Re: EREP Symptom and/or Software Records
>From what we've gleaned so far, EREP is just 'holding the mirror' and some sort of WAIT is causing the CATALOG REDRIVE and is symptomatic of a tuning opportunity. From a sysprog perspective I'm going to review WLM, DB2 buffer use and device busy. Craig Mullins DB/2 series is great and suggests bad SQL is the root cause of a large percentage of DB2 problems. In a message dated 6/18/2017 1:40:12 A.M. Central Daylight Time, haresystemssupp...@comcast.net writes: It's been a few years (retired / consulting now) since I had the duty of looking at EREP records but as with what many others have said, it's about what seems "normal" in your environment, although I've had instances of 0C4 that was occuring regularly, though handled by the piece of software in question without killing anything. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: EREP Symptom and/or Software Records
It's been a few years (retired / consulting now) since I had the duty of looking at EREP records but as with what many others have said, it's about what seems "normal" in your environment, although I've had instances of 0C4 that was occuring regularly, though handled by the piece of software in question without killing anything. As for searching for APARs, you used to be able to copy those symptom strings into the search field such as: PIDS/5695DF105 RIDS/IGG0CLA9 RIDS/IGG0CLX0#L PRCS/00F6 PRCS/ and it was a pretty reliable way to zero in on something. I don't know if IBM's support site search still works that way, but if it does it saves quite a lot of time. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: EREP Symptom and/or Software Records
Most likely cause is SYSZTIOT or SYSZVVDS contention -- so the abend 91A are symptoms of that latent problem (which affects your SLA potentially). Catalog added contention management enhancement in z/OS 1.12 which causes abend 91A for excessive wait and when the wait time threshold is reached the abend "resolves" the serialization. Whether trivial or significant having hundreds per day depends on your specific configuration, but worthy of analysis. Lots of possible causes, particularly for DB2, so pointless to speculate further. On Fri, Jun 16, 2017 at 4:43 PM, Peter Hunkelerwrote: > > >Catalog processing is not my area of expertise, but I think that this > symptom record is the result of a catalog service task being ABTERMed with > code 91A, possibly by a catalog analysis task. I don't remember offhand if > a 91A is also used to terminate a service task when the suspended requesting > task gets ABTERMed. > > > > > So this seems to be part of some CATALOG health checking and recovery > processing. A good thing, basically, but if you have hundreds of those a > day, it may indicate some latent problem. Somewhere in your system setup, > or in some software calling CATALOG services, or even in CATALOG itself. > > > As Jim suggested, you may want to take one or a few dumps and try to find > out why you see so many of them. DAE will suppress duplicate dumps, i.e. > dumps where the symptoms (symptom string) indicate it is likely for the > same cause. You might want to change the SLIP to ACTION=NOSUP,MATCHLIM=5 so > that DAE will not suppress duplicates, until the number of dumps you want > has been taken (MATCHLIM=n). Without the MATCHLIM, each occurrence you > initate a dump being taken, which could kill your preformance. > > > -- > 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 > -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: EREP Symptom and/or Software Records
>Catalog processing is not my area of expertise, but I think that this symptom >record is the result of a catalog service task being ABTERMed with code 91A, >possibly by a catalog analysis task. I don't remember offhand if a 91A is >also used to terminate a service task when the suspended requesting task gets ABTERMed. So this seems to be part of some CATALOG health checking and recovery processing. A good thing, basically, but if you have hundreds of those a day, it may indicate some latent problem. Somewhere in your system setup, or in some software calling CATALOG services, or even in CATALOG itself. As Jim suggested, you may want to take one or a few dumps and try to find out why you see so many of them. DAE will suppress duplicate dumps, i.e. dumps where the symptoms (symptom string) indicate it is likely for the same cause. You might want to change the SLIP to ACTION=NOSUP,MATCHLIM=5 so that DAE will not suppress duplicates, until the number of dumps you want has been taken (MATCHLIM=n). Without the MATCHLIM, each occurrence you initate a dump being taken, which could kill your preformance. -- 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
Re: EREP Symptom and/or Software Records
