Provision the following Lab from https://labs.opentlc.com
Services → Catalogs → All Services → OPENTLC OpenShift 4 Labs → OpenShift 4 Installation Lab
You will get an email with your GUID and your sandbox domain:
-
GUID: 4char
-
toplevel domain: .sandboxNNN.opentlc.com
-
AWS access Credentials
-
bastion host with password
-
Login to the bastion host
ssh <OPENTLC User Name>@bastion.<GUID>.sandbox<SANDBOXID>.opentlc.com sudo -i echo ${GUID}
-
Update to current OS release
sudo yum update -y sudo reboot
NoteYou can also reboot at the end of the setup -
Install the following packages
Some of these packages are available in the EPEL repository, which is replicated in this environemnt. Please make sure EPEL in your repository list.
# sudo yum repolist Loaded plugins: amazon-id, product-id, rhui-lb, search-disabled-repos, subscription-manager This system is not registered with an entitlement server. You can use subscription-manager to register. repo id repo name status ansible-tower-cli/x86_64 Ansible Tower CLI Repository - EL7 x86_64 337 ansible-tower-cli-dependencies/x86_64 Ansible Tower CLI Dependencies Repository - EL7 x86_64 61 epel epel 13.343 rhel-7-server-ansible-2.7-rpms rhel-7-server-ansible-2.7-rpms 21 rhel-7-server-extras-rpms rhel-7-server-extras-rpms 1.229 rhel-7-server-optional-rpms rhel-7-server-optional-rpms 19.636 rhel-7-server-rpms rhel-7-server-rpms
if
epel
is not in this list, run the following commands and try again:$ sudo -i # cd /etc/yum.repos.d/ # fgrep -A4 '[rhel-7-server-rpms]' open_ocp4-ha-lab.repo | sed 's/rhel-7-server-rpms/epel/g' >> open_ocp4-ha-lab.repo # yum clean all & yum repolist
Now install the following packages:
sudo yum -y install python2-boto python2-boto3 jq docker
WarningPlease check that all packages are installed -
Check your host
# cat /etc/redhat-release Red Hat Enterprise Linux Server release 7.6 (Maipo) # curl http://169.254.169.254/latest/dynamic/instance-identity/document/ { "accountId" : "126521742790", "imageId" : "ami-092acf20fad7f7795", "availabilityZone" : "eu-west-1b", "ramdiskId" : null, "kernelId" : null, "privateIp" : "192.168.0.202", "devpayProductCodes" : null, "marketplaceProductCodes" : null, "version" : "2017-09-30", "region" : "eu-west-1", "billingProducts" : [ "bp-6fa54006" ], "instanceId" : "i-0a5109ae466c67f2d", "pendingTime" : "2019-10-31T13:14:24Z", "architecture" : "x86_64", "instanceType" : "t2.large"
NoteMake note of the region, in this case eu-west-1
-
Download AWS Cli for verification
# Download the latest AWS Command Line Interface curl "https://s3.amazonaws.com/aws-cli/awscli-bundle.zip" -o "awscli-bundle.zip" unzip awscli-bundle.zip # Install the AWS CLI into /bin/aws sudo ./awscli-bundle/install -i /usr/local/aws -b /bin/aws # Validate that the AWS CLI works aws --version # Clean up downloaded files rm -rf awscli-bundle awscli-bundle.zip
-
Get OCP installer binaries
OCP_VERSION=4.2.8 wget https://mirror.openshift.com/pub/openshift-v4/clients/ocp/${OCP_VERSION}/openshift-install-linux-${OCP_VERSION}.tar.gz sudo tar zxvf openshift-install-linux-${OCP_VERSION}.tar.gz -C /usr/bin sudo rm -f openshift-install-linux-${OCP_VERSION}.tar.gz /usr/bin/README.md sudo chmod +x /usr/bin/openshift-install
-
get oc CLI tool
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
-
setup bash completion
oc completion bash | sudo tee /etc/bash_completion.d/openshift > /dev/null
-
setup AWS account
cat > credentials << EOT export AWS_ACCESS_KEY="<YOURACCESSKEY>" export AWS_SECRET_KEY="<YOURSECRETKEY>" export REGION=<take region from step (4)> EOT . ./credentials mkdir $HOME/.aws cat << EOF >> $HOME/.aws/credentials [default] aws_access_key_id = ${AWS_ACCESS_KEY} aws_secret_access_key = ${AWS_SECRET_KEY} region = $REGION EOF
-
Test AWS account
aws sts get-caller-identity
-
Create an SSH keypair to be used for your OpenShift environment:
ssh-keygen -f ~/.ssh/sdh-${GUID}-key -N ''
The table below lists the minimum requirements and the minimum number of instances for each node type. This is sufficient of a PoC (Proof of Concept) environments.
