Ed,

please reread my first post!

>If that's the case, then the fault lies with LE and not BCPii.
How is LE to know that it runs under an implicit sub=mstr? That is not how your 
standard Cobol/PL/I program runs. 
What started this off was our general setting of HEAPCHK(ON) on request of 
application development. It seems that one LE enclave terminates which (with 
HEAPCHK on) wants to write a short 'ceedump' to JES2 spool with the heap 
statistics. When that fails, LE issues a U4087 (I think) and tries an IEATDUMP.

>If the allocation for CEEDUMP fails, then the dump should be taken (for
>example) with IEATDUMP.

*We* have an explicit LE statement DYNDUMP that directs *any* dump to HLQ 
TDUMP. BCPII has *chosen* to overwrite that setting by specifying part of the 
LE Default for DYNDUMP, which is the userid of the address space. STC users are 
not allowed to allocate data sets (RACF audit requirement), hence the general 
HLQ of TDUMP (instead of userid). Which HWIBCPII has overwritten.

Either HWIBCPII is LE-enabled (which I find very questionable for a system 
address space started via OS abracadabra), then it will have to set/override 
*all* options that require JES in one way or another or are otherwise harmful 
and not just a select few that are proudly documented in the books to be 
overwritten if allowed, *or* the address space should NOT be LE-enabled at all. 

Barbara

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to