A job is considered roaming as soon as the job has been ordered to scalability server A, and the software delivery (SD) agent registers through scalability server B.
The SD agent roams from scalability server A to scalability server B. The SD agent is automatically registered with scalability server A. A number of jobs have been set up to the agent and are staged in the file database on scalability server A. Then the agent is registered through scalability server B.
When scalability server B reports its new agent to its domain manager, the domain manager triggers the scalability server A. Scalability server A reports the job results for the agent, and the domain manager realizes that this agent has roamed to scalability server B. All jobs that have not yet executed are reset and scalability server B is triggered. Scalability server B picks up the reset jobs.
Additionally, to guarantee that a batch is not broken, the name of the manager that received the jobs, is internally handled by software delivery functions. In the previously shown scenario, the new SD manager will not pick up any job, unless the results have been reported from the old manager, or the jobs have been removed by the administrator, or timed out. During this time new jobs will not be allowed to be setup to the roaming agent.
|
Copyright © 2013 CA.
All rights reserved.
|
|