Configure TRA tra_config.xml Parameters

<tm-config>
...
   <tm-cluster>
      <nodes>
         <node id="1" mgmt-address="10.10.126.63" p2p-address="10.20.110.63"/>
         <node id="2" mgmt-address="10.10.126.64" p2p-address="10.20.110.64"/>
      </nodes>
   </tm-cluster>
...
</tm-config>

For details about Traffic Routing Agent nodes, see the discussion about Traffic Routing Agent requirements.

network-topology-configuration-file="./tra_config_network_topology.xml"

where pod_name is the name of the TRA pod. See the discussion about configuring a containerized Traffic Routing Agent for details about TRA pod names.

• true — To reuse an existing TcpConnector.
• false — (Default.) To create a new TcpConnector whenever one is needed, deleting any existing unused TcpConnectors.

• true — To use huge mtxbuf based pdu cache storage.
• false — (Default.) To use the parking based mechanism for pdu cache storage. When set to false, you can also configure the following entries:
  • pdu-cache-buffer-size — Sets the buffer size, in bytes, to use for parking buffers. A value of zero (0), the default, indicates that the buffer size should be calculated by the TRA as the maximum of the worst case diameter pdu or mdc pdu configured plus storage overhead.
    Note: If the value you specify is too small, it is increased to a size that could store a worst case diameter pdu or mdc pdu configured plus storage overhead.
  • pdu-cache-buffer-pool-size — Sets the number of buffers available for pdu parking. More than one pdu can be stored in a single buffer, depending on the pdu sizes.
  • pdu-cache-use-overflow — Set to true (default) or false. MATRIXX Support recommends that you keep the default setting (true), so that if all PDU cache buffers are in use, the overflow mechanism (dynamically allocated/freed memory) is available for storage requests if more storage is needed.

For example, the PDU entry in this parameter section configures the TRA to use the mtxbuf method for caching pdus.

    <parameters 
        node-poll-interval-clock-ticks="1000"
        pdu-cache-use-mtxbufs="true"
        failed-node-timeout-poll-intervals="100"
        network-topology-configuration-file="./tra_config_network_topology.xml"
        udp-connection-idle-timeout-sec="10"
        udp-connection-max-count="64"
        tcp-connection-max-count="8192"
        product-name="TRA"
        check-network-interface-interval-sec="4"
        />

The pdu entries in this parameter section configure the TRA pool to have individual buffer sizes of 5 million bytes long, to size the buffer pool at 1000 elements, and to allow overflow if all buffers are in use.

    <parameters 
        node-poll-interval-clock-ticks="1000"
        pdu-cache-buffer-size="5000000"
        pdu-cache-buffer-pool-size="1000"
        pdu-cache-use-overflow="true"
        failed-node-timeout-poll-intervals="100"
        network-topology-configuration-file="./tra_config_network_topology.xml"
        udp-connection-idle-timeout-sec="10"
        udp-connection-max-count="64"
        tcp-connection-max-count="2048"
        product-name="TRA"
        check-network-interface-interval-sec="4"
        />

For example:

<parameters
   ....
   default-mdc-response-dmid="3"
   ....
/>

The MDC response domain ID is used when the TRA processes the MDC response packet from the downstream node before transferring it to an upstream peer. If the MtxResponseSysDomain MDC:

• Defines the domain ID, the domain ID value is used by the TRA to include all nodes with the specified matrixx-domain when creating the response node ID map.
  Note: You set the value for matrixx-domain in the pools configuration in the tra_config_network_topology.xml file. For more information, see the discussion about TRA pool node reference.
• Does not define the domain ID, the default-mdc-response-dmid value is used to create the response node ID map.

For more information, see the discussion about self-signed certificate detection and handling in MATRIXX Security.

The max-missed-heartbeat-count parameter specifies the maximum number of missing heartbeat counts from a node. The default value is 20. The missed-heartbeat-above-water-count parameter specifies the number of missed heartbeats above what is specified in max-missed-heartbeat-count before a node in a node set is declared failed. The default value is 1.

The following tra_config.xml excerpt shows these parameters with their default values:

<tm-cluster
   ......
   max-missed-heartbeat-count="20"
   missed-heartbeat-above-water-count="1"
   ......  
        >

For example:

<parameters
   ....
lock-process-memory="off"
....
/>

TRA memory locking is enabled by default.

For more information, see the discussion about applying Traffic Routing Agent configuration changes.