Upgrade Notes

• Prepare new OpenShift 4.3 Cluster
• Install CPD
• Decide whether to keep using your current GitLab instance or whether you move to a new GitLab installation (migration using non GitLab git servers is not supported)

• If you switch to a new GitLab instance - backup and restore all your FSW related repositories in the managed-solutions group to the new GitLab instance using the same group managed-solutions as the target

• Decide whether to keep your MongoDB instance or move to a new instance
• If you switch to new MongoDB instance backup and restore the relevant databases
• Install IBM Financial Services Workbench 2.2 on the new cluster providing connection parameters for the new GitLab and MongoDB instances
• Have all your users (developers and analysts) setup their GitLab access tokens in the Solution Designer
• Have your developers checkout the solutions they need to work on using FSW cli tools (fss)

• Go to your existing GitLab
• Export GitLab projects
• Import GitLab projects to your new GitLab into group managed-solutions
• Alternatively: use GitLab backup and restore functionality to move the data

• Prepare new MongoDB instance
• Restore data from a backup obtained from old instance
• Note new connection details

• Trigger installation of IBM Financial Services Workbench 2.2

• Provide base URL of new Git Provider (or current GitLab if you keep your current) for the automated migration script

• Each user (both developers and analysts) needs to create a Git Token for the current Git Repository in the UserSettings
• Git Provider Admin need to ensure that users have enough access on the Git system
  • to create new solutions: at least developer permissions on the group must be granted
  • to edit the solution content: at least developer permission (on the group or the repository) must be granted
  • to view the solution: at least guest or reporter permission (on the group or the repository) must be granted
  • to delete a solution: at least owner permission on the group must be granted
• Setting up the developer workspaces
  • Run fss upgrade-cli
  • Download the latest version of the cli-config of a Solution from the Solution Designer
  • Run fss setup using the new downloaded file
  • Run fss clone for all relevant solutions

• Using the Solution Controller API:
  • Before you upgrade: Download the template
    • Open https://<NAMESPACE>/swagger-ui.html -> Local Marketplace Controller API
      • Use GET /api/marketplace/v2/templates to retrieve all available templates
      • Download each template using GET /api/marketplace/v2/templates/ {templateId}
  • After the upgrade: Upload the templates
    • Open https://<NAMESPACE>swagger-ui.html -> Local Marketplace Controller API
      • Use POST /api/marketplace/v2/templates/- to upload each template

• Direct copy of s3 bucket from s3 storage to s3 storage
  • Requirement: S3 Client e.g. https://github.com/minio/mc
  • Before upgrade

    # Get credentials from secret e.g. ssob-sdo-minio-secret
    > oc project <HUB NAMESPACE>
    > export GITLAB_SECRET=ssob-sdo-minio-secret
    > export GITLAB_ACCESS_KEY=$(oc get secret ${GITLAB_SECRET} -o jsonpath="{.data.accesskey}" | base64 -d)
    > export GITLAB_SECRET_KEY=$(oc get secret ${GITLAB_SECRET} -o jsonpath="{.data.secretkey}" | base64 -d)
    > export GITLAB_ENDPOINT=$(oc get cm solution-controller-application -o 'go-template={{index .data "marketplace.storage.endpoint"}}')
    > export GITLAB_PORT=$(oc get cm solution-controller-application -o 'go-template={{index .data "marketplace.storage.port"}}')
    
    # Open port forward, if the endpoint is internal like http://<SERVICE>.<NAMESPACE>.svc.cluster.local e.g. http://gitlab-minio-svc.gitlab.svc.cluster.local
    > echo ${GITLAB_ENDPOINT}
    > kubectl -n <NAMESPACE> port-forward svc/<SERVICE> ${GITLAB_PORT}:${GITLAB_PORT}
    > export GITLAB_ENDPOINT="http://localhost"
    
    # Retrieve data
    > mkdir -p data
    > set +o history
    > mc config host add gitlab ${GITLAB_ENDPOINT}:${GITLAB_PORT} ${GITLAB_ACCESS_KEY} ${GITLAB_SECRET_KEY}
    > set -o history
    > mc ls gitlab/marketplace-templates/
    > mc cp --recursive gitlab/marketplace-templates/ data
    > ls -la data/
    

  • After upgrade

    # Get credentials from secret e.g. k5-s3-storage-access
    > oc project <DESIGNER NAMESPACE>
    > export K5S3STORAGE_SECRET=k5-s3-storage-access
    > export K5S3STORAGE_ACCESS_KEY=$(oc get secret ${K5S3STORAGE_SECRET} -o jsonpath="{.data.accesskey}" | base64 -d)
    > export K5S3STORAGE_SECRET_KEY=$(oc get secret ${K5S3STORAGE_SECRET} -o jsonpath="{.data.secretkey}" | base64 -d)
    > export K5S3STORAGE_ENDPOINT=$(oc get cm solution-controller-application -o 'go-template={{index .data "marketplace.storage.endpoint"}}')
    > export K5S3STORAGE_PORT=$(oc get cm solution-controller-application -o 'go-template={{index .data "marketplace.storage.port"}}')
    
    # Open port forward, if the endpoint is internal like https://<SERVICE>.<NAMESPACE>.svc.cluster.local e.g. https://k5-s3-storage.designer.svc.cluster.local
    > echo ${K5S3STORAGE_ENDPOINT}
    > kubectl -n <NAMESPACE> port-forward svc/<SERVICE> ${K5S3STORAGE_PORT}:${K5S3STORAGE_PORT}
    kubectl -n dev-designer port-forward svc/k5-s3-storage 9000:9000
    > export K5S3STORAGE_ENDPOINT="https://localhost"
    
    # Upload data (--insecure is required, as the mc will use localhost and the certificate is created for the service hostname)
    > set +o history
    > mc config host add k5s3storage ${K5S3STORAGE_ENDPOINT}:${K5S3STORAGE_PORT} ${K5S3STORAGE_ACCESS_KEY} ${K5S3STORAGE_SECRET_KEY} --insecure
    > set -o history
    > cd data
    > mc --insecure ls k5s3storage/marketplace-templates --recursive
    > mc ls --recursive .
    > mc --insecure cp --recursive . k5s3storage/marketplace-templates
    > mc --insecure ls k5s3storage/marketplace-templates --recursive

• Run fss upgrade-cli

• Backup your Pipeline resources if necessary (see section Backup and Restore Pipeline resources)
• Delete the instance (name: cluster) of the CRD config.operator.tekton.dev
• Uninstall the community version of this operator from OperatorHub
• Install the OpenShift Pipelines Operator in version 1.0.1 using the OperatorHub
• Restore your Pipeline resources if necessary (see section Backup and Restore Pipeline resources)

#!/bin/bash
rm -f pipelines_backup.yaml
rm -f pipelines_resources_backup.yaml
rm -f pipelines_list
rm -f pipelines_list_resources
exec &> back_up_logfile.txt
echo "START....................................................."
echo "Getting all pipelines and namespaces"
echo "\n"
oc get pipelines --all-namespaces | awk 'NR>2 {print $1 , $2}' > pipelines_list
echo "---" >> pipelines_backup.yaml
echo "extracting every pipeline from all the namespaces"
echo "\n"
cat pipelines_list | while read line 
do
  namespace=`echo $line | cut -d' ' -f1`
  pipeline_name=`echo $line | cut -d' ' -f2`
  echo "back-up of pipeline start.." $pipeline_name
  oc get pipeline $pipeline_name -o yaml -n $namespace >> pipelines_backup.yaml
  echo "---" >> pipelines_backup.yaml
done
echo "Back-up of pipelines in all the namespaces is completed"
echo "\n"
echo "\n"
echo "\n"
echo "Back-up of pipelines resources in all the namespaces is started"
echo "\n"
oc get pipelineresources --all-namespaces | awk 'NR>2 {print $1 , $2}' > pipelines_list_resources
echo "---" >> pipelines_resources_backup.yaml
echo "extracting every pipeline resources from all the namespaces"
echo "\n"
cat pipelines_list_resources | while read line 
do
  namespace=`echo $line | cut -d' ' -f1`
  pipeline_resource_name=`echo $line | cut -d' ' -f2`
  echo "back-up of pipeline resources start.." $pipeline_resource_name
  oc get pipelineresources $pipeline_resource_name -o yaml -n $namespace >> pipelines_resources_backup.yaml
  echo "---" >> pipelines_resources_backup.yaml
done
echo "Back-up of pipelines resources in all the namespaces is completed"
echo "\n"
echo "End....................................................."
            

#!/bin/bash
exec &> restore_logfile.txt
echo "START....................................................."
echo "checking for backup files"
pipeline_backup="pipelines_backup.yaml"
pipelines_resources_backup="pipelines_resources_backup.yaml"
pipelines_list="pipelines_list"
OC_DOMAIN_URL=$1
OC_TOKEN=$2

if [ ! -f "$pipeline_backup" ]
then
    echo "File '${pipeline_backup}' not found."
    exit 0
elif  [ ! -f "$pipelines_resources_backup" ]
then
    echo "File '${pipelines_resources_backup}' not found."
    exit 1
elif  [ -z $1 ]
then
    echo "Openshift URL is not provided."
  echo "Usage: sh restore.sh <Openshift-url-except-namespace> <openshift-token>"
    exit 2
elif  [ -z $2 ]
then
    echo "Openshift Token is not provided."
  echo "Usage: sh restore.sh <Openshift-url-except-namespace> <openshift-token>"
    exit 3
fi
echo "\n"
echo "Started restoring pipelines"
oc apply -f $pipeline_backup
echo "Restoring pipelines completed"
echo "\n"
echo "Started restoring pipeline resources"
oc apply -f $pipelines_resources_backup
echo "Restoring pipeline resources completed"
echo "\n"
echo "Triggering all pipelines to create pipeline runs."
echo "\n"
cat pipelines_list | while read line 
do
  namespace=`echo $line | cut -d' ' -f1`
  pipeline_name=`echo $line | cut -d' ' -f2`
  echo "Triggering pipeline run for the pipeline: " $pipeline_name
  curl -X POST "https://"$namespace"."$OC_DOMAIN_URL"/-/pipelines/v1/"$pipeline_name"" -H "accept: */*" -H "Authorization: Bearer "$OC_TOKEN"" -d "" &
done
echo "\n"
echo "Triggering all pipelines completed"
echo "\n"
echo "End....................................................."