On Dec 29, 2005, at 2:22 AM, Ron and Jenny Hawkins wrote:
Ed,
You could automatically start more initiators :) and if you
implement TMM
then there is no mount request for output (and often for the input
as well).
True but as usual main is issue is tape drive availability. Yes you
can start more inits but there are just so many tape dives
(especially for input). Each installation has to decide itself as to
what is acceptable. Only they can decide if starting more inits or a
plan of action. That is why an exit is probably the best way to
enforce any standards.
Ed
Caleb,
You can use TMM to direct USER tape output to disk, and therefore
there is
no input tape for USER datasets. More user friendly than cancelling
jobs.
Ron
With robotic tape mechanisms mounting tape is probably costs less so
in some situations it maybe almost irrelevant. The issue I see is
throughput for an iebgener step for instance even though it calls for
an input (and possibly an output tape) the step could last for 10
seconds or 30 minutes (or more) it would tie up that specific init
for that amount of elapsed time leaving the other jobs in that job
class waiting. It is in a way allocating resources (tape drives).
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
----------------------------------------------------------------------
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