Type | Count | Operating System | vCPU | RAM (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 |
Jump host |
1 |
RHEL 7.6 |
2 |
4 |
75 |
t2.medium |
For details on production see https://access.redhat.com/articles/4324391
-
prepare Installation:
openshift-install create install-config --dir $HOME/sdh-${GUID}
Use the following answers (replace XXXX and GUID accordningly):
? SSH Public Key [Use arrows to move, type to filter, ? for more help] > /home/mkoch-redhat.com/.ssh/sdh-fb46-key.pub <none> ? Platform aws ? Region <region from above> ? Base Domain sandbox{XXXX}.opentlc.com ? Cluster Name sdh-{GUID} ? Pull Secret [? for help]
Grab the pull secret from the AWS IPI installer page
-
modify/adapt compute nodes regarding SDH requirements in
install-config.yaml
:replace:
[..] compute: - hyperthreading: Enabled name: worker platform: {} replicas: 3 [..]
by:
[..] compute: - hyperthreading: Enabled name: worker platform: aws: type: m4.2xlarge replicas: 3 [..]
NoteYou may save you install-config.yaml for future use Now -
Create the YAML manifests:
openshift-install create manifests --dir $HOME/sdh-${GUID}
-
Disable schedulable masters (Optional)
In OCP 4.2 masters are schedulable by default. If you don’t like it, mark them as not schedulable during the installation:
find "$HOME/sdh-${GUID}/manifests" -type f -name 'cluster-scheduler-*-config.yml' -print0 | \ xargs -0 -r sed -i 's/^\(\s*mastersSchedulable:\s*\)true/\1false/'
-
Create the Ignition configuration files:
openshift-install create ignition-configs --dir $HOME/sdh-${GUID}
-
Modify the worker ignition file to preload kernel modules required for storage and systemd
WarningThis is not supported but saves time, supported is to do it after initial installation by changing the machine sets For use with SAP Datahub the CoreOS nodes need to preload certain kernel modules. This can be done by filling the storage and systemd fields in the ignition file. In the storage field we create a file containing the kernel modules that need to be preloaded, in the systemd section we apply a couple IPtables NAT rules required for SAP Datahub.
cd ${HOME}/sdh-${GUID} mv worker.ign worker.ign.dist jq '.storage = { "files": [ { "contents": { "source": "data:text/plain;charset=utf-8;base64,bmZzZAppcF90YWJsZXMKaXB0X1JFRElSRUNUCg==", "verification": { } }, "filesystem": "root", "mode": 420, "path": "/etc/modules-load.d/sap-datahub-dependencies.conf" } ] }' worker.ign.dist |\ jq -c '.systemd = { "units": [ { "contents": "[Unit]\nDescription=Pre-load kernel modules for SAP Data Hub\nAfter=network.target\n\n[Service]\nType=simple\nExecStart=/usr/sbin/modprobe iptable_nat\nRestart=on-failure\nRestartSec=10\nRemainAfterExit=yes\n\n[Install]\nWantedBy=multi-user.target\n", "enabled": true, "name": "sdh-modules-load.service" } ] } ' > worker.ign
Note-c in the jq command brings the output back in a single line, without -c its readable -
Install cluster
openshift-install create cluster --dir $HOME/sdh-${GUID}
-
verify that changes in worker.ign made it to the system:
-
Verify that the compute nodes are of type
m4.2xlarge
:oc get machines -n openshift-machine-api
sample output:
NAME INSTANCE STATE TYPE REGION ZONE AGE sdh-06d9-5p8xk-master-0 i-0ba81e2443bd3c814 running m4.xlarge eu-central-1 eu-central-1a 30m sdh-06d9-5p8xk-master-1 i-055033ff08f2323ad running m4.xlarge eu-central-1 eu-central-1b 30m sdh-06d9-5p8xk-master-2 i-02928115caba789c3 running m4.xlarge eu-central-1 eu-central-1c 30m sdh-06d9-5p8xk-worker-eu-central-1a-zk4sw i-04099c8f7a803d5c3 running m4.2xlarge eu-central-1 eu-central-1a 29m sdh-06d9-5p8xk-worker-eu-central-1b-82wbq i-0a4a1a504e723700c running m4.2xlarge eu-central-1 eu-central-1b 29m sdh-06d9-5p8xk-worker-eu-central-1c-d99gn i-000d45b2bac8faaa0 running m4.2xlarge eu-central-1 eu-central-1c 29m
-
Verify that the addional kernel modules are in
/etc/modules-load.d/sap-datahub-dependencies.conf
and the servicesdh-modules-load.service
are available on each worker node:for worker in `oc get nodes | awk '/worker/{print $1}'`; do oc debug node/$worker -- chroot /host cat /etc/modules-load.d/sap-datahub-dependencies.conf oc debug node/$worker -- chroot /host systemctl status sdh-modules-load.service done
sample output:
Starting pod/ip-10-0-129-74eu-central-1computeinternal-debug ... To use host binaries, run `chroot /host` nfsd ip_tables ipt_REDIRECT Removing debug pod ... Starting pod/ip-10-0-129-74eu-central-1computeinternal-debug ... To use host binaries, run `chroot /host` ● sdh-modules-load.service - Pre-load kernel modules for SAP Data Hub Loaded: loaded (/etc/systemd/system/sdh-modules-load.service; enabled; vendor preset: enabled) Active: active (exited) since Mon 2019-11-11 10:24:54 UTC; 27min ago Process: 921 ExecStart=/usr/sbin/modprobe iptable_nat (code=exited, status=0/SUCCESS) Main PID: 921 (code=exited, status=0/SUCCESS) CPU: 10ms [...]
