Re: Follow-on: S0C4-11 abend caused by BASSM to address with all X'00' bytes
>> Thanks. What I do not yet understand: The TRNE address is 231A7800, >> where as the address in R15 231A7BB8. Why the difference? > >TRNE is a copy of Translation-Exception Identification, which is >defined in Principles Of Operation. I usually do read the manuals before posting, and I did in this case as well. Unfotunately, I was searching the PoOp (offline in the PDF) for "Translation exception" but got no hint. The missing dash makes the difference. Argrrr Understood now. Slowly derusting my debugging skills. Thanks for all the help with this. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: AW: Re: codepage restrictions on IBM applications
>>I doubt this is still correct information. After all, even z/OS base >>components issue WTOs in mixed case. ZFS is one that comes to my mind, and >>I'm pretty sure there are more but I can't name them without looking up. >> >What about WTOR? You mention ZFS (ITYM zFS; they're not the same.) >and UNIX filesystems are case-sensitive. Yep, zFS not ZFS, although the address space I'm takling about is named ZFS not zFS. And UNIX case sensitivity is not involved here. It is a z/OS program writing mixed case messages using WTOs (and possibly WTORs). Some other components writeing mixed mode messages to syslog: - SMS PDSE support, e.g. IGW040I - z/OS Message Flood Automaiton, CNZZ messages - z/OS SDSF, ISF messages - CICS V4, e.g. DFH0100 -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Follow-on: S0C4-11 abend caused by BASSM to address with all X'00' bytes
>Just my 2 cents: As long as TRAP(ON,SPIE) is set (and you said you cannot turn >it off for various reasons) slip will not get control because (E)SPIE gets >control before slip. So your slip trap probably won't match at all. In another case where the application failed with an S30A-10 sporadically, I had set a SLIP and it matched. But this was not a PGM check, so ESTAE and not ESPIE was in charge. >And your sysmdump may contain a dump, but not the first 0C4. All I'm looking for is a system trace. Hoping to find hints what do look for next. >If it were me, I would a) try to find out if that dump option thing is >customizable I did in parallel to the discussion here and succeeded. I now know how to change the options, but as you might have guessed, I'm working for a large company. And large companies have processes. It will take quite some time to bring those changed options to production, once I could convince engineering to actually do it. >Not sure that you mentioned it - is the problem reproducible at will? Not at will, yet, but it reocurred in production. Application people are still trying to reproduce it in test. Would make everyting much easier. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: S0C4-11 abend caused by BASSM to address with all X'00' bytes
>Is the program CALLed (or "called")? CICS, IMS, DB2, batch? You mentioned >records and database, so I'd take a stab at batch with IMS or DB2. The program is called via EXEC PGM=, and it is using DB2. We live in a complex environment here, i.e. we've got a home grown middle ware layer that application code needs to use for just about anything. Apart from that, there are dozens of applicaition modules loaded in the address space. My job is to find what might have gone wrong causing that writd situation and abend. It's the application people's job to read the code then. I have no knowledge about the business function and business data, so I cannot help with this. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: AW: Re: Follow-on: S0C4-11 abend caused by BASSM to address with all X'00' bytes
>Since z/OS V1.13, you can add CEEOPTS as a JCL Statement Yes, I know. It is not that I would not know *what* to do if I was allowed to do it. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: AW: Re: codepage restrictions on IBM applications
I had a look at manual z/OS V2.1 MVS Assembler Services Guide. It describes the characters "... that will be displayed on the console..." in a table. This tbale contains this table shows a lot of special characters as well as all upper *and* lower case letters. It also say that characters not in the table will be replaced by blanks weh displayed on the console. Unfortunately, the manual does not talk about how the message text is treated when written to the syslog. A quick experiment shows that besides above characters, als accented characters show unchanged in syslog. I can't verify what is displayed on the console. I'm not sure whether the text is inspected by WTO or only later by CONSOLE when displaying on consoles. The later would make sense to me. You want to avoid that strange things happen when strange data is displayed on terminals (consoles). Syslog is actually nothing but a data set consting of records. Should be able to cope with any byte content. Tools such as SDSF which display the syslog would then again make sure only harmless characters are displayed.. Opinions? I'm thinking about sending an RCF asking for clear description of this. --Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Follow-on: S0C4-11 abend caused by BASSM to address with all X'00' bytes
>Make sure that you have your systrace set at maximum, then (not the default >64K, and even 1MB will not be enough, especially on a busy system). I had a look at the size recently but can't remember what it was right now. But I remember It to be of decent size. >Did you get the same set of insufficient dump data? Just for comparison - were >the same addresses involved? Yes, there seems to be some consistency. It was the same NSI address in the PSW in some dumps I lokked at. >Admittedly we never specified RE=11 (or mode=pp) on the slip trap we had >attempted for an 0C4 in my last job, but the slip trap on OC4 only prodcued a >dump when TRAP was set to OFF. Maybe what interfered was the fact that parts >of the code were authorized. Maybe Peter's "middleware infrastructure" is also >authorized, at least in parts. Nothing authorizied involved. What I call "middleware" here is just a bunch of Cobol and Assembler routines that have to be CALLed by application code to perform such basic things as opening, reading, writing, closing data sets. I don't know much about it yet; too new to the company. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Follow-on: S0C4-11 abend caused by BASSM to address with all X'00' bytes
>You are right, I have to eat those words. ESPIE processing translates unresolved program interrupt codes x'10', x'11', x'38', x'39', x'3A', and x'3B' into program interrupt code 4, so an ESPIE established for interrupt code 4 will also get control for any of those access exceptions when they cannot be resolved. Ohhh my. I just asked engineering to set a SLIP as you suggested, Jim, hoping it will match. The above is bad news. I will have them to run the job with LE TRAP(OFF), but I suspect this will not help. I mentioned there it SmartRestart that our appications depend on. SmartRestart surely needs an ESPIE of its own to be able to recognize problems and act accordingly. If this is the case, I'm out of luck with the SLIP. Will have to check whether SmartRestart sets up an ESPIE or not. What is the reasoning behind excluding SLIP from matching when ESPIE is active? Is that information about how ESPIE interferes with ESPIE documented in detail somewhere? The MVS System Commands has only a single line saying that SLIP will not match if ESPIE is active. No mentioning of authorized versus nonauthorized, or what ESPIE does with the pgm check x'10', etc. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Follow-on: S0C4-11 abend caused by BASSM to address with all X'00' bytes
>If you can provide the PTF level of nucleus module IEAVESPI, we could set a SLIP SET,IF,N=IEAVESPI,xxx,xxx) in it at some offset where it has decided that it will be invoking an ESPIE exit. I may consider this and will come back to you in that case. Thanks. >SLIP gets called from RTM. Of course, now that you remind me. How could I forget about this :-( I assume this is true only for error type SLIPs, other SLIPs are recognized in other parts of processing. PER type SLIPs are probably handled by PGM check FLIH, since A PER event is a PGM check event. RTM is not invoked for those; they are not errors. Message ID traps are maybe handled in WTO processing. I may as well be completely wrong with this. SLIP command description has the following note for the COMP= operand: Note: 1. The SLIP action is not taken when the abend completion code is originally a program check (code 0C4) that the system converts to a new value. The following abend completion codes may be originally a program check and converted ones: 11A, 12E, 15D, 15F, 200, 212, 25F, 279, 282, 42A, 430, 57D, 700, 72A, A00, B00, and E00. This is not all too clear for me. Firstly, why is there the text "(code 0C4)" in the first sentence? Is this meant as an example, or is it meant to say the note only applies to 0C4s? Secondly, is the list of abend codes in the second sentence the complete list of completion codes which fall into this category? If so notge 1 would not apply to a PGM check 4 which was percolated and finally lead to an S0C4-4 abend so the abend would be trapped. The same would apply to an PGM 11 converted to a S0C4-11. Can you clarify, please, Jim? >I have not found any such documentation. There is probably room for some documentation improvements in that area. Shall I open an RCF? If so what manual for? -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Follow-on: S0C4-11 abend caused by BASSM to address with all X'00' bytes
-- Peter Hunkeler >SLIP gets control in 4 environments: >-- PER >-- RTM1 (think "FRR") >-- RTM2 (think "ESTAE") >-- MEMTERM How about SLIPs with keyword MSGID=. We've been asked by IBM support in response to a PMR to set a such SLIP to get a dump when a specific message was issues. This is how I learned about the MSGID= type SLIP. Seems to be neither error nor PER type of SLIP. The manual describes SLIP being called as part of WTO procesing. >Completion code applies to the last 3. As to the question, in practice, it >is typically only for some sort of 0C4 (whether PIC 10, PIC 11, PIC 38, >etc) that routines convert that program check to something else, such as >"I tried to access storage identified by the caller, but failed, so >provide a nicer completion code for the user to deal with than the generic >0C4". It could be any program check, it just happens not to be. > >SLIP gets control before recovery routines, so any completion code set by >the recovery routine is not matchable by a SLIP trap. Stupid me. I completely misinterpreted the text in the manual. >I know nothing about SmartRestart, but even LE recommends not using its >ESPIE path (and wisely so, in general). ESPIE is so restrictive that there >is no "surely" here. Are you recommending to run with TRAP(ON,NOSPIE)? The LE Programming Reference manual still says "IBM highly recommends running with TRAP(ON,SPIE) in all environments". -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Follow-on: S0C4-11 abend caused by BASSM to address with all X'00' bytes
>When a WTO message ID matches a SLIP MSGID, we issue a 06F-8 abend >to get into an RTM1 environment (for branch entry WTO) or RTM2 >environment (for SVC entry WTO). This allowed us to avoid the >development cost of creating a new SLIP environment. We retry >from the 06F-8 abend without recording to logrec, so the only place >it is easily visible is in system trace. I see. Setting a MSGID= SLIP kind of stores the message id somewhere for WTO to compare it to the msgid of WTOs, and issue an abend to invoke SLIP when matched. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: How does COBOL detect a recursive call?
