IBM InfoSphere Master Data Management, Version 11.3The mpi_clustercfg table stores clustering information used to provide high availability support to the operational server.
| Attribute | Description |
|---|---|
| noderecno | Clustered node record number |
| caudrecno | Creation of this particular record, from mpi_audhead |
| maudrecno | Last time the record value was modified, from mpi_audhead |
| recstat | Record status; active, inactive, or deleted |
| nodename | Name of clustered node assigned at configuration |
| nodetype | Enumeration specifying the node type. Currently, the only valid type is FULL_TEXT_INDEX. This type is specified in madison.mpi.NodeType in the software development kit. |
| noderole | Flag indicating the role or purpose of this node. Valid options are 0 and NODE_ROLE_PRODUCER_ONLY. A value of 0 means default behavior, which is also the default assignment if nothing else is provided. For Flexible Search, the default means producer and consumer mode. In this role, updates to the operational server place a record on mpi_idxique through the synchronization service handler (producer). The index synchronization service (consumer) then takes the record off mpi_idxique to index. For the NODE_ROLE_PRODUCER_ONLY role, no consumer is ever started. The system only adds records to mpi_idxique for other clusters to consume. |
CREATE TABLE mpi_clustercfg
(
noderecno int NOT NULL,
caudrecno bigint NOT NULL,
maudrecno bigint NOT NULL,
recstat nchar(1) NOT NULL,
nodename nvarchar(255) NOT NULL,
nodetype int NOT NULL,
noderole bigint NOT NULL
)
CREATE UNIQUE INDEX mpi_clustercfg0 ON mpi_clustercfg (noderecno)
CREATE UNIQUE INDEX mpi_clustercfg1 ON mpi_clustercfg (nodename)