Re: REDBOOKS Question
Redbooks once published are not updated. There are no dash numbers or TNLs. A new Redbooks on the same or similar subject has a completely new number. These are only done when a major change to the subject requires it and the lab that develops the product wants a new Redbook. Error should be trapped at the draft stage before publication. If not then they will remain for the life of the Redbook. I have been involved in several Redbooks. I always say the product manual is the definitive source of information on the subject. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +66 811431287 --- On Sun, 7/11/10, Ed Gould ps2...@yahoo.com wrote: From: Ed Gould ps2...@yahoo.com Subject: REDBOOKS Question To: IBM-MAIN@bama.ua.edu Date: Sunday, 7 November, 2010, 4:05 I have forgotten the answer to this question long ago and I am not coming up with any answer to my question to myself.I will pose it to the group and see if anyone on here knows the answer.I am trying to remember how IBM handles it if there is an error (or poorly written or whatever error) happens with REDBOOKS. Are they quietly updated (and if so) is there a new dash number or is there a completely new number assigned or how does IBM handle updates to any REDBOOKS? I know there are errors as long time ago (30+ years) I used to see them with TNL's (we all remember those right?) but I do not remember how IBM handles them currently. Anyone? -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: CPU capping is not working for one Lpar only on CEC?
I am not sure what type of capping you are trying to use. If you want to use hard capping then this will use the weights and it will give you 160/(160+10). This will not be exact. If you look at the PR/SM manual is is usually within 1% but can be over 3^% out. You talk about MSUs, which is soft capping and this uses the rolling 4 hour average. So until your 4 hour average goes over the define you can use over this number of MSUs. You cannot use both hard and soft capping. Looking at the report it says zero for your defined capacity MSUs. So it looks like you are not using soft capping. You have hardware capping set for yes. What are you trying to do. If you are managing the software costs then the MSU 4 hour rolling avaerage is appropriate. If you want to have a maximum capacity for a partition all teh time then you need hard capping. But remember there is a small percentage error rate on this. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +66 811431287 --- On Wed, 3/11/10, Cobe Xu cob...@gmail.com wrote: From: Cobe Xu cob...@gmail.com Subject: CPU capping is not working for one Lpar only on CEC? To: IBM-MAIN@bama.ua.edu Date: Wednesday, 3 November, 2010, 9:59 Hi list, We aim to cap the only active LPAR on the CEC(26 MSU) to 24 MSU. But, I'm a bit confuse when I checked the RMF CPU Activity report as below, which shows that with the interval, SYS2 was able to use up to 25 MSU. (Highlighted) So my questions are: 1. Is this because CPU capping is not working for only one active LPAR on the CEC? If it's the case, any reference? 2. Or, this is related to the WEIGHT value we used to CAP? in our case, we reference our CEC capacity is 26 MSU (about 171 MIPS), target to CAP 24 MSU (160 MIPS). . But, for client's sake, we use MIPS value as the WEIGHT,i.e 160. (as highlighted in the report). And, this mislead the Lpar scheduler that it is over 100% of the CEC. Thus, SYS2 can use as much as it needs. 3. Or any other posibility? Pls shed some light, thanks a lot! PAGE 2 z/OS V1R8 SYSTEM ID SYS2 START 08/17/2010-03.00.00 INTERVAL 000.59.59 RPT VERSION V1R8 RMF END 08/17/2010-04.00.00 CYCLE 0.100 SECONDS MVS PARTITION NAME SYS2 NUMBER OF PHYSICAL PROCESSORS 4 GROUP NAME N/A IMAGE CAPACITY 24 CP 2 LIMIT N/A NUMBER OF CONFIGURED PARTITIONS 5 ICF 2 WAIT COMPLETION NO DISPATCH INTERVAL DYNAMIC - PARTITION DATA - -- LOGICAL PARTITION PROCESSOR DATA -- -- AVERAGE PROCESSOR UTILIZATION PERCENTAGES -- MSU -CAPPING-- PROCESSOR- DISPATCH TIME DATA LOGICAL PROCESSORS --- PHYSICAL PROCESSORS --- NAME S WGT DEF ACT DEF WLM% NUM TYPE EFFECTIVE TOTAL EFFECTIVE TOTAL LPAR MGMT EFFECTIVE TOTAL SYS2 A 160 0 25 YES 0.0 2 CP 01.55.49.238 01.55.53.278 96.52 96.57 0.06 96.52 96.57 SYS6 A 10 0 0 YES 0.0 2 CP 00.00.00.000 00.00.00.000 0.00 0.00 0.00 0.00 0.00 *PHYSICAL* 00.00.01.367 0.02 0.02 -- -- -- TOTAL 01.55.49.238 01.55.54.645 0.08 96.52 96.59 CFP01AH2 A DED 1 ICF 00.59.59.655 00.59.59.725 99.99 99.99 0.00 50.00 50.00 CFP02AH2 A DED 1 ICF 00.59.59.666 00.59.59.708 99.99 99.99 0.00 50.00 50.00 *PHYSICAL* 00.00.00.422 0.01 0.01 -- -- -- TOTAL 01.59.59.321 01.59.59.855 0.01 99.99 100.0 SYS8 -- Cobe Xu Best Regards --- zOS Performance Capacity Analyst E2E Performance Analyst Email: cob...@gmail.com --- -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives
Re: New(ish) redbook on z196
The July copy was a draft version. The final version was published on 15th October. I suggest you download the final version. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +66 811431287 --- On Mon, 1/11/10, Scott Rowe scott.r...@joann.com wrote: From: Scott Rowe scott.r...@joann.com Subject: Re: New(ish) redbook on z196 To: IBM-MAIN@bama.ua.edu Date: Monday, 1 November, 2010, 20:30 I don't know if it's been posted before, but I downloaded the pdf on 7/23. On Mon, Nov 1, 2010 at 2:55 PM, McKown, John john.mck...@healthmarkets.comwrote: Sorry if I'm reposting something that was already posted. http://www.redbooks.ibm.com/Redbooks.nsf/RedbookAbstracts/sg247833.html?Open quote The zEnterprise System consists of the IBM zEnterprise 196 central processor complex, IBM zEnterprise Unified Resource Manager, and IBM zEnterprise BladeCenter(r) Extension. The z196 is designed with improved scalability, performance, security, resiliency, availability, and virtualization. The z196 Model M80 provides up to 1.6 times the total system capacity of the z10 EC Model E64, and all z196 models provide up to twice the available memory of the z10 EC. The zBX infrastructure works with the z196 to enhance System z virtualization and management through an integrated hardware platform that spans mainframe and POWER7 technologies. Through the Unified Resource Manager, the zEnterprise System is managed as a single pool of resources, integrating system and workload management across the environment. This book provides an overview of the zEnterprise System and its functions, features, and associated software support. Greater detail is offered in areas relevant to technical planning. This book is intended for systems engineers, consultants, planners, and anyone wanting to understand the zEnterprise System functions and plan for their usage. It is not intended as an introduction to mainframes. Readers are expected to be generally familiar with existing IBM System z technology and terminology. /quote John McKown Systems Engineer IV IT Administrative Services Group HealthMarkets(r) 9151 Boulevard 26 * N. Richland Hills * TX 76010 (817) 255-3225 phone * (817)-691-6183 cell john.mck...@healthmarkets.com * www.HealthMarkets.com Confidentiality Notice: This e-mail message may contain confidential or proprietary information. If you are not the intended recipient, please contact the sender by reply e-mail and destroy all copies of the original message. HealthMarkets(r) is the brand name for products underwritten and issued by the insurance subsidiaries of HealthMarkets, Inc. -The Chesapeake Life Insurance Company(r), Mid-West National Life Insurance Company of TennesseeSM and The MEGA Life and Health Insurance Company.SM -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html CONFIDENTIALITY/EMAIL NOTICE: The material in this transmission contains confidential and privileged information intended only for the addressee. If you are not the intended recipient, please be advised that you have received this material in error and that any forwarding, copying, printing, distribution, use or disclosure of the material is strictly prohibited. If you have received this material in error, please (i) do not read it, (ii) reply to the sender that you received the message in error, and (iii) erase or destroy the material. Emails are not secure and can be intercepted, amended, lost or destroyed, or contain viruses. You are deemed to have accepted these risks if you communicate with us by email. Thank you. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Storage usage in a job
