CICS ONC RPC operational considerations

The server controller uses EXEC CICS START to start the aliases that run the CICS® programs. CICS limits on the numbers of tasks that can be started may prevent aliases from running as soon as they are started by the server controller. This leads to delays in servicing the client requests, and this may lead to time-outs in the client.

In the XDR routines, storage allocation is done using MVS™ facilities, not CICS facilities. The storage is owned by the RP TCB. If an XDR routine abends, the storage is not freed by the server controller or the alias, nor is it freed by MVS, since the RP task does not end. Repeated abends in XDR routines may lead to shortage of storage that can only be corrected by stopping CICS.

MVS task control blocks (TCBs) used by ONC RPC

The TCB that interacts with TCP/IP for MVS goes into a wait as a result of that interaction. If the wait were on the QR (quasi-reentrant) TCB, all the transactions in the whole CICS region would be delayed. This is avoided by using an extra TCB, the RP TCB, for issuing calls to TCP/IP for MVS.

The RP TCB is used for some processing for every client request, but most of the call processing done by CICS ONC RPC takes place under the QR TCB. The split between the two TCBs is transparent to you for most of your work, but you need to be aware of it for problem determination.

ONC RPC task-related user exit (TRUE)

CICS ONC RPC includes a task-related user exit; this is used to anchor shared storage and to improve CICS ONC RPC’s response to CICS shutdown. CICS ONC RPC does not use a TRUE to pass commands and data to and from TCP/IP for MVS.

Related concepts
ONC RPC concepts
CICS ONC RPC problem determination
Related tasks
CICS ONC RPC recovery procedures
Troubleshooting CICS ONC/RPC
Using messages and codes for ONC RPC
CICS ONC RPC trace information
ONC RPC dump and trace formatting
Debugging the ONC RPC user-replaceable programs
[[ Contents Previous Page | Next Page Index ]]