-
-
Label the pool of worker nodes for use with SAP DataHub:
# oc label machineconfigpool/worker workload=sapdatahub
-
Create the following ContainerRuntimeConfig resource.
# oc create -f - <<EOF apiVersion: machineconfiguration.openshift.io/v1 kind: ContainerRuntimeConfig metadata: name: bumped-pid-limit spec: machineConfigPoolSelector: matchLabels: workload: sapdatahub containerRuntimeConfig: pidsLimit: 4096 EOF
-
Wait until the machineconfigpool/worker becomes updated.
# watch oc get machineconfigpool/worker NAME CONFIG UPDATED UPDATING DEGRADED worker rendered-worker-8f91dd5fdd2f6c5555c405294ce5f83c True False False
-
Verify changed configuration with
for worker in `oc get nodes | awk '/worker/{print $1}'`; do oc debug node/$worker -- cat /host/etc/crio/crio.conf done | grep -i pids_limit
-
Install docker on Jumphost
sudo yum install docker
-
start docker services
sudo systemctl enable docker sudo systemctl start docker
-
Prepare docker for installation from user, i.e. make sure your jumphost user has root-access
sudo usermod -a -G dockerroot mkoch-redhat.com sudo chown root:dockerroot /var/run/docker.sock
Caution/var/run/docker.sock
will beroot:root
after restarting docker daemon. This is a default behaviour because every user of the group dockerroot can become root, by running a priviledged container accessing any root file. -
Log out and back in again to activate the new group
-
Login to docker registry
eval $(aws ecr get-login --no-include-email)
-
store information in Variables
eval $(aws ecr get-login --no-include-email | awk '{ print ( "export DOCKER_LOGIN="$4 ); print ("export DOCKER_TOKEN="$6 ); sub ("^http[s]*://","",$7) ; print ("export DOCKER_REGISTRY="$7)}')
-
create repositories for the docker images in AWS ECR
AWS ECR requires a separate repository with the name of the image for each image before versions of the images can be pushed into AWS ECR
-
create
setup-ecr.yml
with the following content--- - hosts: localhost gather_facts: no connection: local tags: provisioning vars: aws_region: eu-central-1 repo_state: absent ecr_sdh_repos: - com.sap.bds.docker/storagegateway - com.sap.datahub.linuxx86_64/app-base - com.sap.datahub.linuxx86_64/auth-proxy - com.sap.datahub.linuxx86_64/dq-integration - com.sap.datahub.linuxx86_64/elasticsearch - com.sap.datahub.linuxx86_64/flowagent-codegen - com.sap.datahub.linuxx86_64/flowagent-operator - com.sap.datahub.linuxx86_64/flowagent-service - com.sap.datahub.linuxx86_64/fluentd - com.sap.datahub.linuxx86_64/grafana - com.sap.datahub.linuxx86_64/hello-sap - com.sap.datahub.linuxx86_64/init-security - com.sap.datahub.linuxx86_64/kibana - com.sap.datahub.linuxx86_64/kube-state-metrics - com.sap.datahub.linuxx86_64/nats - com.sap.datahub.linuxx86_64/node-exporter - com.sap.datahub.linuxx86_64/opensuse-leap - com.sap.datahub.linuxx86_64/prometheus - com.sap.datahub.linuxx86_64/pushgateway - com.sap.datahub.linuxx86_64/security-operator - com.sap.datahub.linuxx86_64/spark-datasourcedist - com.sap.datahub.linuxx86_64/uaa - com.sap.datahub.linuxx86_64/vflow-python36 - com.sap.datahub.linuxx86_64/vora-deployment-operator - com.sap.datahub.linuxx86_64/vora-dqp - com.sap.datahub.linuxx86_64/vora-dqp-textanalysis - com.sap.datahub.linuxx86_64/vora-license-manager - com.sap.datahub.linuxx86_64/vsolution-golang - com.sap.datahub.linuxx86_64/vsolution-hana_replication - com.sap.datahub.linuxx86_64/vsolution-ml-python - com.sap.datahub.linuxx86_64/rbase - com.sap.datahub.linuxx86_64/vsolution-sapjvm - com.sap.datahub.linuxx86_64/vsolution-spark_on_k8s - com.sap.datahub.linuxx86_64/vsolution-streaming - com.sap.datahub.linuxx86_64/vsolution-textanalysis - com.sap.datahub.linuxx86_64/vsystem - com.sap.datahub.linuxx86_64/vsystem-auth - com.sap.datahub.linuxx86_64/vsystem-hana-init - com.sap.datahub.linuxx86_64/vsystem-module-loader - com.sap.datahub.linuxx86_64/vsystem-shared-ui - com.sap.datahub.linuxx86_64/vsystem-teardown - com.sap.datahub.linuxx86_64/vsystem-ui - com.sap.datahub.linuxx86_64/vsystem-voraadapter - com.sap.datahub.linuxx86_64/vsystem-vrep - com.sap.hana.container/base-opensuse42.3-amd64 - consul - kaniko-project/executor - com.sap.datahub.linuxx86_64/hana - com.sap.datahub.linuxx86_64/sles - com.sap.datahub.linuxx86_64/vsystem-vrep-csi - com.sap.datahub.linuxx86_64/code-server - com.sap.datahub.linuxx86_64/axino-service tasks: - name: Create SAP Datahub Repos ecs_ecr: name: "{{ item }}" state: "{{repo_state}}" with_items: "{{ ecr_sdh_repos }}"
NoteIf you want to use diffferent namespaces for Deployment and SAP Data Modeller follow the steps in the SAP documentation. This is strongly recommended for production environments, because diffrent instance may delete docker images in the registry unintendedly. -
Run the playbook
ansible-playbook setup-ecr.yml -e repo_state=present
-
-
Install helm client
# DESIRED_VERSION=v2.13.1 # curl --silent https://raw.githubusercontent.com/kubernetes/helm/master/scripts/get | \ DESIRED_VERSION="${DESIRED_VERSION:-v2.13.1}" bash
sample output:
Downloading https://get.helm.sh/helm-v2.13.1-linux-amd64.tar.gz
Preparing to install helm and tiller into /usr/local/bin
helm installed into /usr/local/bin/helm
tiller installed into /usr/local/bin/tiller
Run 'helm init' to configure helm.
-
Create according service account
oc create sa -n kube-system tiller
sample output:
serviceaccount/tiller created
-
Add policy:
oc adm policy add-cluster-role-to-user cluster-admin -n kube-system -z tiller
sample output:
clusterrole.rbac.authorization.k8s.io/cluster-admin added: "tiller"
-
Initialize helm:
helm init --service-account=tiller --upgrade --wait
sample output:
Creating /home/mkoch-redhat.com/.helm Creating /home/mkoch-redhat.com/.helm/repository Creating /home/mkoch-redhat.com/.helm/repository/cache Creating /home/mkoch-redhat.com/.helm/repository/local Creating /home/mkoch-redhat.com/.helm/plugins Creating /home/mkoch-redhat.com/.helm/starters Creating /home/mkoch-redhat.com/.helm/cache/archive Creating /home/mkoch-redhat.com/.helm/repository/repositories.yaml Adding stable repo with URL: https://kubernetes-charts.storage.googleapis.com Adding local repo with URL: http://127.0.0.1:8879/charts $HELM_HOME has been configured at /home/mkoch-redhat.com/.helm. Tiller (the Helm server-side component) has been installed into your Kubernetes Cluster. Please note: by default, Tiller is deployed with an insecure 'allow unauthenticated users' policy. To prevent this, run `helm init` with the --tiller-tls-verify flag. For more information on securing your installation see: https://docs.helm.sh/using_helm/#securing-your-helm-installation Happy Helming!