Catalog processing is not my area of expertise, but I think that this symptom record is the result of a catalog service task being ABTERMed with code 91A, possibly by a catalog analysis task. I don't remember offhand if a 91A is also used to terminate a service task when the suspended requesting task gets ABTERMed. I would suggest looking at the syslog to see if there are any catalog messages at the times of these symrecs which might explain the situation. Absent that, I would do SLIP MOD,ID=X91A,DISABLE to get a dump of a 91A abend, and look at that. You may need to open a PMR to catalog for assistance with dump analysis. I would at least look at the SYSTRACE in the dump to see who initiated the 91A ABTERM. Jim Mulder z/OS Diagnosis, Design, Development, Test IBM Corp. Poughkeepsie NY IBM Mainframe Discussion List <IBM-MAIN@LISTSERV.UA.EDU> wrote on 06/15/2017 05:20:21 PM: > From: Turner Cheryl L <cheryl.l.tur...@irs.gov> > To: IBM-MAIN@LISTSERV.UA.EDU > Date: 06/15/2017 11:03 PM > Subject: Re: EREP Symptom and/or Software Records > Sent by: IBM Mainframe Discussion List <IBM-MAIN@LISTSERV.UA.EDU> > > I understand why you may have thought that but no I understand it is > not the slip that spawns the records. But couldn't it be said that > the slip parms are indicating IBM's view of the severity of the > event? I am so new to this that heck I may not be even asking the > questions right. For that I'm sorry. > > For example. Here is one symptom record in particular we are > constantly seeing (there are others but let's use this as an example): > > PIDS/5695DF105 RIDS/IGG0CLA9 RIDS/IGG0CLX0#L PRCS/00F6 PRCS/ > PRCS/ JOBN/DBP1DBM1< > the JOBN changes but they all seem to be DB2 related tasks. All > the other information is the same. > > SYSTEM ENVIRONMENT: > CPU MODEL: 2964 DATE: 166 17 > CPU SERIAL: 0207C7 TIME: 01:06:27.72 > SYSTEM: MBI2 BCP: MVS > RELEASE LEVEL OF SERVICE ROUTINE: HBB77A0 > SYSTEM DATA AT ARCHITECTURE LEVEL: 10 > COMPONENT DATA AT ARCHITECTURE LEVEL: 10 > SYSTEM DATA: || > COMPONENT INFORMATION: > COMPONENT ID: 5695DF105 > COMPONENT RELEASE LEVEL: 220 > SERVICE RELEASE LEVEL:UA82137 > DESCRIPTION OF FUNCTION: CATPROB DATA > > PRIMARY SYMPTOM STRING: > PIDS/5695DF105 RIDS/IGG0CLA9 RIDS/IGG0CLX0#L PRCS/00F6 > PRCS/ JOBN/DBP3DBM1 > > SYMPTOMSYMPTOM DATA EXPLANATION > ------- > PIDS/5695DF105 5695DF105COMPONENT IDENTIFIER > RIDS/IGG0CLA9 IGG0CLA9 ROUTINE IDENTIFIER > RIDS/IGG0CLX0#LIGG0CLX0#L ROUTINE IDENTIFIER > PRCS/00F6 00F6 RETURN CODE > PRCS/ RETURN CODE > JOBN/DBP3DBM1 DBP3DBM1 JOB NAME > > THE SYMPTOM RECORD DOES NOT CONTAIN A SECONDARY SYMPTOM STRING. > FREE FORMAT COMPONENT INFORMATION: > > And then there appears to be a snap dump of storage on each one. > > Nothing on IBMLINK matching anything that I can think to search on > from the fields. In the syslog we see IBM slip trap x91A taken > about the time of each record. > 2017166 01:06:27.72 0284 IEA989I SLIP TRAP ID=X91A > MATCHED. JOBNAME=CATALOG , ASID=0086. > > And there are sometimes 100s of this particular symptom records on a > given lpar, per day. > Slip settings are: > ID=X91A,NONPER,ENABLED > ACTION=NODUMP,SET BY CONS INTERNAL,RBLEVEL=ERROR,COMP=91A > > 91A > > Explanation: A request to abnormally end the catalog address space (CAS) > service task was issued either through the MODIFY CATALOG,RESTART command, > or through catalog analysis task processing. > > System Action: The system re-drives the catalog request currently in > process. > > We are not issuing a MODIFY CATALOG RESTART command at the time of > any of the logrecs being cut. SO might there something wrong with > the catalog process that all these redrives are necessary? Is it > normal behavior? So many questions and I'm clueless, unfortunately. > > So what I guess I was trying to wrap my head around is: if there > isn't a need to take a dump, etc. (as specified in the SLIP setting) > then why have logic to cut 100's of symptom records at all for that > particular issue? And if we're cutting 100's of records - is it > really a problem? And like Ed said, it's noise, and I don't know > enough to know it's a proble
Re: EREP Symptom and/or Software Records
Cheryl, So, IBM supplies many standard Slip Suppressions. the ACTION would be NODUMP. This is due to various events in z/OS that will just issue them, 91A, 133, D37, B37, E37 just because dumps are not really needed and "everyone" just knows what they are. I consider these annoyance and just part of the process. As for Logrec, IBM uses the philosophy, just in case - create a logrec record. Sometimes with storage sometimes without storage. This is incase you need to go back and look for symptoms about an event. If the Logrec data is kept longer than a day then you would be able be able to go and see if there are system control block information that could be helpful in diagnosing the issue. What I usually go by are the following guidelines 1) How often is the event occurring -Original Message- >From: Turner Cheryl L <cheryl.l.tur...