> I'm a bit OCD about trying to make all my code RENT,REUS. My main way to > think of this is "would this still run correctly if it were burned into ROM?" > I try to make the answer to that YES. Put all your RENT modules into an APF authorized load library. No AC(1) required! The code will be loaded into subpool 252 key 0 storage. It will blow up with S0C4-4 if not really reentrant. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: RFE to add new sizes to AVGREC and IDCAMS Define
>I'd be happier with dispensing with AVGREC entirely and being allowed to code: SPACE=(1,(1M,1000K)) * SI (decimal) prefixes, please. Ohh, no. Please not another new syntax. The Early-morning-after-heavy-New-Years-Eve-party design of AVGREC is stupid enough. The first subparameter of SPACE= specifies the *unit*: number of bytes (aka block), tracks (TRK), or cylinders (CYL). I will never understand how IBM could think AVGREC is the way to go. Why not just add KB, MB, GB, TB, ... to TRK and CYL as space unit specification? SPACE=(GB,(1,3)) gives you 1GB primary and 3GB secondary. Simple to understand and remember. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: RFE to add new sizes to AVGREC and IDCAMS Define
>Once documented, this parameter would be more trouble to remove than it's >worth. Don't like it? Don't use it. But also don't carry it forward in any new >extension. I use it almost exclusively. Do I like it? Well, I got used to it. I did not say it should be removed. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: BPXAS sysout class
>Is there any way for me to get a bit bucket sysout class for BPXAS? My STCCLASS parm in JES2 is set to a regular output class and it looks like BPXAS uses that. I need something that will definitely work, since change BPXAS can put you in the pain cave. I just stumbled over this; has it been solved? Could not see anything in IBM-Main's archives. You might convert the BPXAS procedure to a started job by adding a JOB statement to it. On the Job statement you can then code MSGCLASS=. Something like //BPXAS JOB MSGCLASS=X,MSGLEVEL=(1,1) //BPXAS EXEC PGM=IEFIIC,REGION=0M,PARM=(,,&GETWORK,BPXPRJRW' Warning: I have not tested this (dont' have access to a system where I could). -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: BPXAS sysout class
> I'm a bit wary of testing this change. A broken initiator process would be > deadly. I think there is not too much risk to test this after IPL. The important UNIX process have been run. Make the change, shutdown idle BPXAS, login to a shell, issue some commands, see what happens. Backout the change if anything goes wrong. For the next IPL, if you're not in a SYSPLEX, you can prepare alternate members for MSTJCLxx and JESx (selectable via system symbols, selectable via LOAD member) to have alternative PROCLIB concatenations with the original and the modified BPXAS procedures, resp. But, yes, if the reular automated deletion via JES purge command solves the problem, sure go this way. Easier and no risk. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: PSF/AFP use wit COBOL
>We are in the process of moving 30 or so COBOL programs that use AFP from VSE >to z/OS. While we have a handle on the changes required to use PSF we are >having a somewhat difficult time compiling and 'LinkEditing' the programs.At >link time we get a ton of what appear to be C C++ unresolved referenced >messages. z/OS does not provide AFP support in the sense of programming interfaces that a program can call to generate AFP documents. Don't know if VSE has this, but I assume, not. This is why I think you've been using an add-on product to provide this. Its probably those (C/C++) routines that yor're missing on z/OS now. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Would HiperDispatch likely delay heavy multitasking job?
Environment is: - 6 * z13 model 712 - some 10 z/OS LPARs on each CEC - overall LCP to PCP ratio ~ 4- 12-way production sysplex has two 12 LCP LPARs on each CEC - prod LPARs have a ~38% share based on their weight => 4 vertical high and 1 vertical medium CPs - Neither CEC nor prod LPARs re overly busy (RMF shows some 40% - 60%) - CPU Activity Report shows Job in question is DB2 reorg utlility job which runs some 30 subtasks in parallel runs for 2 hours. RMF III (60s intervals) shows that the job mostly has a good workflow (80%+), is heavily using CP (often 80%-90%) and at the same time is heavily delayed for CP (40%-60%). The job is seen to use up to 230% CP. Question I've been asked: Would the job run even faster, if it was moved to a higher goal, higher importance service class. I understand that with HiperDispach=YES, WLM will assign work to the CP nodes it bulids, so the tasks will be dispatched on the same small set of LCPs, which (for the vertical high CPs) will always be dispatched on the same PCPs. Question: Is the assignment to dispatcher nodes by address space? In other words, would the 30+ subtasks of the reorg job all be assigned to the same node? Is so, the job would not be able to run more CPs in parallel at any point in time than there are LCPs in the node (highds, mediums, and unparked lows), right? Any thoughts? -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Would HiperDispatch likely delay heavy multitasking job?
>For more on these issues, you can see our presentations from the last SHARE on >our website at http://watsonwalker.com/publications/presentations/. And if >you're a Tuning Letter subscriber, look at our 2015 No 4 issue for both the >article that Mike mentioned and another article on how HiperDispatch works. Thanks. I'll have a look at your presentations as well. I was studying an excellent papaer from Robert Vaupel. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Would HiperDispatch likely delay heavy multitasking job?
>Cheryl's tuning letter fall 2015 described a case where they implemented MSU capping ... [snip] Ooops, I just saw the some text vanished from my mail... sorry for the incpmpleted data.. I wanted to add that the CPU Activity Report shows the system (LPAR) being busy between ~30% and ~50% (physically), and the whole CEC is nowhere near running at the physical (12CP ) capacity limits. We're running with Group Capacity Limit, and are also nowhere near being capped. So there should be enough spare capacity for the system (LPAR) to to use. But it does not. The 7 vertical low CPs are mostly parked or unparked for only a few percent. I wonder why MVS is not using more CPs and more Capacity for that job. But then there is the interesting comment from Greg: RMF might be fooling me. I understood RMF to count a CP delay sample when it finds a ready work unit on the WUQ (dispatcher queue). My (probably incorrect) thinking is that a ready WU should get access to a CP in this situation (spare capacity / parked CPs). This is why I'm interested in the mechanism that influence WU queueing and selection when HiperDispatch is on, i.e. when there is one WUQ per processor node. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Would HiperDispatch likely delay heavy multitasking job?