-
Check that the tiller pod is running:
$ oc get pods -n kube-system
NAME READY STATUS RESTARTS AGE
tiller-deploy-dbb85cb99-szjtt 1/1 Running 0 3m59s
-
Create Project for SAP DH
$ oc new-project sdh Now using project "sdh" on server "https://api.cluster-d217.sandbox1789.opentlc.com:6443". You can add applications to this project with the 'new-app' command. For example, try: oc new-app django-psql-example to build a new example application in Python. Or use kubectl to deploy a simple Kubernetes application: kubectl create deployment hello-node --image=gcr.io/hello-minikube-zero-install/hello-node
-
Add required priviledges
oc adm policy add-scc-to-group anyuid "system:serviceaccounts:$(oc project -q)" oc adm policy add-scc-to-group hostmount-anyuid "system:serviceaccounts:$(oc project -q)" oc adm policy add-scc-to-user privileged -z "vora-vsystem-$(oc project -q)" oc adm policy add-scc-to-user privileged -z "$(oc project -q)-elasticsearch" oc adm policy add-scc-to-user privileged -z "$(oc project -q)-fluentd" oc adm policy add-scc-to-user privileged -z "default" oc adm policy add-scc-to-user privileged -z "vora-vflow-server"
New for SAP DH 2.7
oc adm policy add-scc-to-user hostaccess -z "$(oc project -q)-nodeexporter" oc adm policy add-scc-to-user privileged -z "vora-vsystem-$(oc project -q)-vrep"
sample output:
$ oc adm policy add-scc-to-group anyuid "system:serviceaccounts:$(oc project -q)"
securitycontextconstraints.security.openshift.io/anyuid added to groups: ["system:serviceaccounts:sdh"]
$ oc adm policy add-scc-to-group hostmount-anyuid "system:serviceaccounts:$(oc project -q)"
securitycontextconstraints.security.openshift.io/hostmount-anyuid added to groups: ["system:serviceaccounts:sdh"]
$ oc adm policy add-scc-to-user privileged -z "vora-vsystem-$(oc project -q)"
securitycontextconstraints.security.openshift.io/privileged added to: ["system:serviceaccount:sdh:vora-vsystem-sdh"]
$ oc adm policy add-scc-to-user privileged -z "$(oc project -q)-elasticsearch"
securitycontextconstraints.security.openshift.io/privileged added to: ["system:serviceaccount:sdh:sdh-elasticsearch"]
$ oc adm policy add-scc-to-user privileged -z "$(oc project -q)-fluentd"
securitycontextconstraints.security.openshift.io/privileged added to: ["system:serviceaccount:sdh:sdh-fluentd"]
$ oc adm policy add-scc-to-user privileged -z "default"
securitycontextconstraints.security.openshift.io/privileged added to: ["system:serviceaccount:sdh:default"]
$ oc adm policy add-scc-to-user privileged -z "vora-vflow-server"
securitycontextconstraints.security.openshift.io/privileged added to: ["system:serviceaccount:sdh:vora-vflow-server"]
-
As a cluster-admin, allow the project administrator to manage SDH custom resources.
# oc create -f - <<EOF kind: ClusterRole apiVersion: rbac.authorization.k8s.io/v1 metadata: name: aggregate-sapvc-admin-edit labels: rbac.authorization.k8s.io/aggregate-to-admin: "true" rbac.authorization.k8s.io/aggregate-to-edit: "true" rules: - apiGroups: ["sap.com"] resources: ["voraclusters"] verbs: ["get", "list", "watch", "create", "update", "patch", "delete", "deletecollection"] --- kind: ClusterRole apiVersion: rbac.authorization.k8s.io/v1 metadata: name: aggregate-sapvc-view labels: # Add these permissions to the "view" default role. rbac.authorization.k8s.io/aggregate-to-view: "true" rules: - apiGroups: ["sap.com"] resources: ["voraclusters"] verbs: ["get", "list", "watch"] EOF
sample output:
clusterrole.rbac.authorization.k8s.io/aggregate-sapvc-admin-edit created clusterrole.rbac.authorization.k8s.io/aggregate-sapvc-view created
SDH Observer is comtainer which patches datahub deployment contexts to run properly on OpenShift. It monitors the deployment and make the changes when appropriate.
For more information see : https://access.redhat.com/articles/4324391#deploy-sdh-observer
-
Switch to project sdh:
oc status In project sdh on server https://api.cluster-d217.sandbox1789.opentlc.com:6443 You have no services, deployment configs, or build configs. Run 'oc new-app' to create an application.
