resmgr fails authorization running from a subtask in cics

[mainbld]

Hello out there,
I inherited some code where I work in cics that was using an estae exit in a subtask to perform retry logic. I have verified that this wasn't working. I need to do something more robust and am switching over to use a resmgr.

--All my libraries are APF authorized.-- checked it many times.

The init pltpi pgm is APF auth. and yes lked with ac=1
The subtask is APF auth. ditto.

In the subtask I am using modeset mode=sup before the 'resmgr' call. I blow up with a 047 abend. If I turn off the modeset, I get rc 52 from the resmgr call. This is the same reason.

This code works outside of CICS.

I am new to cics and trying a variety of things; unsuccessfully. What is the best retry method I should use in cics, if resmgr is not what I should be using.

Thank you,
jeff

 

[stevexff]

What was the problem with the ESTAE? Can't we just fix that instead?

 

[mainbld]

What I found out through research is that CICS turns off the jscbauth bit after phase I of plpti processing. Nobody is APF authorized in cics (after phase I) so I cannot setup a resmgr. My intent was to attempt a task-level resmgr (not at the address space level). During phase I, it is possible to run a authorized pgm in cics, but I wouldn't have the tcb for my subtask at that point. So...

I am in agreement and back to the estae. What is strange is that I get a rc 0 creating the estaex ct..., and then blow it up on purpose trying to get rtm to branch to my retry address. It doesn't go there. Yet the same code does work outside of cics. I will be looking at some dumps this morning. I have to believe that Cics is front-ending my estae some how. I am planning to keep all my recovery in my cics subtask program.

I am open for any suggestions.
Thank you for getting back to me.

 

[stevexff]

It's been a while, but from memory CICS already has its own ESTAE.

 

[mainbld]

I got my estaex working in cics just fine. done!