The daily edition.

Scope: Rocket Enterprise Server 12.0 for Linux/UNIX, using PostgreSQL.

An IMS TM region normally has a local message queue. In a Performance and Availability Cluster (PAC), Rocket instead supports persisting that queue in a PostgreSQL-based Database File Handler (DBFH) datastore. Multiple Enterprise Server regions in the PAC can share this single queue, providing the basis for IMS TM redundancy and scale-out. The queue is therefore not copied between regions; it is externalized into common persistent storage. IMS Support (PAC) Configuring the IMS Message Queue

This distinction matters operationally. The documented architecture shifts continuity away from transferring or draining a private region-local queue and toward keeping the shared PostgreSQL queue available to the PAC’s regions. Rocket does not document a separate queue-draining control for this configuration; queue sharing is the availability mechanism described in these pages.

The configuration has several dependencies. A PAC and its Primary Scale-Out Repository must already exist, and the IMS TM queue datastore must use PostgreSQL. Rocket recommends installing PostgreSQL on a server separate from Enterprise Server. PostgreSQL’s max_prepared_transactions setting must be greater than zero, and the PostgreSQL XA switch module must be built because the queue participates through an XA resource. The Enterprise Server environment also requires ES_DB_FH=Y, with MFDBFH_CONFIG pointing to the DBFH configuration file. Configuring IMS TM Message Queue for Database File Handling

In ESCWA, install the PAC into an IMS-enabled region, configure its database connection and datastore, and add the XA resource using the PostgreSQL switch module and the documented xa_open connection string. Under IMS Compatibility → Configuration → TM, set Queue Name to the datastore URL, in the documented form:

sql://server-name/datastore-name?folder=/folder-name

Queue placement is a static configuration choice: changing Queue Name, Application Path, MFS Path, MFS character options, or Buffer Count after startup requires a region restart. By contrast, queue-capacity and maintenance controls—including Maximum Blocks, Active Keypoint Frequency, transaction run length, allocation cushion, and flush aggregation—can be changed while the region is running and take effect immediately. Configuring the IMS Message Queue

An important boundary is that the documented PAC support applies to IMS TM; Rocket states that IMS DB is not currently supported in a PAC. IMS Support (PAC)

Takeaway: IMS TM availability in a PAC depends on every participating region using the same PostgreSQL-backed DBFH queue, with XA and prepared transactions configured correctly.