Re: OT? Java application servers.
I mostly agree with Kirk, but I disagree with this bit (or at least the implication): Although [WebSphere Application Server is] not in any way light- weight (in terms of resources or administration) WebSphere Application Server for z/OS will almost certainly outperform any other Java EE server on z/OS given equal CPU resources. (That's a pretty safe bet, because it's the only JEE server extensively and repeatedly benchmarked and tuned specifically for z/OS.) Also, in terms of administration, compared to what? None of the other JEE servers have both ISPF and Web administrative panels (and highly praised ones), none of the others have SMP/E installation and maintenance (including for a matching JVM), none of the others handle HTTP server setup on z/OS so well or at all (much less with the Apache-derived IBM HTTP Server V7), none of the others explicitly support RMF and SMF, none of the others explicitly support z/OS problem determination facilities Heck, none of the others spend much (or any) time even documenting installation, maintenance, and operations procedures on z/OS. You get the idea. But if Kirk is arguing that WAS for z/OS is the most functionally rich JEE server for z/OS, yes, that's clearly true. - - - - - Timothy Sipples Resident Architect (Based in Singapore) STG Value Creation and Complex Deals Team IBM Growth Markets E-Mail: timothy.sipp...@us.ibm.com -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Delete all members of a PDS that is allocated
In listserv%201006231937408428.0...@bama.ua.edu, on 06/23/2010 at 07:37 PM, gsg gsg_...@yahoo.com said: Does anyone know of a way to delete all of the members of a PDS which is allocated by job scheduler software? Yes; STOW DCB,,I is the one least likely to break. But if that software has the PDS open then you may have issues. We currently use an assembler program that does a reset(I think), but management wants us to not use assembler. I can pretty much guaranty decreased reliability and increased maintenance cost, but it's not my dog. I know that IEHPROGM can delete a specific member, When another job has the PDS allocated? I don't think you can wild card it. I believe that the most recent IDCAMS let's you wildcard the member name. -- Shmuel (Seymour J.) Metz, SysProg and JOAT ISO position; see http://patriot.net/~shmuel/resume/brief.html We don't care. We don't have to care, we're Congress. (S877: The Shut up and Eat Your spam act of 2003) -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Delete all members of a PDS that is allocated
In aanlktilff16w3azr3ycvb5pfguqddt8th-shz6aaj...@mail.gmail.com, on 06/25/2010 at 02:29 PM, George Henke gahe...@gmail.com said: Years ago Bell Labs Management required all application programming be done in BAL. They were running BPS? -- Shmuel (Seymour J.) Metz, SysProg and JOAT ISO position; see http://patriot.net/~shmuel/resume/brief.html We don't care. We don't have to care, we're Congress. (S877: The Shut up and Eat Your spam act of 2003) -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: instream data
In listserv%201006262246290980.0...@bama.ua.edu, on 06/26/2010 at 10:46 PM, Paul Gilmartin paulgboul...@aim.com said: I do not mean to conflate converter operations with interpreter operations. That doesn't change things; it would be an integrity exoposure either way. If the external generator were forked into a separate address space and setuid submitter, the security exposures would be limited to DoS via long-running or output-spewing generators. Aside from that, Mrs. Lincoln, how was the play? Even as they now recognize /* Rexx ...? In principle, but it would take more code. How does one pass a handle of a PDS member to /bin/sh? Extend the syntax. Use a pipe. There are ways to do it. This instantly gets into trouble with the VTAM TSO half-ass^H^H^Hduplex terminal handling. ObMarkTwain You mean the missing functionality that's been there for decades? Cf. the infuriating RUNNING/INPUT toggle in 3270 OMVS. That's the OMVS application code, not the VTIOC services. Similarly, the behavior of ISPF is due to ISPF code, not TSO. I vaguely recall there there is code on the CBT tape that runs full duplex in TSO. It's not rocket science. -- Shmuel (Seymour J.) Metz, SysProg and JOAT ISO position; see http://patriot.net/~shmuel/resume/brief.html We don't care. We don't have to care, we're Congress. (S877: The Shut up and Eat Your spam act of 2003) -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: instream data
-Original Message- From: IBM Mainframe Discussion List [mailto:ibm-m...@bama.ua.edu] On Behalf Of Shmuel Metz (Seymour J.) Sent: Saturday, June 26, 2010 10:50 PM To: IBM-MAIN@bama.ua.edu Subject: Re: instream data In listserv%201006252231432049.0...@bama.ua.edu, on 06/25/2010 at 10:31 PM, Paul Gilmartin paulgboul...@aim.com said: Wouldn't it be great if the C/I could run shell scripts (or Rexx, or Perl or ...)? No; it would be a serious integrity breach. Pardon my immense ignorance, but how would such an obviously useful extension be a serious integrity breach? I just do not see the connection. A looping Rexx script submitting infinite new jobs would be just as much of an annoyance as a shell or perl script doing the same thing, but neither rises to the level of integrity breach as far as I can see; punishable programmer stupidity, yes, but not integrity breach. Please explain. Peter This message and any attachments are intended only for the use of the addressee and may contain information that is privileged and confidential. If the reader of the message is not the intended recipient or an authorized representative of the intended recipient, you are hereby notified that any dissemination of this communication is strictly prohibited. If you have received this communication in error, please notify us immediately by e-mail and delete the message and any attachments from your system. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Percen tage of co de execute d that is user writt e n was Re: Delete al l members of a PDS t hat is all oc ated