@irs.gov> >Sent: Jun 15, 2017 2:20 PM >To: IBM-MAIN@LISTSERV.UA.EDU >Subject: Re: EREP Symptom and/or Software Records > >I understand why you may have thought that but no I understand it is not the >slip that spawns the records. But couldn't it be said that the slip parms are >indicating IBM's view of the severity of the event? I am so new to this that >heck I may not be even asking the questions right. For that I'm sorry. > >For example. Here is one symptom record in particular we are constantly >seeing (there are others but let's use this as an example): > >PIDS/5695DF105 RIDS/IGG0CLA9 RIDS/IGG0CLX0#L PRCS/00F6 PRCS/ > PRCS/ JOBN/DBP1DBM1< the JOBN > changes but they all seem to be DB2 related tasks. All the other > information is the same. > > SYSTEM ENVIRONMENT: > CPU MODEL: 2964 DATE: 166 17 > CPU SERIAL: 0207C7 TIME: 01:06:27.72 > SYSTEM: MBI2 BCP: MVS > RELEASE LEVEL OF SERVICE ROUTINE: HBB77A0 > SYSTEM DATA AT ARCHITECTURE LEVEL: 10 > COMPONENT DATA AT ARCHITECTURE LEVEL: 10 > SYSTEM DATA: || > COMPONENT INFORMATION: > COMPONENT ID: 5695DF105 > COMPONENT RELEASE LEVEL: 220 > SERVICE RELEASE LEVEL:UA82137 > DESCRIPTION OF FUNCTION: CATPROB DATA > >PRIMARY SYMPTOM STRING: >PIDS/5695DF105 RIDS/IGG0CLA9 RIDS/IGG0CLX0#L PRCS/00F6 >PRCS/ JOBN/DBP3DBM1 > >SYMPTOMSYMPTOM DATA EXPLANATION >------- >PIDS/5695DF105 5695DF105COMPONENT IDENTIFIER >RIDS/IGG0CLA9 IGG0CLA9 ROUTINE IDENTIFIER >RIDS/IGG0CLX0#LIGG0CLX0#L ROUTINE IDENTIFIER >PRCS/00F6 00F6 RETURN CODE >PRCS/ RETURN CODE >JOBN/DBP3DBM1 DBP3DBM1 JOB NAME > >THE SYMPTOM RECORD DOES NOT CONTAIN A SECONDARY SYMPTOM STRING. >FREE FORMAT COMPONENT INFORMATION: > >And then there appears to be a snap dump of storage on each one. > >Nothing on IBMLINK matching anything that I can think to search on from the >fields. In the syslog we see IBM slip trap x91A taken about the time of each >record. >2017166 01:06:27.72 0284 IEA989I SLIP TRAP ID=X91A MATCHED. >JOBNAME=CATALOG , ASID=0086. > >And there are sometimes 100s of this particular symptom records on a given >lpar, per day. >Slip settings are: >ID=X91A,NONPER,ENABLED >ACTION=NODUMP,SET BY CONS INTERNAL,RBLEVEL=ERROR,COMP=91A > >91A > > Explanation: A request to abnormally end the catalog address space (CAS) > service task was issued either through the MODIFY CATALOG,RESTART command, > or through catalog analysis task processing. > > System Action: The system re-drives the
Re: EREP Symptom and/or Software Records
I understand why you may have thought that but no I understand it is not the slip that spawns the records. But couldn't it be said that the slip parms are indicating IBM's view of the severity of the event? I am so new to this that heck I may not be even asking the questions right. For that I'm sorry. For example. Here is one symptom record in particular we are constantly seeing (there are others but let's use this as an example): PIDS/5695DF105 RIDS/IGG0CLA9 RIDS/IGG0CLX0#L PRCS/00F6 PRCS/ PRCS/ JOBN/DBP1DBM1< the JOBN changes but they all seem to be DB2 related tasks. All the other information is the same. SYSTEM ENVIRONMENT: CPU MODEL: 2964 DATE: 166 17 CPU SERIAL: 0207C7 TIME: 01:06:27.72 SYSTEM: MBI2 BCP: MVS RELEASE LEVEL OF SERVICE ROUTINE: HBB77A0 SYSTEM DATA AT ARCHITECTURE LEVEL: 10 COMPONENT DATA AT ARCHITECTURE LEVEL: 10 SYSTEM DATA: || COMPONENT INFORMATION: COMPONENT ID: 5695DF105 COMPONENT RELEASE LEVEL: 220 SERVICE RELEASE LEVEL:UA82137 DESCRIPTION OF FUNCTION: CATPROB DATA PRIMARY SYMPTOM STRING: PIDS/5695DF105 RIDS/IGG0CLA9 RIDS/IGG0CLX0#L PRCS/00F6 PRCS/ JOBN/DBP3DBM1 SYMPTOMSYMPTOM DATA EXPLANATION ------- PIDS/5695DF105 5695DF105COMPONENT IDENTIFIER RIDS/IGG0CLA9 IGG0CLA9 ROUTINE IDENTIFIER RIDS/IGG0CLX0#LIGG0CLX0#L ROUTINE IDENTIFIER PRCS/00F6 00F6 RETURN CODE PRCS/ RETURN CODE JOBN/DBP3DBM1 DBP3DBM1 JOB NAME THE SYMPTOM RECORD DOES NOT CONTAIN A SECONDARY SYMPTOM STRING. FREE FORMAT COMPONENT INFORMATION: And then there appears to be a snap dump of storage on each one. Nothing on IBMLINK matching anything that I can think to search on from the fields. In the syslog we see IBM slip trap x91A taken about the time of each record. 2017166 01:06:27.72 0284 IEA989I SLIP TRAP ID=X91A MATCHED. JOBNAME=CATALOG , ASID=0086. And there are sometimes 100s of this particular symptom records on a given lpar, per day. Slip settings are: ID=X91A,NONPER,ENABLED ACTION=NODUMP,SET BY CONS INTERNAL,RBLEVEL=ERROR,COMP=91A 91A Explanation: A request to abnormally end the catalog address space (CAS) service task was issued either through the MODIFY CATALOG,RESTART command, or through catalog analysis task processing. System Action: The system re-drives the catalog request currently in process. We are not issuing a MODIFY CATALOG RESTART command at the time of any of the logrecs being cut. SO might there something wrong with the catalog process that all these redrives are necessary? Is it normal behavior? So many questions and I'm clueless, unfortunately. So