AWS Fargate is a serverless compute engine for containers that work with both Amazon Elastic Container Service (ECS) and Amazon Elastic Kubernetes Service (EKS). For a highly-available CAS deployment, running the Apereo CAS backed with the Hazelcast Ticket Registry can be a great option, especially when run as a Docker container in an Amazon ECS environment to take advantage of ECS member discovery. In this blog post, we will take a quick look at the minimum configuration required to make this deployment scenario possible.
Our starting position is as follows:
Once you have configured the CAS WAR Overlay with the Hazelcast Ticket Registry module, we can begin by configuring Hazelcast to allow for member discovery options. Whether running in Amazon EC2 or ECS, Hazelcast support in CAS can handle auto-discovery automatically via a dedicated extension.
At a minimum, our CAS configuration for an Amazon ECS would look as follows:
cas.ticket.registry.hazelcast.cluster.discovery.enabled=true cas.ticket.registry.hazelcast.cluster.discovery.aws.security-group-name=service:cas-service-development cas.ticket.registry.hazelcast.cluster.multicast-enabled=false cas.ticket.registry.hazelcast.cluster.discovery.aws.port=-5701
Be sure to adjust the member address range to assist with auto-discovery.
Note that we do not need to specify an
iam-role setting; This setting only affects deployments on Amazon EC2. If you are deploying CAS in an Amazon ECS environment, the role is automatically fetched from the task definition that is assigned to run CAS. You need to make sure the role assigned to the task definition is configured with the correct and sufficient set of permission to let the deployment proceed.
If you prefer, you can also use an access/secret key combination without relying on role definitions:
To troubleshoot, you can always adjust the logs to investigate errors:
<Logger name="com.hazelcast" level="trace" additivity="false"> <AppenderRef ref="console"/> <AppenderRef ref="file"/> </Logger>
I hope this review was of some help to you and I am sure that both this post as well as the functionality it attempts to explain can be improved in any number of ways. Please feel free to engage and contribute as best as you can.