-
Deploy SDH observer
OCPVER=4.2 INSECURE_REGISTRY=false oc process -f https://raw.githubusercontent.com/miminar/sdh-helpers/master/sdh-observer.yaml \ NAMESPACE="$(oc project -q)" \ BASE_IMAGE_TAG="${OCPVER:-4.2}" \ MARK_REGISTRY_INSECURE=${INSECURE_REGISTRY:-0} | oc create -f -
For installing SAP DataHub you need your need your S-User account and password.
-
Download SAP DataHub binaries & unzip on jumphost
-
Go to SAP Software Download Center, login with your SAP account and search for DATA HUB 2 or access this link.
-
Download the SAP Data Hub Foundation file, for example:
DHFOUNDATION07_2-80004015.ZIP (SAP DATA HUB - FOUNDATION 2.7)`
. -
Unpack the installer file and change to this directory. Type
install.sh -h
to verify the installer options$ unzip DHFOUNDATION07_2-80004015.ZIP $ cd SAPDataHub-2.7.152-Foundation $ ./install.sh -h
-
-
Set Environment Variables to define Namespace and verify docker registry
echo $DOCKER_REGISTRY export NAMESPACE=sdh
-
Mirror the SDH images to the local registry
On the local disk is not enough space to mirror everything so repeat the following steps until everything is uploaded:
-
Preload images
$ ./install.sh -b -a
Noteif you receive an error with no basic auth credentials
you may need to login to AWS ECR registry:eval $(aws ecr get-login --no-include-email)
-
-
It is possible to cleanup some images that are uploaded from time to time:
for i in $(docker images | awk '/'$DOCKER_REGISTRY'/ { print $1":"$2 }'); do docker inspect $i --format='{{.Size}} {{.RepoTags}}'; done | sort -n
take the largest image and make sure it is uploaded. Then remove it from the local disk:
docker rmi <names of largest image>
-
When the upload has stopped due to diskspace errors like
write /var/lib/docker/tmp/GetImageBlob391058538: no space left on device 2019-11-08T14:48:30+0000 [ERROR] Image pulling failed, please see logs above!
delete all existing locally cached images and re-run the preload
$ docker rmi -f $(docker images | awk '{ print $3}' | uniq ) $ ./install.sh -b -a
NoteThis takes a couple of hours -
Make sure that workers can access the ECR registry.
To access the ECR Registry you have to attach a sufficent access policy to the worker role.
sdh_worker_role=$(aws iam list-instance-profiles | jq -r '.InstanceProfiles[] | select(.InstanceProfileName | test("worker-profile")) | .Roles[] | select(.RoleName | test("worker-role")) | "\(.RoleName)"')
Now check if the worker has the
AmazonEC2ContainerRegistryPowerUser
role attached:aws iam list-attached-role-policies --role-name $sdh_worker_role { "AttachedPolicies": [ { "PolicyName": "AmazonEC2ContainerRegistryPowerUser", "PolicyArn": "arn:aws:iam::aws:policy/AmazonEC2ContainerRegistryPowerUser" } ] }
if you don’t see the role, run:
aws iam attach-role-policy --role-name $sdh_worker_role --policy-arn arn:aws:iam::aws:policy/AmazonEC2ContainerRegistryPowerUser
-
Figure out Installation parameters
check storage class:
[mkoch-redhat.com@clientvm 1 ~]$ oc get storageclass NAME PROVISIONER AGE gp2 (default) kubernetes.io/aws-ebs 2d22h
for Amazon EBS is fine.
