RE: Performance issue - CLRLIB vs RNMOBJ vs DLTLIB / Timeslice

Steve,

Some time in V4, IBM introduced Dynamic Priority Scheduling as an improved
way to optimize interactive and batch workloads.  I included timeslice in
this, but that's not correct.  Timeslice still does what it always did, but
with today's fast CPUs, it doesn't offer much help.  Timeslice sets the MAX
time a thread in a job is allowed to process A TRANSACTION. It does not mean
that the job gets the to execute exclusively for that space of time.  On the
old CISC boxes, it was more important.  There's a lot of information on this
in Chapter 14 of the Work Management guide 
http://publib.boulder.ibm.com/iseries/v5r1/ic2924/books/c4153063.pdf
that helps explain how to appropriately calculate an effective timeslice.

As to why tape can be faster, depending on your storage controller, your
tape subsystem can restore the access path faster than the system can
rebuild it.  I'm not really an expert on the new hardware, so I can't
comment more.

hth,

Eric DeLong
Sally Beauty Company
MIS-Project Manager (BSG)
940-898-7863 or ext. 1863



-----Original Message-----
From: Steve Johnson [mailto:sjohnson@xxxxxxxx]
Sent: Wednesday, January 21, 2004 10:00 AM
To: midrange-l@xxxxxxxxxxxx
Subject: RE: Performance issue - CLRLIB vs RNMOBJ vs DLTLIB / Timeslice


Eric and all,
 
>> timeslice...  It doesn't work the way it used to, so I
doubt it really helps much.

What has changed to affect the way the system responds to timeslice value
changes?

>>Are there many access paths that are being rebuilt?

Yes, and this is probably the biggest issue as several have pointed out.
Restoring from tape might be worth checking into, but why would a restore
from tape be faster?  When I hear "tape" I automatically think it will be
slower.  My ignorance is showing, again...   I guess the answer is, "well
that depends".  I am on a model 830 and the tape drive is a 3580-001.

>> Do you really need the entire library copied?

We do eventually need to be able to trim down and dummy up the test data for
Sarbanes-Oxley compliance, and that will reduce the time it takes for the
environment data replication.  

Someone else suggested SAVCHGOBJ from PROD and restore that to the TEST env
but due to changes in test/training/development environments we need to
start with a fresh copy of all objects.

The other suggestion from someone else to do RSTLIB SAVLIB(MYLIB) DEV(TAP69)
MBROPT(*ALL) ALWOBJDIF(*ALL) 
RSTLIB(TESTLIB) would work if happen to need it the day after a full backup
is done.

Thanks for all comments/suggestions.    

Steve



_______________________________________________
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.