what I guess I was trying to wrap my head around is: if there isn't a need to take a dump, etc. (as specified in the SLIP setting) then why have logic to cut 100's of symptom records at all for that particular issue? And if we're cutting 100's of records - is it really a problem? And like Ed said, it's noise, and I don't know enough to know it's a problem or not and sometimes how to go about diagnosing. So I was hoping to get some help (which I have) in how to handle these and others going forward. Thanks for your and others responses, though. It's much appreciated and I'm taking it all in as much as I can. -Original Message- From: IBM Mainframe Discussion List [mailto:IBM-MAIN@listserv.ua.edu] On Behalf Of Peter Hunkeler Sent: Thursday, June 15, 2017 4:20 PM To: IBM-MAIN@listserv.ua.edu Subject: AW: Re: EREP Symptom and/or Software Records >But I still
AW: Re: EREP Symptom and/or Software Records
>But I still can't get my head around, why cut 100's of symptom/software >records a day at all for a particular problem, if we're just going to ignore >them - abend or not. But I'll try to let that not keep me awake at night. I may well be wrong with my interpretation of the above statement and a similar one in you initial post. Anyway, here I go... I seem to understand that you got the impression that it is all those SLIPs that are responsible for the logrec entries, that is, the logrec records are written because of a SLIP. This is not the case. A problem arises in the software, and this may lead to an ABEND (SVC 13) being issued either by the software explicitly, or by some service routine that was called. Logrec records are a consequence of this. SLIPs are set to perform an action when events, such as ABENDs, and a lot more, occur. The logrec records are written independently. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: EREP Symptom and/or Software Records
> On Jun 14, 2017, at 8:47 AM, Turner Cheryl Lwrote: > > was hoping you'd chime in John (and I appreciate the responses from Skip, > Lizette and EdG, as well). Since I am the person that does the maintenance > and OS upgrades now, I was taking it upon myself to be a bit proactive (or > creating busy work?) and looking at the EREP reports for potential problems > where there may be PTFs available but maybe we just didn't have them on at > the time due to our maintenance windows. We are still in the process of fine > tuning which reports to generate and we are unloading the reports to GDGs . > > So I will summarize the advice as this: Look at them, as you have time. > Decide if any of them are a true problem or something worth investigating > further, check out IBMLINK and/or look for a way to fix it. Open a PMR to > IBM/vendor if really unsure. > > But I still can't get my head around, why cut 100's of symptom/software > records a day at all for a particular problem, if we're just going to ignore > them - abend or not. But I'll try to let that not keep me awake at night. > > Thanks everyone. I don’t know about others but I like a clean ship. Its a few minutes a day routine. If you spot anything that looks unusual then its worth a look. Unusual is different for each person. If I saw more than 2 that was unusual to me. The research was done with the daily look at any hyper APARS. The number of hypers was small (usually). I hated to call IBM for a problem that was already addressed. Generally talking with IBM was not an issue except for thing like CBPDO and other system distributions. The CBPDO calls turned to longish calls and were on average the least productive. Other places at IBM were superb and it was 5 minute at most, unless we forget the PSF people who were thorough and understood traces better than I. When it came to level 3 the phone calls were longish but except for a couple of instances relatively painless. When I got into a disagreement with level 2, I just called the duty manager and explained the issue and he seemed to always come out on my side. This clean ship as I called it, showed in the fact we almost always never had an outage due to software. We were on the bleeding edge with DASD and always had to keep DFDSS up to date. That was the other reason I kept the system up to date. Ed -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: EREP Symptom and/or Software Records
