Frequent Advisor
Re: Controlled Starting and Stopping OO
[ Edited ]

Title: Starting and Stopping Operations Orchestration

Document ID:

KM1107935

Product: Operations Orchestration Version: 9.00,7.60,7.51,9.01

Summary: Details on Starting and Stopping Operations Orchestration

 

What is the HP best practice for starting and stopping HP Operations Orchestration Central, RAS, Scheduler, and Cluster?

 

Solution

While there currently is no formal white paper specific to the starting and stopping of the Operations Orchestration services, Support has determined a preferred start order for the services.

Prerequisites:

    1. Database: The database should be operating prior to starting OO Services.
    2. Shared Repository: If you are using a cluster, the Shared Repository must be available via SMB, NFS, or whatever solution is being used.

Start Order:

    1. Cluster: If this is a cluster node member, bring up the Clustering (shared repository) service.
    2. RAS: The RAS should be brought up when the repository is available.
    3. Central: The Central relies on RAS, so should be brought up next, within 5 minutes (There is a RAS-Central co-dependence).
    4. Scheduler: With RAS & Central up, Scheduler could now potentially launch scheduled flows, so it should be brought up next.
    5. Load Balancer: If you have multiple nodes in the cluster, when enough are up to warrant load balancing, bring up the Load Balancer.

The specific implementation is left to the customer, but many customers have used rc files on Unix systems created using symlinks to the provided .sh scripts for the HP Operations Orchestration services with success.