As we export the UI via Open Shift routes the name for the cert-domain is like this:
vsystem-{namespace}.{wildcard_domain}
, so in our case usevsystem-sdh.apps.sdh-${GUID}.sandboxNNN.opentlc.com
-
So the following parameters should used to kick-off installation
CautionPlease do not use checkpoint store in the installation. To use checkpoint store you need to create an S3 storage bucket and assign access to the S3 storage as described in the apendix to all cluster nodes including the bastion host (for checking) ./install.sh -i -a --enable-kaniko=yes \ --pv-storage-class="gp2"
for an unattended install you can type something similar to this:
./install.sh -a -i \ --pv-storage-class="gp2"\ --enable-kaniko=yes\ --vora-system-password 'MyPassw0rd!' \ --vora-admin-username redhat \ --vora-admin-password 'MyPassw0rd!' \ --enable-checkpoint-store no \ --cert-domain ${CERT_DOMAIN}
sample output log:
[...] No SSL certificate has been provided via the --provide-certs parameter. The SAP Data Hub installer will generate a self-signed certificate for TLS and JWT. Please enter the SAN (Subject Alternative Name) for the certificate, which must match the fully qualified domain name (FQDN) of the Kubernetes node to be accessed externally: vsystem-sdh.apps.cluster-d217.sandbox1789.opentlc.com SAP Data Hub System Tenant Administrator Credentials Provide a password for the "system" user of "system" tenant. The password must have 8-255 characters and must contain lower case, upper case, numerical and on of the following special characters . @ # $ %% * + _ ? ! It cannot contain spaces. Please enter a password for "system" user of "system" tenant: R3dh4t1! Please reenter your password: SAP Dat Hub Initial Tenant Administrator Credentials Provide a username and password for administrator user of "default" tenant. The username must have at least 4 and at most 60 characters Allowed characters: alphabetic(only lowercase), digits and hyphens Username is not allowed to begin/end with hyphens and cannot contain multiple consecutive hyphens Please enter a username for default tenant: redhat Do you want to use the same "system" user password for "redhat" user of "default" tenant? (yes/no) yes Do you want to configure security contexts for Hadoop/Kerberized Hadoop? (yes/no) no 2019-11-07T11:56:05+0000 [INFO] Configuring contexts with: python2.7 configure_contexts.py -a -n --set Vora_JWT_Issuer_NI.default --set Vora_Default_TLS_Configuration_NI.default secret/vora.conf.secop.contexts created secret/vora.conf.secop.contexts labeled 2019-11-07T11:56:06+0000 [INFO] Vora streaming tables require Vora's checkpoint store\n Enable Vora checkpoint store? (yes/no) no # ###### Configuration Summary ####### installer: ASK_FOR_CERTS: '' AUDITLOG_MODE: production CERT_DOMAIN: vsystem-sdh.apps.cluster-d217.sandbox1789.opentlc.com CHECKPOINT_STORE_TYPE: '' CHECKPOINT_STORE_TYPE_RAW: '' CLUSTER_HTTPS_PROXY: '' CLUSTER_HTTP_PROXY: '' CLUSTER_NO_PROXY: '' CONSUL_STORAGE_CLASS: '' CUSTOM_DOCKER_LOG_PATH: '' DIAGNOSTIC_STORAGE_CLASS: '' DISABLE_INSTALLER_LOGGING: '' DISK_STORAGE_CLASS: '' DLOG_STORAGE_CLASS: '' DOCKER_REGISTRY: 126521742790.dkr.ecr.eu-central-1.amazonaws.com ENABLE_CHECKPOINT_STORE: 'false' ENABLE_DIAGNOSTIC_PERSISTENCY: 'yes' ENABLE_DQP_ANTIAFFINITY: 'yes' ENABLE_KANIKO: 'yes' ENABLE_NETWORK_POLICIES: 'no' ENABLE_RBAC: 'yes' HANA_STORAGE_CLASS: '' IMAGE_PULL_SECRET: '' PACKAGE_VERSION: 2.6.102 PV_STORAGE_CLASS: '' TILLER_NAMESPACE: '' USE_K8S_DISCOVERY: 'yes' VALIDATE_CHECKPOINT_STORE: '' VFLOW_AWS_IAM_ROLE: '' VFLOW_IMAGE_PULL_SECRET: '' VFLOW_REGISTRY: 126521742790.dkr.ecr.eu-central-1.amazonaws.com VORA_ADMIN_USERNAME: redhat VORA_FLAVOR: '' VORA_VSYSTEM_DEFAULT_TENANT_NAME: default VSYSTEM_LOAD_NFS_MODULES: 'yes' VSYSTEM_STORAGE_CLASS: '' ###################################### [...]
-
While the installation is running watch all pods coming up with
oc get pods --namespace=sdh -w
-
Test cluster health using Helm test (info is printed by installer):
$ helm test <watch and use output from installer>
-
(Optional) Manually confirm consul cluster is healthy.
-
kubectl exec vora-consul-0 consul members --namespace=sdh | grep server
OpenShift allows you to access the Data Hub services via routes as opposed to regular NodePorts. For example, instead of accessing the vsystem service via https://master-node.example.com:32322
, after the service exposure, you will be able to access it at https://vsystem-sdh.wildcard-domain
. This is an alternative to the official guide documentation to Expose the Service From Outside the Network.