Phil, You do not say what kind of batch job it is. Is it normal qsam/vsam or is it DB2? The CPU percentage (CPU time /elapsed time) is over 70%. This is very high for a non DB2. Even for DB2 it would have to find most of the data already in bufferpools. The fialure is insufficienty storage to loadf a module. The module name is a sample module for LE(searhced GOOGLE for it). So it is probably not big. Also you have paging at 228K according to your joblog. I would assume your working set is large. I would not expect it to be a real storage problem It is almost certainly a virtual storage problem in the jobs region. Try increasing your region size. Also are you sure you are not looping on the same record - hence the high CPU percentage. Maybe you build an in memory table and the larger volumes are blowing the region. I would recommend fixing the virtual storage AND the high CPU percentage. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +66 811431287 --- On Wed, 20/10/10, Mike Schwab mike.a.sch...@gmail.com wrote: From: Mike Schwab mike.a.sch...@gmail.com Subject: Re: Storage usage in a job To: IBM-MAIN@bama.ua.edu Date: Wednesday, 20 October, 2010, 4:42 120T = 120,000 4K pages over 480,000,000 bytes On Tue, Oct 19, 2010 at 9:10 PM, Phil Smith p...@voltage.com wrote: We're doing some load testing, and running out of storage, but we can't figure out how to tell WHAT storage. Shortly before it blew off, the REAL column in SDSF showed 120T. We're pretty sure that wasn't 120 terabytes; it seems to be T for Thousand. But that isn't a lot of memory?! Here's the job output when it failed - does this offer any clues? Any hints, or pointers to doc, are gratefully accepted! EW4000I FETCH FOR MODULE EDCU FROM DDNAME -LNKLST- FAILED BECAUSE INSUFFICIENT STORAGE WAS AVAILABLE. SV031I LIBRARY ACCESS FAILED FOR MODULE EDCU, RETURN CODE 24, REASON CODE 26080021, DDNAME *LNKLST* --TIMINGS (MINS.)-- -PAGING COUNTS STEPNAME PROCSTEP RC EXCP CONN TCB SRB CLOCK SERV WORKLOAD PAGE SWAP VIO SWAPS RUN 3632 474037 51 345.99 .00 482.8 558935 BATCH 288K 0 0 0 HUDIUNIR ENDED. NAME- TOTAL TCB CPU TIME= 345.99 TOTAL ELAPSED TIME= 482.8 HASP395 HUDIUNIR ENDED -- ...phsiii Phil Smith III p...@voltage.com Voltage Security, Inc. www.voltage.com -- Mike A Schwab, Springfield IL USA Where do Forest Rangers go to get away from it all? -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Batch loop in SYSplex
Mohd, Just some basic questions. Is this a one off event or have you had several times? Is it always the same job? What sort of VSAM dataset is it? How big is it? Do you have RACF erase on scratch turned on for it? If so, and the data set is big, the delete will take a long time writing binary zeros over it. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +66 811431287 --- On Mon, 11/10/10, Mohd Shahrifuddin fud...@bayss.com wrote: From: Mohd Shahrifuddin fud...@bayss.com Subject: Batch loop in SYSplex To: IBM-MAIN@bama.ua.edu Date: Monday, 11 October, 2010, 2:57 Dear All, Environment: We in z/OS 1.9, CICS, IMS and VSAM environment. Running Parallel SYSPLEX 4 LPAR. Lately we apply new PTF RSU1006 for all three(z/OS,IMS and CICS). Encounter Problem: We have using our own develop Job Schedule running in LPAR A. Our Batch job we route to running in LPAR B. IMS is Sharing all LPAR. Before PTF we not encounter any problem but after complete rooling IPL all LPAR we encounter problem. The problem is the batch job running in LPAR B cannot complete just only to delete define VSAM file. The worst the job using alot of CPU look like loop job. If we run the batch job using the same LPAR (LPAR A) with the job schedule is not encounter any problem. Help: If any of lister face same problem please advice us what the problem. If any cause PTF make the situation problem. Thank You in advance, Mohd Shahrifuddin. Shanghai, China. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: SORT question
Frank, I did not plan for SORT to read the records. SORTIN is dummy. The exit reads the records directly and inserts the records into sort. I have written many such E15 exits to process SMF data. The E15 does the read from its own input file. It then only reads the records until the higher date is reached. It then stops reading any more records and signals to sort that the input is finished. This will only read the required records. Terry Draper series Performance Consultant w...@btopenworld.com mobile: +66 811431287 --- On Wed, 29/9/10, Frank Yaeger yae...@us.ibm.com wrote: From: Frank Yaeger yae...@us.ibm.com Subject: Re: SORT question To: IBM-MAIN@bama.ua.edu Date: Wednesday, 29 September, 2010, 20:30 Terry Draper wrote on IBM Mainframe Discussion List IBM-MAIN@bama.ua.edu wrote on 09/29/2010 02:43:19 AM: Maybe try using an E15 exit to read the input file. Stop passing the records when you reach the date higher than required.. I am not sure how I would pass the date to the E15. Could it be a parm. Or it could be read from a separate single record input file on first pass through the E15. Any comments on this? This would probably have the opposite effect of what was requested (better performance due to NOT reading records once the date is found). Although, you can stop passing the records to the exit by having the exit pass back RC=8 early, ALL of the records will still be read; some will just not be passed to the exit. But since an exit is not really needed in this case, and using an exit involves additional overhead, this wouldn't accomplish anything. Frank Yaeger - DFSORT Development Team (IBM) - yae...@us.ibm.com Specialties: JOINKEYS, FINDREP, WHEN=GROUP, ICETOOL, Symbols, Migration = DFSORT/MVS is on the Web at http://www.ibm.com/storage/dfsort -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: SORT question
Maybe try using an E15 exit to read the input file. Stop passing the records when you reach the date higher than required.. I am not sure how I would pass the date to the E15. Could it be a parm. Or it could be read from a separate single record input file on first pass through the E15. Any comments on this? Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +66 811431287 --- On Tue, 28/9/10, Frank Yaeger yae...@us.ibm.com wrote: From: Frank Yaeger yae...@us.ibm.com Subject: Re: SORT question To: IBM-MAIN@bama.ua.edu Date: Tuesday, 28 September, 2010, 18:13 David Bettern on IBM Mainframe Discussion List IBM-MAIN@bama.ua.edu wrote on 09/28/2010 08:14:26 AM: I think the real question was how to tell sort to stop reading anymore input records when it reaches the first record that's greater than the selection criteria instead of reading through the rest of the file but not selecting any of those records. I'm not sure we have a way to do that but I'm sure Frank will weigh in shortly. The only way to tell DFSORT to stop processing records is with STOPAFT=n. But in this case, we can't know what n is ahead of time. So I don't see a way to do it in one pass. Frank Yaeger - DFSORT Development Team (IBM) - yae...@us.ibm.com Specialties: JOINKEYS, FINDREP, WHEN=GROUP, ICETOOL, Symbols, Migration = DFSORT/MVS is on the Web at http://www.ibm.com/storage/dfsort -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: SORT question
