| Concepts |
Installation |
|
This section covers all of the basic principles of HA technology, and specifically the concepts that VCS uses. This includes the VCS HA framework, how LLT and GAB contribute to the framework through communication, membership management and atomic broadcast. The Management of resources by Agents, under the control of the VCS HA daemon, had. This section covers the design and build principles of HA systems, so that the student will understand the requirements of VCS for correct installation.
|
This section prepares the student for the installation process, identifying the mechanism and procedure. All necessary information required for the installation is discussed and presented.
|
| Preparing Services for High Availability |
Online Configuration |
|
This section takes the student through the process of creating and testing the application on all cluster nodes before integration of the application into the cluster. The student creates VxVM disk groups and volumes, file systems, mount points, performs IP and NIC configuration and testing to ensure all is well before VCS is involved.
|
The application is brought under cluster control by through online configuration. This section presents the CLI approach to online configuration. The whole process is presented through a worked example. The workshop also includes the online configuration through the Java Administration console as an alternative approach.
|
| Starting and Stopping the Cluster |
Offline Configuration |
|
In preparation for performing offline configuration, the course presents the starting and stopping of the cluster to allow the distribution of manually constructed configurations. How to stop the cluster leaving all applications running and how to start the cluster daemons stale are included in this section.
|
This section shows how the student can introduce a hand produced configuration either from template or cut/paste text editing and subsequent introduction into the cluster and distribution to all nodes as they join the cluster.
|
| Sharing Network Resources |
Event Notification |
|
This section teaches the student how to build more scalable service group configurations when OS resources are shared by applications. So when multiple applications have a dependency upon a NIC, for example, by using a Parallel service group with NIC resources and Phantoms, and the Application group uses a Proxy, the solution can be much more scalable and efficient.
|
This section teaches the student how to configure notification of cluster events. This is through several different components: The NotifierMngr resource, Resource owner and Group owner attributes, and through triggers.
|
| Cluster Response to Resource Faults |
Cluster Communication |
|
This section shows the students how VCS responds to resource faults, and how to moify the cluster's tolerance, and monitoring features to tune them to application needs.
|
This section reviews how the had, GAB and LLT cooperate to provide the cluster communication channels and cluster membership management.
|
| Response to System and Communication Failures |
I/O Fencing: |
|
This section demonstrates VCS's response to system and communication failures. It describes the state of Jeopardy and how service groups will become AutoDisabled.
|
This section describes the mechanism and theory behind IO Fencing. The lab involves the reconfiguration of the cluster to operate in an IO Fenced environment.
|
| Example Application: NFS |
Example Application: Samba |
|
Sets up a realistic HA NFS environment.
|
Sets up a realistic HA samba service.
|
| Reconfiguring Cluster Membership |
Service group Interactions |
|
Prepares the student for adding and removing nodes to/from a cluster.
|
This section describes and demonstrates group dependencies that can be explicitly and implicitly defined in the cluster through deployment of Service Group links and also through special resources such as FileOnOff and ElifNone resources.
|
| Network Resources |
|
|
This section introduces the MultiNIC resources MultiNICA and MultiNICB, and there corresponding IPMultiNIC and IPMultiNICB resources for performing interface failover before service group failover. This section compares and contrasts the two approaches, and builds a solution in the lab exercises using MultiNICB.
|
|
