On 07-Jan-2016 13:41 -0700, Troy Hyde wrote:
On 07-Jan-2016 13:38 -0700, Troy Hyde wrote:
On Thursday, January 07, 2016 1:04 PM broehmer wrote:
Has anyone heard of PDM not working if the library being worked
on has too many objects in it?
Using option 12 to view the library results in the screen
returning right back to the same screen. It won't step into the
library. DSPLIB works fine and shows the objects but using PDM is
a no go.
Once or twice a year I have someone come into my office and tell
me PDM doesn't work and it's invariably because of too many
objects in a library.
I believe PDM has a 16gb limit on its working storage. If the
number of objects in the library exceed this limit it get the
error.
If you use WRKOBJPDM and hit <F4>, then subset your list you may
be able to get what you're looking for.
Found the IBM page describing the problem. Should have included it
with my last message.
[http://www.ibm.com/support/docview.wss?uid=nas8N1013215]
TechNote Reference #: N1013215 Historical Number: 513987411
Abstract: When working in PDM, an error message PDM0371 may be issued.
According to the symptom description [per followups to\by the OP
stating that no message was issued], other than the word "may", the
above TechNote document suggests there exists either a defect for the
list not being presented, or a defect for the msg PDM0371 not being
presented; the word "will" in the text conspicuously appears to override
the word "may" in the abstract:
"Reason: PDM has a 16MB segment in which it stores the names of the
objects and libraries that it is working with. If it exceeds this limit,
it *will fail with the message PDM0371*. This can occur when the list of
libraries or objects being worked with is very long or when there are
large objects in the list. Damaged objects can also affect the 16MB
segment. The size of this segment is a design limitation of PDM and
cannot be changed."
Other issues with the above quoted text:
• "large objects" surely is twaddle or badly characterizes; the size
of the stored [fixed-length 30-byte] "names" is static, unchanging no
matter the size of the objects
• "cannot" is hyperbole; "will not" is more apt, per the
subject-matter is software, which of course inherently *can* be changed;
a choice was made to not enhance the QUYLIST program and ecosystem
• The effect of "Damaged objects" is not on the segment, but on how
the code respond to unexpected conditions per condition handlers;
notably, if reference to an object yields a MCH0601, the code could
react generically as if the /list space/ was breached rather than more
appropriately detecting the condition was specific to an
object-reference *other than* a failed reference to the list-space itself.
p.s. Had the PDM code been created\owned within the IBM Rochester
[MN] lab, I am confident that the feature would have been corrected
[eventually] instead of the current effect whereby the issue was closed
as Permanent Restriction (PRS) [and a TechNote document created to
maintain a record of that APAR closure\response].
midrange.com