-
Look up the
vsystem
service:# oc project sdh # switch to the Data Hub project # oc get services | grep "vsystem " vsystem ClusterIP 172.30.227.186 <none> 8797/TCP 19h
-
create the route
# oc create route passthrough --service=vsystem # oc get route NAME HOST/PORT PATH SERVICES PORT TERMINATION WILDCARD vsystem vsystem-sdh.wildcard-domain vsystem vsystem passthrough None
-
(Optional) Expose the SAP Vora Transaction Coordinator for external access:
# oc create route passthrough --service=vora-tx-coordinator-ext # oc get route NAME HOST/PORT PATH SERVICES PORT TERMINATION WILDCARD vora-tx-coordinator-ext vora-tx-coordinator-ext-sdh.wildcard-domain vora-tx-coordinator-ext tc-ext passthrough None vsystem vsystem-sdh.wildcard-domain vsystem vsystem passthrough None
Note
|
if you want to create a different hostname instead of the auto-generated use the option --hostname=vora-tx-coordinator.wildcard-domain
|
-
(Optional) Expose the SAP HANA Wire for external access
# oc create route passthrough --service=vora-tx-coordinator-ext --port=hana-wire --dry-run -o yaml | \ oc patch -o yaml --local -p '{"metadata":{"name":"hana-wire"}}' -f - | oc create -f - # oc get route NAME HOST/PORT PATH SERVICES PORT TERMINATION WILDCARD hana-wire hana-wire-sdh.wildcard-domain vora-tx-coordinator-ext hana-wire passthrough None vora-tx-coordinator-ext vora-tx-coordinator-ext-sdh.wildcard-domain vora-tx-coordinator-ext tc-ext passthrough None vsystem vsystem-sdh.wildcard-domain vsystem vsystem passthrough None
You can now access the SDH web console at https://vsystem-sdh.wildcard-domain
Note
|
Exposing via NodePorts is possible, too, but for OpenShift exposure using routes is preferred |
SAP Data Hub installer allows to specify "AWS IAM Role for Pipeline Modeler" when AWS ECR Registry is used as the external registry. However, due to a bug in Data Hub, the Modeler cannot use it. In order to use AWS ECR Registry for Data Hub, one can follow the instructions at Provide Access Credentials for a Password Protected Container Registry by using the AWS_ACCESS_KEY as user and the AWS_SECRET_KEY as password:
-
Create a secret in DataHub
-
create the following secret file with AWS credentials:
# cat >/tmp/vsystem-registry-secret.txt <<EOF username: "$AWS_ACCESS_KEY" password: "$AWS_SECRET_KEY" EOF
NoteThe quotes around user and password are important -
Log in to SAP Datahub (https://vsystem-sdh.apps.sdh-${GUID}.sandboxNNN.opentlc.com/) with tenant "default" and user and password you chose during installation
-
click the System Managemt tile
-
click the Application Configuration & Secrets button above the search bar.
-
Click the Secrets tab, and then click the Create icon.
-
For the secret name, enter
vflow-registry
. -
Browse to select and upload the secret file
vsystem-registry-secret.txt
that you previously created. -
Click Create.
-
-
Apply the newly created secret to the application configuration:
-
To open the configuration settings, click the Application Configuration & Secrets button above the search bar.
-
In the Configuration tab, find the following parameter:
Modeler: Name of the vSystem secret containing the credentials for Docker registry
. -
Enter
vflow-registry
, which is the name of the secret that you previously created.
-
-
In SAP Data Hub System Management, start the Modeler application:
-
Launch the SAP Data Hub System Management application and open the Applications tab.
-
Select the Modeler application in the left pane, and click the Create an Application button in the upper right.
-
To clean up the environment, do the following:
-
Log in to your bastion VM.
-
cleanup AWS registry
-
Delete all images from registry by running the following shell script
!/bin/bash for r in $(aws ecr describe-repositories | awk '/repositoryName/ {print $2}' | tr -d '\",'); do echo "Cleaning up repository $r" for i in $(aws ecr list-images --repository-name $r | awk '/imageDigest/ {print $2}' | tr -d '\",'); do set -x aws ecr batch-delete-image --repository-name $r --image-ids imageDigest=$i set +x done done
-
delete the repositories from the registry
# ansible-playbook setup-ecr.yml -e repo_state=absent
-
delete ECR policy from worker nodes
aws iam detach-role-policy --role-name sdh-fb46-vkszx-worker-role --policy-arn arn:aws:iam::aws:policy/AmazonEC2ContainerRegistryPowerUser
-
-
Delete the cluster:
openshift-install destroy cluster --dir=${HOME}/sdh-${GUID}
-
Delete all of the files created by the OpenShift installer:
rm -rf ${HOME}/.kube rm -rf ${HOME}/sdh-${GUID}
Delete your environment from https://labs.opentlc.com.
This concludes the SAP DataHub lab.