Thank you Bruce. A coworker had downloaded the Logrec Viewer many moons ago, but I wasn't aware of it or that we had it. Your post reminded him to tell me about it and confirm it still works :) We are not currently exploiting PFA. -Original Message- From: IBM Mainframe Discussion List [mailto:IBM-MAIN@listserv.ua.edu] On Behalf Of Bruce Hewson Sent: Thursday, June 15, 2017 1:20 AM To: IBM-MAIN@listserv.ua.edu Subject: Re: EREP Symptom and/or Software Records Hello Cheryl, check out the Logrec Viewer exec - https://www.ibm.com/systems/z/os/zos/tools/downloads/logrec-viewer.html also you can monitor the PFA_LOGREC_ARRIVAl_RATE healthcheck to tell you when you get a burst of LOGREC events. Regards Bruce On Tue, 13 Jun 2017 16:24:34 -0500, Turner Cheryl L <cheryl.l.tur...@irs.gov> wrote: >Greetings. >Our former sysprog, who paid attention to the more finer system details, has >left the building for greener pastures. So now we seem to have to step up our >game. However, I'm not sure what to do or how. > >We are running several EREP reports to see what software or symptom records >are being cut per LPAR (mostly just HISTORY reports for now). We are finding >that a lot of records are being cut at the time an IBM supplied SLIP trap is >taken (for example X13E, X47B, X91A). Some of these records can exceed >hundreds on a given day. > >What should we/I be doing? Reporting them to IBM? I just don't understand why >IBM would set the SLIP yet cut a symptom or software record too. We can't be >the only shop seeing these. Yet I've tried to research a few on IBMLINK but >can't find any hits for known problems. Or maybe there is a way to turn of >the creation of the software/symptom record? Though I can't wrap my head >around that either, thinking why are they then being cut at all if it's not >anything to look into? > >Any schooling you can give, would be most appreciated! But please, be gentle. >I'm out of my element. Many thanks to you all. > >-- >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 -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: EREP Symptom and/or Software Records
Hello Cheryl, check out the Logrec Viewer exec - https://www.ibm.com/systems/z/os/zos/tools/downloads/logrec-viewer.html also you can monitor the PFA_LOGREC_ARRIVAl_RATE healthcheck to tell you when you get a burst of LOGREC events. Regards Bruce On Tue, 13 Jun 2017 16:24:34 -0500, Turner Cheryl Lwrote: >Greetings. >Our former sysprog, who paid attention to the more finer system details, has >left the building for greener pastures. So now we seem to have to step up our >game. However, I'm not sure what to do or how. > >We are running several EREP reports to see what software or symptom records >are being cut per LPAR (mostly just HISTORY reports for now). We are finding >that a lot of records are being cut at the time an IBM supplied SLIP trap is >taken (for example X13E, X47B, X91A). Some of these records can exceed >hundreds on a given day. > >What should we/I be doing? Reporting them to IBM? I just don't understand why >IBM would set the SLIP yet cut a symptom or software record too. We can't be >the only shop seeing these. Yet I've tried to research a few on IBMLINK but >can't find any hits for known problems. Or maybe there is a way to turn of >the creation of the software/symptom record? Though I can't wrap my head >around that either, thinking why are they then being cut at all if it's not >anything to look into? > >Any schooling you can give, would be most appreciated! But please, be gentle. >I'm out of my element. Many thanks to you all. > >-- >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
Re: EREP Symptom and/or Software Records
Hopefully DAE has kicked in and you are just getting symptom records. IPCS DAE(3.5) you can get a little insight into who's waving their hands. http://www-01.ibm.com/support/docview.wss?uid=isg1OA08663 In a message dated 6/14/2017 8:47:57 A.M. Central Daylight Time, cheryl.l.tur...@irs.gov writes: But I still can't get my head around, why cut 100's of symptom/software records a day at all for a particular problem, if we're just going to ignore them - abend or not. But I'll try to let that not keep me awake at night. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: EREP Symptom and/or Software Records
Turner Cheryl L wrote: I was hoping you'd chime in John (and I appreciate the responses from Skip, Lizette and EdG, as well). Since I am the person that does the maintenance and OS upgrades now, I was taking it upon myself to be a bit proactive (or creating busy work?) and looking at the EREP reports for potential problems where there may be PTFs available but maybe we just didn't have them on at the time due to our maintenance windows. We are still in the process of fine tuning which reports to generate and we are unloading the reports to GDGs . So I will summarize the advice as this: Look at them, as you have time. Decide if any of them are a true problem or something worth investigating further, check out IBMLINK and/or look for a way to fix it. Open a PMR to IBM/vendor if really unsure. But I still can't get my head around, why cut 100's of symptom/software records a day at all for a particular problem, if we're just going to ignore them - abend or not. But I'll try to let that not keep me awake at night. Well, I think we likely don't know whether they are truly problems or not, so we cut the records so we have first failure data capture for an actual problem. Back when first we