Global FCM panel
In DB2 for LUW (Linux, Unix, and Windows) environments, when intra-partition parallelism is active, the fast communication manager (FCM) provides for communication between agents working on the same request. The configurable database manager parameters for the FCM are as follows:
-
fcm_num_anchors —This number increases when a query grows more complex.
-
fcm_num_buffers —This number increases to match any increase in fcm_num_rqb .
-
fcm_num_connect —This number increases when the number of partitions increases
-
fcm_num_rqb —This number increases when the number of partitions increases or when a query grows more complex.
FCM is started when you start the instance. The Global FCM panel on the instance home page provides the following information:
-
Buffers moved—The total of buffers sent and received by the named source and target partitions. In the FCM Analysis drilldown, this value is shown on the Internode Summary tab in the Buffers Total column.
-
Congestion—A Yes/No switch indicating data traffic for this connection is congested. In the FCM Analysis drilldown, this value is shown on the Internode Summary tab in the Connection Status column.
-
Buffer shortage—A Yes/No switch indicating that the number of available buffers is getting low. In the FCM Analysis drilldown, this value is shown on the Resources tab in the Buffers LWM (low watermark) column.
-
CE shortage—A Yes/No switch indicating that the number of available connection entries is getting low. In the FCM Analysis drilldown, this value is shown on the Resources tab in the Buffers LWM (low watermark) column.
-
MA shortage—A Yes/No switch indicating that the number of available message anchors is getting low. In the FCM Analysis drilldown, this value is shown on the Resources tab in the Msg Anchor LWM (low watermark) column.
-
RB shortage—A Yes/No switch indicating that the number of available request blocks is getting low. In the FCM Analysis drilldown, this value is shown on the Resources tab in the RQ Block LWM (low watermark) column.
The following illustration is an example of the Global FCM panel:
Intra-partition parallelism
Intra-partition parallelism is the name for the dividing up of a query execution into several parts such as index creation, database loading, or SQL queries--operations usually considered as parts of a single task. When intra-partition parallelism is turned on, FCM helps these tasks run in parallel within a single database partition.
Note: In single-partition environments, the INTRA_PARALLEL database manager configuration parameter must be set to Yes before FCM can be used.