the "kubeconfig" file: /home/generic_emea_mkoch/.kube/configESC[0KESC[71G Prerequisite Check Result Name Current Value Result Error Message KUBECONFIG /home/generic_emea_mkoch/.kube/config + (passed) Kubernetes Server Version 1.20.0 + (passed) Choose "Retry (r)" to retry the Prerequisite Check. Choose "Next (n)" to continue. Choose action Retry(r)/Next(n)/<F1> for help: n Execute step Collect Input *************************************************************************** * Choose whether you want to run the deployment in typical or expert mode * *************************************************************************** You can run the deployment either in typical or expert mode: - Typical Mode If you choose "Typical Mode", the option is performed with default settings. As a result, you only have to respond to a small selection of prompts. - Expert Mode If you choose "Expert Mode", you are prompted for all parameters. > 1. Typical Mode 2. Expert Mode Choose action <F12> for Back/<F1> for help possible values [1,2]: 2 ************************ * SLC Bridge Namespace * ************************ Enter the Kubernetes namespace for the SLC Bridge. Choose action <F12> for Back/<Tab> for completion/<F1> for help Namespace: sap-slcbridge ************************ * Administrator User * ************************ Specify the name of the administrator user for the SLC Bridge Base. Choose action <F12> for Back/<F1> for help User Name: admin ******************************* * Administrator User Password * ******************************* Define the password of the administrator user admin Choose action <F12> for Back/<F1> for help Password of User admin: Confirm: *********************************************** * Service Type of the SLC Bridge Base Service * *********************************************** In order to access the SLC Bridge Base, the UI Port needs to be exposed. This is accomplished by defining a Kubernetes service. Kubernetes offers multiple service types. SAP currently supports the following service types. You have to select one of them. - Service Type "LoadBalancer" is suitable if your Kubernetes cluster comes with a controller for this service type. For example, this is the case for all hyperscaler platforms. - Service Type "NodePort" is suitable if your Kubernetes cluster runs on premise and the cluster nodes can be reached from your network > 1. Service Type LoadBalancer 2. Service Type NodePort Choose action <F12> for Back/<F1> for help possible values [1,2]: 2 ************************ * Proxy Settings * ************************ Do you want to configure Proxy Settings for the Pods running in the cluster? This is necessary if the Pods in the cluster are running behind a proxy. Configure Proxy Settings: n Choose action <F12> for Back/<F1> for help possible values [yes(y)/no(n)]: n Execute step Show Summary ************************ * Parameter Summary * ************************ Choose "Next" to start the deployment with the displayed parameter values or choose "Back" to revise the parameters. SLC Bridge Namespace Namespace: sap-slcbridge Image Registry User Image registry user name: user-q5j0lq SLP_BRIDGE_REPOSITORY_PASSWORD Enter Logon Information S-User Name: S0000000000 IMAGES_SAP_SUSER_PASSWORD KUBECONFIG Path to the "kubeconfig" file: /home/generic_emea_mkoch/.kube/config Choose whether you want to run the deployment in typical or expert mode 1. Typical Mode > 2. Expert Mode Administrator User User Name: admin Administrator User Password Service Type of the SLC Bridge Base Service 1. Service Type LoadBalancer > 2. Service Type NodePort Proxy Settings Configure Proxy Settings: n Choose "Next" to start the deployment with the displayed parameter values or choose "Back" to revise the parameters. Choose action <F12> for Back/Next(n)/<F1> for help: ESC[1G Choose action <F12> for Back/Next(n)/<F1> for help: n Apply Secret Template (secret-slcbridge.yml)... Execute step Master secret Apply Secret Template (secret-nginx.yml)... Execute step Nginx secret Execute step Wait for Kubernetes Object SLCBridgeNamespace Execute step Wait for Kubernetes Object SLCBridgeServiceAccount Execute step Wait for Kubernetes Object DefaultsMap Execute step Execute Service Execute step Wait for Kubernetes Object ProductHistory Execute step Wait for Kubernetes Object MasterSecret Execute step Wait for Kubernetes Object NginxSecret Execute step Wait for Kubernetes Object SLCBridgePod Execute step SL Container Bridge ************************ * Message * ************************ Deployment "slcbridgebase" has 1 available replicas in namespace "sap-slcbridge" Service slcbridgebase-service is listening on any of the kubernetes nodes on "https://node:30713/docs/index.html" Choose action Next(n)/<F1> for help: n Execute step Get User Feedback ****************************** * Provide feedback to SAP SE * ****************************** Dear user, please help us improve our software by providing your feedback (press <F1> for more information). > 1. Fill out questionnaire 2. Send analytics data only 3. No feedback Choose action <F12> for Back/<F1> for help possible values [1,2,3]: 3 Execute step Service Completed
Code language: HTML, XML (xml)

