-
Feature Request
-
Resolution: Unresolved
-
Undefined
-
None
-
2.3
-
False
-
-
False
Feature Overview
Managing Capacity with Instances (scaling your mesh on AAP on OCP) guide: http://nightlies.testing.ansible.com/docs_7d5d0dae81129579/automation-controller//4.3.0/html/administration/instances.html
The current process to scale mesh on OpenShift deployments of AAP requires that
- A secret be created to use the default execution env on remote exec nodes
- Modifying the ee_pull_credentials_secret parameter in the Operator
- Once you've added the instance, download an "Install Bundle" and use a separate machine that can access the remote execution node you are attempting to add to run an ansible-playbook that can then install receptor on that new remote execution node.
This whole process feels very cumbersome for the user to accomplish.
Proposal
Have the AAP installation via Operator include:
- A default credential that can access the OCP cluster from where AAP was installed
- A default job template (similar to what we do with Demo Job Template which already exists with the AAP install) that can:
- create the the ee-pull-secret on the OCP cluster for you
- modify the ee_pull_credentials_secret parameter in the Operator for you
- includes the contents of the Install Bundle playbook within this job template
NOTE: I have these tasks already written that can be leveraged for creating and modifying the ee_pull_credentials_secret_parameter in the operator.
By doing the above you can have automation controller do all the steps within controller itself and not need a secondary machine that can access the OCP cluster nor have access to the remote exec node you are trying to add.
The only intervention from the end user would be to associate to the job template to the OCP inventory and the remote exec node they wish to add.