>>Is so, the job would not be able to run more CPs in parallel at any point in >>time than there are LCPs in the node (highds, mediums, and unparked lows), >>right? >I am confused by this question. Of course, regardless of the other conditions >that you've asked about, a job can never run more tasks in parallel "at the >same point in time" than there are Logical CPs available, which can't be more >than the number of physical CPs available at that instant. Sorry for not being clear enough. I understand that with HiperDispatch on, WLM creates "CP nodes" based on the topology. The LPAR in question has 12 LCPs, 4 highs, 1 medium and 7 lows. This would lead to 2 nodes with 2 highs each, the medium in one of the nodes and the lows spread (3 & 4?). WLM then creates a work unit queue (WUQ, or dispatcher queue) for each node. Work units (WU) are then assigned to a node, so only CPs from that node will select WUs from that node. I'm assuming that the assignement to a node is an address space attribute and not a WU atttibute (would neglect the positice effect processor caches for which HiperDispatch was invented). If all tasks of an address space are in the same node, there can never be more WU running in parallel than there are (L)CPs in the node. In my example, 2 highs plus the medium (for one node), up to 4 unparked (if unparked) low CPs. That would explain why RMF sees often CP delay even though there is plenty of spare capacity. If all this is true, would givint the job higher "prioritiy" really help? Caution: This is the first time I care to look into this in that depth. Maybe I'm all wrong. Happy to learn the truth. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: AW: Re: Would HiperDispatch likely delay heavy multitasking job?
>You have to dig into queuing theory, arrival rate, arrival pattern, and service duration of the work to understand why RMF is reporting CPU delay samples yet WLM does not unpark processors to run it. I think I have an idea of this even without knowing the theories. >HiperDispatch is a balancing act between providing the best processor efficiency vs dispatch responsiveness. If the average CPU utilization of the 3 CPUs is not relatively high *and* the delay for work to get dispatched is small, such that the cost/delay of using the vertical low CPUs is believed to not be worth it, then the vertical low CPUs will not be unparked. After reading and thinking about HiperDispatch in depth, this was kind of my conclusion and the reason I started this thread. I still have the pre HiperDipatch way of how dispatching worked in mind and had to lear it is all completely different now. That is why I started to doubt higher priority would help. I think now it would not or not much. As always with questions about complex things where the questioner has limited knowledge, it is difficult to find out what exactly to ask. >You can always try turning HiperDispatch off and see what happens. Not really. This is a rather large environment and it is only very few jobs that show this symptom. I doubt I will be allowd to switch it off, and more yet, I trust, overall, it would hurt than not. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: OMVS command history
The shell (/bin/sh to be specific) writes all commands into its history file .sh_history. This is kept across sessions. Command line editing when in a ssh (telnet/rlogin) shell session will use this file, as well as the "fc" and "history" shell commands. When using TSO/OMVS (the 3270 interface), the OMVS TSO command processor is the interface between the shell and your 3270 terminal. It is OMVS that keeps a history of everything entered *in memory* only. Once you logout, OMVS command processor will terminate and release its storage. There goes the OMVS command history. Note that .sh_history is still feeded when logged in through OMVS, but OMVS has no access to this file. HTH -- Peter Hunkeler On 25 May 2015 at 12:55, Ken MacKenzie wrote: > Hi All, > > Something's puzzled me for some time and I wonder if it is a quirk at this > site. > > When I go into OMVS and enter a command, I can then press F12 to recall > it. Repeatedly pressing F12 brings back older commands. That's great but > when I go out and come back in, pressing F12 does nothing until I've > entered a command. > > I notice that there is a file called .sh_history.mysuerid which contains > my command history. > > My question, then, is why is my command history kept if I can't use it? Or > is there something set up incorrectly here? > > > Ken MacKenzie > Pramerica Systems Ireland Limited > is a private company limited by shares > incorporated and registered in the Republic of Ireland with registered > number 319900 > and registered office at 6th Floor, South Bank House, Barrow Street, > Dublin 4, Ireland. > > -- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN > -- Mike Shorkend m...@shorkend.com www.shorkend.com Tel: +972524208743 Fax: +97239772196 -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: JES2 Exit 23
Long time ago (MVS/ESA 3.x in the late 80ies), I wrote JES2 and PSF exits to provide (part of) the accounting information as well as some other stuff to PSF. It was a combination of JES Exits 3 and 23, as well as PSF exits 1 and 2. At that time, I did not want to take the burden to extend any JES2 control blokc held on the spool, and since it was sufficient, I used some user fields in the JCT so save acoounting information in JES2 exit 3. Later, JES2 exit 23 aquired some storage (getmain) to store this data for PSF exits 1 and 2. PSF exit 2 (job end) was in charge of releasing the storage after use. This is from a quick refresh of my memory. Note also that I have not verified if there is an easier way nowadays. I'll send you some code snippets offlist. -- Peter Hunkeler Von: Lizette Koehler An: IBM-MAIN@LISTSERV.UA.EDU Betreff: Re: JES2 Exit 23 Datum: 31.05.15 07:49 I have not found any definitive info on how to extract the Accounting information on a JOB producing output and getting it to the APSUX01/APSUX01P exit. You might want to contact IBM directly via an SR for assistance. What version of PSF are you running? Are you using Assembler or C Language? Have you looked at the IEFJESCT to see what is available in the JES2 Communications Area? Lizette > -Original Message- > From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] > On Behalf Of Lizette Koehler > Sent: Saturday, May 30, 2015 5:28 AM > To: IBM-MAIN@LISTSERV.UA.EDU > Subject: Re: JES2 Exit 23 > > If you are not aware, there is also a JES2 List that might be helpful as well. > IBMMAIN or JES2 will be good. > > To join JES2 - use this URL > JES2 http://listserv.vt.edu/cgi-bin/wa?A0=jes2-l > > Have you looked at the JES2 EXITS manual? SA22-7534-13 z/OS JES2 > Installation Exits > > Lizette > > > > > -Original Message- > > From: IBM Mainframe Discussion List [mailto:IBM- > m...@listserv.ua.edu] > > On Behalf Of Markus Haselbach > > Sent: Saturday, May 30, 2015 12:38 AM > > To: IBM-MAIN@LISTSERV.UA.EDU > > Subject: JES2 Exit 23 > > > > Hallo, > > I?m for the first time involved with JES2 exits. The issue is > > printing with PSF. I have to write accounting information from the Job > > Card on the Job Header Page. This will be done in PSF Header Exit > > APSUX01. The question is how to get the information there. I would > > like to pass the accounting field by putting it in the JSPA in JES2 > > Exit 23 HASX23A. Can somebody give me a hint, how to get the Jobcard in > HASX23A, or an other solution? > > Best regards > > Markus Haselbach > > Credit-Suisse AG > -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: AW: Re: OMVS command history
>>but OMVS has no access to this file. >What about the history command? Well, as I wrote, the OMVS command processor has no access to the .sh_history file (because it has not been programmed that way). You can of course use the histroy shell command. But this is not related to what the OP asked about. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Cache Fast Write (CFW) - Data in the CFW is *not* mirrored, is it?
My GDPS mate asked us to make sure that cache fast write (CFW) on out (PPRC) mirrored boxes is disabled because it is not compatible with HyperSwap. This statement presumes that data in the CFW is not mirrored, is it? Up to now, I got somewhat inconsistent informtion about CFW in current DASD subsystems. Trying to get a clear view with the help of specialist here on the list. I understand that CFW can be switched on/off via IDCAMS. While this can be protected via RACF, this is not bullet proof. Some software (DFSort?) might write its own channel programm to set the switch. I therefore asked our DASD provider if CFW can be permanently disabled on the box. Got a "Yes, but nobody does it" Talking to someone else, I was told that modern DASD subsystems always use CFW; the user has no influence in this (??) Finally I haven't been able to figure out whether data in the CFW cache is mirrored *before* of *after* the data has been hardened to the platter. Appreciate any insight. Thanks in advance -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: AW: Re: AW: Re: OMVS command history
>>Well, as I wrote, the OMVS command processor > >There's more to OMVS than just the shell. The phrase "but OMVS has no >access" clearly refers to more than the shell. Not sure what you're referring to, but I was talking about the OMVS *TSO command processor*, only. And no, that statement you cited does neither refer to a shell nor to anything more than what I wrote: The OMVS TSO Command processor has not been programmed to read/write/care for any command history file any UNIX shell program might use. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Cache Fast Write (CFW) - Data in the CFW is *not* mirrored, is it?