started to do that, there were probably only a few of these records a day. But systems...well..."got bigger and faster." I'd suggest getting familiar with what's out there, as a side research project, and then deciding for yourself whether there is preventive value in spending time on things like this. There might be, if you find unusual records or patterns of activity, but doing this by IEB-EYEBALL takes an awful lot of time so post-processing them is probably a better way to do that. The value in doing the research at all is in being able to skip over things you don't care about when you have to look at the data to try to find a problem. Likewise, you will find that if you clear out the list of dumps to be suppressed that many of the dumps that result won't be worth pursuing. (RTM, in particular, was or is often a victim.) Anyway, that's my opinion. -- John Eells IBM Poughkeepsie ee...@us.ibm.com -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: EREP Symptom and/or Software Records
I was hoping you'd chime in John (and I appreciate the responses from Skip, Lizette and EdG, as well). Since I am the person that does the maintenance and OS upgrades now, I was taking it upon myself to be a bit proactive (or creating busy work?) and looking at the EREP reports for potential problems where there may be PTFs available but maybe we just didn't have them on at the time due to our maintenance windows. We are still in the process of fine tuning which reports to generate and we are unloading the reports to GDGs . So I will summarize the advice as this: Look at them, as you have time. Decide if any of them are a true problem or something worth investigating further, check out IBMLINK and/or look for a way to fix it. Open a PMR to IBM/vendor if really unsure. But I still can't get my head around, why cut 100's of symptom/software records a day at all for a particular problem, if we're just going to ignore them - abend or not. But I'll try to let that not keep me awake at night. Thanks everyone. -Original Message- From: IBM Mainframe Discussion List [mailto:IBM-MAIN@listserv.ua.edu] On Behalf Of John Eells Sent: Wednesday, June 14, 2017 9:10 AM To: IBM-MAIN@listserv.ua.edu Subject: Re: EREP Symptom and/or Software Records Turner Cheryl L wrote: > Greetings. > Our former sysprog, who paid attention to the more finer system details, has > left the building for greener pastures. So now we seem to have to step up > our game. However, I'm not sure what to do or how. > > We are running several EREP reports to see what software or symptom records > are being cut per LPAR (mostly just HISTORY reports for now). We are finding > that a lot of records are being cut at the time an IBM supplied SLIP trap is > taken (for example X13E, X47B, X91A). Some of these records can exceed > hundreds on a given day. > > What should we/I be doing? Reporting them to IBM? I just don't understand why > IBM would set the SLIP yet cut a symptom or software record too. We can't be > the only shop seeing these. Yet I've tried to research a few on IBMLINK but > can't find any hits for known problems. Or maybe there is a way to turn of > the creation of the software/symptom record? Though I can't wrap my head > around that either, thinking why are they then being cut at all if it's not > anything to look into? > > Any schooling you can give, would be most appreciated! But please, be gentle. > I'm out of my element. Many thanks to you all. If you care to research these, look up the abends and see whether you really care. Using 13E as an example: Explanation: The task that created a subtask issued a DETACH macro for that subtask, specifying STAE=NO, before the subtask ended. This may or may not be an error, depending on the intent of the user. Consequently, the system does not abnormally end the task issuing the DETACH macro. With no proof to back my assumptions, I nonetheless suspect that, if you look at these, you will find that most or all of those are from application programs that summarily shot their subtasks rather than signaling them to complete or waiting for them to complete. You can discuss such things with the application owner, but you might find most of them reluctant to "fix" the problem when what they are doing is working from their perspective and the net ill effect is "noise" in EREP reports. I will observe, though, that the system programmer action for ABEND13E could use some updates... -- John Eells IBM Poughkeepsie ee...@us.ibm.com -- 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
Re: EREP Symptom and/or Software Records
