The focus of System Automation is
to automate the availability of IT resources. This availability is
accomplished by starting and stopping resources automatically and
in the correct sequence.
Resources are located on a system, which is referred to as node in
the context of a cluster. Resources controlled by System Automation
can be applications, services, mounted disks, network addresses or
even data replication. Basically anything on a node which can be monitored,
started and stopped with help of commands or through an API. For each
resource Tivoli® System Automation
provides an availability state and offers a way to start and stop
them.
The main characteristics of all System Automation family
products are:
Automation
System Automation family
products help you to describe proper start up and shut down dependencies
between resources and resource groups. Once the resources are described
in an automation policy, the operator can bring up or shut down an
application in an reliable way.
One important functional aspect
of IBM® Tivoli System Automation family product is
the ability to automate and coordinate the start up and shut down
of resources. This is accomplished with help of relationships between
resources defined in an automation policy.
Availability and High availability
System Automation family
products offers availability and high availability (HA) functions
for its resources. System Automation products
constantly monitor resources and react on undesired or unexpected
availability state changes. In the case of a failure, resources are
restarted based on the information given by the policy.
The following
figure shows an example where the resource DB can be started
on three different nodes of the HA cluster.
Figure 1. Cluster with 3 nodes. The resource DB
can run on each node.
Availability: Requires to keep resources online or offline.
The desired state is defined in the policy.
High availability: Requires to define redundancy. If a resource
has an unexpected outage it might not be possible to restart it on
the same node with the same configuration. Therefore products in the System Automation product
family offer the capability to establish a cluster of nodes. In a
cluster, multiple servers are connected together and act as one big
virtual system. A cluster provides all services to the end user, that
are defined on the nodes. For example, if a node fails, the corresponding
applications can be moved to another node within the cluster.
The System Automation
product family consists of the following products:
System Automation for z⁄OS
System Automation for Multiplatforms
System Automation Application Manager
The following figure visualizes a typical setup of a heterogeneous
multi-tiered application. Figure 2. Sample setup of a heterogeneous multi-tiered application
The figure shows 3 different tiers represented by a cluster
running on different operating systems: Linux®, AIX®, z/OS® Sysplex.
A setup like this leads to pain points like
Managing complexity: As multiple systems and operating systems
are used, a broad skill set is required.
Monitoring: Many resources run on different systems in different
clusters. Administrators need to have a good overview of the current
status of their systems.
Automation: Resources have a well defined relationship to each
other. It is required to start and stop the resources in this order.
High availability: Application failures can have impact on related
systems. A constant service of the application needs to be assured.
