After I intercept a message (WTOR) the AOF rules call an OPS/REXX which issued an OPSREPLY to respond to the WTOR reply automatically. Yesterday, I had two return code=12 (WTOR reply not found). I look in the log (OPSLOG & MVS SYSLOG), and the reply came out on the console later after CA-OPS/MVS intercept the message. Most of the times the "REPLY" is available right away.
I have these questions about trying to fix the skip of reply: Q1. Why sometimes, replies are not available right away? Q2. What is the difference between specifying "DELAY" versus "WAIT»?
Delay = OPSREPLY command processor is delay, what does that means?
Is all OPSREPLY delay for a unique reply message id?
My reply message (ex: ONL909E) is issued multiple time for different IBM MQ manager channel or queues, will it stop reply for all of them?
Wait = CA-OPS/MVS won't reply before x sec have past, will my OPS/REXX wait or do I need to add CMDWAIT on my OPSREPLY command?
Q3. In the entire book "CA-OPS/MVS command and function" there is all lot of "SYSWAIT" parameters, but nowhere it tells what it does or means, no definition?
z/OS CA OPS/MVS
1) OPSREPLY Delay(n) is an unconditional wait - no reply will be attempted until the specified seconds have elapsed; Wait(n) is a conditional wait in that OPSREPLY checks each second for a matching WTOR, when found the reply occurs, if after the number of seconds expires and no matching WTOR is found then a bad RC is issued.
a) just the OPSREPLY with the DELAY- - each OPSREPLY command is unique.
b) see above response on how OPSREPLY WAIT works, but the REXX exec will not continue to the next instruction until OPSREPLY is done.
2) SYSWAIT keyword - (Optional) For a cross-system request, the SYSWAIT keyword specifies the number of seconds that CA OPS/MVS should wait for the remote system to receive and respond to the request. The default is the value of the MSFSYSWAIT parameter; you can specify any number of seconds from 1 to 600. 3) Either way of coding will work - if this is issued from a MSG rule NOOUTPUT is assumed; as a good coding practice you should specify the complete syntax of keywords in place of the accepted abbreviations. And your example did not include msg.replyid, be careful to not create a local variable with the stem name of an OPS/MVS event variable - do not use a local var of replyid as event variable name msg.replyid would be changed. 4) Use ADDRESS OPER when you are capturing the WTOR, as it is more direct/efficient since the executing code has all the information it needs. Use OPSREPLY after the WTOR/s have been issued if you can not or do not need to when the WTOR is issued. Example - use OPSREPLY when shutting down an application that requires a response/command to a WTOR that is always outstanding, like IDMS or IMS.. Change the message rule to use this statement: ADDRESS OPER "R "msg.replyid",yourintendedreply"