Turner Cheryl L wrote: Greetings. Our former sysprog, who paid attention to the more finer system details, has left the building for greener pastures. So now we seem to have to step up our game. However, I'm not sure what to do or how. We are running several EREP reports to see what software or symptom records are being cut per LPAR (mostly just HISTORY reports for now). We are finding that a lot of records are being cut at the time an IBM supplied SLIP trap is taken (for example X13E, X47B, X91A). Some of these records can exceed hundreds on a given day. What should we/I be doing? Reporting them to IBM? I just don't understand why IBM would set the SLIP yet cut a symptom or software record too. We can't be the only shop seeing these. Yet I've tried to research a few on IBMLINK but can't find any hits for known problems. Or maybe there is a way to turn of the creation of the software/symptom record? Though I can't wrap my head around that either, thinking why are they then being cut at all if it's not anything to look into? Any schooling you can give, would be most appreciated! But please, be gentle. I'm out of my element. Many thanks to you all. If you care to research these, look up the abends and see whether you really care. Using 13E as an example: Explanation: The task that created a subtask issued a DETACH macro for that subtask, specifying STAE=NO, before the subtask ended. This may or may not be an error, depending on the intent of the user. Consequently, the system does not abnormally end the task issuing the DETACH macro. With no proof to back my assumptions, I nonetheless suspect that, if you look at these, you will find that most or all of those are from application programs that summarily shot their subtasks rather than signaling them to complete or waiting for them to complete. You can discuss such things with the application owner, but you might find most of them reluctant to "fix" the problem when what they are doing is working from their perspective and the net ill effect is "noise" in EREP reports. I will observe, though, that the system programmer action for ABEND13E could use some updates... -- John Eells IBM Poughkeepsie ee...@us.ibm.com -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: EREP Symptom and/or Software Records
So Logrec has two types of sysprogs using it the ones that want to know everything that is going on - read it hourly (i use to do that) the ones that only look at it when it is associated with a problem. (i do that now.) If your shop has a lot of issues with software and/or hardware, then reviewing logrec is helpful. There is a tool called LOGGER that that helps with reviewing logrec online. I am not sure where that REXX is but I think you should be able to find it in an IBM Tools area. If you do not have a lot of software or hardware issue, then you might want to revisit what should be done about logrec. I have it in a offloaded GDG and rarely look at it. When I had physical cartridge tape, I needed to see how bad the errors were on the tapes. Now that i have virtual tape, I hardly look at it at all. Lizette -Original Message- >From: Turner Cheryl L <cheryl.l.tur...@irs.gov> >Sent: Jun 13, 2017 2:24 PM >To: IBM-MAIN@LISTSERV.UA.EDU >Subject: EREP Symptom and/or Software Records > >Greetings. >Our former sysprog, who paid attention to the more finer system details, has >left the building for greener pastures. So now we seem to have to step up our >game. However, I'm not sure what to do or how. > >We are running several EREP reports to see what software or symptom records >are being cut per LPAR (mostly just HISTORY reports for now). We are finding >that a lot of records are being cut at the time an IBM supplied SLIP trap is >taken (for example X13E, X47B, X91A). Some of these records can exceed >hundreds on a given day. > >What should we/I be doing? Reporting them to IBM? I just don't understand why >IBM would set the SLIP yet cut a symptom or software record too. We can't be >the only shop seeing these. Yet I've tried to research a few on IBMLINK but >can't find any hits for known problems. Or maybe there is a way to turn of >the creation of the software/symptom record? Though I can't wrap my head >around that either, thinking why are they then being cut at all if it's not >anything to look into? > >Any schooling you can give, would be most appreciated! But please, be gentle. >I'm out of my element. Many thanks to you all. > -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: EREP Symptom and/or Software Records
