Worker Groups in OO10


[PDF]Worker Groups in OO10 - Rackcdn.comhttps://b6b45000d3362c7b69f8-0a7250d8bde19e2e09b93fa07c907bb0.ssl.cf5.rackc...

151 downloads 719 Views 597KB Size

HP OO 10 OnBoarding Kit – Community Assistance Team

Worker Groups in OO10 This tutorial demonstrates the usage of worker groups in HP Operations Orchestration 10. x. Version10.x introduces the new concepts Worker Groups and Group Aliases. Benefitting from these concepts, the flow author and the administrator can control where specific operations are run. This is important in larger installations which, potentially, have multiple RAS servers in several data centers and across different network zones. Overview The OO component that executes a flow step is called a worker. Both the Central and the RAS contains a worker and additionally the RAS contains a component to handle the communication with Central. An important part of achieving high availability and increase scalability is to group workers into logical collections: worker groups. Group Aliases let you separate between assigning an operation to a worker during authoring time and in the runtime environment. We will explain the worker group functionality by demonstrating an implementation of the following use case. The Infrastructure Manager requires a flow for restarting Windows servers. The specification includes the following:  The same flow should be used both for servers in the internal network and for other servers that reside in the DMZ.  Keep an audit trail of the runs on a specific server in the backend. The Central Admin requires that the content will be flexible and support changes in RAS deployment topology without the need for content changes and deployment of a new content pack each time workers are installed\removed or worker groups are rearranged. In the following sections we explain how to meet these requirements:  How the author can use the group aliases and RAS override features in order to dynamically route the operation to the desired worker group, and avoid using hardcoded group names.  How the administrator can assign worker groups to group aliases used by the operations in the flow.

HP Operations Orchestration HP Operations Orchestration (HP OO) is a next generation IT Process Automation solution that is designed from the ground up to increase automation adoption whether in a traditional data center or hybrid cloud environment. OO 10 Community Onboarding Kit This tutorial is part of the onboarding kit created for the OO community in order to make it even faster to learn OO 10. It is for both new OO 10 users and for existing OO 9 users who want to know what’s new. The kit contains four persona-based learning tracks for administrators, flow authors, integrators, and end users. Each track provides knowledge assets of the following types:       

Videos Tutorials Quick Guides Presentations Direct links to relevant manuals Sample Content Best Practices

Worker Groups in OO10

Studio Project In order to implement the described use case the author should perform the following activities using OO Studio: 1. Define group aliases 2. Author the flows and define steps to be run by these group aliases

Define the Group Aliases In the screenshot below we have created the two group aliases for the use case.

So we have group aliases for:  Workers in the DMZ (to connect to servers in this network segment)  Workers in the Backend Network (for the audit trail) The Central administrator will map these group aliases to worker groups – a task described later in this document.

The Main Flow Let’s assume that according to the IT naming policy, servers that reside in the DMZ network segment have the string “dmz” as part of their FQDN. Therefore, the first step of the main flow checks the server FQDN variable (defined in the flow inputs), in order to understand which worker group that should run the operation as shown below:

HP Operations Orchestration, February 2014

Worker Groups in OO10

This is done by utilizing the out-of-the-box String Occurrence Counter operation and as you can see, the Check Domain step has three responses: DMZ, INTERNAL, and failure. Let’s take a closer look at the operation:

We have configured the responses in the following way:  If the string “dmz” exists 0 times, the operation result will be 0 and the response will be INTERNAL.  If the string exists more than 0 times, the operation result will be greater than 0, and the response will be DMZ.  We left the failure response as is.

HP Operations Orchestration, February 2014

Worker Groups in OO10

Each of the above responses is wired to the next relevant step; Restart Internal Server and Restart DMZ Server. Both steps use the Restart Server subflow but the difference is that the Restart DMZ Server step has the rasGroup input populated with the DMZ_GROUP group alias name as shown here:

While in the Restart INTERNAL Server step, this input variable is populated with an empty constant value (equivalent to null). The final step in the main flow runs the Logger subflow, which is explained later on.

Restart the Server (Subflow) The first step (using the out-of-the-box Check Null operation) determines if the rasGroup variable is populated.

If it is not populated, the default worker group should be used. Thus we do not need to do anything special before running the Restart Server operation and this will run on the default worker group. If the rasGroup variable is populated, we want to change the group alias used in the next (Restart Windows) operation. We did this by using the out-of-the-box Set Flow Variable operation:

HP Operations Orchestration, February 2014

Worker Groups in OO10

As you can see, we populated the overrideNRAS variable from the rasGroup variable (which gets the DMZ group alias name at runtime). When we open the Restart Windows Server operation, we can see that overrideNRAS is used by default for the Override Group field.

Logger (Subflow) In the previous section, we described how to dynamically set the worker group alias according to the FQDN of the server that we want to restart. As described in the Overview section, one of the requirements in the use case is to log our runs to a separate server, which resides in a backend network segment.

HP Operations Orchestration, February 2014

Worker Groups in OO10

In this case, we do not need to determine the worker group dynamically. It is enough to do it directly in the relevant step. This is done by dragging the worker group alias onto the Group Alias icon of the operation (marked in red).

Our project is ready. The next steps are to configure Central and deploy the content pack.

Configure Central In order to deploy the new content, the Central Administrator performs the following steps: 1. Verify that the necessary worker groups exist and the correct workers (RAS servers) are assigned. 2. Deploy the content pack. 3. Map the group aliases to the relevant worker groups in the environment. Assign Workers to Worker Groups We have two worker groups: DMZ_RAS_GROUP and BackEnd_GROUP. 1. We assign the DMZ RAS servers to the DMZ group and the BackEnd RAS servers to the BackEnd group. 2. If a group does not exist, we create it using the same Assign to Group combo box.

3. Next, we deploy the content pack, and the group aliases are created.

HP Operations Orchestration, February 2014

Worker Groups in OO10

Map Worker Groups to Group Aliases 1. After the content pack is deployed, and the group aliases are created, we go to the Content workspace, and choose the Group Aliases tab under Configuration Items.

2. Next, we assign the worker groups to the relevant group aliases.

We note that for legacy purposes the default worker group is called “RAS_Operator_Path”.

Check the Central Server Environment Remember that you can assign workers to more than one worker group. This enables you to build the groups in order to allow the flow to continue, while making sure that the correct worker group is assigned, as explained in the next section.

Run the Flow Let’s observe which worker groups are used when we run the flow. Since the server does not belong to the DMZ, Restart INTERNAL Server is chosen. Thus, the group override is not used and the worker group that is picked up is the default one.

HP Operations Orchestration, February 2014

Worker Groups in OO10

As you can see, the Write to File operation uses the BackEnd_WRK_GROUP as required:

We are always interested in your questions and feedback, so please post them on the OO community forums: https://hpln.hp.com/node/21/og/forum/37

HP Operations Orchestration, February 2014