The question stated I have a large file that is sorted in ascending order by a date field, MMDD. So I must assume it is already in sequence od date. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +66 811431287 --- On Wed, 29/9/10, Staller, Allan allan.stal...@kbm1.com wrote: From: Staller, Allan allan.stal...@kbm1.com Subject: Re: SORT question To: IBM-MAIN@bama.ua.edu Date: Wednesday, 29 September, 2010, 15:05 At E15 time, the file has not yet been sorted. snip Maybe try using an E15 exit to read the input file. Stop passing the records when you reach the date higher than required.. I am not sure how I would pass the date to the E15. Could it be a parm. Or it could be read from a separate single record input file on first pass through the E15. Any comments on this? /snip -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: SQA/ESQA HAS EXPANDED INTO CSA/ECSA
Matan, The display you gave is for real storage use. You need the virtual storage view for the common areas. RMF monitor III has this display. By itself an overflow from SQA/ESQA to CSA/ECSA is not a problem. That is so long as you have ample room in CSA/ECSA to accommodate it. If the overflow is for SQA to CSA you are talking about areas with limited size below the line. Just increases the size of either may reduce your private area below the line by 1 megabyte. So in this case I would do serious investigation of why the increase. Most use of common has been move above the line. If the overflow is from ESQA to ECSA then your ECSA is probably big enough to handle it. If you do not have a lot of free space in ECSA already then increase at the next IPL. If you have an address space that requires a large extended private (such as DB2 with SAP) then be careful not to reduce extended private too much. If you continue to have an overflow then you may want to also increase ESQA (although this is not essential). Note that some overflow pages may hang around until the next IPL even though the total ESQA requirement has gone down. This is because the requester of the overflow pages is still using them. So the message in itself is not a problem. It should trigger you to investigate more as detailed above. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +66 811431287 --- On Tue, 21/9/10, Matan Cohen matancohen...@gmail.com wrote: From: Matan Cohen matancohen...@gmail.com Subject: SQA/ESQA HAS EXPANDED INTO CSA/ECSA To: IBM-MAIN@bama.ua.edu Date: Tuesday, 21 September, 2010, 9:34 Hi, When I started the websphere on the z/9 i noticed this message : *IRA103I **SQA**/**ESQA** HAS EXPANDED INTO **CSA**/**ECSA** BY 887 PAGES* entering to the Omegamon for MVS show this : Major Area Real Minor Area Not/Fix Fixed Total === High Private 1,117M 1,109M 8M 1,117M === Extended Private 4,091M 4,013M 79,664K 4,091M (ELSQA) 92K 76,808K 76,900K --- Extended Common 104,884K CSA 42,448K 16,860K 59,308K FLPA 12K 12K PLPA 20,560K 288K 20,848K SQA 1,996K 12,552K 14,548K Read/Write Nuc 488K 488K Read-only Nuc 9,680K 9,680K === Common 2,132K Read-only Nuc 112K 112K Read/Write Nuc 52K 52K SQA 332K 332K PLPA 1,184K 1,184K CSA 352K 100K 452K --- Private 55,580K V=V 48,936K 6,628K 55,564K (LSQA) 5,836K 5,836K System Area 16K 16K --- Abs Zero Frame 24K 24K 24K === Available 187,724K DataOnly Spaces 446,216K DataOnly Sp Mgmt 5,876K Shared Fixed 664K Shared Pageable 144,948K Page Table 5,656K Local Quad 3,312K BDF 24K TDF 24K SQA Reserved 20K DAT Off Nucleus 16K --- Total Storage 6G we have performance problem with the WAS but in HIGH CPU meaning I wander if i should care for this message and increase the SQA or consider other steps, any advise on this issue will be appreciate . -- best regards, matan cohen MF System Administrator. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: CPU Spikes completely hangs sessions;No logons: need inputs
Amit. Firstly you should have RMF monitor III setup so that it can see the address spaces in all the LPARs in the Sysplex. Make sure RMF address spaces are at high priority. You can then see who is getting the CPU and how much the LPAR is using. Have the TSO userids of selected (trusted not to abuse the priority) performance people set to a single period high priority service class. Alternatively you can get someone reset to a high priority at problem times. It is useful to have some TSO users stay logged on nearly all the time. Not sure of the best options to achieve this. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +66 811431287 --- On Mon, 6/9/10, amit amitpdu...@gmail.com wrote: From: amit amitpdu...@gmail.com Subject: CPU Spikes completely hangs sessions;No logons: need inputs To: IBM-MAIN@bama.ua.edu Date: Monday, 6 September, 2010, 11:59 hi, i have come across similar times where in my services/STCs demand a lot of CPU and thus resulting soft capping. some times even when msu's are being increased, via system tools though Ops can see that my stcs are taking more CPU, i am unable to logon to the LPAR..so are other users. is there a way that i can monitor/ view the STC of the affected LPAR from a diff one, in sysplex. secondly, how do i prioritize my TSO ID in that LPAR so that it may logon and check/action to trap the issue. My questions might sound foolish or/and inappropriate, but can assure, are not just for Fun. eager to learn, if possible some help/experience shared. TIA, Amit -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: JES2 vs. JES3
Having worked with both JES2 and JES3 (JES3 first), I would say it depends on what functionality you require. JES3 has some multisystem capability that you may like. If there is no requirement for this then definitely go JES2. JES2 can handle multisystems adequately. JES2 has tended to get support before JES3. Many JES3 systems have been converted to JES2. Someone new should definitely look to use JES2. Now I will be shot down by the JES3 bigots. All I say is have have no axe to grind for either. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +66 811431287 --- On Sun, 5/9/10, Thompson, Steve steve_thomp...@stercomm.com wrote: From: Thompson, Steve steve_thomp...@stercomm.com Subject: JES2 vs. JES3 To: IBM-MAIN@bama.ua.edu Date: Sunday, 5 September, 2010, 23:13 -Original Message- From: IBM Mainframe Discussion List [mailto:ibm-m...@bama.ua.edu] On Behalf Of Clark Morris Sent: Sunday, September 05, 2010 4:55 PM To: IBM-MAIN@bama.ua.edu Subject: Re: O/T IBM to Ship World's Fastest Computer Chip SNIPPAGE As someone who was in a field where you can't get a consensus on whether JES2 is better than JES3 and who is a follower of transportation issues (and a member of Transport Action Atlantic), I doubt a reporter would be able to determine easily which side of an argument is flat out wrong, even with some hours of research. SNIPPAGE I created a new thread out of this because I think this is a bit more important a topic -- so why let it get lost in O/T IBM blah blah? Why is JES2 better than JES3? Why is JES3 better than JES2? Why would JES3 be preferred over JES2? Regards, Steve Thompson -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: WLM
