Re: What part of z/OS is the OS?
Daniel A. McLaughlin([EMAIL PROTECTED])@Tue, Aug 29, 2006 at 08:24:54AM -0400: So which part of Windows is the OS? File handling I/O GUI Obviously it's a sum of its parts, as is Z/OS, Z/VM, Linux, Unix...and so on. If a component is removed and renders it useless. it would seem to be, IMHO, that it is germane to the OS. Can you work without JES, or a like function? Not very well. This comment pretty much encapsulated my thoughts. The Windows GUI isn't part of the OS *functionality*, but has become part of the OS *code base*. I could certainly digress into the technical reasons this is undesireable, and what about it was prompted by lousy ethics, but I'll resist. With respect to MVS, when I see these blurry lines I have to wonder which parts actually provide the core OS functionality. I tend to think of that functionality along similar lines as those Lindy and others mentioned ... The software that 1) enables the devices to be used at all 2) provides the API for accessing those resources 3) schedules the access to the hardware That doesn't eliminate all the fuzziness, but I don't like to stray to far from that when defining the term OS. Where IBM might like for their customers to think of the OS as the entire package solution, it seems reasonable to me to discard that notion long enough to ask which parts actually provide that core functionality. After four years in operations, I've pushed myself pretty hard to understand the system beyond batch and online systems (and beyond what our other operators learn, for that matter), but I still don't know many of the components, yet. So, I wonder along with Lindy which parts are the OS, and which are not. Wade Curry Sr. Implementation Mgr., ATT San Diego, CA -- 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
Re: What part of z/OS is the OS?
Daniel A. McLaughlin([EMAIL PROTECTED])@Tue, Aug 29, 2006 at 08:24:54AM -0400: So which part of Windows is the OS? File handling I/O GUI Obviously it's a sum of its parts, as is Z/OS, Z/VM, Linux, Unix...and so on. If a component is removed and renders it useless. it would seem to be, IMHO, that it is germane to the OS. Can you work without JES, or a like function? Not very well. I meant to address one more thing in my previous post. When you ask which parts would disable the system when removed, it isn't as practical as it may seem. The parts that your business relies on might not be part of the OS. If your mission-critical wintel application /shudder requires the GUI, does that make the GUI part of the OS? If you were to remove the programs that comprise the GUI would the computer become unuseable? Windows (in it's current incarnation) probably would fail to be useable without the GUI, but that still wouldn't indicate that it is part of the OS in my opinion. This removal = uselessness criteria would be useful if the issues were not conflated by the users' definition of useless and the mixing of OS and non-OS functionality in the system software. (Each case would need to be evaluated separately. I'm not making any judgement with that statement.) Wade Curry Sr. Implementation Mgr., ATT San Diego, CA -- 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
Re: How to Submit and monitor a Job?
Michael Knigge([EMAIL PROTECTED])@Tue, Aug 08, 2006 at 07:08:36AM -0500: Steve, What is it you need to do once it ends? Do *you* just need to know that it has run? Do you need to do something once it runs? Do you need to know how it completed? Well, currenty I only need to submit the job and need the information that it ended. But I guess in the near future I need more: is it running, which step and how it ended/abended. I could use the internal reader, but the problem is that I don't get back the Jobid of the job and so I can't check (even not by looking into SDSF) if the submitted job has started. I don't know if this is an installation-specific feature or not... I'm able to add a parm to the jobcard like this: NOTIFY=SYSUID I will then get console messages telling me when it has completed and what the CC was. An actual userid can be used, of course. Using the symbolic allows the message to go to whomever submitted the job. In addition, using ROUTE XEQ will let me know when the job was submitted, and when it is complete, using console messages once again. Those only provide the very minimal information that you need at this point. It's a starting place, though. Wade Curry -- 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
Re: Using Java in batch on z/OS?
