Bill, One of the main problems I have seen with CKD volumes for Linux running under z/VM is that the LINUX FICON drivers cannot request sequential pre fetch. If there is a skip-sequential type of activity occurring it will tend to defeat the sequential detection algorithms in the storage.
There is an all or nothing command that allows you to set all the IO for a mini-disk to be Seq pre fetch, or just leave it as normal. Sort of like cracking nuts with a sledge hammer. It has to be set at boot time. While I have not had a chance to study or benchmark it, I'm inclined to think that the cache algorithms in the storage controllers will do a better job handling Linux/Oracle IO patterns to a Fibre Channel LUN, rather than a FICON volume because many of the cache hints that z/OS workloads provide are not there for Linux on Z. Ron > -----Original Message----- > From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf Of > Bill Fairchild > Sent: Tuesday, March 22, 2011 11:28 AM > To: [email protected] > Subject: Re: [IBM-MAIN] junking CKD; was "Social Security Confronts IT > Obsolescence" > > Lynn, > > I signed in to LinkedIn and was unable to find the reference, so thanks for > including the URL for the missing LinkedIn reference. I read through that > reference and saw 16 comments, only one of which mentioned very bad throughput > for CKD disks. No technical explanation was given for the bad throughput; > i.e., was it hardware limitations in CKD, software limitations in Oracle, > etc.? This comment was posted 23 days ago, ca. 28 years after IBM first > announced ECKD and forward-thinking users began planning to junk their CKD by > going to ECKD. > > I remember many problems with ECKD when first implemented, just as there were > many problems with OS/360 in the late 1960s, with MVS in the late 1970s, with > the new VSAM catalog structure in the early 1980s, etc. But all those > problems were addressed, code and hardware were redesigned, etc. I still > stand by my earlier post - most users DID junk their CKD several decades ago. > Why was this one shop still using CKD 28 years after CKD's replacement first > appeared? Is this installation also running with a 28-year-old version of > their operating system? I also sent this same question-comment into the > LinkedIn thread. > > Bill Fairchild > Rocket Software > > -----Original Message----- > From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf Of > Anne & Lynn Wheeler > Sent: Tuesday, March 22, 2011 12:17 PM > To: [email protected] > Subject: Re: junking CKD; was "Social Security Confronts IT Obsolescence" > > [email protected] (Bill Fairchild) writes: > > It was. ECKD was announced in the early- to mid-80s, which was 25+ > > years ago, which is several decades ago. Not all users respond > > quickly. > > re: > http://www.garlic.com/~lynn/2011e.html#31 "Social Security Confronts IT > Obsolescence" > > oh, the missing linkedin URL reference: > http://lnkd.in/ajGuA2 > > reference was to "CKD" disks didn't change ... controllers added some extra > stuff (eckd) ... originally for "Calypso" ... the 3880 controller speed- > matching buffer ... allowing 3380/3880 3mbyte connenction to > (370/168 2880) 1.5mbyte channels ... which had enormous problems (that > wouldn't/don't exist w/FBA). > > old email discussing calypso (eckd) and how bad the problems were (several > severity ones in the field): > http://www.garlic.com/~lynn/2007e.html#email820907b > > above also mentions the dismal prognosis of ever getting MVS to support FBA > (I've periodically mentioned in the past about being told that even if I > provided MVS with fully integrated & tested FBA support, I still needed $26M > business case to cover education and pubs ... and I couldn't use lifecycle > savings ... only incremental new sales). > > past posts with references to calypso/eckd: > http://www.garlic.com/~lynn/2004o.html#7 Integer types for 128-bit addressing > http://www.garlic.com/~lynn/2007e.html#40 FBA rant > http://www.garlic.com/~lynn/2007f.html#0 FBA rant > http://www.garlic.com/~lynn/2008q.html#40 TOPS-10 > http://www.garlic.com/~lynn/2009k.html#44 Z/VM support for FBA devices was Re: > z/OS support of HMC's 3270 emulation? > http://www.garlic.com/~lynn/2009p.html#11 Secret Service plans IT reboot > http://www.garlic.com/~lynn/2010e.html#36 What was old is new again (water > chilled) > http://www.garlic.com/~lynn/2010h.html#30 45 years of Mainframe > http://www.garlic.com/~lynn/2010n.html#14 Mainframe Slang terms > > -- > virtualization experience starting Jan1968, online at home since Mar1970 > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, send email to > [email protected] with the message: GET IBM-MAIN INFO Search the archives > at http://bama.ua.edu/archives/ibm-main.html > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to [email protected] with the message: GET IBM-MAIN INFO > Search the archives at http://bama.ua.edu/archives/ibm-main.html ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