I would also look at the presentations on the IBM WLM web page. These tell you a lot more than the manuals. For production CICS I would always go with response time goals. I know it uses a bit more CPU, but you get better controls. For development, I would probably use velocity (especially if in same Sysplex as production). Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +966 556730876 --- On Wed, 3/2/10, Kelman, Tom thomas.kel...@commercebank.com wrote: From: Kelman, Tom thomas.kel...@commercebank.com Subject: Re: WLM To: IBM-MAIN@bama.ua.edu Date: Wednesday, 3 February, 2010, 18:14 I am in the process of completely redesigning our WLM policy, so I'm going through pretty much the same as you. Although, I do have some experience in designing one shortly after WLM appeared on the scene. What I'm trying to get a handle on is new functionality and subsystems that have been introduced since I originally set up a WLM. What WLM manual are you looking at? Definitely get the MVS Planning Workload Management manual. Also, I would recommend the System Programmer's Guide to Workload Management Redbook. You might also want to review any performance manuals available for whatever subsystems you have (DB2, IMS, CICS, MQ, etc.). As far as response time vs. velocity goals are concerned, response time goals are a whole lot easier to deal with. If you have CICS, for example, you probably have a good idea how fast you want the transactions to finish. That and IMS are excellent candidates for response time goals. Also, response time goals stay the same between operating system and hardware changes. Velocity goals need to be reevaluated and possibly changed. However, long running tasks (long batch jobs, forever running STCs, etc.) require velocity goals, but you can set up a response time goal service class for some batch jobs. For example, in my first WLM I had set up a response time goal of 60% witching 15 minutes for short running jobs. It worked well, but you do need to know, and be able to control, your batch environment to do that. Also, when setting up a response time goal, use percentage response time, not average response time. IMHO the average response time goal is worthless. Tom Kelman Enterprise Capacity Planner Commerce Bank of Kansas City (816) 760-7632 -Original Message- From: IBM Mainframe Discussion List [mailto:ibm-m...@bama.ua.edu] On Behalf Of gsg Sent: Wednesday, February 03, 2010 11:15 AM To: IBM-MAIN@bama.ua.edu Subject: Re: WLM What are the benefits on controlling via a response time goal vs. a velocity goal? Our system has been pretty constrained lately and I'm looking for ways to improve it, but I'm not that familiar with WLM. By the way, I am looking at the WLM manual was well. I need to get the MVS Planning Workload Management one though. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html * If you wish to communicate securely with Commerce Bank and its affiliates, you must log into your account under Online Services at http://www.commercebank.com or use the Commerce Bank Secure Email Message Center at https://securemail.commercebank.com NOTICE: This electronic mail message and any attached files are confidential. The information is exclusively for the use of the individual or entity intended as the recipient. If you are not the intended recipient, any use, copying, printing, reviewing, retention, disclosure, distribution or forwarding of the message or any attached file is not authorized and is strictly prohibited. If you have received this electronic mail message in error, please advise the sender by reply electronic mail immediately and permanently delete the original transmission, any attachments and any copies of this message from your computer system. * -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: IBM countersues Neon over zPrime accelerator
Everyone keeps talking about the impact on the IBM software bill. I would be very worried about the impact of my non-IBM software bill. Even if IBM allowed the use of speciality engines (which I would not expect them to do), other software suppliers may not. I think the zPrime product could be bad news for everyone, even the users of the product. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +966 556730876 --- On Wed, 3/2/10, Mark Zelden mark.zel...@zurichna.com wrote: From: Mark Zelden mark.zel...@zurichna.com Subject: Re: IBM countersues Neon over zPrime accelerator To: IBM-MAIN@bama.ua.edu Date: Wednesday, 3 February, 2010, 14:04 On Wed, 3 Feb 2010 09:03:33 +0100, R.S. r.skoru...@bremultibank.com.pl wrote: Mark Zelden pisze: [...] I don't have a stake either way, but if I were rooting for Neon and they won this battle, IBM would still be free to change the licensing rules or not charge less for special engines, or they could just change the code and break zPrime for Neon. So the only benefit seems to be temporary for anyone using the software. Long term, it could hurt everyone else. Another scenario: Neon wins, IBM tries to break zPrime and is sued for trust practices. Then LOSES anti-trust trial and is obliged to keep zPrime working. Effect: everyone can lower his HW and SW bills. zPrime working doesn't do anyone any good if specialty engines don't get a price break (IBM can't be forced to sell them cheaper). That is the whole point of them. So the effect is more likely to be no more specialty engines (why have them at the same price). That hurts everyone who uses them today** that doesn't run zPrime. ** Everyone that uses them enough to get a price advantage. It doesn't help to purchase a specialty engine at 1/4 the price of a general engine and only run it at 10% utilization. But ZAAPZIIP in z/OS 1.11 (and the support added to z/OS 1.9 1.10 via OA27495) help. Mark -- Mark Zelden Sr. Software and Systems Architect - z/OS Team Lead Zurich North America / Farmers Insurance Group - ZFUS G-ITO mailto:mark.zel...@zurichna.com z/OS Systems Programming expert at http://expertanswercenter.techtarget.com/ Mark's MVS Utilities: http://home.flash.net/~mzelden/mvsutil.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: DFSORT memory used paging
My response is not directly to this posting, just my comments to the original question. I do not have the original post. First, I have to ask what is more important to you? Is it SORTs or your online systems? I will assume online. We had problems with large sorts impacting our online. CICS is usually storage protected, so it was not impacted. We run Websphere Application Server and this was badly affected. WAS is not storage protected and uses a lot of memory. So it was a prime target for page steals. After the sort started we had high page-in rates for WAS. This was an issue. We looked at DFSORT options and decided we wanted to stop the page steals.We changed the EXPOLD parameter to zero, The default is MAX. This means that now DFSORT will only try to use available memory and not steal unreferenced pages from other work. If you are looking to make SORT fly then you may want to leave DFSORT to steal pages. But we want our on-line systems to perform and so DFSORT is a low priority. At the time of our problem we were storage constrained. Not paging, but low available frame queue. We added more memory and now have a lot free. We still do not allow DFSORT to steal pages though. On paging, I come from the school that I do not want to page. If I could restrict to very low priority work then maybe a little paging. The cost of memory means that you get more benefit from adding memory than messing around with the paging subsystem. Yes I want a paging subsystem that can handle large paging, but I do not want to use it. Its there as insurance. Also beware of free memory depletion when you take an SVCDUMP. This can hit you the same way. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +966 556730876 --- On Tue, 19/1/10, Staller, Allan allan.stal...@kbm1.com wrote: From: Staller, Allan allan.stal...@kbm1.com Subject: Re: DFSORT memory used paging To: IBM-MAIN@bama.ua.edu Date: Tuesday, 19 January, 2010, 13:19 Page out is asynchronous and will only affect paging performance to the extent that the channel/device is not available for demand page-in. snip By %USE I meant page slots used, meaning increase in page out. Why does workflow drop from %use? too much juice wasted on paging out/in. The issue is not page config, but increase in paging out/in drop in workflow. /snip -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