Check SAP LifeCycleBridge is Running

If everything went well you can see the resources like this:

$ oc -n sap-slcbridge get all NAME READY STATUS RESTARTS AGE pod/slcbridgebase-6cd8b94579-4l72q 2/2 Running 0 24m NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE service/slcbridgebase-service NodePort 172.30.122.31 <none> 9000:30578/TCP 24m NAME READY UP-TO-DATE AVAILABLE AGE deployment.apps/slcbridgebase 1/1 1 1 24m NAME DESIRED CURRENT READY AGE replicaset.apps/slcbridgebase-6cd8b94579 1 1 1 24m
Code language: HTML, XML (xml)

connect to the bridge

  1. If you are on AWS and have chosen LoadBalancer the installer prints the URL how to access the SLCB Bridge and your are done
  2. If you chose node port, the Service is exposed on the given port on any node of the cluster, so you need to get exposed node port and pick an IP address of one of the nodes and point your browser to: https://<IP>:<NodePort>/docs/index.html
    Example:
    • Get IP: 
         $ oc get node -o wide sdi-worker-1
      NAME STATUS ROLES AGE VERSION INTERNAL-IP EXTERNAL-IP OS-IMAGE KERNEL-VERSION CONTAINER-RUNTIME
      sdi-worker-1 Ready sdi,worker 14d v1.19.0+9c69bdc 10.19.20.160 Red Hat Enterprise Linux CoreOS 46.82.202101131942-0 (Ootpa) 4.18.0-193.40.1.el8_2.x86_64 cri-o://1.19.1-2.rhaos4.6.git2af9ecf.el8
    • Get Port:
         $ oc get svc -n "${SLCB_NAMESPACE:-sap-slcbridge}" slcbridgebase-service -o jsonpath=$'{.spec.ports[0].nodePort}\n'
      30578
    • To access the bridge point your browser to:
      https://10.19.20.160:30578/docs/index.htm
  3. if you have choosen NodePort, but the Nodes of your cluster are behind a firewall and not reachable with the above URL you can use OpenShift Portforwarding to access the installer. Run the following command:
    oc port-forward svc/slcbridgebase-service 9000
    Now you can point your browser to https://localhost:9000/docs/index.htm to access the installer

If everything worked well, you should see this page now:

SDI installation

Finally you can start the SDI Installation

  1. Now go back to the Maintenance Planner browser Tab

     

    Note: The bridge has to be opened in an active window all the time while working with (MP)

    Enter the hostname and port from the previous step.
    image

  2. Click Next and the Deploy
    image
  3. If you see the following
    image
    switch back to the SLCB browser tab. You will see this:
    image
    Click OK
  4. Now you need to enter Your S-User credentials and click next
    image
  5. Select “SAP DATA INTELLIGENCE 3 – DI Platform Full” and click next
    image
  6. Enter the OpenShift Namespace where SDI should run in. In our case it is sdi and click next
    image
  7. Select Advanced Installation and click Next
    image
  8. Enter a password for the System Tenant Administrator
    image
  9. Enter the Default Tenant name:
    image
  10. Enter the Default Tenant Adminstrator name and password:
    image
  11. As our cluster has direct access to the internet we do not need to
    set proxies. If this is different in you environment read the SAP
    documentation

     

    image
  12. disable backup

     

    Do not configure backup: SAP Note 2918288

    […​] the object storage infrastructure NooBaa cannot be used as
    backup media, if Vora is used.

    To disable backup be sure to remove the check mark (picture is
    default)

    image
  13. enable Checkpoint store

     

    ensure the checkmark is set
    image

    select S3 Compatible object store
    image

    Use the name and credentials for checkpoint store created earlier,
    the endpoint for NooBa S3 is always
    http://s3.openshift-storage.svc.cluster.local

    imageimage
    image
    image

    Even if your cluster is setup correctly, the validation takes a
    while. Don’t worry. If it fails, double check that you use http and
    not https. With private certificates this may not work

  14. Continue with the defaults on the next screens

     

    Use the default storage class for persistent volumes
    image

    no special docker log path
    image

    Enable Kaniko
    image

    You do not need a differnt container image repo for modeler for demo
    purpose
    image

    Enable load Kernel Modules, although this has been already handled
    by the installer. It will work without, but does not harm anything.

    image

    Leave defaults

    image
    image
    image
    image
    image
  15. Change the clustername to something useful
    image

     

    Next is a summary of the installation parameters.