In 1407858731-1277604868-cardhu_decombobulator_blackberry.rim.net-406249...@bda026.bisx.prod.on.blackberry, on 06/27/2010 at 02:14 AM, Ted MacNEIL eamacn...@yahoo.ca said: I have, and shall continue, refrained from ad hominem attacks in this discussion. That's a welcome change from your usual. All I've really said is if a user/developer is productive and effective in one language, why are we going to force them to learn/use another. You said much more than that. -- Shmuel (Seymour J.) Metz, SysProg and JOAT ISO position; see http://patriot.net/~shmuel/resume/brief.html We don't care. We don't have to care, we're Congress. (S877: The Shut up and Eat Your spam act of 2003) -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: OT? Java application servers.
On Sun, Jun 27, 2010 at 2:10 AM, Timothy Sipples timothy.sipp...@us.ibm.com wrote: I mostly agree with Kirk, but I disagree with this bit (or at least the implication): Although [WebSphere Application Server is] not in any way light- weight (in terms of resources or administration) WebSphere Application Server for z/OS will almost certainly outperform any other Java EE server on z/OS given equal CPU resources. (That's a pretty safe bet, because it's the only JEE server extensively and repeatedly benchmarked and tuned specifically for z/OS.) Timothy, I really appreciate your thoughtful participation on this list, and I think that we really do agree on most things. But the concept of light-weight is not one of them. So I'll take the bet you mention: please point me to a benchmark that shows Websphere (and all of its adress spaces) using fewer resources (CPU, memory) than a little Tomcat started task running a modest JSP or servlet application at modest transaction rates.I didn't mention disk space requirements, since that would just be cruel. We could also look at just non-zAAP CPU resources if you want. Also, in terms of administration, compared to what? None of the other JEE servers have both ISPF and Web administrative panels (and highly praised ones), none of the others have SMP/E installation and maintenance (including for a matching JVM), none of the others handle HTTP server setup on z/OS so well or at all (much less with the Apache-derived IBM HTTP Server V7), none of the others explicitly support RMF and SMF, none of the others explicitly support z/OS problem determination facilities Heck, none of the others spend much (or any) time even documenting installation, maintenance, and operations procedures on z/OS. You get the idea. Again, I agree that WAS is robust, but that is not the same thing as light-weight. Also, this question was not just about bloated JEE app servers.Will anyone else who has experience with administrating WAS on z/OS and also knows Tomcat help me out here? How about this: we'll have a half-hour webinar where in the first 15 minutes I'll demonstrate how to install Tomcat on z/OS and install a web app into it, and then you can do the same for WAS. Again, I'm not saying that WAS is not a fantastically robust Java EE app server. It is. Kirk Wolf Dovetailed Technologies http://dovetail.com But if Kirk is arguing that WAS for z/OS is the most functionally rich JEE server for z/OS, yes, that's clearly true. - - - - - Timothy Sipples Resident Architect (Based in Singapore) STG Value Creation and Complex Deals Team IBM Growth Markets E-Mail: timothy.sipp...@us.ibm.com -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu 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 lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Delete all members of a PDS that is allocated
On 06/26/2010 09:16 PM, Ted MacNEIL wrote: But here we are talking about a case of manipulating directory entries and internal content of an Operating-system-specific construct, a PDS. I can do that with REXX and ISPF Library Management Services -- no HLASM on my part. This prompted me to check latest ISPF manuals for LM services to see if I had missed something new. I see the the member delete service that has always been there, but there is still no PDS reset function (like in the CBT PDS utility). I wouldn't call deleting all individual members via ISPF LM services a viable production approach except for a PDS with a small number of members, or perhaps for a one time exercise. Going that route starts off requiring an order of magnitude more processing resources than a reset for even a small PDS and gets exponentially worse as the number of members increases. My unquoted remarks after the quoted sentence were: Higher-level languages like COBOL, that are deliberately designed to be machine-independent, provide no syntax to specify actions on that level. It just can't be done without depending at some