Configure core group memory utilization

Use the administrative console to control the maximum amount of heap memory available for the underlying core group transport to allocate.

We should understand the following information:

• Other factors, such as the number of network interface cards on a machine, how the Network interface card is used, and network speed, can affect the memory that the high availability manager requires to efficiently handle network messages

• The amount of memory that the high availability manager controls while sending network messages varies. Typically, the high availability manager only retains control of this memory for very short lengths of time.

• The default setting of 100 MB is typically sufficient for most network deployment topologies.

• The maximum value for this setting is 250 MB.

• In core groups that contain both v6.x and v7 processes, all memory settings must be equal. The default memory setting for v6.x processes is 10 MB. Therefore we must adjust either the v6.x or the v7 setting to make both settings equal.

The high availability manager consumes its allocated heap memory when it sends interprocess messages on behalf of other services that use the high availability manager functionality. For example, the heap memory might be consumed while sending memory-to-memory replication data, or highly available routing data from one core group member to another core group member.

The value specified for the Transport memory size property acts as a flow control mechanism that might affect the speed at which data is replicated or the speed at which routing data is made available to other core group members.

Therefore, we might want to perform this task if we are seeing large numbers of any of the following Distribution and Consistency Services (DCS) congestion messages in SystemOut.log:

DCSV1051W, a high severity congestion event for outgoing messages
DCSV1052W, a medium severity congestion event for outgoing messages
DCSV1054W, a medium severity congestion event for incoming messages

Under extreme workloads, these messages might still occur on a properly tuned system.

Mixed-version environment: If we are running in a mixed cell environment, we must continue to use the following two memory configuration attributes to configuring core group memory utilization for the v6.x cells:

• The Transport buffer size property, which is set for each individual core group member. To specify a value for this property, in the administrative console, navigate one of the following paths:

  • If the core group member is an application server, click...
    Servers > Server Types > WebSphere application servers > server > Core group service

  • If the core group member is a node agent, click...
    System Administration > Node agents > node_agent_name > Core group service

  • If the core group member is a deployment manager, click...
    System Administration > Deployment manager > Core group service

• The IBM_CS_DATASTACK_MEG custom property.

  To specify a value for this custom property, in the administrative console, click...

  Servers > Core Groups > Core group settings > core_group_name > Custom properties

mixv

Change the amount of memory available for in-flight messages and network communication buffers

1. In the administrative console, click...
   Servers > Core Groups > Core group settings > core_group_name
   

2. Change the setting for the Transport memory size property.

   This property, specifies in megabytes, the maximum amount of heap memory that can be allocated to the high availability manager. The maximum value that can be specified for this setting is 250.

   For transitioning users: This property includes the functionality of the Transport buffer size property that exists for v6.x cells.trns

3. Click OK.

4. Click Review to review our changes.

5. Select Synchronize changes with Nodes, and then click Save to save the changes.

6. Restart all members of the core group.

After the servers restart, they all run with the new memory settings.

IBM recommends using the High Performance Extensible Logging (HPEL) log and trace infrastructure . We view HPEL log and trace information using the logViewer .

Related:

Core groups (high availability domains)