  16. Kick of the installation Kick of the installation. After
    installation the following screen appears
    image

Caution

Note your System ID. In this example it is 11bw3dz

Post Installation work

Get Access to the SDI console

We have configured the sdi-observer to make the route to the admin interface available. You can check this with the following command:

# oc rollout status -n sdi-observer -w dc/sdi-observer
Code language: PHP (php)

If sdi-observer has exported the route correctly, you will get:

replication controller "sdi-observer-2" successfully rolled out
Code language: JavaScript (javascript)

You can double check with

# oc get routes -n sdi
Code language: PHP (php)

you will get a an output like:

NAME HOST/PORT PATH SERVICES PORT TERMINATION WILDCARD vsystem vsystem-sdi.apps.cluster-1251.1251.example.opentlc.com vsystem vsystem reencrypt/Redirect None

You can now access the SDI management console at https://vsystem-<SDI_NAMESPACE>.apps.<cluster_name>.<base_domain>. In this example it is https://vsystem-sdi.apps.cluster-1251.1251.example.opentlc.com

Configure the Connection to Data Lake

  1. Login to the SDI Console at https://vsystem-<SDI_NAMESPACE>.apps.<cluster_name>.<base_domain> use the tenant default, user defaultadmin and the password from the installation procedure

     

    image

    image

  2. Click Connection Management tile
    image
  3. Click on +
    image
  4. Enter the following values and click Test Connection:

     

    Parameter Value
    Connection Type SDL
    Id DI_DATA_LAKE
    Object Storage Type S3
    Endpoint http://s3.openshift-storage.svc.cluster.local
    Access Key ID from above (see storage-credentials.txt)
    Secret Access Key from above (see storage-credentials.txt)
    Root Path from above (see storage-credentials.txt)
    image

  5. If connection test is successful click on Create

     

    image

  6. done

     

    image

SDI validation

Finally you can do the following steps to validate that your installation of SAP Data Intelligence
has been successful. The process is described at Sap Data Intelligence Installation Guide
Now go back to the Data Intelligence Launchpad (or login like in the previous step)

Define a pipeline

  1. Launch the Modeler running by clicking the Modeler tile (be patient, it takes a while)
    image
  2. Enter com.sap.demo.datagenerator in the search field and click on Data Genrator.
    image
  3. Save the configuration:
    image
  4. Start the graph.
    image
  5. Check that the status changes to Running (this may take several minutes).
    image
  6. Open the Terminal user interface, by right-click the Terminal operator and selecting OpenUI.
    image
  7. Once the Data Generator is running you will get something like the following, otherwise an error is shown
    image
  8. Stop the graph once you observe its output in the Terminal.
    image

Check Your Machine Learning setup

  1. To create an ML scenario, open ML Scenario Manager from the SAP Data Intelligence Launchpad
    image
    Be patient the first start takes a while
  2. Click Create and

     

    a. enter a name for your scenario and

    b, Optional: a business question that you want to answer (such as How strong will the weekly sales be in our store?)

    c. Click Create.

    image

    The details for your scenario appear and your scenario is added to the list of ML scenarios on the overview page.

  3. On the details page for your scenario, click Create in the Notebooks tab to create a new Jupyter Notebook.

     

    a. In the Create Notebook dialog box, enter a unique name for your notebook.

    b. Optional: Enter a description of your notebook.

    c. Click Create to create your Jupyter notebook.

    image

    The creation takes a while until your notebook opens in JupyterLab.

  4. When you open your notebook for the first time, you are prompted to select your kernel. Choose Python 3.
    image
  5. In your JupyterLab notebook, copy the following code into a cell and run it:

     

    import sapdi
    from hdfs import InsecureClient
    client = InsecureClient('http://datalake:50070')
    client.status("/")
    image

    Check that the code runs without errors.

    The code should return JSON similar to the following:

    {'pathSuffix': '',
     'type': 'DIRECTORY',
     'length': 0,
     'owner': 'admin',
     'group': 'admin',
     'permission': '777',
     'accessTime': 0,
     'modificationTime': 1576237423061,
     'blockSize': 0,
     'replication': 1}
    image

You are done. You have successfully installed and tested SAP Data Intelligence

By Markus Koch

I joined Red Hat in October 2010 and since 2017 I work as a Partner Enablement Manager in the EMEA Partner Team, helping our partners to understand, sell and implement SAP solutions with Red Hat products.
I participate in the upstream SAP Ansible Automation project. Previous stations in my career include Sun Microsystems and EMC².

One reply on “How to install SAP Data Intelligence on Red Hat OpenShift”