DB2 and PAV
I have a situation where we are writing to a Shark DASD from DB2. I know DB2 will start multiple write engines to write the data from the buffer pool. However all the writes will go to the same volume. My questions are about PAV. 1. The DASD controller will serialise writes to the same extent - as specified in the define extent CCW. What does DB2 specify for the extent? Is the the DASD extent or just the area it is writing to? If so what area is specified? I cannot find this documented anywhere. The many write engines will be trying to build the single table. 2. We may have many write engines started to the same DB2 table on a single volume. There may be many of these.What is the maximum number of PAVs that WLM will give it? The reason for the question is to evaluate the benefit (if any) of running multiple JOBs which will all contribute to the data Db2 will write to the single tablespace (i.e. dataset). Terry Draper zSeries Performance Consultant -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: WLM Imp1
I see no technical reason to not put anything in importance 1. The only thing that matters is the relative importance. Importance 1 does not mean anything special. HOWEVER! I can see why you would not want to much there. If you put many of your important workloads in importance 1 and you found that one workload was extra special, then what do you do? I can understand designing around importance 2 to discretionary and then see what really has a strong business need for importance 1. This is one of those It Depends questions. There is no answer which is right for everyone. Just understand the implications and make your own decisions. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +966 556730876 --- On Tue, 24/11/09, R Hey sys...@yahoo.com wrote: From: R Hey sys...@yahoo.com Subject: WLM Imp1 To: IBM-MAIN@bama.ua.edu Date: Tuesday, 24 November, 2009, 8:23 Hi, It's been recommended NOT to use Imp 1 in WLM. My client has been using Imp1 for online CICS. I’m planning to change it to use Imp2 buy changing all Imp to Imp+1. Can you think of any potential problems/issues? TIA, Rez -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: DB2 and PAV
Bill and John, Thanks for your reply. You said that eight was the maximum for MA. Did you mean Multiple Allegiance, i.e. accesses from multiple systems. Does this also apply when all I/Os are form one system? I understand how things work, its just the undocumented bits that are difficult. It was easier to get these when I was in IBM doing this job. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +966 556730876 --- On Tue, 24/11/09, John Baxter john.bax...@atcoitek.com wrote: From: John Baxter john.bax...@atcoitek.com Subject: Re: DB2 and PAV To: IBM-MAIN@bama.ua.edu Date: Tuesday, 24 November, 2009, 15:59 Bill is correct in that DB2 will minimize the extent sizes for write operations, therefore increasing the likelihood of multiple, parallel operations to the same device. But also remember that DB2 will try to maximize the effectiveness of write I/O's, endeavoring to externalize as many pages as possible in one I/O operation (which may contain many chained CCW's). Another point to remember is that if the aliases are managed by WLM, a given DASD volume may become starved, and IOSQ builds up because of the long WLM process cycle. There are threshold controls for the BP's, and generally one tries to set the pageset level thresholds fairly low to discourage write bursts and maintain a steady trickle of writes. Obviously it would be ideal to strike a balance between write efficacy and keeping the write I/O's from overly bunching-up. Another strong recommendation is to implement HyperPAV aliasing, which we found to almost totally mitigate the concerns mentioned above. These aliases are assigned to I/O operations as needed and released for reuse (within the same LCU or associated CSS, if configured) after the extent-level I/O has completed. The maximum number of aliases assigned to any one base address can be very high (and vendor-dependent, I believe). This area of DB2 performance management is challenging but extremely interesting and needs to pull together expertise from the DBAs, sysprogs and DASD specialists. John Baxter -Original Message- From: IBM Mainframe Discussion List [mailto:ibm-m...@bama.ua.edu] On Behalf Of Bill Fairchild Sent: Tuesday, November 24, 2009 7:26 AM To: IBM-MAIN@bama.ua.edu Subject: Re: DB2 and PAV 1. DB2 tries very hard to minimize the size of the extent defined in the Define Extent CCW for each I/O that it does to a Shark-type DASD that supports Multiple Allegiance (MA). E.g., if an I/O references only one track, then that channel program will have only that one track within the defined extent. To verify, run GTF and trace I/Os to that one volume from DB2, then look at the defined extents that will appear in the Prefix CCW for SSCH trace records and in the Define Extent for I/O interrupt trace records. 2. I don't know for sure, but I imagine that WLM will allow the device to have as many PAVs as its controller microcode support. E.g, for the 2105s available in 2000 when I last looked at this issue, the maximum number of simultaneous I/Os allowed by the microcode to any one device was eight regardless of how many PAVs were assigned. Eight was the maximum MA level. The number of PAVs can be less than, equal to, or greater than the max MA number, but any more than the max will result in queued I/O requests if the workload produces simultaneous I/Os fast enough. Bill Fairchild Software Developer Rocket Software 275 Grove Street · Newton, MA 02466-2272 · USA Tel: +1.617.614.4503 · Mobile: +1.508.341.1715 Email: bi...@mainstar.com Web: www.rocketsoftware.com -Original Message- From: IBM Mainframe Discussion List [mailto:ibm-m...@bama.ua.edu] On Behalf Of Terry Draper Sent: Tuesday, November 24, 2009 2:32 AM To: IBM-MAIN@bama.ua.edu Subject: DB2 and PAV 1. The DASD controller will serialise writes to the same extent - as specified in the define extent CCW. What does DB2 specify for the extent? Is the the DASD extent or just the area it is writing to? If so what area is specified? I cannot find this documented anywhere. The many write engines will be trying to build the single table. 2. We may have many write engines started to the same DB2 table on a single volume. There may be many of these.What is the maximum number of PAVs that WLM will give it? Terry Draper zSeries Performance Consultant -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html The information transmitted is intended only for the addressee and may contain confidential, proprietary and/or privileged material. Any unauthorized review, distribution or other use of or the taking of any action in reliance upon this information is prohibited. If you receive this in error, please contact
Re: DB2 and PAV