McKown, John : Fri, Jun 02, 2006 at 01:28:50PM -0500: [mailto:[EMAIL PROTECTED] On Behalf Of Hunkeler Peter They are not useless. Real shops are saving real dollars! Just ask any shop that has zAAPs and runs WebSphere or SAP on z/OS. IMHO, it's just plain stupid marketing driven nonsense. z/OS MVS has proven to be superior in managing very diverse workloads with a single pool of processors. As of now it has to manage three processor pools, three work queues,... I wonder what the next speciality engine (what a terribly misleaging term) will be. Wrong approach, IBM. I must respectfully disagree. It is not stupid marketting by IBM. It is smart marketting by IBM. Why? Because although IBM started the tiered pricing it is now endemic in the market. IBM is trying to allow companies to legally upgrade their CPU power without getting hit for software upgrade fees by other companies. /snip However, I do agree that the entire tiered pricing is a millstone on the zSeries. I understand the why of it. But something better is required so that a company can run cost effective software while the software vendors make a reasonable profit. Unfortunately, I don't like anything that I've seen as a proposed replacement. /... if you (John) or anyone else really does understand the why of tiered pricing (beyond just having an efficient way to extract money) I'd really like to hear it. That sounded sufficiently sarcastic that I probably need to say that this is a serious question. Did/does IBM perceive a benefit to the *customer* as a result of tiered pricing... past or present? I've been in the mainframe world for 4 years, and that lack of exposure/experience is showing, I'm afraid. Frankly, pricing per MIPS sounds fundamentally unethical to me. But since I've never been involved in purchasing a mainframe, I can't say I see the big picture. I still see it as the purchase of a large asset, like a house. No one would charge you $10,000 less if you promise to brick-off the door to the second or third bedroom. And actually, I'm assuming the price/MIPS is a recurring fee. Perhaps it's just an up-front pricing scheme to induce a company to upgrade with at least a partial CP? Wade Curry Sr. Implementation Mgr ATT Services, Inc. San Diego, CA -- 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
Pricing model (Was: Using Java in batch on z/OS?)
R.S.([EMAIL PROTECTED])@Fri, Jun 02, 2006 at 10:25:01PM +0200: IMHO the truth is in the middle. It is wise to encourage people to *start using mainframe* - that's why the price for the special engines is low. It is also wise to keep current incomes from legacy users unaffected - that's why regualr CP's are still expensive and subject to software fees (MSU base). The idea is clear: if you want to run Linux, then you have a choice, you are not doomed to mainframe. So, IBM want to make z/Platform more attractive. Effect: IFL. Similar idea for JAVA workload, since JAVA is (more or less) platform independent. Similar approach for DB2 modern workload. Can be done on unix or windows. COMPETITION. Obviously old fat stinking IMS/CICS/COBOL/VSAM users pay the bucks at regular prices, because they have no choice. Migration is hard, expensive and failure-prone. They pay for regular CPs and MSU-based MLC. I don't know what MLC and MSU stand for, but I assume they refer to the hardware pricing scheme. You have included the issue of software fees, though, which I know even less about. I'm surprised by the fat old stinking label. At my shop we have IMS, CICS, COBOL, and VSAM. I was under the impression that the vast majority of mainframe installations use either IMS or CICS, and probably both COBOL and VSAM. There's no question that they are old, and ubiquity doesn't necessarily prove they don't stink. I've just never heard a mainframer describe them in that fashion, so I'm hoping you'll elaborate. At the very least, I find it difficult to describe what an installation with a modern workload would look like. I guess an installation that exclusively used VM and Linux might qualify as at least non-traditional. Is that configuration common at all? What other options are there for z/OS shop, though? - even a newer one? I can imagine an installation leaving out COBOL in favor of C/C++. I would guess, though, that it is not incredibly popular or cheap to buy an installation without either IMS or CICS. Writing entirely new online transaction/data connection frameworks for new applications seems like more expense and work than most companies would want to take on when there are already two well-known products available. However, these 'traditional' users also think how to reduce the costs. One of the ideas is to move some workload to special engines. IMHO JAVA is not the best choice for batch processing. For sure, it is technically senseless to migrate from existing (*working*), usually COBOL-based batch to JAVA. It is a signal to IBM: something goes wrong, maybe we should tune the pricing model. Wade Curry -- 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