point on code written in Assembler - either a local utility or some vendor utility. The LM management Services callable from REXX are of course vendor supplied utilities probably written in Assembler, not part of the REXX language itself, which simply re-makes my point that to do machine-dependent type things you have to go outside the machine-independent high-level language. The issue in this case becomes one of whether you use Assembler-implemented utilities that your installation writes and maintains or Assembler-implemented utilities from some other source. I don't know of any IBM utilities or non-Assembler interfaces included with z/OS that do a PDS RESET, but perhaps they exist. This capability certainly exists in freebie utlities on the CBT collection, or non-free vendor product alternatives if management insists that all utilities used have full maintenance support. -- Joel C. Ewing, Fort Smith, ARjremoveccapsew...@acm.org -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Claude Shannon biography in German
For those of you who 1) are interested in the history of computing and 2) read German, there is an enormously interesting new book available: Axel Roch. Claude E. Shannon - Spielzeug, Leben und die geheime Geschichte seiner Theorie der Information, Berlin: Gegenstalt-Verlag, 2010. Roch is the man who recently elucidated the history of the [monitor] mouse in surprising ways, tracing its use back to pre-PC air-traffic control terminals; and he writes very well indeed. John Gilmore Ashland, MA 01721-1817 USA _ The New Busy is not the too busy. Combine all your e-mail accounts with Hotmail. http://www.windowslive.com/campaign/thenewbusy?tile=multiaccountocid=PID28326::T:WLMTAGL:ON:WL:en-US:WM_HMP:042010_4 -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Assembler programs was Re: Delete all members of a PDS that is allocated
On 27 Jun 2010 07:30:35 -0700, in bit.listserv.ibm-main you wrote: On 06/26/2010 09:16 PM, Ted MacNEIL wrote: But here we are talking about a case of manipulating directory entries and internal content of an Operating-system-specific construct, a PDS. I can do that with REXX and ISPF Library Management Services -- no HLASM on my part. This prompted me to check latest ISPF manuals for LM services to see if I had missed something new. I see the the member delete service that has always been there, but there is still no PDS reset function (like in the CBT PDS utility). I wouldn't call deleting all individual members via ISPF LM services a viable production approach except for a PDS with a small number of members, or perhaps for a one time exercise. Going that route starts off requiring an order of magnitude more processing resources than a reset for even a small PDS and gets exponentially worse as the number of members increases. My unquoted remarks after the quoted sentence were: Higher-level languages like COBOL, that are deliberately designed to be machine-independent, provide no syntax to specify actions on that level. It just can't be done without depending at some point on code written in Assembler - either a local utility or some vendor utility. Until IBM provides a language or variant such as a systems flavor of C/C++ that has access to all of the facilities (including the peculiar linking conventions for some JES exits, any management that does not keep access to assembler expertise is playing with fire. Assembler can be largely self taught by looking at generated code from the HLL of choice although a course on Macros was helpful. Clark Morris The LM management Services callable from REXX are of course vendor supplied utilities probably written in Assembler, not part of the REXX language itself, which simply re-makes my point that to do machine-dependent type things you have to go outside the machine-independent high-level language. The issue in this case becomes one of whether you use Assembler-implemented utilities that your installation writes and maintains or Assembler-implemented utilities from some other source. I don't know of any IBM utilities or non-Assembler interfaces included with z/OS that do a PDS RESET, but perhaps they exist. This capability certainly exists in freebie utlities on the CBT collection, or non-free vendor product alternatives if management insists that all utilities used have full maintenance support. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: instream data
On Sun, 2010-06-27 at 09:53 -0400, Farley, Peter x23353 wrote: -Original Message- From: IBM Mainframe Discussion List [mailto:ibm-m...@bama.ua.edu] On Behalf Of Shmuel Metz (Seymour J.) Sent: Saturday, June 26, 2010 10:50 PM To: IBM-MAIN@bama.ua.edu Subject: Re: instream data In listserv%201006252231432049.0...@bama.ua.edu, on 06/25/2010 at 10:31 PM, Paul Gilmartin paulgboul...@aim.com said: Wouldn't it be great if the C/I could run shell scripts (or Rexx, or Perl or ...)? No; it would be a serious integrity breach. Pardon my immense ignorance, but how would such an obviously useful extension be a serious integrity breach? I just do not see the connection. A looping Rexx script submitting infinite new jobs would be just as much of an annoyance as a shell or perl script doing the same thing, but neither rises to the level of integrity breach as far as I can see; punishable programmer stupidity, yes, but not integrity breach. Please explain. Peter I am also quite curious as to why/how this would be a breach of integrity. The closest that I have come up with is if the script were allowed to do things such as open/read/write a standard file WHILE EXECUTING AS PART OF THE CONVERTER/INTERPRETER. That might allow improper file access if the initiator is running RACF trusted or with a high security ID. So I guess that I would like the ability to run a special purpose scripting language to enhance JCL creation. Or looping, how I would like some looping at times! -- John McKown Maranatha! -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
