Change the default replication buffer size for pipeline models

The replication of pipeline models in the Batch Tracking application is dependent on the CMX Replication Server.

The size of the transfer message buffer (MAX_XFRQ_MESSAGE_SIZE) is defaulted to 12 MB, and expressed in Bytes as 12 * 1024 * 1024.

To change the default replication buffer size

The required buffer size (Bytes) can be estimated using the following procedure:

1. Collect the following information for the pipeline model:

   • Number of Batch Records

   • Number of Anomaly Records

   • Number of Segments

   • Number of Fragments per Segment

   • Number of Nodes

2. Collect the individual record size for the following types (in Bytes):

   • Individual Record Size for Batches

   • Individual Record Size for Anomalies

   • Individual Record Size for Fragments

   • Individual Record Size for ETAs

     Note: Individual record sizes may vary between systems. For each system, the individual record sizes are available in the %DataRoot%\RealTime\database\INTEL_NT\types.txt file.

3. Perform the following calculations:

   • Total Batch Size: Number of Batch Records * Individual Record Size for Batches

   • Total Anomaly Size: Number of Anomaly Records * Individual Record Size for Anomalies

   • Total Fragment Size: Number of Segments * Number of Fragments per Segment * Individual Record Size for Fragments

   • Total ETA Size: (Number of Batch Records + Number of Anomaly Records + (Number of Segments * Number of Fragments per Segment)) * Number of Nodes * Individual Record Size for ETAs

   • Buffer Size in Bytes: Total Batch Size + Total Anomaly Size + Total Fragment Size + Total ETA Size

4. Change MAX_XFRQ_MESSAGE_SIZE to the newly calculated value as follows:

   1. Navigate to the following location:

      %DataRoot%\RealTime\config\settings

   2. Open or create a copy of the RealTime_cmxrep.json file.

   3. Add a key-value entry to the file with the following format:

      "RealTime": {

      "MAX_XFRQ_MESSAGE_SIZE": <Buffer Size in Bytes>

      }

   4. Ensure any backup procedures are completed prior to restarting RealTime.

   5. Restart RealTime.

Example:

The following example shows how to calculate the buffer size for a sample pipeline model:

5. Collect the following information for the pipeline model:

   • Number of Batches: 10

   • Number of Anomalies: 20

   • Number of Segments: 100

   • Number of Fragments per Segment: 2

   • Number of Nodes: 120

6. Collect the individual records size for the following types (in Bytes):

   • Individual record size for Batches: 352

   • Individual record size for Anomalies: 264

   • Individual record size for Fragments: 320

   • Individual records size for ETAs: 376

7. Perform the following calculations:

   • Total Batch Size: 10*352 = 3520

   • Total Anomaly Size: 20*264 = 5280

   • Total Fragment Size: 100*2*230 = 64000

   • Total ETA Size: (10+20+(100*2))*120*376 = 10377600

   • Buffer Size in Bytes: 3520+5280+64000+10377600 = 10450400

8. Change MAX_XFRQ_MESSAGE_SIZE to the newly calculated value as follows:

   6. Navigate to the following location:

      %DataRoot%\RealTime\config\settings

   7. Open or create a copy of the RealTime_cmxrep.json file.

   8. Add an entry to the file with the following format:

      "RealTime": {

      "MAX_XFRQ_MESSAGE_SIZE": 10450400

      }

   9. Complete required backup procedures.

   10. Restart RealTime.

For more information on CMX replication configuration settings and how to modify the settings, refer to the AVEVA Enterprise SCADA RealTime Administrator Guide, Volume 1.