JS7 JobScheduler

Space shortcuts

Page tree

Versions Compared

Old Version 11

changes.mady.by.user Andreas Püschel

Saved on

compared with

New Version Current

changes.mady.by.user Alan Amos

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

Users have a number of options to organize for organizing their scheduling clusters.

  • Agent Cluster: The operational layer of a Director Agent and a number of Subagents acting as an Agent Cluster.
  • Subagent Clusters: The functional layer that groups a number of Subagents to different Subagent Clusters

This includes to use of a smaller number of operational Agent Clusters with a higher number of Subagent Clusters and vice versa.

For information on how to set up Agent Clusters and Subagent Clusters see see JS7 - Management of Agent Clusters

...

One Cluster per Subagent

When adding a Subagent, the JOC Cockpit can be instructed to add a Subagent Cluster that which holds the single Subagent.

  • This allows to assign a job to assigned to an individual Subagent should if the job is to be executed only with this Subagent only.
  • The name of the Subagent Cluster corresponds to the Subagent ID.

...

  • fixed-priority: execute jobs with the first Subagent and switch to the next Subagent only if the first Subagent becomes unavailable (active-passive cluster).
  • round-robin: execute each the next job on the next next Subagent (active-active cluster).

...

Director Agent as Subagent Cluster Member

Users can add the a Director Agent to their Subagent Clusters:

  • The Primary and Secondary Director Agents each include a Subagent that can be used to execute jobs.
  • The Director Subagent can be added to any Subagent Clusters.
  • Users can decide not to use the Director Agent for job execution but for orchestration of Subagents only in order to . This will guarantee the highest level of robustness.

...

  • A single or a few Agent Clusters can be used that with each holds holding a larger number of Subagent Clusters.
    • Each Subagent Cluster selects a number of Subagents.
    • The same Subagents can occur in any number of Subagent Clusters.
    • The same Subagents cannot occur in more than one Agent Cluster.
    • This scenario can be applied for load sharing between server nodes that act for a number of clustered applications.
  • A larger number of Agent Clusters can be used that hold a single or a small number of Subagent Clusters.
    • Agent Clusters are autonomous and are not related to each other. 
    • Each Agent Cluster is assigned an individual Director Agent that is optionally clustered.
    • This scenario is preferably applied in a situation when communication between a Director Agent and Subagents should be limited to a few servers, for example for use of , when Agent Clusters are used in separate network zones.

...