The Oracle WebLogic Server Kubernetes Operator (the “operator”) is designed to fulfill a similar role to that which a human operator would fill in a traditional data center deployment. It contains a set of useful built-in knowledge about how to perform various life cycle operations on a domain correctly.
Human operators are typically responsible for starting and stopping environments, initiating backups, performing scaling operations, performing manual tasks associated with disaster recovery and high availability needs and coordinating actions with other operators in other data centers. It is envisaged that the operator will have similar responsibilities in a Kubernetes environment.
It is important to note the distinction between an operator and an administrator. A WebLogic Server administrator typically has different responsibilities centered around managing the detailed configuration of the WebLogic domains. The operator has only limited interest in the domain configuration, with its main concern being the high-level topology of the domain; for example, how many clusters and servers, and information about network access points, such as channels.
Human operators may manage more than one domain, and the operator is also designed to be able to manage more than one domain. Like its human counterpart, the operator will only take actions against domains that it is told to manage, and will ignore any other domains that may be present in the same environment.
Like a human operator, the operator is designed to be event-based. It waits for a significant event to occur, or for a scheduled time to perform some action, and then takes the appropriate action. Examples of significant events include being made aware of a new domain that needs to be managed, receiving a request to scale up a WebLogic cluster, or receiving a request to perform a backup of a domain.
The operator is designed with security in mind from the outset. Some examples of the specific security practices we follow are:
The operator is designed to avoid imposing any arbitrary restriction on how WebLogic Server may be configured or used in Kubernetes. Where there are restrictions, these are based on the availability of some specific feature in Kubernetes; for example, multicast support.
The operator learns of WebLogic domains through instances of a domain Kubernetes resource. When the operator is installed, it creates a Kubernetes Custom Resource Definition. This custom resource definition defines the domain resource type. After this type is defined, you can manage domain resources using
kubectl just like any other resource type. For instance,
kubectl get domain or
kubectl edit domain domain1.
Schema for domain resources is here.
The schema for the domain resource is designed to be as sparse as possible. It includes the connection details for the Administration Server, but all of the other content is operational details about which servers should be started, environment variables, and details about what should be exposed outside the Kubernetes cluster. This way, the WebLogic domain’s configuration remains the normative configuration.