When a software product produces a MLWTO but, due to any performance problem, it isn't able to generate the end line, the Operating System forces it automatically and it is visible in the SYSLOG. For Example: ... .M 4000000 IDAB 2018198 01:06:09.64 STC22688 00000094 PXM0107 MODIFY TEXT = .D 149 00000094 START(xxx)DB2(nnnn)AU .D 149 00000094 HLVL(SYXP.CASOL)DST(D .E 149 00000094 MESSAGE TIMED OUT - MESSAGE COMPLETION FORCED ------- forced by OS ------- ... In this case product CA PTXMAN (XMANAGER) generated an MLWTO but, due to any specific problem (that should be investigated in its area), it wasn't able to finish the MLWTO with the end line message and the Operating System itself forced this end line message after some time to avoid any additional loop problem.
It usually happens that this end line message forced by the Operating System is not visible in the OPSLOG, that will look like the following: ... Time Job Name WTO ID OPSF ----+----1----+----2----+----3----+----4----+-- 01:06:09 PTXMAN 34B615DD 2408 PXM0107 MODIFY TEXT = 01:06:09 PTXMAN 34B615DD 2408 START(xxx)DB2(nnnn)AUTO(Y)INT(0100)DUR(0000)DYN 01:06:09 PTXMAN 34B615DD 2408 HLVL(your.hlqs)DST(vvvvvv)SDE(N)SQE(Y)CHV(N)P ...
How can OPSMVS manage this exception in MLWTO if the end line is not kept in OPSLOG?
Environment
Z/OS - CA OPS/MVS - other vendor products generating MLWTO in Syslog
Resolution
From the IBM documentation :
Controlling automatic ending of Multi-line WTO messages
z/OS MVS Planning: Operations SA23-1390-00
If a program issues a multi-line WTO message but does not end the message by issuing an endline, the system will hold the message in a staging area until an endline is received. Unended multiline messages will not be delivered to a console until an endline is received. The system monitors unended multi-line WTOs and will detect when a time interval threshold has been exceeded. This threshold specifies the number of seconds that the system will wait before truncating an unended multiline WTO that has not received a connecting WTO. When this time interval threshold is reached, the system ends the message automatically. To end a multi-line WTO message when it detects that no data line or endline has been issued for the message after an interval of 30 seconds, the system issues the following endline: MESSAGE TIMED OUT - MESSAGE COMPLETION FORCED
This end line forced by the Operating System is not routed thru the subsystem and does not originate from the job that issued the MLWTO, so OPS does not see it and cannot report it into the OPSLOG. These are issued purely for CONSOLE tasks benefit, so that buffers can be freed.
From an OPS perspective, if a product issues MLWTO's that don't finish, there is a problem with that product and it should be brought to the vendors attention. Meanwhile, to have OPSMVS automation working correctly with these MLWTOs, it is possible to create a global variable when the first line of the WTO is intercepted and schedule a dynamic TOD rule for a time that it should have been completed by.
For any specific problem in this area Please open a case with CA Support