A dynamic cluster is a server cluster that uses weights to balance
the workloads of its cluster members dynamically, based on performance information
that is collected from the cluster members.
Before you begin
Complete the following prerequisites before creating a dynamic cluster:
- Create a node group. See Creating or editing a node group
for more information.
- You must have configurator administrative privileges to create a dynamic
cluster.
- Verify that the application placement controller is enabled. The application
placement controller is enabled by default. The application placement controller
enables the autonomic capabilities of dynamic clusters. To enable the application
placement controller, click Operational policies > Autonomic managers >
Application placement controller. For more information about the properties
that you can change on the application placement controller, see Monitoring and tuning the application placement controller
- Decide if you are going to use vertical stacking. Vertical stacking
can improve bottleneck conditions by enabling the placement controller to
start more than one instance of the dynamic cluster on a node.
About this task
A dynamic cluster is an application deployment target that
can expand and contract depending on the workload in your environment. Dynamic
clusters work with autonomic managers, including the application placement
controller and the dynamic workload manager to maximize the use of your computing
resources. Dynamic clusters are required for many of the WebSphere® Extended
Deployment autonomic functions, including high availability and service policies.
Procedure
- Create the dynamic cluster. In the administrative console,
click Servers > Dynamic clusters > New.
- Enter dynamic cluster information.
In this step you give a name to the dynamic cluster, specify the node
group to which the dynamic cluster is mapped, and define other dynamic cluster
properties.
- Select a dynamic cluster template. You can either build
servers that are based on a default template, or you can use an existing server
from your environment as the template. The
displayed templates are valid for the platform that corresponds to the node
group that you selected for your dynamic cluster. The template also corresponds
to the version of your deployment manager.
- Specify dynamic cluster-specific properties.
- Define the minimum number of cluster instances. The default
minimum number of instances is one instance and the maximum default is no
limit on instances.
- Define the maximum number of cluster instances. The default
value has no limit on the number of cluster instances.
- Determine whether to enable vertical stacking. When
you configure vertical stacking, more than one dynamic cluster instance can
start on the same node.
- Confirm the dynamic cluster creation. Click Finish > Save to
save the changes to the master configuration.
- Select the mode of operation. In the administrative console, click Servers
> Dynamic clusters. Select the dynamic clusters that you want to modify.
Choose the operational mode and click Set mode.
Important: To
use dynamic application placement, click Automatic or Supervised as
the mode of operation.
If you have
nodes that run on z/OS® systems, use dynamic clusters in supervised or
automatic mode. If you want to prevent WebSphere Extended Deployment from
automatically starting a cluster member on a logical partition (LPAR) that
is hosting a dynamic cluster when the LPAR processor is too busy, then you
must define the cpuUtilizationThreshold custom property. In this scenario, WebSphere Extended
Deployment starts the cluster member only if transaction demand requires additional
cluster members and the processor utilization on the LPAR is less than the
cpuUtilizationThreshold value. See cpuUtilizationThreshold custom property for more information.
Results
The application placement controller starts
a cluster instance on each node for the dynamic cluster. Cluster instances
starts one server on each node concurrently across all of the nodes in the
node group for your cluster. When all of the instances for the dynamic cluster
are started, the cluster instances for another dynamic cluster in your environment
can begin to start.
When the dynamic clusters start,
at least one instance of each dynamic cluster in your environment becomes
available as soon as possible. Multiple instances on the same node can start
concurrently if you have multiple processors on the same node. For example,
if you have two processors on a node, two instances can start concurrently.
The application placement controller continues to start instances evenly across
the nodes for all the dynamic clusters until the minimum number of instances
for each dynamic cluster is reached.
Example
Use the following placement scenarios as a guideline for your
dynamic cluster settings:
- To specify that exactly two servers are started when the dynamic cluster
is running: Click Keep multiple instances started at all times. Set
the Number of instances value to 2, and set the maximum number
of instances. Click Limit the number of instances that can start and
set the Number of instances value to 2.
- To limit the number of started servers to five, and to stop the servers
when no activity occurs, click Stop all instances during periods of inactivity.
Set the maximum number of instances by clicking Limit the number of instances
that can start and set the Number of instances value to 5.
- To keep at least one instance active at all times, and to support an unlimited
number of instances to start, click Keep one instance started at all times and Do
not limit the number of instances that can start.
What to do next
Remember: To change your server ports after the server has
been created, you must change the ports on each server instance. You cannot
change the ports through the dynamic cluster template. Because multiple servers
to be on the same node by configuring vertical stacking, the ports must be
unique for each server instance.