Page History
Table of Contents |
---|
Introduction
- The
...
Cluster Roles
Roles include the Active Controller, the Standby Controller and the Witness.
Graphviz |
---|
digraph structs {
# compound=true;
rankdir=LR;
node[shape="record"];
ActiveWitness [label="<f0>Primary Witness\n Active ",shape="rectangle",style="filled",fillcolor="limegreen",fontname="Arial",fontsize="12pt"]
node[shape="record"];
ActiveController [label="<f1>Primary Controller\n Active ",shape="rectangle",style="filled",fillcolor="lightskyblue",fontname="Arial",fontsize="12pt"]
node[shape="record"];
StandbyController[label="<f2>SecondaryController\n Standby ",shape="rectangle",style="filled",fillcolor="lightgrey",fontname="Arial",fontsize="12pt"]
subgraph normal_operation {
fontname="Arial";
fontsize="12pt";
ActiveWitness:f0 -> ActiveController [label=" hold connection ",fontname="Arial",fontsize="10pt"]
ActiveWitness:f0 -> StandbyController [label=" hold connection ",fontname="Arial",fontsize="10pt"]
ActiveController:f1 -> StandbyController [label=" hold connection ",fontname="Arial",fontsize="10pt"]
StandbyController:f2 -> ActiveController [label=" hold connection ",fontname="Arial",fontsize="10pt"]
}
} |
xx
...
- cluster architecture describes:
- cluster operation
- node distribution
- site distribution
Anchor | ||||
---|---|---|---|---|
|
- Download: JS7_JobScheduler_Cluster_Architecture.pptx
- Download: JS7_JobScheduler_Cluster_Architecture.pdf
- Use the links above if the document below does not display in your browser.
|
High Level Architecture
JS7 includes the following products:
- The JOC Cockpit, which is used to manage the JS7 - Inventory of scheduling objects, to monitor and control workflow execution and to access the JS7 - History.
- The Controller, which is used to orchestrate Agents, to deploy workflows and to pass orders for workflow execution to Agents.
- Agents, which are dedicated to a Controller and are used to execute workflows and jobs. Agents can act autonomously for workflows that execute all included jobs with the same Agent.
JS7 products can be clustered:
- JOC Cockpit and Controller can each implement a passive cluster with an active node and a standby node.
- Agents can be operated standalone or as instances of a JS7 - Agent Cluster which implements active-passive or active-active scheduling.
Controller Cluster Operation
A Controller Cluster offers restart capabilities for fail-over and switch-over:
Controller Cluster of Two Nodes
Limited high availability is provided from a cluster using two nodes:
Controller Cluster of Three Nodes
A number of disaster scenarios are covered when using a cluster of three nodes:
Controller Cluster of Four Nodes
High availability for most disaster scenarios when loosing more than one node is covered by a cluster using four nodes:
Director Agent Cluster Operation
The Director Agent Cluster works similarly to a Controller Cluster:
Site Redundancy
Operating nodes for JS7 products from separate sites is a means to increase redundancy.
Redundancy of Two Active/DR Sites
Use of an Active Site and a DR Site allows the distribution of nodes for switch-over between sites:
Redundancy of Three Active Sites
Use of three Active Sites offers improved redundancy for fail-over and switch-over: