---------- Forwarded message ----------
From: Shai Hess <mfnetd...@mfnetdisk.com>
Date: Mon, Jul 30, 2012 at 1:24 PM
Subject: improve IO scheduling to MFNetDisk emulated 3390 disks
To: shai.h...@gmail.com
**
HI,
Many performance improvement lastly change the timing issue of MFNetDisk.
The throughput increase and IO request elapse time decreased.
This situation change the scheduling issue dramatically.
High priority jobs start to control the disk per UCB (with CIOD which is
like IBM PAV, no problems). Low IO priority wait more time for IO.
For example if you run many SUBSYSTEM IO with high IO priority per second
per one IO device (High IOQ time in RMF) and you try to run VTOC list using
ISPF 3.4, the listing of the VTOC may take long time because IOS serve the
high priority first and it may be taken long time before the ISPF 3.4 will
received IO service time.
To eliminate this condition I make some changes to the MFNetDisk IOS to
give more but small IO services time to low priority job like TSO user to
eliminate timeout condition and to response in fair time and not for
unexpected long time as it may be when the IOQ per UCB is very high.
So, PTFID=211 is ready for download. The change is in MFNetDisk MF.
The PTFID description:
Improve IO scheduling for MFNetDisk disk emulation to enable low job IO
priority to run as well in fair time.
This eliminates cases where many high IO priority jobs control IO
scheduling and low IO priority jobs are delayed.
July 30, 2012
Thanks,
God bless you.
Shai Hess, MFNetDisk product.
--
Thank.
God bless you,
Shai
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