Ted, I agree with you about SLA's etc. But this is an investigation about how we partition a huge tablespace that is reaching the size limit. The tablespace will be unavailable to the users during the copying and we are trying to find the best copy processing to reduce the outage. The current suggested process writes to each partition and fills it before going to the next. Hence the questions about concurrent write I/O. Sometimes you do need to get down into the details. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +966 556730876 --- On Tue, 24/11/09, Ted MacNEIL eamacn...@yahoo.ca wrote: From: Ted MacNEIL eamacn...@yahoo.ca Subject: Re: DB2 and PAV To: IBM-MAIN@bama.ua.edu Date: Tuesday, 24 November, 2009, 19:13 This area of DB2 performance management is challenging but extremely interesting and needs to pull together expertise from the DBAs, sysprogs and DASD specialists. Too many people get hung up on the technical aspects. Rather than worrying about: How many write engines? How is the Defined Extent used/sized? How many (HIPER)PAVs are allocated/usable? VSAM chaining? ETC? The real issue is: Are my Service Levels being met? AND: Is my resource consumption appropriate? If the answer to the above two questions is YES -- relax. If NO, then use your understanding of the technology to fix it. - Too busy driving to stop for gas! -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Ron Ferguson's VSAM Book
There is also an IBM ITSO Redbook called VSAM Demystified. Published in 2003. Has good rating from feed backs. In practise I have not found VSAM too difficult to tune. You just need to research how it works first. I have got major improvements from a little work. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +966 556730876 --- On Tue, 6/10/09, Jim Marshall jim.marsh...@opm.gov wrote: From: Jim Marshall jim.marsh...@opm.gov Subject: Ron Ferguson's VSAM Book To: IBM-MAIN@bama.ua.edu Date: Tuesday, 6 October, 2009, 12:15 PM Looked in my archives and came up with Virtual Sequential Access Method The Complete Source Book for VSAM File Structures By Ronald Ferguson Seventh Printing 1992 no code assigned as is done today Ron used to give these to anyone who took his VSAM training when he ran SIS (Software Information Services). Many times at SHARE when he was presently the Mainstar (SIS Follow on firm) booth would have some. You may want to check with Mainstar if they are available. Much of the content is in the VSAM Manager software product manual. Jim Marshall -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: db2 data sharing - group attachment
Munif, DB2 aplications can only connect to a DB2 in the same image. You cannot shut down one member of the DB2 data sharing group and still access DB2 data from that image. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +966 556730876 --- On Wed, 30/9/09, Munif Sadek munif.sa...@gmail.com wrote: From: Munif Sadek munif.sa...@gmail.com Subject: db2 data sharing - group attachment To: IBM-MAIN@bama.ua.edu Date: Wednesday, 30 September, 2009, 8:29 AM Hi I have started two db2 data sharing members on two different z/OS images in a parallel Sysplex and can connect to them via group attachment name (SSID) in a batch job. My problem is If i shut down one of the member I am not able to use group attachment name on that z/OS image to *automagically* send my request across z/OS image to the other DB2 member. what's the missing bit here and how much work is involved to achieve this. we are z/OS 1.9 and DB2 9 system. Any pointers in the right direction is highly appreciated. Best regards Munif -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: What happens in a DS8000 Metro Mirror (PPRC-Sync) config when the relationship between a volume pair is broken?
Jan, I suggest you look in the Redbook at this address: http://www.redbooks.ibm.com/abstracts/sg246787.html Book number is: SG24-6787 Title: DS8000 Copy Services for IBM Systems z It is dated February 2009 and so is up to date. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +966 556730876 --- On Sat, 29/8/09, Jan Vanbrabant jan.vanbrab...@telenet.be wrote: From: Jan Vanbrabant jan.vanbrab...@telenet.be Subject: What happens in a DS8000 Metro Mirror (PPRC-Sync) config when the relationship between a volume pair is broken? To: IBM-MAIN@bama.ua.edu Date: Saturday, 29 August, 2009, 1:33 PM Hi, 1° What happens in a DS8000 Metro Mirror (PPRC-Sync) config when the relationship between a volume pair is broken, say, because of network problems? This network outage could for example last for hours. How does the system keep track of the source updates changes in the mean time, in order to resynchronize the volume pair later on? With the IBM FlashCopy SE feature, there is something like the copy repository where only the capacity needed to save pre-change images of the source data is allocated. Is there also kind of such a “copy repository” with Metro Mirror? If yes, has it to be sized? Or is the storage server capable of handling managing this by itself? Where can I find more info on this? Have any benchmarks or tests been done in this perspective? I would appreciate it utmost if you would route me to appropriate reading material? 2° I assume (dare hope) that the resynchronization process can be (or is) prioritized the right way in order to not perturbate the real usually ongoing I/O response time? Cheers,Jan -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Can CICS region share more than one processor
Don, Prior to L8 TCBs, DB2 SQL requests executed on threads. Each thread was a seperate TCB. So the QR TCB and each thread TCB could execute in parallel. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +966 556730876 --- On Wed, 19/8/09, Don Deese don_de...@cpexpert.org wrote: From: Don Deese don_de...@cpexpert.org Subject: Re: Can CICS region share more than one processor To: IBM-MAIN@bama.ua.edu Date: Wednesday, 19 August, 2009, 3:48 PM Hi Mohammad, I don't understand the relevance of your question within the context of Dr. Merrill's posting. Dr. Merrill was illustrating the concurrent execution of multiple TCBs (both the QR TCB and L8 TCBs) that would not have been possible prior to the OTE design. It is somewhat irrelevant as to whether DB2 CPU time was charged to a QR TCB or to an application TCB in past. The fact that the DB2 processing would have been executed in series off QR TCBs, but now executes in parallel with L8 TCBs is the important point that Dr. Merrill was making. Keep in mind that Dr. Merrill was responding to the OP query Can CICS region share more than one processor. FWIW, I have data from CPExpert users that show 10, 20, 30 or more L8 TCBs executing concurrently and using more than 100% of CPU (namely, they are using multiple CPUs concurrently for more than 100% of the time). This would not have been possible prior to the OTE design. As Dr. Merrill pointed out, there are 22 TCB types with CICS/TS 4.1, and several of these TCB types can have multiple TCBs executing concurrently. The number of current TCBs can be controlled by the MAXxxxTCBS in the SIT or in the JVMPROFILE Resource Definition. There are, of course, limiting factors inherent in the environment (for example, CICS monitors the amount of available MVS storage and will not attach new TCBs from the JVM TCB pool if storage is severely constrained). Regards, Don ** Don Deese, Computer Management Sciences, Inc. Voice: (804) 776-7109 Fax: (804) 776-7139 http://www.cpexpert.org ** At 09:16 AM 8/13/2009, you wrote: Thanks Dr. Merrill for your illustrative example but I do have a question about it. Since L8 TCBs are used to execute DB2 code as well, what part of 10,298 seconds is for DB2 ? Since DB2 related code never executed on QR TCB anyway, that portion of CPU usage is moot for this discussion. The real question is how much of this CPU now runs on L8 TCB which used to run on QR TCB due to the aggressive OTE exploitation ? Regards Mohammad On Wed, 12 Aug 2009 15:18:23 -0500, Barry Merrill ba...