In a recent note R.S. said:

> Date:         Sun, 21 Aug 2005 21:52:00 +0200
>                                                 ...   Remember: the
> goal is to install ControlM, not to create lowercase userid. However
> thread author asked about lowercase useid, but he really needed help to
> install ControlM. He didn't express it, maybe even didn't realised it,
> nevermind.
>
> Remember "supervisor state and CICS application program thread" ? It was
> very similar situation. Supervisor state was not needed and didn't
> solved the real problem.
>
> IMHO we should pay attention to the real nature of the problem, not
> literally the question asked.
>
And z/OS itself frequently loses sight of the real problem.  An example:

Background:  Our test systems share some DASD with our production system.
However, not all migration tape drives are connected to all test systems.
Therefore we run HSM only on the production system.

The problem:  I submit a batch job to one test system.  Shortly, I
notice that one step is stalled:

    B00474 *0041 ARC0055A REPLY 'GO' OR 'CANCEL'

This oughtn't be a problem.  I log on to the production system and do
DSLIST Info on the data set.  It swiftly recalls, and I return to the
test system:

    /0041,GO

Alas, Conway's Law asserts its pernicious influence:

     IEE600I REPLY TO 0041 IS;GO
     ARC0050A DFSMSHSM IS NOT ACTIVE - START DFSMSHSM
    *0042 ARC0055A REPLY 'GO' OR 'CANCEL'

To what purpose?!  The real problem is to allocate the data set; starting
HSM is a subordinate task; irrelevant once the data set has been otherwise
recalled.  As I envision the current logic:

    if migrated then

        Retry:
        if HSM inactive then
            WTOR ARC0050A
            if reply=='GO' then goto Retry
            else fail
            endif
        endif

    endif

    ... continue with allocation ...

The logic could be made right simply by moving the "Retry:" label from
"if HSM inactive" to "if migrated", allowing detecting that the data
set is no longer migrated.  But z/OS has focused too literally on the
question of activating HSM, and forgotten the goal of allocating the
data set.

-- gil
--
StorageTek
INFORMATION made POWERFUL

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Reply via email to