A good Mainframe(ZOS) oriented JAVA discussion group
Hello; Does anyone have a suggestion to a discussion group that is focused primarily on JAVA as implemented on the ZOS Platform? Thank you in advance; Sincerely; Kenneth J. Kripke kkri...@mindspring.com -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Delete all members of a PDS that is allocated
I just have to ask this one question. How difficult is it to get the number of directory blocks from a PDS in a _ program (where the blank can be filled in with COBOL, PL/1, REXX, etc., but not HLASM/ASM) and then open the data set with RECFM=U and then write an initial directory block followed by n empty blocks? With that question asked, if it is not so difficult, wouldn't that allow you to clear a PDS right rapidly? The only access you would have to have for the data set is update, because you aren't deleting it or [re-]allocating it. Regards, Steve Thompson -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
How to find the OMVS UID in use
I have a problem on a system (ACF2 based) where some users are unable to use the OMVS command, for example, properly from TSO. The message indicates that the MAXPROCUSER number is too low but this should not be the case. What I want to know is which UID this user is running with. If there an easy way to find the UID that a user ends up with when logged on. I can easily find the PID, but that doesn't help. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: How to find the OMVS UID in use
Anthony Fletcher wrote: I have a problem on a system (ACF2 based) where some users are unable to use the OMVS command, for example, properly from TSO. The message indicates that the MAXPROCUSER number is too low but this should not be the case. What I want to know is which UID this user is running with. If there an easy way to find the UID that a user ends up with when logged on. I can easily find the PID, but that doesn't help. It's not clear precisely what you're after, but the 'id' command sounds like it will fill the need: id -u user_name might do it. -- Kind regards, -Steve Comstock The Trainer's Friend, Inc. 303-393-8716 http://www.trainersfriend.com * To get a good Return on your Investment, first make an investment! + Training your people is an excellent investment -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Delete all members of a PDS that is allocated
On Sun, 27 Jun 2010 15:00:14 -0400, Thompson, Steve wrote: How difficult is it to get the number of directory blocks from a PDS in a _ program (where the blank can be filled in with COBOL, PL/1, REXX, etc., but not HLASM/ASM) and then open the data set with RECFM=U and then write an initial directory block followed by n empty blocks? Don't forget the keys. And what about DS1LSTAR? And what about PDSE? Other than that, I believe you need only overwrite the first block. -- gil -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Delete all members of a PDS that is allocated
On Sun, 27 Jun 2010 09:30:10 -0500, Joel C. Ewing wrote: the CBT PDS utility). I wouldn't call deleting all individual members via ISPF LM services a viable production approach except for a PDS with a small number of members, or perhaps for a one time exercise. Going that route starts off requiring an order of magnitude more processing resources than a reset for even a small PDS and gets exponentially worse as the number of members increases. Not exponentially. Quadratically, if I understand the technique. -- gil -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: instream data
On Sun, 27 Jun 2010 14:10:06 EDT, Ed Finnell wrote: In a message dated 6/27/2010 12:27:56 P.M. Central Daylight Time, joarmc writes: like the ability to run a special purpose scripting language to enhance JCL creation. Or looping, how I would like some looping at times! Yeah, just get your XEDIT guy to code it up. 6000 jobs later, the system comes back! There was a time, not very long ago, when ad-hoc use of INTRDR was deprecated (controlled?) because they were a finite resource. I suspect one could still create pretty much havoc by OPENing INTRDRs up to a system limit, and feeding each the content of an SVC dump (perhaps throttled to a few dozen bytes/sec) and watching the SKIPPING FOR JOB CARD messages scroll by on the operators' console. -- gil -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: zVM training