>Where's Ron when we need him ?. Yep, but what does this help me? -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Is EXCP data from SMF 30 interval records repeated in step end records?
Please bear with me; I never had to dig into this detail on step end EXCP SMF reporting. I'm analysing a perfomance issue with long running (35 hours) a batch job. Looking at JESYSMSG, I stumbled across information that our IEFACTRT exit writes at step end. There is one step that seamed to do millions of I/Os against STEPLIB. STEPLIB is a two DSN concatenation, however, I see dozens of STEPLIB related I/O numbers Trying to understand, I started to have a detailed look at the SMF 30 records for the step. We do have interval recording, so I can see one subtype 1, a bunch of subtype 2, followed by one subtype 3 for the lat interval, and finally 1 subtype 4 records at the end of the step. Q1: The I/O counts in the interval records are *not* cumultive, but reflect the count since the previous record has been cut, correct? This is how I read the doc. Q2: The EXCP information that was written to the interval records is repeated in the step end record, correct? Q3: IEFACTRT will only see subtype 4 records, correct? Q4: If Q1-Q3 are correct, this explains why IEFACTRT listing I/Os by DDname and device can shoe dozens or hundreds of EXCP numbers for a two DSN STEPLIB. Thanks for confirmation or correction where due. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Is EXCP data from SMF 30 interval records repeated in step end records?
>Do you also see subtype 5 (Job termination)? Yes, but they are not of interest for the question at hand. >What is your DDCONS and INTERVAL setting? DDCONS is what I was missing, well actually, I forgot about it. Thanks for the reminder. We do have DDCONS(NO) -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Is EXCP data from SMF 30 interval records repeated in step end records?
>Do you have SAS and MXG at your shop? If so, then the MXG.SOURCLIB will >have information on what you are looking at. No, we don't. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Is EXCP data from SMF 30 interval records repeated in step end records?
>DDCONS does not do I/O, it only consolidates SMF I/O records, costing (much) >CPU and elapstime to >produce the records, but it will not add I/O counts to >the step. Yep, I understand it accumulates all the interval counts for a given DDname and device and writes only one entry with the total instead of merely repeating the individual entries from the interval records. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: OMVS command history
Thanks for this post. I was about to suggest to (mostly) forget about TSO OMVS
and start logging in via ssh. You've taken the burden off me :-)
>To edit your command history, press esc-k. Press additional k keys to get
>older commands. After you have the command you want, you will be in vi edit
>mode. After editing, press enter to submit.
May I add to the above? No matter whether you've setup to use emacs or vi as
your shell command line editor, the following apploes to shell command line
editing (I hope it really applies equally to emacs, since I only ever use vi).
Whenever the shell presents you the command prompt, think of it as if you were
editing the .sh_history file, and are about to add a new line to the end of the
file. You are in input mode. Whatever you type will be in that line. Hitting
the enter key will a) terminate that line and b) have the shell process that
command line.
Hitting ESC while at the shell prompt (remember you're in input mode) will take
you out of input mode. Vi calls this new mode the "command mode". Keys you type
will be interpreted as editor commands, while the current file in the editor is
still the .sh_history line. You now move up in this file with "k" and down with
"j". So this positions your "one line window view" (the command line) in the
file.
Once you see the command you've been looking for, hit enter to re-run it as if
it had just been entered (see above), or, use any editor command to move around
on the command line ("h" for one character to the left, "l" for on to the
right), or any editor command to start changing that line ("i" for insert
before the cursor, "a" for insert after).
There is a notable difference between command line editing and file editing.
Hitting enter whil in insert mode starts a new line in the later case. In the
former case it has the shell execute the command. So there is no need to ESCape
from insert mode into command mode, once you're done with command line editing,
torun the command.
Finally, don't be afraind from learing that bit of vi Kirk suggested. I'm an
old MVS guy and I did it when vi became available on MVS (OS/390 V1.x). I've
never regretted.
--
Peter Hunkeler
--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Statement of Integrity and UNIX.
>It used to be the case that a root user, or a user with access to root, might >be able to gain access to any >resource, if there happened to be another user, >with an OMVS segment, who had access to the resource. This is the reason BPX.DAEMON exists. What you wrote is true *if* BPX.DAEMON is *not* defined at all. When BPX.DAEMON is defined, which is higly recommended, then a process running with (e)uid=0 needs to also have READ access to BPX.DAEMON to be able to make an (MVS) identity switch to anyone (having an OMVS segment). -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
How comes SDSF HOLD panel shows one line per OutputGroup instead of per job?
This is embarrassing but I have to admit I don't have a clue what I did, and more importantly, how to back out. In SDSF's hold panel, I suddenly see one line for each output group (i.e. SYSOUT DD) for all of my jobs (well at least for this prefix setting). I don't see this woth other prefix settings. Those jobs display as usual with one line per job. I know I had some fingercheck error, but didn't notice exactly what it was. Just hit RESET and continued. Now this is my punishment ;-) RTFM and searching SDSF's help didn't give me a hint what I might have done. Any idea? -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: How comes SDSF HOLD panel shows one line per OutputGroup instead of per job?
>Any chance of a screen print/trap to show this phenomena? Yes, I could, but further investigation revealed that somehow the sysout dds for the jobs in question were somehow split into separate JES2 output groups. So it is obvious now, why each SYSOUT is display on a line of its own: They're independent JES2 output groups So the question has changed: What did I do to split these jobs from the initial single output group to one per SYSOUT? I have searched the log but could not yet find anything useful. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: How comes SDSF HOLD panel shows one line per OutputGroup instead of per job?
Solved! It was SMP/E who fooled me. I'm sorry to not have recognized this before posting. I had an SMP/E job doing some LIST processing for one zone. All DDs except from the CSI were dynamically allocated. I repeated the LIST statement in the same step to list from some more zones. SMP/E allocated and deallocated for each LIST, which lead to a set of new output groups each time (SMPLOG, RPT, LIST, OUT). I never seem to have done this before, or, my memory is failing -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: How comes SDSF HOLD panel shows one line per OutputGroup instead of per job?
>The multiple output groups are caused when you do a SET BOUNDARY within a >step. If you use a different step for each SET BOUNDARY, then a single output >entry appears for the job. Yep, this is what I did: multiple lists each for a different zone. I added DD statements to this job for the SMPLIST, LOG, OUT and RPT to get all output in a single output group. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Problem with FTP into a PDS
I'm not a big fan of the "Irisch commands" (OEDIT, OGET, etc.), but for the problem at hand, OGETX with the SUFFIX() operand might be useful. All suffices must be the same, and the remaining name must be a valid member name. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: SDSF and Carriage Control
>Ah! Here it is!: >asa - Interpret ASA/FORTRAN carriage control But it's an incomplete implementation. Doesn't support the "-" (space 3 lines) ASA CC. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Q: FTP Exits
>I've been asked to find a way to check FTP usage in our z/OS >environment (2.1 at this point) to see if users are specifying IP >addresses or names to be resolved. IP packets contain IP addresses and ports, not DNS names. The receiver cannot tell whether the user behind the client specified a hostname (to be resolved by a DNS lookup) or an IP address directly. If the former, then the client software must resolve the hostname via DNS lookup in order to start a TCP session via IP address. With HTTP, the client sends how the server was addressed to the server. So the web server knows whether the user entered an IP address or a hostname. This is the basis for multihoming, aka. VirtualHosts in web servers. I don't know but doubt that this information is part of the FTP protocol. If not, what you desire seems not possible to me. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: SDSF and Carriage Control
>IOF does and has for at least 14 years support carriage control emulated >display (both RECFM=A or M), except for overprinting. Pardon my ignorance, but what is IOF? -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: AW: Q: FTP Exits
>Never underestimate the power of an annoyed developer. There is a way to get >to FTP's PARM= value. I'd never dare to. My point was that if it is not part of the protocol definition, a client may or may not send additional information. However clever a developer might be, it would only help if user had no way but to use his invention. That surely is not given for FTP. I'm gonna have a look at some Wireshark traces, and at the FTP RFC(s). -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: SDSF and Carriage Control
>Is "-" part of the ASA standard or is it an IBMism? --
Peter Hunkeler
>It's pretty hard to find a definitive document. I'd have greater trust in one
>that wasn't published by IBM.
