On Oct 28, 2007, at 10:03 PM, Ed Long wrote:
Hi everyone.
Sorry for the delay in replying to the several most helpful
suggestions for which I am most grateful.
Root cause identified and solved.
I had in fact assembled and linked the exit correctly into
user.lpalib.
USER.LPALIB is #1 in the LPALSTXX concatenation.
LPALSTXX (actually LPALSTDB) is being selected at IPL time.
SYS1.LPALIB is in fact referenced about 2/3's of the way down in
the LPALSTXX member.
BUT, at least on z/OS 1.5, a copy of SYS1.LPALIB is always, and
under the covers, prepended to the LPALST so that it gets selected
first even if its included in the list.
As a result, the dummy IEEU29 was being used. It, of course, does
nothing, and does it quietly.
So, I renamed the bogus IEEU29 to IEEU29BG (for bogus), reipled
and away go my SMF datasets.
The default IEEU29 appears to also have a hard restriction on DSN
length of 4 characters after the period.
Ed:
Check to see which version of IEFU29 you have. There was at least one
bug I found in the CBIPO version and it was at least 5 years before
we hit the bug. I don't have the source for it, but IIRC I found a
really old reference to the bug on IBMLINK. Not sure how to tell you
if you have the most recent version but look at the assembled output
and see if there is a change that occurred in or before 2000 +- 5
years (probably -5 years). IIRC it was a small change. If you have
the most recent version take a look at CBTTAPE.ORG for a newer version.
Ed
----------------------------------------------------------------------
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