-snip-- After all what is a consultant anyway. Just someone who has read the manual because nobody else wants too and more often than not its the client's manual anyway or through the client's internet. Well, I'm sure there are some out there like that. But I take offense at your implication. Many times a consultant is someone who's walked the path before and learned how to do it right. All too often it's the managers making the purchase decision who don't do their homework about the consultant(s) they hire, so they choose based on price, or on how good a salesperson the consultant is, instead of on proven skills and value provided. -unsnip-- I too resent the implications of that statement, having spent 30 years as a RACF administrator (amont all the other hats I had to wear). While there are a great number of people who will pass themselves off as experts and try to become consultants, there's just no substitute for experience. I can't say that I understand all the nuances of RACF with respect to CICS and DB2, and I don't pretend to have those aspects of experience, I've worked with RACF in doing things like enforcing DSNAME standards, DR considerations, password and USERID authentication, etc. Granted these are only the first steps in defining an effective security environment, they're already more than many shops consider in their overall standards set. (If they have any :-) ) I don't even PRETEND to be perfect; just helpful. I'm expensive, but I think my overall experience is worth it, considering that I also have a reasonable background in DR and helping to define standards that make a shop more idiot-proof AND more secure. snip-- Anybody can talk it, fake it, in the classroom, -unsnip--- I deny that utterly and completely. When you work as a consultant, you're expected to walk the walk and talk the talk and prove anything you say. Failure gets you a contract revision: out to the sidewalk. If I'm taking a class and the instructor tells me that actions A B and C will have this effect and I go home and try it, my results better be what he tells me to expect, or he's going to have some tall splainin to do. I expect ANY instructor to know exactly what he/she is talking about, or I'm going to make that instructor VERY UNCOMFORTABLE, to say the least. When I went to OS/360 school, 40 years ago, the instructors were the guys that actually wrote the code. If I asked a detailed question, they sat down with me, after class, and showed me that what they told us was true. They don't do that now, more's the pity, but they will get good accurate answers to student questions. That ain't faking it. Bottom line: Consultants and Instructors are not always the buffoons you make them out to be; a few of us actually know what we're talking about. Management might not believe that, but those of us in the trenches know it all too well. And most of us know how to seperate the BSers from the genuine article. Rick -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Delete all members of a PDS that is allocated
I was unaware of Mr. Smith's involvement. Rick - Ed Finnell wrote: In a message dated 6/25/2010 5:31:07 P.M. Central Daylight Time, rfocht...@ync.net writes: with Bruce Leland, the original author, for about 25 years. Author(s) Mike Smith and Bruce Leland. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu 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 lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Delete all members of a PDS that is allocated
On 6/27/2010 3:00 PM, Thompson, Steve wrote: How difficult is it to get the number of directory blocks from a PDS in a _ program (where the blank can be filled in with COBOL, PL/1, REXX, etc., but not HLASM/ASM) and then open the data set with RECFM=U and then write an initial directory block followed by n empty blocks? With that question asked, if it is not so difficult, wouldn't that allow you to clear a PDS right rapidly? The only access you would have to have for the data set is update, because you aren't deleting it or [re-]allocating it. The intrinsic problem is that to do it correctly (for a PDS) the program needs to be authorized, as DS1NOBDB needs to be reset. DS1LSTAR could be set correctly by having the write flag on, and the correct CCHHR in the DCB (of the EOF at end of directory). An IEBCOPY compress in place would also reset DS1LSTAR, but I've never tried whether it resets DS1NOBDB (normally it would have no need to, as a compress doesn't change the member count). Gerhard Postpischil Bradford, VT -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Delete all members of a PDS that is allocated
-snip Management may know something you don't. Management may know that very few of their staff knows BAL. And if a program blows up in the middle of the night (or any other time) no one may be available to debug it. I have seen this happen myself as I was the one that was called in. For support. I lobbied after that for no more assembler and was successful at getting this approved. --unsnip-- Been in a similar situation myself. Made appropriate corrections and notified the progrmmer what the problem was and how to fix it. When he ignored that advice and the problem recurred, he got called on the carpet. After several more occurences, requiring me to come in and affect the fix, he was promoted to the sidewalk and I became responsible for the program. End of problem. Aft4er that, management would accept any BAL program, providing that I wrote it and was willing to provide assistance if it failed. Track records mean a lot! :-) Rick -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Delete all members of a PDS that is allocated