While reading RFC 959 (FTP) for some other reasons, I stumbled across the
following text:
3.1.1.5.2. CARRIAGE CONTROL (ASA)
The file contains ASA (FORTRAN) vertical format control
characters. (See RFC 740 Appendix C; and Communications
of the ACM, Vol. 7, No. 10, p. 606, October 1964.) In a
line or a record formatted according to the ASA Standard,
the first character is not to be printed. Instead, it
should be used to determine the vertical movement of the
paper which should take place before the rest of the
record is printed.
The ASA Standard specifies the following control
characters:
Character Vertical Spacing
blank Move paper up one line
0 Move paper up two lines
1 Move paper to top of next page
+ No movement, i.e., overprint
Then looking at RFC 740 (NETRJS) I read in appendix C
Carriage Control
The carriage control characters sent in a printer channel by NETRJS
conform to IBM's extended USASI code, defined by the following table:
CODE ACTION BEFORE WRITING RECORD
Blank Space one line before printing
0 Space two lines before printing
- Space three lines before printing
+ Suppress space before printing
1 Skip to channel 1
2 Skip to channel 2
3 Skip to channel 3
4 Skip to channel 4
5 Skip to channel 5
6 Skip to channel 6
7 Skip to channel 7
8 Skip to channel 8
9 Skip to channel 9
A Skip to channel 10
B Skip to channel 11
C Skip to channel 12
So, I conclude that the "-" indeed is an IBM extension.
--
Peter Hunkeler
--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: vary ,console with an IEF238D outstanding.
>Can't know for sure it's a bug. There might be a very good reason why an >exclusive ENQ is necessary, but I don't see it. Especially considering it's >apparently not necessary for similar operations... If you don't mind the effort, open a PMR, describe the situaiton including the ENQ contention you see and ask why it is necessary. I'm sure, IBM will look into this and maybe find out that the exclusive ENQ in not needed. Or maybe it is needed, but then they should explain why, so we can understand another bit of it. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: SMF Type 30
>My belief is that the IBM developers know what a task is and would not have >included "task" in the field name if the field applied to anything other than >a task. They sure know and they sure know that a UNIX processs is just a standard MVS task with some additional attributes. If a program running as a UNIX process does, for example, a local spawn(), it creates a new process in the very same address space. But this process needs to be tied to a TCB in order to be dispatchable, so a new task is created as part of spwn() processing. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: LE Condition Handler
> Retry is not possible, though, and if the operator issues a second cancel > command the non-APF program cannot trap that one, but the first one can be > handled. And this very fact disqualifies an x22 recovery routine as the solution for such an important task as writing data out to disk. Been there seen that: Operators issuing "C SOMEJOB" some 50 times within 10 minutes or so (no joke, and that's not the whole story) against some hanging address space which did not get cancelled. With this experience in mind, I would never consider an ESTAE exit to be able to reliably do lengthy actions on a x22 abend. And I consider writing important data out to disk to be an endless lengthy operation. Alternatives: - Could you change the code to use a coupling factility structure to keep that index data? - Could you change the code to use a system logger logstream? -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Anyone Using...PSF-Managed, Channel-Attached Printers?
A former employer of mine was heavily using a couple of ESCON channel attached Océ printers under PSF. They changed the interface to TCP/IP around 2011 (roughly). Still heavy PSF users but only with TCP/IP attached printers. --Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: AW: Re: SMF Type 30
>I believe I have seen it discussed that "task" is more closely akin to >"thread". Threads also need TCBs under the hood. Processes and threads are both ways to perform multitasking, so both need TCBs. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: ENQ rname_addr description
> The RNAME must be from 1 to 255 bytes long, and can contain any hexadecimal
> character from X'00' to
X'FF'.
To me a character is a character, thus a "hexadecimal character" would be one
of {C'a', ..., C'f', C'A', ..., C'F', C'0', ..., C'9' }. Why not "... 255 bytes
long, and can contain bytes of any value, i.e. x'00' to x'FF'"
--
Peter Hunkeler
--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Dataset in use on shared DASD
> D GRS,C This is of no value when there is no contention. Nothing will be returned. You might have been thinking of this command: D GRS,RES=(SYSDSN,data.set.of.interrest) But even this command will at most tell you something about whether the dataset is currently allocated or not. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: FTP - how get RDW and ASCII
> Not Unix, but... Not UNIX either, but... //ICONV EXEC PROC=EDCICONV, // INFILE=?FRED.INFILE?, // OUTFILE=?FRED.OUTFILE?, // FROMC=?IBM-037?, // TOC=?ISO8859-1? -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: How to output afp datastream (was How to output arp datastream)
> Is there any way I can output the afp data stream to a qsam file format？and > transmit the dataset to other platform! Not sure what your problem is, and what you consider a "QSAM file format". An AFP datastream, when created on z/OS is nothing more than a sequence of variable length records. Rendering software, such as ISIS Papyrus, write the records using plain normal z/OS, i.e. QSAM. When is comes to send the data to some other platform, just use FTP in *binary* mode. When you receive AFP datastreams generated on non-z/OS platforms, you must "reblock" the data. This "AFP" speak and a bit misleading. What it does is to split the byte stream received into variable length records. I may be of more help, if you can be more specific about what your problem is. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: How to output afp datastream (was How to output arp datastream)
> Actually, we want outsource the printing jobs to other vendor and they > require only arp format that why we want keep a AFP copy from our batch job You again write "arp" and not AFP. Is this intentionally? If so, I fist have to understand what "arp" is? In the meantime, I continue to think we're talking about AFP. In AFP, the datastream which contains the documents to be printed does not normally contain all the resources required to print the data. Fonts, overlays, page segments, from definitions are usualy kept in AFP resource libraries. You either setup some processes to send those resources to your partner, or you licensse the IBM ACIF utility to combine the datastream and the resource into a portable datastream. To capture the original data stream, you simpla change the DD statement in the jobs generating the datastrems. I assume the jobs currently write to JES (//anyname DD SYSOUT=..). Change this to //anyname DD DISP=(NEW,CATLG),DSN=your.dsn.goes.here,RECFM=VBM,LRECL=32756. You teh have the original datastream in a data set on your system, which will also be your copy to keep. Next, if required, ACIF this dataset. Next, send either the original or the ACIFed data set to your vendor. You can do this with any file transfer method tha both end support. Just make sure the transfer is *binary*" in any case. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
ISPF 3.2 (allocate) does not honor SDB
It seems I'm having a senior moment. For years, I used to allocate data sets in ISPF 3.2, leaving the block size field empty. The system then calculated the block size based on RECFM anf LRECL. The result was half track blocking. At my new employer, I get the block size set to maximum (for RECFM & LRECL), so FB-80 leads to 32720. This is z/OS V2.1. This is only via ISPF 3.2; in batch, I still get half track blocking. What am I missing? -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: ISPF 3.2 (allocate) does not honor SDB
>What type of dataset? PS, PO, PDSE, ...? The optimal block size for PDSE >(FB-80, DSNTYPE=LIBRARY) is 32720. Yes, its about PDSEs. Kind a makes sense, sure. However I have been using PDSEs for a long time and don't seem to remeber to have seen this. Must have been blind (meaning I didn't care to look to carefully), obviously. Did this change with z/OS V2.1? -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: AW: Re: ISPF 3.2 (allocate) does not honor SDB
>No.
>Check the archives ("Default System BLKSIZE for PDSE" in Oct 2006)
>"Partitioned Data Set Extended Usage Guide" (
>http://www.redbooks.ibm.com/abstracts/sg246106.html?Open ) Figure 10-32 shows
>a PDSE created 2004-11-04 with SMS.IND=R (SDB) and block size 32720.
How embarrassing. I understand I'm at an age where one starts to forget.
However, I don't understand I just never recognized this behaviour. Anyway,
thanks for the pointer to that thread. Interesting reading.
