Checkpointing of the TRS database and the directory database

When a network node is added to an APPN network, it must complete a full topology exchange with the other nodes in the network. In large APPN networks, full topology exchanges can involve the transfer of large amounts of data, affecting performance by slowing down the CP-CP session activation process.

When a node that has just joined the APPN network attempts to establish sessions, it has to monitor the location of all resources that it wishes to establish sessions with by issuing a broadcast search or a directed search to a central directory server.

By checkpointing the TRS and directory services databases, the topology and resource location information can be saved and used if VTAM® is restarted.

Checkpointing saves the databases to the checkpoint data sets at the time either of the following conditions occurs:

The VTAM operator issues a MODIFY CHKPT command.
The VTAM operator issues a HALT or HALT QUICK command to terminate the VTAM network node.

This process of saving the databases is also called database hardening.

Note: The databases are not saved if the VTAM operator does the following actions:

Issues a HALT CANCEL command to terminate the VTAM network node
Cancels VTAM by canceling VTAM partition

When a database is checkpointed, it is written to a checkpoint data set. This data set is read by VTAM when it is restarted and the information is placed in the databases. If checkpointing is not used, both the topology and directory services databases have to be completely rebuilt. When checkpointing is used, only those changes to resources and topology since the last checkpoint need to be monitored.

To specify which of the checkpointed databases are to be loaded at startup time, you can use the INITDB start option. By default, both the topology and directory services databases are loaded. You can also specify topology only, directory services only, or neither.