NUMTCB

Specifies how many worker tasks (TCBs) are started simultaneously to process incoming connections from a client.

For schedulers, 5.

For user servers, 1.

1

35

Each worker task has its own DD statement Tnnnnnnn in the JESx job log. For every worker task started, the main task writes message TAU0010I to the MAINTASK DD statement:

TAU0010I 14:47:07.563 Subtask 1 started A(TCB) : 008B33B0 A(ECB) :
2236C0E0
TAU0010I 14:47:07.563 Subtask 2 started A(TCB) : 008B3120 A(ECB) :
2236C144
TAU0010I 14:47:07.563 Subtask 3 started A(TCB) : 0089EE78 A(ECB) : 2236C1A8

This message contains the address of the TCB (Task Control Block) and ECB (Event Control Block) for the worker task. When a worker task is posted for an incoming request, message TAU0011I is written to the MAINTASK DD statement:

TAU0011I 15:10:17.883 Subtask 1 posted A(TCB) : 008B33B0 A(ECB) : 2236C0E0

When all worker tasks are busy processing client requests and no worker task can be posted for a new client request, TAU0009E is written to the MAINTASK DD statement and the main task waits for one of the worker tasks to become available again:

TAU0009E 15:10:20.882 No free ECB available. Waiting for next free ECB.

If this message appears too often, the number of worker tasks NUMTCB should be increased.

Note: Each subtask requires private storage to send and receive data (see BUFFERSIZE), and the storage required for the application. Storage is allocated when the worker task is established at scheduler address space startup.

The number of tasks is calculated by the formula: NUMTCB + 4 (CMDTASK + SRVTASK + MSGTASK + LISTENER).

Do not set NUMTCB higher than 33 (when using BUFFERSIZE="10M", as the size of the private region will be insufficient otherwise.