Re: DSLIST Include Additional Qualifiers
Paul Gilmartin([EMAIL PROTECTED])@Sat, May 20, 2006 at 09:42:58AM -0600: In ISPF DSLIST, I enter as Dsname Level the pattern, USER.**.CSI, and deselect Include Additional Qualifiers. The display presented shows such as: USER.XYZ.GLOBAL.CSI *VSAM* USER.XYZ.GLOBAL.CSI.DATA VOL999 USER.XYZ.GLOBAL.CSI.INDEXVOL999 Does this make sense? Why is it showing me additional qualifiers when I excluded them? Also, while some of the options in the entry panel are sticky, apparently saved in the profile, two of the more tedious options, Confirm ... Delete, seem to be turned on every time I enter DSLIST. Is there any way to turn them off and make them stay off? Hi Gil, I'll address the sticky 'confirm on delete' issue first, since it has been addressed by others already. I can't say that I know how to make it stick, but I happen to agree with you that if _I_ decide I want _my_ 'confirm on delete' setting to stick, then there ought to be an option to make it persist. Others can go ahead and cringe thinking about it if that makes them feel better, I suppose. As for the additional qualifiers, I have noticed a couple things about that. One is that the setting *does* stick. The second is that it only seems to stick in new started screens. What I mean is that, when ISPF starts up, the first screen will not respect that setting when I use the DSLIST command. This is also true of any other screens that I may start via a command in my logon CLIST. Any other screens that are started after that *will* respect that setting. I don't know if this is due to how the LOGON proc and scripts are put together, or anything else. I have taken to starting only one screen, though. I run ESP in that first screen, and set PF13 to start 2 or 3 other screens and name them for me. Kind of a kludge, but it was quick, easy, and worth it since I use DSLIST and personal dataset lists *a*lot*. I've never changed my 'confirm on delete' setting. You might want to just check to see if the same feature/bug is affecting that setting too. Wade Curry Sr. Implementation Mgr ATT Services, Inc -- 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
Re: Submitting job using ReXX
Tim Hare([EMAIL PROTECTED])@Sun, Apr 30, 2006 at 08:30:52AM -0400: //S2C9011 JOB (S2C9,XX,TI),'', // MSGCLASS=T, // NOTIFY=SYSUID, //*TYPRUN=SCAN, // CLASS=X //* //PROCLIB JCLLIB ORDER=(PROD.PROC.LIB,PROD.PROC.LIB) //* //JOBLIB DD DSN=POITS.BATPROD.LOAD,DISP=SHR // DD DSN=PODBM.SCIC.RESLIB,DISP=SHR //* I don't see a comma after the first JOBLIB DD statement. I think that is what matches this error message you received: The JESYSMSG in the spool says EXPECTED CONTINUATION NOT RECEIVED Sorry, Wade - that's incorrect. You don't put commas between two DD statements, they are two independent statements. A comma at the end of the line is used to indicate continuation of one statement. The error message means, essentially, that somewhere a new statement starts before the old one completed. If MSGLEVEL=(1,1) usually you can look through the JESJCL listing for the statement mentioned in the error message and quickly determine that it has a comma after the last parameter, and that there's a JCL statement (DD, EXEC, OUTPUT, etc.) after it. Heh... thank you for a gracious correction, Tim. A case of neuro-synaptic flatulence on my part. I'm not very experienced yet, but I know better. oops. Wade Curry Sr. Implementation Mgr ATT Services, Inc. -- 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
Re: PDS Directory Question
Ted MacNEIL([EMAIL PROTECTED])@Fri, Apr 28, 2006 at 12:00:00AM +: There are tools that can expand PDS directories, but I have not had great success with them. In the past, I would have recommended converting to a PDSE; unfortunately there have been all sorts of integrity issues with them ever since IBM removed the requirement for them to be SMS-Managed. And, that requires a re-allocation, anyways. I've been in a *very* large shop for about four years. The responsibilities are extremely segregated for these systems, so I rarely get to see the sysprog side of life. (ESP, ReXX, MVS application support and upgrades are part of my usual fare). So, owing to my lopsided exposure to mainframes, I am hoping that you'll explain these PDSE issues. I've been using PDSEs extensively for personal datasets and haven't experienced any problems as a result of SMS interaction (or any other). If the PDSE datasets are SMS managed, does this alleviate the problems you've seen? Wade Curry Sr. Implementation Mgr. ATT Services, Inc. -- 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