---snip-- I just have to ask this one question. How difficult is it to get the number of directory blocks from a PDS in a _ program (where the blank can be filled in with COBOL, PL/1, REXX, etc., but not HLASM/ASM) and then open the data set with RECFM=U and then write an initial directory block followed by n empty blocks? With that question asked, if it is not so difficult, wouldn't that allow you to clear a PDS right rapidly? The only access you would have to have for the data set is update, because you aren't deleting it or [re-]allocating it. -unsnip--- That might work, as long as you have a mechanism for manipulating the keys in the directory also. The last used block in the directory has a key of X''. Rick -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Assembler programs was Re: Delete all members of a PDS that is allocated
On 27 Jun 2010 15:20:32 -0700, in bit.listserv.ibm-main you wrote: ---snip-- Until IBM provides a language or variant such as a systems flavor of C/C++ that has access to all of the facilities (including the peculiar linking conventions for some JES exits, any management that does not keep access to assembler expertise is playing with fire. Assembler can be largely self taught by looking at generated code from the HLL of choice although a course on Macros was helpful. --unsnip--- I disagree. Looking at others' code might help to learn the usage of the instruction set, but it won't teach efficient programming techniques. Rick Looking at good generated code from decently coded HLL programs can give a clue. However, what amount of knowledge is adequate for maintaining JES exits and simple tools from the CBT tape? More important does the person believe in testing in a safe environment and reading the appropriate manuals (and knowing whether they understand them)? When you start getting into cross memory services, I don't know that I would want to touch that code let alone things that get into locks so knowing your limitations is important. Clark Morris -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Assembler programs was Re: Delete all members of a PDS that is allocated
On Sun, 2010-06-27 at 21:02 -0300, Clark Morris wrote: On 27 Jun 2010 15:20:32 -0700, in bit.listserv.ibm-main you wrote: ---snip-- Until IBM provides a language or variant such as a systems flavor of C/C++ that has access to all of the facilities (including the peculiar linking conventions for some JES exits, any management that does not keep access to assembler expertise is playing with fire. Assembler can be largely self taught by looking at generated code from the HLL of choice although a course on Macros was helpful. --unsnip--- I disagree. Looking at others' code might help to learn the usage of the instruction set, but it won't teach efficient programming techniques. Rick Looking at good generated code from decently coded HLL programs can give a clue. However, what amount of knowledge is adequate for maintaining JES exits and simple tools from the CBT tape? More important does the person believe in testing in a safe environment and reading the appropriate manuals (and knowing whether they understand them)? When you start getting into cross memory services, I don't know that I would want to touch that code let alone things that get into locks so knowing your limitations is important. Clark Morris I think that there is a difference between having a normal (ain't no such beastie) application programmer and an old style sysprog. I think sysprogs need HLASM. I am not convinced that it is necessary for applications people to really know HLASM or even z architecture. Today's COBOL is much better than in the past. I can see needing the speed of assembler in embedded type applications. But that is not z/OS's forte. Commercial programming and COBOL go together like pancakes and maple syrup. But other languages are coming up for the webified world. I like PHP. PHP using DB2 with perhaps some COBOL stored programs is, IMO, likely one of the best ways to talk via the web. But others may reasonable disagree. -- John McKown Maranatha! -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Delete all members of a PDS that is allocated
At 3:55 PM -0500 on 06/27/2010, Paul Gilmartin wrote about Re: Delete all members of a PDS that is allocated: On Sun, 27 Jun 2010 09:30:10 -0500, Joel C. Ewing wrote: the CBT PDS utility). I wouldn't call deleting all individual members via ISPF LM services a viable production approach except for a PDS with a small number of members, or perhaps for a one time exercise. Going that route starts off requiring an order of magnitude more processing resources than a reset for even a small PDS and gets exponentially worse as the number of members increases. Not exponentially. Quadratically, if I understand the technique. -- gil If I remember correctly, the index blocks at the start of PDS are filled with the directory entries and when you delete one member, all of the subsequent entries shift forward (and shift from block to block). Thus when you delete in alpha order (as opposed to reverse alpha order), you remove the first member entry and all the others shift to fill the gap. Assuming that the same number of members fill each block (due to being the same length) then the number of write operations is computed by this method: T = Total number of Members M = The number of members per block For I = 1 to T Compute the Sum of the Ceiling of I/M (Ceiling meaning I/M rounded up to the next integer). -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: OT? Java application servers.
