DFS0769I ttt SELECTIVE DISPATCHING - rrrr
Explanation
Selective dispatching was in effect since the last checkpoint because of a resource shortage. ttt identifies the TCB under which selective dispatching was in effect. rrrr identifies the resources which had the shortage.
- TCB
- Possible causes of selective dispatching
- CTL
-
- High volume of communications activity.
- Insufficient number of SAPs defined for the size of the terminal network.
- A disruptive event that caused communications processing to be interrupted and a flood of queued communications work.
- IMS internal problems.
- ALM
-
- High volume of LU 6.2 communications activity.
- A disruptive event that caused LU 6.2 communications to be interrupted and a flood of queued LU 6.2 work.
- An application program coded to maintain allocated LU 6.2 conversations for an extended period.
- IMS internal problems.
- OIM
-
- OTMA experienced a high volume of OTMA CM1 or CM0 input transactions.
- Insufficient number of pre-allocated SAPs are defined for the OTMA client.
- OTMA communication was interrupted because either IMS was restarted or OTMA was stopped and restarted.
- An OTMA client application is coded to keep CM1 sessions for an extended period of time.
- IMS experienced internal problems, such as long wait for a common resource, a long wait for I/O to the log, or a wait for a latch.
When an IMS TCB is in selective dispatching, only priority units of work (for example, the MTO terminal) and those already holding resources can be processed until the resources become available again. This could degrade response time for some terminal users.
The SAP resource can cause selective dispatching when it is. in short supply. SAPs are IMS internal control blocks that are required to process IMS units of work, communication requests (terminals) and LU 6.2 requests. If the supply of available SAPs becomes low, IMS attempts to allocate additional SAPs up to a maximum amount. If no SAPs are available to process a unit of work, IMS invokes selective dispatching for SAPs. When this occurs, IMS units of work can be processed only when another unit of work terminates and releases its SAP. This can cause degraded performance of the IMS system.
System action
Continues processing.
Operator response
This message indicates a possible resource shortage. If this message occurs frequently, or is accompanied by performance degradation, the system programmer should be informed.
System programmer response
This message does not necessarily indicate that there is a problem with IMS. Selective dispatching can occur for legitimate reasons (for example, during an IMS checkpoint, or due to quick peaks in communications activity. If this message occurs frequently or is accompanied by IMS performance problems such as poor terminal response time, then additional analysis is advisable.
If the TCB in selective dispatching is CTL, the number of available dynamic SAPs can be increased by changing the SAV= parameter on the IMS startup JCL or in the DFSPBxxx members.
If the TCB in selective dispatching is ALM, the problem could be an application program that does not deallocate its LU 6.2 conversation for a long period. This can occur when an application program does not receive an immediate LU 6.2 response to CONFIRM.
If the TCB is OIM, the selective dispatching applies to OTMA and the reason for using selective dispatching was probably transient. To avoid selective dispatching for OTMA, you can adjust the number of pre-allocated SAPs and the maximum number of SAPs that IMS can allocate dynamically by specifying the DSAP and DSAPMAX parameters in the OTMA client descriptor in the DFSYDTx PROCLIB member and restarting IMS.
A dispatcher statistics log record (type X'45' subcode X'0F') is written at each checkpoint. These log records should be printed and analyzed for selective dispatching problems. Additional statistics records (type X'45') might be useful in indicating high storage pool utilization. Any problem reports related to selective dispatching should include a copy of the IMS log that contains the checkpoint intervals during which the problem occurred or a printed copy of the X'45' records for those checkpoints.