On Sat, 29 Aug 2015 01:12:46 -0500, Tim Hare wrote:
>Did not see this caveat, and I had some big troubles from this, so I have to
>mention this:
>
>IF you allocate an INTRDR under TSO, you need to find a way to limit how many
>concurrent users use it. Unless things have changed since I retired, internal
>readers are _not_ an unlimited resource, there are only so many defined in the
>system. If you have a lot of concurrent users of your submit process and they
>use up _all_ of the internal readers, then regular SUBMIT commands can start
>failing .
>
Does "under TSO" make a difference?
I believe I've read here that newer releases of z/OS have much relaxed that
limit.
It was probably established in an era when it was reasonable to assme that no
shop would have more than a few hundred card readers.
Still, it's reasonable to assume that JES has some remaining limit on the number
of concurrent interfaces it can support. If not, one might encounter a limit of
TIOTs or JFCBs or SPOOL capacity or ...
DoS is hard to prevent. No resource is unlimited.
/* Rexx -- I won't try this -- (untested) */
do I = 1
say I BPXWDYN( 'alloc rtddn(DD) sysout writer(INTRDR) msg(2)' )
S.1 = '//'X'I 'JOB Invalid statement' /* Ad lib. Use junk MSGCLASS. */
'EXECIO 1 DISKW' DD '(stem S.' /* Ensure OPEN. */
end I
BPXWDYN bypasses the DYNAMNBR limit that ALLOCATE enforces.
... And watch the PURGED messages stream by in SYSLOG.
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