--
Peter Hunkeler
--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: REXX Exec Abend S106
>However, since you are running them using TSO CALL command, your programs are >already being ATTACHed. Therefore using ATTACHMVS would not give any added >benefit in terms of storage cleanup. I think you'll need to get a dump and >figure out what storage isn't being freed. Storage is very often obtained in subpool 0, and this subpool is owned by the job step task. So it will only be freed at job step end. So neither ATTACH of ATTACHMVS would help. Does the program beind called for each member use QSAM to open the data set? Does it FREEPOOL the buffer pool after close or use RMODE31=BUFF on the DCBE ? If not, the buffers will slowly eat up the available storage, since they are not automatically freed by CLOSE. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Ported tools and z/OS 2.2
> Maybe they are not available until the GA Date for z/OS V2.2 which is Sept > 30, 2015 A few months ago, we asked IBM this very question. I've changed to another employer in the meantime, thus I don't have access to the original answer anymore. However, it was basically that Perl, PHP and OpenSSH were removed from maketing. Suggestion was to download from Rocket Software. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: AW: Re: Ported tools and z/OS 2.2
>Peter's post is only partially correct. It is true that Perl and PHP have >been moved to Rocket. I stand corrected. But in my defense I wrote I don't have access to those documents anymore. I still think it was three components that were removed from the ported tools. >Also, the Apache web server is still available from IBM. More than that, Apache will replace the current Domino Go baesd HTTP server in z/OS V2.2. So Apache will be a base element apart from also being available as part of the ported tools. As per z/OS V2.2 docs, the base element is Apache V9, and the poerted tools version is 8.5. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
What are SMF30 fields SMF30HVO and SMF30HVH telling me?
I've got a new job which forces me to think about such things as the topic at hand. Please bear with me. I (think) I understand how virtual and real storage, below, above the line and bar work. In course of analysing a production problem related to paging, I'm searching the SMF30 haystack for hints on what's going on. I'm looking at the storage section and am having difficulties to understand what SMF30HVO and SMF30HVH are telling me. I thought I had understood, but as I see HVH values to be significantly higher than the HVO values, I'm confused. I've searched the archives and Google but could not find useful information. >From the description of the fields in the SMF manual, I would have expected >HVO to be at least as high as HVH. Anyone to educate me of point me the some good doc on this? -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: What are SMF30 fields SMF30HVO and SMF30HVH telling me?
> Might it be OA48512? Looks so. Thanks -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
CICS suffering from Page-In delay: Page Stealing versus Paging Out
A word in advance: I have subscribed to the MXG-L list but have not yet granted access, so I'm asking here first. I'm busy working on a problem some of CICS region are suffering from periodically. There are times when some CICS regions are paging-in some MB of storage and are thus suffering from the paging delay. Since this is about page-in, the storage that is paged-out has been in use before. It is not too clear yet what this storage is and what is causing it to be referenced again in that massive way. That's being investigated and is not directly my question here, although any hint is welcome as well. I'm trying to understand why that amount of storage has been paged-out from those CICS regions, and when this has happened. looking at the SMF30 subtype 2 for the CICS regions a indeed can see intervals with large number of pages being paged-out (SMF30PGO). I do not see any non-zero values indicating pages had been stolen (SMF30PST). It has been my understanding (which may be wrong, of course) that page steaqling happens when the Available Frame Queue goes below the min threshold. SRM/RSM will them steal pages and page them out. I would expect those numbers to appear in SMF30PST. And if this number is 0 for an address space, this one was not a donor in a low AFQ situation, right? We do see the AFQ going low sometimes, and it was my assumption that those CICSs were victims by having pages stolen. This does not seem to be the case. What then causes the pages to be paged-out (SMF30PGO)? Is CICS doing PGOUT by itself? -- Peter Hunkeler -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Copy a directory structure with copying files.
I'm adding some pax options as follows: a) I want the target to inherit all source attributes:==> pax -rwE -pe -XCM source target b) I want the target to keep source attributes, except for ACLs (if an) which should be set baset on the target environment's ACLs:==> pax -rwE -ppx -XCM source target I like to see what's been copied, therefore the -E option And don't forget that you may need appropriate rights, depending on what you intend to do. E.g. setting extended attributes (APF, program controlles, sharelib) require access to BPX.FILEATTR profiles. copytree can be much lower than pax. Both are much slower that an IDCAMS REPRO, which is my preferred option *if* the part being copied is contained in a separate file system. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: CICS suffering from Page-In delay: Page Stealing versus Paging Out
> There is an option in the WLM to make a service class memory critical. Yes, I know and I have verified that we do *not* used it. Why? Good question that I will have to raise. I'm very astonished the CICSs are not set storage critical. But I'm new to this installation and don't know historical information available, yet. Apart from this, I have to live with the current setup and need to find answers to my previous questions. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Copy a directory structure with copying files.
I'm sorry for the bad formatting. Trying again I'm adding some pax options as follows: a) I want the target to inherit all source attributes: ==> pax -rwE -pe -XCM source target b) I want the target to keep source attributes, except for ACLs (if an) which should be set baset on the target environment's ACLs: ==> pax -rwE -ppx -XCM source target I like to see what's been copied, therefore the -E option And don't forget that you may need appropriate rights, depending on what you intend to do. E.g. setting extended attributes (APF, program controlles, sharelib) require access to BPX.FILEATTR profiles. copytree can be much lower than pax. Both are much slower that an IDCAMS REPRO, which is my preferred option *if* the part being copied is contained in a separate file system. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: IEAVAPE2/IEAVPSE2 another address space
>> What the big deal I didn't want to use my real name never saw that as one of >> the requirements for >> Using IBMMAIN >> > My real name is Joseph Reichman What a big deal?? Do *you* like to communicate with people who think they have a need to hide their real personality? I definitely do not. There are people who post on lists without mentioning their name. Bad enough. But signing in the name of someone else? And who can we be sure Josef Reichman isn't just another name you've arbitrarily chosen? I had not replied to any post of you since long time, and I will continue to refrain from doing so. This a one-time exception. You don't seem to be honest, neither about you, nor about you programming intent. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
RMF III Frame Occupancy vs. Frames WSET
Cross-posted from MXG-L (and extended) The RMF guide describes the "Working Set" to include frames backing Dataspaces and Hiperspaces. I assume the "Frame Occupancy" does only include address space frames. Is this correct? The RMF manual is not clear about this. My assumption is based on the fact that I do see entries with much higher "Active Frames WSET" than "Frame Occupancy", and I do know that the job is using Dataspaces. Are shared Dataspace and Hiperspace frames counted for the onwing address space, only? Assume address space A (AS-A) creates some Dataspace (DS-A) and shares it with AS-B. AS-A stores 2MB of data into DS-A. At that time 500 frames are assigned to the Dataspace and counted in AS-A's WSET number. Assume some frames get stolen from the Dataspace. Some time later, AS-B accesses the all data in DS-A, i.e. all 500 frames are needed, leading to page-in of the once stolen pages. To whom are those frames accounted? True or False? - The ones not stolen had been counted for AS-A when AS-A initially wrote into DS-A. - The ones not stolen are *not* counted for AS-B when accessing DS-A. - The ones now paged-in are counted for the *owing* AS-A. - The ones now paged-in are *not* counted for AS-B when accessing DS-A. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
RMF STORF "Frame Occupancy Total"
Cross-posted from MXG-L Looking at RMF III STORF reports, I'm trying to understand what the report is telling me, especially the frame occupancy. It confuses me that field "Frame Occupancy Total" (short FOT hereafter) is described as being the *sum* of fields "Frame Occupancy Active" (short FOA) plus "Frame Occupancy Idle" (short FOI). Here is why... I understand at each sample RMF sees a job either active (using CPU) or idle (waiting for something). It looks at the number of frames and adds that to the corresponding field (and calculates the average over the interval). FOI is a subset of the FOA frames, right? if this is correct, then I'd expect FOA to always be higher than FOI (unless a lot of new frames have assigned to the job just after a sample and the job goes idle before the next sample). If FOI is a subset of FOA, what is the reason FOT = FOA + FOI? -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: z/OS version by Sysres Module