Re: Submitting job using ReXX
Rashmi Nijaguni Mogali([EMAIL PROTECTED])@Wed, Apr 26, 2006 at 03:39:07PM +0530: The job looks something like this: //S2C9011 JOB (S2C9,XX,TI),'', // MSGCLASS=T, // NOTIFY=SYSUID, //*TYPRUN=SCAN, // CLASS=X //* //PROCLIB JCLLIB ORDER=(PROD.PROC.LIB,PROD.PROC.LIB) //* //JOBLIB DD DSN=POITS.BATPROD.LOAD,DISP=SHR // DD DSN=PODBM.SCIC.RESLIB,DISP=SHR //* I don't see a comma after the first JOBLIB DD statement. I think that is what matches this error message you received: The JESYSMSG in the spool says EXPECTED CONTINUATION NOT RECEIVED Also I am not getting the notification messages. Not sure what's causing the NOTIFY statement to be ignored. Is it possible that you have told ISPF to block your broadcast messages, holding them for later? My response is not as deep and complex as the other answers you've received, but then neither am I. ;-) Wade Curry Sr. Implementation Mgr ATT Services, Inc. San Diego, CA -- 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
Re: IBM-MAIN and punches out of order (Was: Anquish of JCL)
Chris Mason([EMAIL PROTECTED])@Sun, Apr 16, 2006 at 05:34:50PM +0200: Ted, You'll have to try harder with sorting out this timing issue with your ISP. It's quite impossible to follow a prize fight with the punches out of sequence. Chris Mason I have noticed that the replies are often out of order on this list. The funny thing is that I'm subscribed to several mailing lists, and all of them are ordered and (sub)threaded properly. The only lists I've ever had this problem with is this one and the tso rexx list. I'm using Courier IMAPs with 2 or 3 different mail readers, so I'm sure it's not the mail reader itself. I haven't really examined this closely, but I'm wondering if there is something slightly funky going on with the list software's handling of the message headers. Any other observations among the rest of you? Wade Curry Sr. Implementation Mgr ATT Services, Inc. -- 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
Re: Anquish of JCL (Was: Migrating me from linux/bsd to zOS)
Ted MacNEIL said: However it's reason d'etre is indicated by its name, short for catenate. It's purpose is to take multiple files and catenate all the input into a single output stream. And, catenate is intuitive? To a certain degree it is. Just like it's obvious to you that when you include two DD lines in a single DDNAME you get concatenation. Same thing. One of the key unix concepts is everything is a file, including standard input, standard output, etc. With that perspective, what should happen when you concatenate two text files into a third? And what should happen if you concatenate a text file to standard output? The issue isn't whether or not you agree with the alleged benefits everything behaving like a file. The point is that perspective makes the system more understandable for lot of people. My point was that all OS's have a LIST command. I was told that LIST was an invalid description of what the command did. What word would you pick? I don't know what Gil would pick, but cat works fine for me. Unix does have a list command. It is abbreviated to ls and it lists the directory contents. That also seems pretty straight forward to me. And, although my position on MVS v. Unix is _not_ that unix is better, I still have to say that there are *plenty* of utilities, panels, programs, subsystems, whatever, in an MVS environment that have names that give *no* clue as to what they do. I saw UNIX before I saw TSO (not before I saw JCL)! What's with these strange names? That has been my complaint since 1976! And, what kind of word is catenate? Concatenate is a valid english word! I learned it as concatenate myself, but I did find catenate in several on-line dictionaries. They are synonymous. It seems that familiarity and an easy learning curve are overly valued. Unfamiliar things don't always make sense at first, and so they are commonly avoided. If the learning curve is steep, then that will likely be avoided, too. The question is, does a familiar easily learned program *always* mean it's better? Was it worth it to you folks to learn your native languages? Human languages differ dramatically from each other, and are very complex. I still rather like using them to