On 23-Oct-2015 13:17 -0500, Jim Franz wrote:
To apply ptf *immed need to execute Execute ENDJS DTALIB(*ALL) first.
SI51856
FWiW, if contemplating apply that maintenance separately instead of
with C4143710, then consider instead applying the more recent PTF
SI56344 that includes the additional fixes from APAR SE61555.
To be accurate, the _Special Instructions_ for that specific PTF, do
not actually include that text "Execute ENDJS DTALIB(*ALL)".
While that text does appear [several times in fact], within the PTF
Cover Letter for SI51856, under a heading of "Special Instructions", the
inclusion of that text is only included as part of the _Superseded_
special instructions.
For reference only, the more complete text includes three invocations:
"If you are applying this PTF immediately, [then] you must do the
following:
1) Execute ENDJS DTALIB(*ALL) to end the monitor.
2) Load and Apply this PTF.
3) Execute STRJS DTALIB(*ALL) to start the monitor."
That list of instructions only appear under\after the text noting
what follows is "SUPERSEDED SPECIAL INSTRUCTIONS". Thus the
instructions exist for each of several named superseded PTFs, not the
PTF SI51856 itself. So whenever any one of the superseded PTFs is not
yet applied, and the intention is to APY(*IMMED) of the PTF SI51856,
then there is the need to perform those special instructions.
So if the [latest of the PTFs for which those special instructions
are included] PTF SI48560 is already applied, then according to the PTF
cover letter [correctly or incorrectly], there are no special
instructions applicable for APYPTF SELECT(SI51856) [irrespective of
immediate or delayed apply].
We will have couple long running jobs running.
Is this a bad idea to end AJS, install ptf, start AJS with scheduled
jobs running..?
Not expecting jobs to crash, but also don't want AJS to freak out..
I have almost zero experience with the Advanced Job Scheduler
feature, so my reply legitimately could be classified as a SWAG.
I would expect that given those three steps are the typical
recommended operations [identical, in fact, for several of the
superseded PTFs], and given they are offered without any further
explanation or warnings about ill-effects, then issuing those requests
probably should be inferred to be effectively harmless, even when issued
during normal operations; excepting I suppose that one might also expect
that no jobs could be started using the feature at the time of and
between end and start, that no jobs that were scheduled to start while
the feature is ended would actually start, and that any scheduled job
that did not start due to the feature being unavailable would be started
[according to any established rules about exceptions for jobs that
/missed/ their start] after the feature becomes available again per
having restarted the feature with the Start Job Scheduler (STRJS)
command request. In other words, though not explicitly stated, the
intended implication is probably that the three steps would be issued
consecutively, each issued presently if not immediately following the prior.
I posit that had the instructions not been deemed somewhat mundane
with regard to allowing them to be requested during normal operations,
then one would expect that the instructions instead would have
instructed to quiesce the system, rather than just ending the JS Monitor
via the End Job Scheduler (ENDJS) command request
midrange.com
