David, The problem with a "real 3390-9" was rotation latency. IOSQ time was more of a symptom then a feature.
The 3390-9 in IBM livery had three logical 3390 CKD tracks on a physical track and spun at one third of the speed of the 3390-3. Average latency was 21ms, compared to 7ms on the mod 3s. Not so bad for long, chained sequential IO, but pretty ugly for and random access cache misses. The OP point of all the hardware being the same is perfectly valid. Mod 3s, 9s, 27s, 5s and As all spin at 7200, 10K or 15K RPM depending on the underlying disk drives (or don't spin on SSD). The chance of IOSQ and need for PAV really depends on your calculator and Little's Law :-) Ron > -----Original Message----- > From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf Of > Jousma, David > Sent: Monday, January 23, 2012 7:18 AM > To: [email protected] > Subject: Re: [IBM-MAIN] Going from mod-3 to mod9 > > Iosq? Really? This is raid dasd, isn't "slow" 3390-9's a thing of > the past? > > _________________________________________________________________ > Dave Jousma > Assistant Vice President, Mainframe Services [email protected] > 1830 East Paris, Grand Rapids, MI 49546 MD RSCB2H p 616.653.8429 f > 616.653.2717 > > > -----Original Message----- > From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf Of > Staller, Allan > Sent: Monday, January 23, 2012 10:09 AM > To: [email protected] > Subject: Re: Going from mod-3 to mod9 > > <snip> > 1. We just had 3390-9's configured into our EMC box. This box also contains > 3390-3's. Given that the hardware is the same throughout and all other things > being equal, is there any decrease in response time on the mod-9's? > </snip> > > Beware of IOSQ! From the viewpoint of the Operating System, you now have > 3 times as much data behind the actuator on Mod-9's as Mod-3's. If the > Operating system *thinks* the device is busy, the IO is queued off the UCB and > never even tried until it comes to head of queue. Multi-system access will > exacerbate this even more. > > <snip> > 2. I have a VSAM file (with extended addressability) that currently spans 27 > mod-3's. My plan is to move it to 10 mod-9's using IDCAMS REPRO to back it up > then doing a DELETE/DEFINE, (changing the SPACE parameter from CYL(2500 150) > to CYL(9500 150)), then REPRO from the backup. Am I setting myself up for a > failure? > </snip> > > I do not believe there will be an issue. Obviously, you already have all of > the SMS stuff in place. I would, however suggest REPRO from "original" to > "original.new" and after successful repro operation, alter "original" to > "original.old" and then alter "original.new" to original". > In this way, the original is always available for a momentary recover if any > problems are discovered with the target of the repro. > > This e-mail transmission contains information that is confidential and may be > privileged. It is intended only for the addressee(s) named above. If you > receive this e-mail in error, please do not read, copy or disseminate it in > any manner. If you are not the intended recipient, any disclosure, copying, > distribution or use of the contents of this information is prohibited. Please > reply to the message immediately by informing the sender that the message was > misdirected. After replying, please erase it from your computer system. Your > assistance in correcting this error is appreciated. > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, send email to > [email protected] with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