@mxg.com wrote: In the old days, a CICS subsystem's capacity was limited by the amount of CPU TCB time needed for that single QR TCB. Based on my analysis when OTE was brand new, of the CPU time consumed by each of these new CICS TCBs, I planned this post to argue that going to OTE didn't help much, because most of the CICS CPU time was still being spent under the QR TCB. I could NOT have been more wrong! Analyzing new CICS/TS 4.1 Open Beta data from a VERY aggressive OTE exploiter site shows (from their SMF 110, subtype 2 Dispatcher Statistics segments, MXG CICDS and CICINTRV datasets): Total TCB CPU in Dispatcher Records = 13,080 seconds Total TCB CPU in QR TCB = 2,776 seconds Total TCB CPU in L8 TCB = 10,298 seconds Total TCB CPU in all other TCBs = 6 seconds Aha, you say, OTE still doesn't help; the CPU time just moved from the QR TCB to the L8 TCB, so the capacity limit just moved from one TCB to the other, right? Wrong again. While the QR TCB can attach only a single TCB, these new TCBs can attach multiple TCBs; in fact, the SMF data shows that the L8 TCB attached a maximum of 22 TCBs, each of which is a separate dispatchable unit. So, it REALLY does look like that these multiple OTE TCBs do eliminate the old one-TCB CICS capacity limitations, and does indeed spread your CICS time across MANY TCBs. (Total SRB time in the Dispatcher Records was only 65 seconds.) Barry Merrill Herbert W. Barry Merrill, PhD President-Programmer Merrill Consultants MXG Software 10717 Cromwell Drive Dallas, TX 75229 ba...@mxg.com http://www.mxg.com admin questions: ad...@mxg.com technical questions: supp...@mxg.com tel: 214 351 1966 fax: 214 350 3694 -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Can CICS region share more than one processor
Tommy,' You say you are VSAM. There is a sub-task for some VSAM functions which is an option within a CICS address space. I do not think it takes a lot away from the main task, but make sure you are using this option. You could set up a TOR and route to multiple AORs and for VSAM read only files, these could be shared between AORs. For update files create a FOR and ship the requests from the AOR to the AOR. You could ship all VSAM requests to the FOR if you want. This will add to the overall CPU but will split the load across several tasks. It should not require application code changes. You need to decide how to split the load between the AORs and implement in the TOR the appropriate routing. You can use CICSPLEX SM to control the routing or create your own routing exit. I thought most users use the TOR/AOR structure now. to avoid your problem. I would also look at data tables to try to reduce the CPU load for VSAM accesses. Or you could go out and buy a big engine z10. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +966 556730876 --- On Thu, 13/8/09, Tommy Tsui tommyt...@gmail.com wrote: From: Tommy Tsui tommyt...@gmail.com Subject: Re: Can CICS region share more than one processor To: IBM-MAIN@bama.ua.edu Date: Thursday, 13 August, 2009, 11:40 PM Our data access for applications is VSAM only without DB2. We have not implement OTE yet that means we only has QR TCB. As you know, we have to re-write many user program either switch to cicsplex with RLS or DB2 access. On Fri, Aug 14, 2009 at 1:01 AM, Terry Draperw...@btopenworld.com wrote: Tommy, Can I ask a couple of fundamental questions? What is the data access for the applications? Is it DB2, DL/1, VSAM or something else. If DB2 (and I think DL/1) these will already be running on threads and these use their own TCBs, 1 per thread. If so I cannot understand your problem. Also do you have a TOR and AORs structure. If not I suggest you go that way. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +966 556730876 --- On Wed, 12/8/09, Tommy Tsui tommyt...@gmail.com wrote: From: Tommy Tsui tommyt...@gmail.com Subject: Can CICS region share more than one processor To: IBM-MAIN@bama.ua.edu Date: Wednesday, 12 August, 2009, 2:44 PM Hi , We hit a problem that our cics cannot utilized more than one CPU processor and IBM recommend our shop upgrade to CICSplex .Except this, is there any other way to solve this problem? any comment will be appreciated best regards -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Can CICS region share more than one processor
Tommy, Can I ask a couple of fundamental questions? What is the data access for the applications? Is it DB2, DL/1, VSAM or something else. If DB2 (and I think DL/1) these will already be running on threads and these use their own TCBs, 1 per thread. If so I cannot understand your problem. Also do you have a TOR and AORs structure. If not I suggest you go that way. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +966 556730876 --- On Wed, 12/8/09, Tommy Tsui tommyt...@gmail.com wrote: From: Tommy Tsui tommyt...@gmail.com Subject: Can CICS region share more than one processor To: IBM-MAIN@bama.ua.edu Date: Wednesday, 12 August, 2009, 2:44 PM Hi , We hit a problem that our cics cannot utilized more than one CPU processor and IBM recommend our shop upgrade to CICSplex .Except this, is there any other way to solve this problem? any comment will be appreciated best regards -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: SDSF DA fields CPU/L/Z/
David, On the SDSF DA panel the fields CPU/L/Z are as follows: CPU ( SDSF help calls this MVS busy) this is the percentage busy the the z/OS operating system thinks it is busy. This includes both the time the logical CP is running plus the time the logical CP is not in a wait state and is waiting for PR/SM to dispatch it. This is because another LPAR is running. L (SDSF help call this LPAR busy) this is the logical CP is running. This is the correct CPU utilisation of the LPAR. Z (SDSF help calls this zAAP busy) this is the zAAP busy percent if you have one. To explain the first two let me sketch a diagram: = time A B C Logical CP running === Logical CP running If during the whole time z/OS has work to run the MVS busy would be the whole time (A + B + C). LPAR busy would be the whole time less the part (A + C). The part is when other LPARs have control the real CPU (C). The bigger the difference between MVS busy and LPAR busy the more your work is being delayed by other LPARs. If this is you main production system them you should look at increasing this LPAR's weight. You owe me a pint for this quick course. Just joking. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +966 556730876 --- On Wed, 5/8/09, David Speake david.spe...@bcbssc.com wrote: From: David Speake david.spe...@bcbssc.com Subject: SDSF DA fields CPU/L/Z/ To: IBM-MAIN@bama.ua.edu Date: Wednesday, 5 August, 2009, 12:02 AM SDSF DA displays a line of information above the 'COMMAND INPUT' line. SDSF DA SYSJ (ALL) PAG 0 CPU/L 68/ 34 LINE 1-1 (1) Can anyone point me to an IBM manual for citation as to exactly what CPU/L 68/ 34 in the above mean. The Help panel shows CPU/L/Z/ 26/ 26/ 0 | Percentage of time the CPU is busy, MVS, LPAR and zAAP views. The zAAP view would, I guess, reflect busy % for all CPUs (if any) designated as zAAP processors by PR/SM. But what is the definition/distinction for CPU/L (MVS/LPAR) views. I am looking at z/OS: Resource Measurement Facility Performance Management Guide, B.4.2 Value of LPAR CPU management and I cannot form a good definition. An explanation would be appreciated. A definition citation even more so. David -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: WLM Service Class assignment Order