> Is there a module within SYSRES dataset's which can help me to determine the z/OS version ? > This Question is just for the Knowledge sake and not trying solve any problem. "D IPLINFO" gives you that and more. Or look at CVT field CVTPRODN. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: PL/I and optional parameters
> I have a PL/I subroutine call that has 297 optional parameters. That is the > function has a parameter list with up to 100 tuples of parameters. The first > tuple is required, and tuple 2 through 100 are optional. [snip] > Is there a shorter way of declaring 297 optional parameters in PL/I? You may want to look at the "LIST attribute" of an ENTRY declaration in the Enterprise PL/1 Language Reference. It might require the Enterprise edition of the complier, I'm not sure it was supported on elder versions of the comiler. There are builtin funrctions that help you find how many optionals have been specified and where they are. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: tcpip reserved ports
> Ah! That did the trick. I was playing around with the D,TCPIP,TCIP,... > command and could not find what I needed, thanks. You can also get the list via console command: D TCPIP,TCPIP,N,PORTL -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: OMVS mount failure
> I think in TSO/ISPF Option 6 you can enter BPXMTEXT EF096058 "BPXMTEXT" is a rexx that resides in SYS1.SBPXEXEC. Provided you have this library in the SYSPROC or SYSEXEC contcatenation of your TSO forground or batch JCL, you can issue "TSO BPXMTEXT " from any TSO command prompt, ISPF command line, or other REXX program. Note that it is also a z/OS UNIX shell command, so simply use "bpxmtext " in a shell session (watch the case!). It is sometimes worth giving it a try when you get some 8 character reson code or err info code in some context which does not look like z/OS UNIX in the first place, such as TCP/IP messages. Can tell a sample fro the top of my head but I remember to have been able to quickly answer the question "what the h* is this message telling me". It may, however, give you falls answers in case the reason code happens to look like a z/OS UNIX cod when it it not. So be carefull. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: AW: PL/I and optional parameters
> The earlier post by Peter Hunkeler was referring to the LIST attribute of > ENTRY, described here: > http://publibz.boulder.ibm.com/cgi-bin/bookmgr_OS390/BOOKS/IBM3LR60/6.10.5 Yep, this is what I referred to. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Subsys not down while Shutdown
> Thanks Lucas & Lizette. I have terminated the Task from list and performed > the Load. > Now is systems is up and running fine. Great. Now that this has been sorted out, you should invest some time to find out what ought to be brought down before an IPL what not. Automation (if you're using automation) should know and do what is necessary, but you should be prepared for the case when something does not work as intended. There are tasks (STCs) that you better try to stop in the proper way, or you might get into troubles at next startup (e.g. DB2 which might need some long lasting recovery). On the other hand, there are some tasks which you really don't need to care about, and there are tasks which cannot be stopped at all. The list is depending on your setup and the products you use. It is also useful to have a list of tasks to start, and in what order, in case you need to perform an IPL manually. Again automation may fail to complete the startup for whatever reason. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: BPXWUNIX and sort
>The problem that I am experiencing is that the data is being sorted using the >ASCII collating sequence (i.e., numerics precede alphabetics). I understand the UNIX sort shell command always sorts according to the *collating sequence* defined in the current locale. Is't not ASCII and it's not EBCDIC, it's not a sort in the binary content of the bytes. I had not looked into this in detail before but it seems that the C-locale (the default) defines that digits collate before upper case and upper case before lower case. Other locales do otherwise, e.g. De_CH seems to define digits collate before letters, and lower case and upper case letters collate equally, i.e. ther is no diference between 'a' and 'A'. You should be able to see the difference is you add a stem containing one variable "LC_COLLATE=..." to your bpxwunix call. Now with all this said, I can't come up with a solution to your problem, namely sorting the stem in strict binary sequence. -- ßph -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: CICS 5.3 and transaction isolation questions
>Got my answer from the CICS list. I'm not subscribed to the CICS list, yet the answer interests me (and maybe others on this list). Do you mind posting it here so it will be in the archive? -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: curious: Popularity & use of C on z/OS.
>https://www.javatpoint.com/features-of-java Everything written on any web page is true, or what? C is procedural programming language and as such it can be compared to COBOL, PL/1, Pascal, etc. But comparing a procedural language to an object oriented language just makes no sense to me. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: IEFA111I at IPL Time for z/OS V2.3
Lizette,do you consider this message to indicate a potential problem? As I read it, I think it is purely informational. Does it matter where the values come from? I read it as kind of debug help information in case you run into a problem that could be related to one setting or the other. We're not on V2.3 yet, so I can't check myself, but I would expect the message to appear in the joblog of any and all jobs, started tasks, and TSO sessions. Interesting piece: Even IBM KC shows only DGDBIAS, and non of the other parameters. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: SYSPLEX distance
> I'm still not clear on how a 'geographically dispersed sysplex' (original > definition, not 'GDPS') would work. You say "original definition". I seem to remember, but might be wrong, that the term GDPS was coined when sysplexes were al contained within a single building or in buildings near by. GDPS was taking sysplexes with members in data centers up to a few kilometers apart. Apart from the longer distance between members, they were sysplexes as usual. No XRC involved. --Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: DFSORT for HTTP logs - RECFM and BLKSiZE?
>I do believe that I should be able to point DFSORT at any file (Unix or not) >where performing the functions of DFSORT make sense, and not be required to >specify in JCL the physical attributes of LRECL, RECFM and most especially >BLKSIZE. These should all be derivable from the physical dataset at open time. When accessing a z/OS UNIX file with JCL PATH allocation, OPEN will setup everything to read the data from the file system and return it as if it was a data set record when READing thereafter. The difficulty for this process is to invent RECFM, LRECL, and BLKSIZE, because all of this is unknown to a UNIX file. How would OPEN know whether all "records" in the file are of same length without reading the whole file, so it can set RECFM=F? How would OPEN know how long the longest record is without reading the whole file, so it can set LRECL? DFSMS Using Data Set says the defaults are: RECFM=U, LRECL=80, BLKSIZE=80. So it is up to the program to setup the DCB as expected, or to specify all of this in JCL (or SVC99). Just my $0.02 -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: AW: Re: IEFA111I at IPL Time for z/OS V2.3
>I've arranged to have the System Messages documentation updated. Thanks, Scott, for having the doc updated. Saves me an RCF. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Only very old matches when searching the archives
I'm trying to search IBM-Main's archive but am getting only matches from 2007 and elder. Even searching for words I can see in current threads do not show up. Is anyone else having the same problem this? -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Only very old matches when searching the archives
>So, since you're finding information from 2007, I'll assume you're in the >right archive, (a). Yep, I've verified this. >How are you searching? Via the web, e.g., from the Search box or Search link >on the right-hand navigation here: >https://listserv.ua.edu/cgi-bin/wa?A1=ind1801&L=IBM-MAIN Via web. I treid the different places where you can start searching. Same retsiult. >Or by sending commands to the listserv server address? I will try this. >Via the web, from what I can see, you will be presented with the 100 oldest >entries that match your search term. You can advance to 100 newer entries that >match by selecting More Hits at the bottom of the list. And you can keep >advancing forward, 100 entries at a time, by continuing to select More Hits. This can well be. I don't seem to be able to change that, I mean to start from today and seach backwards. The sort field says "date/time" most recent first", but that may be missleading. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: AMASPZAP usage
>Of course I know the offset and old data. What I cannot predict is disk address and in fact cannot understand why AMASPZAP wants me to provide unnecessary information. I think you may misunderstand the parameters. The offset on the VER and REP statements is the byte offset from the beginning of the physical record, aka block, you told SPZAP to look at. How would SPZAP know which block to work on, if you don't tell it? The CCHHR is exactly this information. -- Peter Hunkeler -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Only very old matches when searching the archives
>How many hits does it say that there are for your search?
Well, that is part of the mystery. I was searching for words that must appear
in post, such as "storage", and I got "No Matches".
In the meantime, I tried a search via mail as show below. Instant reply with a
long list of hits, with an overview at the tops, instructions to get all post
in detail, and more detailed information from each post at the end.
An example of an advanced search string:search ('storage' or 'initialize' or
'clear') in ibm-main since 2015 where sender contains phun
I'll use this interface in the future.
