Hi All

My TSM software notification email this morning has in it an entry for
IC70809.  This describes a situation where a scheduler using schedmode
polling contacts the server after the start time of a schedule, but still
within the window.

The doc says that a randomization delay is not applied in this case,
however the client behaviour is that a randomization delay IS applied.

The suggested path for this APAR is that the doc be changed to reflect
reality.  There is nowhere for the user community to have any input on this
so I'm raising it here.  My strong preference would be for the behaviour to
be changed to reflect the doc.  Some schedules have a necessarily long
window and for clientaction schedules it is not possible to reduce the
window below a day.  If there is some issue with the client hanging up, or
a parameter issue that stops the scheduler from running correctly and need
to be restarted it is frustrating to have wait for a randomization delay
after the start time of the schedule.  I have even had cases where the
restart of a scheduler process was within the schedule window, but the
randomization delay pushed the start of the backup outside of the window.

So I vote for a fix to the code rather than a fix to the doc.  Is anyone
else of the same opinion?

BTW I first saw this issue with a 5.1 client a few years back, but as 5.1
was way out of support at that time did not follow it up.  I wonder when
the client behaviour changed, or if it has ever been right?


Regards

Steve

Steven Harris

TSM Admin
Paraparaumu, New Zealand