Kirk Wolf writes: So I'll take the bet you mention: please point me to a benchmark that shows Websphere (and all of its adress spaces) using fewer resources (CPU, memory) than a little Tomcat started task running a modest JSP or servlet application at modest transaction rates. As always, I speak only for myself Ah, so you're allowed to design the benchmark to demonstrate a particular piece of software in the best possible light? (Though I still have my doubts on CPU.) Tomcat isn't even a Java Enterprise Edition-compliant server! Shall we also include z/TPF in the benchmarking while we're at it? :-) I stand behind the phrase almost certainly. And let me elaborate on it: in the real world, across the vast majority of the sort of real world JEE workloads that mainframe customers want and/or need to run, WAS z/OS is almost certainly going to outperform anything else. But sure, if you want to pick a single application profile where the application server isn't actually *doing* much, I suppose anything is possible. May I add that most people also care about price-performance. That's been very neatly addressed: the System z Solution Edition for WebSphere. How about this: we'll have a half-hour webinar where in the first 15 minutes I'll demonstrate how to install Tomcat on z/OS and install a web app into it, and then you can do the same for WAS. As long as we're setting groundrules, how about I set some more? (And I'll even quadruple your time. :-)) You must install Tomcat using SMP/E. You must maintain it using SMP/E. You must demonstrate problem determination using as many standard z/OS facilities as possible. You must show us the RMF and SMF output from Tomcat which illustrates sub-application granularity. Oh, and show us how you administer it using Tomcat-specific ISPF panels supplied by the Tomcat community. I'd also like you to demonstrate clustering (including application deployment and change management), load balancing, failover (within and across a Parallel Sysplex), and Workload Manager configuration details -- and how easy it is to administer all that with Tomcat. (And of course you'll be demonstrating Tomcat's scripting capabilities so you can show how z/OS system programmers can automate every management task using their favorite z/OS automation tool, because after all automation helps make things much easier to administer and less prone to error.) Also please show how much EBCDIC-toleration Tomcat has. And I'd like to see support for RRS, z/OS security of any sort (RACF, ACF2, TopSecret -- take your pick), high performance connectors to CICS Transaction Server and IMS Transaction Manager (since just a few of us might have one or both of those) I don't think I'm suggesting anything radical here. All these attributes relate directly to ease-of-administration (and also performance in some cases, such as SMF/RMF). If you have a mainframe you may not need need *all* of these attributes, but chances are excellent you'll need some of them. Look, I never said Tomcat was bad. Far from it. I think it's wonderful! (I also like Apache Geronimo with Tomcat or Jetty.) I think it would be wonderful if more mainframe shops ran it. Tomcat is also different than WAS: they're aimed quite differently. But you implied that WAS did not offer performance or ease-of-administration, and I disagree, strongly, so I'm saying so -- perhaps a bit too strongly. :-) For the sort of things most mainframers want to do, WAS z/OS sure as heck offers both performance and ease-of-administration, and how. But let me make this all less abstract. For those of you who would like to see IBM's actual training WAS z/OS materials to judge for yourself, here they are: http://www.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/PRS3422 Those training materials cover *much* more than just standing up one trivial application once, so please do take note when drawing comparisons. If you'd like to attend the class in person, unfortunately you just missed one starting in Dallas today (June 28). There's another one in Gaithersburg, Maryland, USA, from September 21 to 23, 2010: http://www.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/PRS1778 - - - - - Timothy Sipples Resident Architect (Based in Singapore) STG Value Creation and Complex Deals Team IBM Growth Markets E-Mail: timothy.sipp...@us.ibm.com -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html