--Peter Hunkeler
--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Only very old matches when searching the archives
>Did you check both archives? There is a current on and older one. No, I tried searching IBM-MAIN because I knew the post I'm looking for is not older than two years. Anyway, as said in a previous post, I'll be using the "command line interface", i.e. searching via mail sinc it has much better search options. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Only very old matches when searching the archives
>This is WAD. You are probably not using the search tool properly. Can you tell me how to use it properly. Starting with the oldest matches from 2005 and repeatedly clicking on "more hits" is not useful if you look for a term that appears very often. But I understand from previous answers, this is how it works. Fine. Done with the web interface. Switching to command line interface (as I already mentioned). >Look at 'Advanced Options' in 'Search Archives' on the IBm-MAIN website. The advance options page has a "String" field and a "Sort by" selection drop down. Not sure who this would help (and, btw, I did use this). -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Extended format internals
>Is is possible to print the content of additional 32-byte sufix using >AMASPZAP or maybe other tool? Would ADRDSSU's PRINT command be of help here? It content of tracks, block by block. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: codepages
>I have an issue where a customer is not forthright in providing us a the codepage of data coming into a TCP STC on z/OS ..We bring the data in via a Scoket-Read,. Is there any way I can query the datastream and find out what codepage it is ? There was a lot of advice about how to *guess* on the encoding of the data. I for one have a strong opinion on who is responsible to unambiguously declare all attributes of data: The owner of the data. In your case this is the customer. When it comes to text data, the code page the data is encoded with is one of those attributes, maybe the most important one. I would never take responsibility for correct processing of text data if I'm *not* told the code page the text was encoded with. Sure, you can take a look an guess, and this might be easier in some geographies. Here in Europe, there are too many similar yet different code pages in use. You might not look at a snippet of data containing some of those variant characters. Or, more likely, you see those characters. How do you decide whether the glyph show with code page A or the one show with code page B is the correct one? Your customer needs to understand its his/her responsibility to tell you. -- Peter Hunkeler -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Dataclass question
>I probably would not have DSORG=PO sharing the same MGMTCLAS with batch files. Interesting differentiation. I do not contradict per se, but attributes should be different for DSORG=PS versus DSORG=PO, and DSORGP=PO-E. >I agree COND IMMED or CONDITIONAL is more appropriate for PDS. No, space release is not appropriate for PDSs in any form. PDSs are still limited to 16 extents (if memory serves me right), and each new member as well as each update of existing members will write at the end of the data set's allocated space. After space was releases, each write will add another secondary extent. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: AW: Re: Dataclass question
>What is your concern with secondary extents on a PDS? I don't have a problem with secondary extents as such. It is only that I fear the 16 extent limit may me reached sooner with space release. Extent reduction may help, but only if the device has low "extent" allocation activity. Otherwise, chances are lower to get an adjacent extent. >I don't think my "more appropriate" is an absolute in the same way as "not >appropriate for PDSs in any form." Yeah, my statement was probably too absolute :-) I just don't see a problem with highly overallocated space of PDSs. YMMV, though. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: codepages
>Browse could be useful for a file too large to fit in main storage (Nowadays? >Really!?) ITYM virtual storage rather than main storage. VIEW/EDIT want to be able to read the whole file into the extended private area (below the bar, that is). So this limits the size of the data set to some 1.4 - 1.5 GiB. VIEW/EDIT should really learn to work in a window mode, so larger data sets can be viewed/edited. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Question regarding IPCS RUNCHAIN: Why does it not run the chain?
Long time not done, I'm writing some REXX code to format data in an IPCS
session. The code successfully formats some header data but fails to run a
chain of control block. IPCS RUNCHAIN calls the separate REXX routine which
formats one such control block. So far so good.
I can't seem to find what I'm doing wrong, but RUNCHAIN stops after the first
round. The link address is at offset 0, and I see it is not zero. I have
reduced the REXX named in the RUNCHAIN to "say hello;return" but still no joy.
What am I missing"
/* REXX */ /*some REXX and ICPS statements here*/ "RUNC" "AD(" DSHHEAD ")
LI(0) EXEC(( %I$CCRCBH" "X" ASIDhex DSPName TRACE "))"
and the REXX subroutine I$CCRCBH starts with /* REXX */ Say "Hello" return
Sample output seen in IPCS: DSH: Data Space Header at address 0. +
INIT. 00072A8F E36E28D6ALET. 01FF003C ORIG.
+0010 SIZE. 37112810 FREE. 37112810 HEAD. 0338 TAIL.
37112608+0020 HSTA. 0030 HSTL. 00C2 DMPT.
LIST 0338. ASID(X'01E4') LENGTH(X'1000') AREA 1 block processed
There is a single "Hello" on the terminal before returning to IPCS. The real
code behaves the same: It correctly displays a single instance of the control
block, then terminates,
-- Peter Hunkeler
--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: codepages
>>VIEW/EDIT should really learn to work in a window mode, so larger data sets >>can be viewed/edited. > >Or use storage above the bar. I'm not a fan of this. How much would you allow it to use? Loading a huge data set into the editor would allocate a huge number frames, just to drop them shortly (whatever that means) thereafter. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: codepages
>>>Browse could be useful for a file too large to fit in main storage >>>(Nowadays? Really!?) >> >> ITYM virtual storage rather than main storage. >> >OK. Picky, picky! I didn't mean to be picky. What I meant (but did not write) was the fact that large main storage does not help as long as TSO is not AMODE64 aware. >> VIEW/EDIT want to be able to read the whole file into the extended private >> area (below the bar, that is). So this limits the size of the data set to >> some 1.4 - 1.5 GiB. >> >Well, yes. How often is that a problem? Every now and then, do I have to look for the needle in a days worth of operlog, which was offloaded into a data set. It is extremely helpful when I can work with VIEW. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Question regarding IPCS RUNCHAIN: Why does it not run the chain?
>Peter, I don't see what's wrong, but I suggest running it with an IPCS LIST >command in the EXEC(()) first to assure the rexx is not misrepresenting what's >at offset 0. I will try tomorrow to see if it makes a difference. However, as I wrote the actual code does format the expected values which I retrieve via EVAL and EVALSYM. EVALSYM is used to retrieve the address behind the "current" address in symbol X. EVALs start at this address. But as I write this, I would be rather surprise when code inside the EXEC(( )) would influence the surrounding RUNCHAIN. The ADDRESS, LINK and CHAIN parameters are all RUNCHAIN parameters. An additional LIST would write something to the output which is not wanted. I wanted but forgot to mention that a similar pair, although CLISTs, does work as expected. However, there is a CBFORMAT inside the EXEC(( )) CLIST. Maybe this really makes the difference. Curious to see what happens when I add a LIST tomorrow. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Advice for WLM tuning of OMVS forked address spaces
>As you mentioned,it is important that the work that newly arrives into the >system is properly classified and has not overly aggressive goals. However, >it is equally important that the other "important" is correctly classified >with a goal that is aggressive enough to protect against other work. This last >item is "usually" what contributes most to such unwanted effects. WLM makes it settings based on historical events, right? So, the service class that OMVS work is assigned to might currently have a higher dispatching priority when that bunch of jobs is submitted. It will take WLM a few policy adjustment cycles to correct that. Also, if many such jobs are submitted in parallel, WLM might have to start batch initiators for the jobs, and then at least two BPXAS initiators per job for the UNIX processes for at least a couple of jobs. Then, depending on how quick the UNIX work terminates, BPXAS inits may be reused for the other job's UNIX processes. I assume there is more UNIXish work being done under the ssh process, so another bunch of BPXAS may be required. You might want the customer to have a look at the operlog and find out if there is a spike of INIT / BPXAS being started. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: codepages
>In what way is TSO not AMODE64 aware? I run AMODE64 programs under TSO all the time. The talk is about ISPF VIEW/EDIT, and these programs do not do AMODE64 as far as I understand. You can of course alway run your own AMODE64 code in any address space. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: Question regarding IPCS RUNCHAIN: Why does it not run the chain?
>The point is that debugging the chaining issue with a list command might be >easier than with the rexx. Good point. This is why I had replace the actual code with a simply SAY, so that I know the code was called. But, I agree, one step at a time is better. So, I misunderstood. You did not want the LIST inside the REXX but instead of the REXX. That now makes perfect sense. Will see tomorrow. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