communicate with other humans. We have a bazillion people out there (give or take several scads) who are familiar with GUI interfaces, and think those programs that have them are intuitive mostly because of familiarity. Menus can be very useful, but sometimes they can become a stifling hindrance - a real maze of options, dialogs, and windows. (How do you shut down your Windows computer? Click on the [START] button.) (How many people think that PhotoShop's menus and dialogs are intuitive to learn?) Computers are powerful and complex enough to warrant learning a more complex and powerful language if you want to do real work. So intuitive seems to me rather like a moving target depending on the user and the goals. Plus, it just isn't the most compelling feature to me. One old joke in unixland is You *still use* vi ?? Response:No, I *learned* vi. I have no intention of giving up using vi, because it is powerful. The effort to learn it has paid off and I no longer think about the keystrokes and commands. (Except when I accidentally use ISPF editor commands :-) )The fact that it wasn't intuitive didn't inhibit me because I saw an experienced user do some powerful stuff with it and I expected a benefit from sticking to it. So, I prefer the flexibility of the unix command line. I prefer the clearly superior logging and job control of MVS. I prefer the price of entry-level unix systems. I prefer the outstanding mainframe throughput and reliability. I'm sure I could come up with better examples. If we're going to compare unix and MVS, though, these are the kinds of features I think are more important. And let's be honest with ourselves. Who else besides a mainframer and a unix guy would ever actually (seriously) assert that JCL, shell, TSO, vi... are intuitive? Wade Curry ATT Services, Inc. Sr. Implementation Mgr. o _ _ _ _o /\__ \\o (_)\__/o (_) _ \__(_) (_)/_ \_| \_|/'\/ (_)(_) (_) (_)(_) (_)' _\o_ -- 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
Re: Anquish of JCL (Was: Migrating me from linux/bsd to zOS)
R.S. said: Darren Evans-Young wrote: [...] vi used to be the only editor included on standard Unix installations. Now, the vendors include other editors. However, if you have to boot from the CD to recover a system in single user mode, vi is usually the only editor available. You better know if you need to edit a file like /etc/vfstab to get a system bootable again. [...] IMHO there also ed. Presonally I prefer ed, because I can work with it. I cannot use vi. I use computers for years, people say, I'm not idiot. I met many editors, some of them were unconvenient, but I swear, I cannot do anything in vi. Anything except getting beeping termial. Radoslaw, I think you'll get a kick out of this: http://www.gnu.org/fun/jokes/ed.msg.html Wade Curry Sr. Implementation Mgr ATT Services, Inc. o _ _ _ _o /\__ \\o (_)\__/o (_) _ \__(_) (_)/_ \_| \_|/'\/ (_)(_) (_) (_)(_) (_)' _\o_ -- 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
Re: Anquish of JCL (Was: Migrating me from linux/bsd to zOS)
McKown, John said: -Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Richard Tsujimoto On a few of our unix boxes, we actually have ISPF installed for a few crippled mainframers who have to do work on unix. Personally, I feel *when in Rome, ...* Well, if we're going to be off-topic, I've just discovered kate on my Linux system. Now THAT is a nice little editor. Side pane has a directory listing from which I can click on a file. That file is then brought into the editor. I can have multiple files in the editor in a single pane or split (vertical or horizontal) pane view. On the bottom is a terminal pane where I can enter UNIX commands, including other applications (such as staying in a DB2 interactive session). Did I mention that the editor has syntax highlighting for many file types and, if needed, you can create you own syntax highlighting via an XML encoded parameter file? I haven't tried that yet. I find it is still useful to know vi because of those times that I am remotely logged in. Console-based editors can be much faster in that situation. Also, most folks I know who use vi (and that's quite a few) actually do their best to get vim. This is a version of vi that is friendlier, does syntax highlighting, code function folding, splits the screen between multiple files, and lots of other stuff. I use it on Windows because I can't stand notepad. http://www.vim.orgIf you dare :-) Wade Curry Sr. Implementation Mgr ATT Services, Inc. o _ _ _ _o /\__ \\o (_)\__/o (_) _ \__(_) (_)/_ \_| \_|/'\/ (_)(_) (_) (_)(_) (_)' _\o_ -- 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