-
when first deploying a hub cluster no apps should be included. Comment out the apps repository in
0-bootstrap/single-cluster/kustomize.yaml
-
There are several necessary components for EVERY RHACM install in 1-infra and 2-services that MUST be deployed. They are as follows: 1- infra:
- argocd/consolenotification.yaml
- argocd/namespace-sealed-secrets.yaml
- argocd/namespace-openshift-acm.yaml
- argocd/namespace-openshift-acm-observability.yaml
- argocd/namespace-openshift-rhacm-credentials.yaml
- argocd/namespace-policies.yaml
- argocd/namespace-sso.yaml
- argocd/namespace-sso-integration.yaml
-
Other items may be deployed as needed (according to comments and READMEs) but you should comment out the things you do not need 2-services:
- argocd/operators/cert-manager.yaml
- argocd/instances/cert-manager.yaml
- argocd/instances/sealed-secrets.yaml
- argocd/operators/external-secrets.yaml
- argocd/operators/openshift-pipelines.yaml
- argocd/operators/openshift-acm.yaml
- argocd/instances/openshift-acm-instance.yaml
- argocd/instances/openshift-acm-observability-instance.yaml
- The script to modify infra currently only works on MacOS
- Creating RoleBindings and ArgoCD instances may show a 'no resource found' error message, but ultimately the command does work - you can check inside the cluster or simply run the command again to verify.
- Committing tokens to GitHub will cause an issue (token found in commit has been revoked) - be sure that you remove any tokens, or add them to a GitIgnore before committing
- Git set and unset scripts will not work for any forks other than direct forks of the upstream repo
- This issue is currently being resolved but ultimately this step must be performed manually
- RHACM health check is not correct for ACM
- Azure healthchecks show sync ok even while the provision job is running. Might be better to have this app show as syncing until the job is finished
- AWS Healthcheck shows healthy and synced before the cluster is provisioned. This causes sync wave issues that need to be accounted for in following jobs
- We need to document EXACTLY what (if any) values Helm/Kustomize expects for each app
- the ClusterSet is created and clusters are added, but they cannot be moved between ClusterSets (as the label is configured at time of provisioning)
- The clusterset may not show up in RHACM until the ClusterSet object is created - we need to create the clusterset in RHACM
-
create.sh script (make_template.sh script) now correctly adjusts the values.yaml file for aws
-
Submariner now waits for a cluster as well as a submariner deployment prior to applying a NAT patch
-
RHSSO Sync waves adjusted
-
We need to be careful of elastic IP limits in AWS and region constraints in Azure. Might be best to start with a decently clean slate before we film demos
-
Infra script not tabbing values cirrectly
ComparisonError
rpc error: code = Unknown desc = `kustomize build /tmp/https___github.com_otp-demo-test_otp-gitops/0-bootstrap/single-cluster/1-infra --enable-alpha-plugins` failed exit status 1: Error: error converting YAML to JSON: yaml: line 34: did not find expected '-' indicator
- resolved by tabbing infra deployments correctly