1. Configuring Elasticity in Dynamic Clusters for Oracle WebLogic Server
  2. On-Demand Scaling

3 On-Demand Scaling

Oracle WebLogic Server supports on-demand scaling, which is the ability to scale a dynamic cluster up or down by manually adding or removing running dynamic server instances as needed. On-demand scaling is the simplest and most basic form of incorporating elasticity into a WebLogic domain.

What is On-Demand Scaling?

On-demand scaling allows you to manually add or remove running dynamic server instances from an active dynamic cluster as needed.For example, if the average user-request backlog in dynamic cluster members is trending up, indicating a need for higher processing capacity, you can add running dynamic server instances to the dynamic cluster. When the backlog for user-requests drops substantially, you can shut down idle dynamic server instances.

When you expand the cluster, the size must not exceed the limit set in Maximum Dynamic Cluster Size. WebLogic Server adds only running dynamic server instances up to the Maximum Dynamic Cluster Size limit. Also, when you shrink the cluster, the number of running dynamic servers cannot be below the limit set in Minimum Dynamic Cluster Size.

To perform on-demand scaling, you must have a dynamic cluster configured. See Dynamic Clusters in Administering Clusters for Oracle WebLogic Server.

You can perform on-demand scaling using the Fusion Middleware Control or WLST, as described in the following sections:

Note:

Before performing on-demand scaling of a dynamic cluster, make sure that the scaling operation will not have adverse effects on the rest of the system. For example, before a scale up, it may be necessary to adjust JDBC connection pool sizes and Work Manager parameters to avoid exceeding database capacity. Before a scale down, make sure that any inflight work has a chance to finish before the dynamic server instances are shut down. Also, if you are using JMS or JTA, do not use a scale down approach that lowers the dynamic cluster size configuration setting. See Best Practices for Using Clustered JMS Services in Administering JMS Resources for Oracle WebLogic Server.

Parent topic: On-Demand Scaling

On-Demand Scaling Using Fusion Middleware Control

You can perform on-demand scaling using Fusion Middleware Control.

  1. From the WebLogic Domain dropdown menu, select Environment, then select Clusters.
  2. In the Clusters table, select the row of the dynamic cluster you want to scale up or scale down.
  3. Click the Scale Up/Down button.
  4. On the Scale Up/Down page, you can review current dynamic cluster settings. In the Desired Server Count field, use the arrows to set the number of dynamic server instances you want running in this dynamic cluster. Click Scale Up/Down to activate this setting. A confirmation message is displayed to notify you that your scale up or scale down operation was successful.

For more information about using Fusion Middleware Control to manage WebLogic Server, see Administration in Administering Oracle WebLogic Server with Fusion Middleware Control.

Parent topic: On-Demand Scaling

On-Demand Scaling Using WLST

You can manually scale a dynamic cluster up or down using the WLST commands.

To manually scale your dynamic cluster using WLST:

  • Use the scaleUp command to increase the number of running dynamic servers in your dynamic cluster. When you use the scaleUp command and enable the updateConfiguration argument, WLST will increase the size of the dynamic cluster by the specified number of dynamic servers and start these server instances.

  • Use the scaleDown command to decrease the number of running dynamic servers in your dynamic cluster. When you use the scaleDown command and enable the updateConfiguration argument, WLST will gracefully shut down the specified number of running dynamic servers and remove them from the dynamic cluster.

    Note:

    The scaleDown command lowers the cluster dynamic cluster size setting when its updateConfiguration option is enabled, which in turn can abandon JMS persistent messages and JTA transactions that are associated with retired servers. If you are using JMS or JTA, do not use scaleDown with its updateConfiguration option enabled. See Best Practices for Using Cluster Targeted JMS Services in Administering JMS Resources for Oracle WebLogic Server.

See scaleUp and scaleDown commands in WLST Command Reference for Oracle WebLogic Server.

See Expanding or Reducing Dynamic Clusters in Administering Clusters for Oracle WebLogic Server.

Example 3-1 demonstrates how to use WLST to scale a dynamic cluster named DynamicCluster up by two running dynamic server instances and then scale it back down by one running dynamic server instance.

Example 3-1 On-Demand Scaling Using WLST

wls:/offline> connect('weblogic','weblogic','t3://localhost:7001')
Connecting to weblogic server instance running at t3://localhost:7001 as username weblogic...

Successfully connected to Admin Server 'AdminServer' that belongs to domain 'mydomain'.

wls:/mydomain/serverConfig/> scaleUp("DynamicCluster", 2, true ,true)
Remote ScaleUp started successfully after 38 seconds. Waiting for 2 servers to reach the running state.  The timeout is 600 seconds.

All servers are now running.
wls:/mydomain/serverConfig/> scaleDown("DynamicCluster", 1, true ,true)
Remote ScaleDown started successfully after 0 seconds.
The servers were stopped successfully.

Parent topic: On-Demand Scaling