> On Jun 13, 2017, at 4:24 PM, Turner Cheryl Lwrote: > > Greetings. > Our former sysprog, who paid attention to the more finer system details, has > left the building for greener pastures. So now we seem to have to step up > our game. However, I'm not sure what to do or how. > > We are running several EREP reports to see what software or symptom records > are being cut per LPAR (mostly just HISTORY reports for now). We are finding > that a lot of records are being cut at the time an IBM supplied SLIP trap is > taken (for example X13E, X47B, X91A). Some of these records can exceed > hundreds on a given day. > > What should we/I be doing? Reporting them to IBM? I just don't understand why > IBM would set the SLIP yet cut a symptom or software record too. We can't be > the only shop seeing these. Yet I've tried to research a few on IBMLINK but > can't find any hits for known problems. Or maybe there is a way to turn of > the creation of the software/symptom record? Though I can't wrap my head > around that either, thinking why are they then being cut at all if it's not > anything to look into? I was doing this in the 1990’s. I would look through the report to see if there were possibly duplicates. If there was I would fire up ibmlink and do a search on the modules and abend code. I almost always get a hit. I would research the apartment and look up the pff number and it would go in ASAP (scheduled IPL every weekend). Sometimes a system dump would show up as well. Kept an eye on those as well. It kept our system up and running and no strange things ever happened. I also kept a eagle eye on PSF bends as we were ECS and always needed the latest software available. I also was on a first name basis of level 2/3 for PSF issues. It was a lot of work because IMO PSF was bug ridden when we got it. Ed > > Any schooling you can give, would be most appreciated! But please, be gentle. > I'm out of my element. Many thanks to you all. > > -- > 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
Re: EREP Symptom and/or Software Records
Those IBM supplied SLIP traps are generally there to suppress irrelevant SVC dumps: SL SET,C=13E,ID=X13E,A=NODUMP,END Imagine what would happen if you got a dump for each S13E in addition to the LOGREC record? The truth is that some abends relating to 'improper' termination cleanup are just part of MVS noise. They seldom mean anything at all, but they are after all abends, so they are acknowledged and recorded. But unless you (or a vendor) needs to pursue the cause, nothing else needs to be done. There are also NODUMP or NOSVCD SLIP traps for S013, Sx37, and other common 'user error' abends. If you need a dump for such an error, you can override the SLIP trap, but in most cases, the cause--and the solution--are obvious. . . J.O.Skip Robinson Southern California Edison Company Electric Dragon Team Paddler SHARE MVS Program Co-Manager 323-715-0595 Mobile 626-543-6132 Office ⇐=== NEW robin...@sce.com -Original Message- From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf Of Turner Cheryl L Sent: Tuesday, June 13, 2017 2:25 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: (External):EREP Symptom and/or Software Records Greetings. Our former sysprog, who paid attention to the more finer system details, has left the building for greener pastures. So now we seem to have to step up our game. However, I'm not sure what to do or how. We are running several EREP reports to see what software or symptom records are being cut per LPAR (mostly just HISTORY reports for now). We are finding that a lot of records are being cut at the time an IBM supplied SLIP trap is taken (for example X13E, X47B, X91A). Some of these records can exceed hundreds on a given day. What should we/I be doing? Reporting them to IBM? I just don't understand why IBM would set the SLIP yet cut a symptom or software record too. We can't be the only shop seeing these. Yet I've tried to research a few on IBMLINK but can't find any hits for known problems. Or maybe there is a way to turn of the creation of the software/symptom record? Though I can't wrap my head around that either, thinking why are they then being cut at all if it's not anything to look into? Any schooling you can give, would be most appreciated! But please, be gentle. I'm out of my element. Many thanks to you all. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
EREP Symptom and/or Software Records
Greetings. Our former sysprog, who paid attention to the more finer system details, has left the building for greener pastures. So now we seem to have to step up our game. However, I'm not sure what to do or how. We are running several EREP reports to see what software or symptom records are being cut per LPAR (mostly just HISTORY reports for now). We are finding that a lot of records are being cut at the time an IBM supplied SLIP trap is taken (for example X13E, X47B, X91A). Some of these records can exceed hundreds on a given day. What should we/I be doing? Reporting them to IBM? I just don't understand why IBM would set the SLIP yet cut a symptom or software record too. We can't be the only shop seeing these. Yet I've tried to research a few on IBMLINK but can't find any hits for known problems. Or maybe there is a way to turn of the creation of the software/symptom record? Though I can't wrap my head around that either, thinking why are they then being cut at all if it's not anything to look into? Any schooling you can give, would be most appreciated! But please, be gentle. I'm out of my element. Many thanks to you all. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN