ibm-ecosystem-engineering / azure-arm-templates Goto Github PK

• Login to azure
• Login to cluster
• Delay ARM Template

Alternately, don't use a YAML file, just pipe into oc apply command.

The script contains a append to a file for the activemq.service. If the script is re-run, the original file is still present and the lines are duplicated when applied.

Items to add:

• Primary VNet, subnets and NSG
• DR VNet, subnets and NSG
• VNet peering
• Bastion host at primary site
• Bastion host at DR site
• Firewall at primary site
• Firewall at DR site
• VMs for Nodes - Primary, HA, DR and Standby
• Data disks for VMs
• Key Vault primary region
• Key Vault in DR region
• Server certificate created and stored in both key vaults
• Application Gateway at primary site
• Application Gateway at DR site
• API Management service at primary site
• API Management service at DR site
• Script to deploy and configure Safer Payments software
• DR site backup file storage CIFS (secure) mounted to node3
• Add multiple public IP addresses to firewalls (for multi-AZ)
• Add firewall application rules
• Add firewall outbound rules for server updates, etc
• Add firewall network rules
• User defined routing for outbound traffic via firewall (except bastion)
• Add insights to all resources
• Add tag option
• main template parameters to set small/medium/large sizes
• Create user interface definition

The following line,

PSQL_NAME=$(log-output ${PSQL_HOST} | sed 's/.postgres.database.azure.com//g')

Should be,

PSQL_NAME=$(echo ${PSQL_HOST} | sed 's/.postgres.database.azure.com//g')

Log message is a duplicate of the catalog source rather than relating to the operator group.

Marketplace listing should include https://ibm.biz/int-licenses in the tooltip for the license.

The azuredeploy.json template does not reference the API version on line 302

"principalId": "[reference(parameters('managedIdName')).principalId]"

This will fail if using an existing managed identity and should be updated to the following:

"principalId": "[reference(parameters('managedIdName'), '2018-11-30').principalId]"

Timeout is currently 90 minutes. Given that the script will wait for 90 minutes for the last step, the overall timeout should be greater.

Currently deploys as a pod which may cause problems if pod fails to start or is accidentally deleted.

Currently users need to enter these

Current script outputs no resources for namespace ... which is as it should work while waiting for subscription, but may cause problems with users.

For CP4I deployment, the operator installation sometimes fails with the operator and CSV not being associated. To fix this, it is necessary to delete the CSV from the CLI. Should add a step that deletes the CSV and retries the operator readiness check if the operator is not ready with a timeframe like 10 minutes.

Images in the OMEnvironment yaml embedded in the deploy-oms.sh script currently has hard coded images. These need to be variable to be able to create different options.