Jacky, With response time goals in CICS, the CICS STC service class is only used at start up. After that WLM builds its own own groups to put CICS address spaces in. Depending on the CICS transactions within an address space and how they are achieveing their goals, the CICS address space as a whole is put in the appropriate group. WLM can only manage CICS at the address space level. What you find is that all your CICS address spaces move together if you distribute the same transactions across all your CICS. You can see this in RMF II where all the AORs may have the same dispatching priority and maybe your TORs have another priority. There is a presentation on the IBM WLM web site describing this. http://www-03.ibm.com/servers/resources/servers_eserver_zseries_zos_wlm_pdf_cidwlm_pdf_cidwlm.pdf Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +966 556730876 From: Jacky Bright jacky.bri...@gmail.com To: IBM-MAIN@bama.ua.edu Sent: Thursday, 4 June, 2009 2:04:14 PM Subject: WLM Service Class assignment Order Can anyone tell me how WLM decides which service class to assign to CICS Subsystem ? I have following Classification Rules defined in system. For the type CICS following are the entries : Qualifier Qualifier Starting Service Report # type name position Class Class - -- -- - 1 SI RTGCICS CICS1TP RTG3T 2 . TN . M* CICS1TP IASP 2 . TN . V* CICS1TP NDSAP 1 SI R1%CICS CICSSLA and For the type STCS following are the entries : Qualifier Qualifier Starting Service Report # type name position Class Class - -- -- - 1 SYG ABCRL SYSSTC 2 . TN . R1%CICS CICSSTC RCICS 2 . TN . E%CICS CICSSTC ECICS 2 . TN . R%CICS CICSSTC FCICS In my SDSF for DA panel I can see that for R*CICS region the Service Class is CICSSTC. But confused whether there is any significance for the entry defined for type CICS. i.e. CICSSLA service Class. JAcky -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Global mirroring
We have synchronous from A to B and asynchronous from A to C. Sounds just what you want to do. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +966 556730876 --- On Tue, 2/6/09, Michael Saraco michael.sar...@baer-consulting.com wrote: From: Michael Saraco michael.sar...@baer-consulting.com Subject: Global mirroring To: IBM-MAIN@bama.ua.edu Date: Tuesday, 2 June, 2009, 5:52 PM The question I have is. Can I use Global mirroring to copy from site A to Site B and also at the same time copy from site A to site C? Or do you have to go from site A to B then from site B to site C? -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: System data sets and 3390-27
Brad, I have seen some responses to your question, which raised questions. I will ignore the RVA, which works differently with its virtualisation. The latest DASD does reserve all the space for a specified volume. So allocating a 1 cylinder data set on a mod 27 would waste nearly all the space. With some of the copy functions you can get away with less for the copy, but not the original volumes. The data is striped across the raid DASD, so the real I/Os for a specific volume can be to different RAID DASD in parallel. So a mod 27 is not an issue here. One thing that would be an issue is if you did real RESERVEs and did not convert them to systems wide enqueues. You would be RESERVEing a lot of data from other systems. My major concern is the actual performance of a mod 27, irrespective of the type of data. If you put a lot of active data on a mod 27 then it will need more PAVs than a mod 3 or mod 9. I know you could have WLM managed dynamic PAVs. But the WLM addition of extra PAVs will lag behind an increase in I/O activity to a volume. The delays will be seen in IOSQ time. One solution is HIPERPAVs, which are only on the latest IBM DASD - these allocate PAVs from a pool as required rather than assigning to specific volumes. I do not think EMC have HIPERPAVs and do not know if they have committed to them and if so when. I think I would play safe for your new environment and put busy system data sets on smaller models than 27. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +966 556730876 --- On Mon, 18/5/09, Brad Wissink bjwi...@iastate.edu wrote: From: Brad Wissink bjwi...@iastate.edu Subject: System data sets and 3390-27 To: IBM-MAIN@bama.ua.edu Date: Monday, 18 May, 2009, 8:39 PM We are moving from a Shark with no PAV's to a EMC SAN with PAV's. We are going to use 3390-27's. I have read most of the archives on 3390-27's but I did not see anything on running your sysres or system data sets from a 27? We are thinking about going totally with 3390-27's, which means we must run with the sysres, jes checkpoint and spool, catalogues, RACF database, etc on the 27's. Is anyone doing that? What do we need to watch out for? -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Any Batch Pipes experience?
I seem to remember that there was a JOB restructuring function available with Batch Pipes. Not aware that it was used much. Or if its still around. The problem with that product was that it relied on history. As soon as you change a JOB, for the next N (not sure how many) runs it does no optimization. This is while it relearns the new JOB structure. I think any JCL change to the JOB would stop the optimization. Thus the batch window will run longer for this period of relearning. Also there were several things it could not detect. Me, I would go with a manual process to fix the big gains. Terry Draper zSeries Performance Specialist --- On Thu, 19/3/09, Spencer, Mike mike_spen...@bmc.com wrote: From: Spencer, Mike mike_spen...@bmc.com Subject: Re: Any Batch Pipes experience? To: IBM-MAIN@bama.ua.edu Date: Thursday, 19 March, 2009, 11:26 AM Ah, to keep with the 80's theme, way, dude if you have lots of time on your hands. Yes, the user could insert another JOB card, but with thousands of jobs executing, and not being a consultant, I have better things to do than waste my company's dollars trying to manually streamline my batch processes. And of course there are no IO savings by inserting another JOB card. MAINVIEW Batch Optimizer provides the ability to pipe around DB2 or IMS steps to negate any negative impact. It's all a very simple process to implement. Michael Spencer BMC Software -Original Message- From: IBM Mainframe Discussion List [mailto:ibm-m...@bama.ua.edu] On Behalf Of Nemo Sent: Wednesday, March 18, 2009 12:59 PM To: IBM-MAIN@bama.ua.edu Subject: Re: Any Batch Pipes experience? On Wed, 18 Mar 2009 07:45:09 -0500, Spencer, Mike wrote: Batch Pipes can only move data between two different jobs. It cannot move data between steps because in a Batch Pipes world, there is no way to get multiple steps running in parallel for the pipe to work. no way?!? All the user needs to do is insert another JOB card (non-duplicate name is a plus but not a requirement if JES' duplicate jobs are allowed to execute concurrently). That sounds like a way to me. MAINVIEW Batch Optimizer from BMC will run steps in parallel, entire jobs in parallel, and optimizes QSAM and native VSAM I/O processing among other items. Does BMC's MAINVIEW Batch Optimizer have a clue if more than one of those parallel job steps are updating the same DB2 table (and thus potentially causing damage)? Some steps have implicit serialization requirements. How does BO cope with those? -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Any Batch Pipes experience?
I just did some research on the old faithful Google for Batchpipes. I found in Wikipedia, in the History section: BatchPipes Version 1 was developed in the late 1980s and early 1990s simply as a technique to speed up MVS/ESA batch processing. In 1997 the functionality of BatchPipes was integrated into a larger IBM product - SmartBatch (which incorporated two BMC Corporation product features: DataAccelerator and BatchAccelerator). However SmartBatch was discontinued in April 2000. Subsequently BatchPipes Version 2 was released, incorporating BatchPipes Version 1 and some additional features from SmartBatch: BatchPipePlex and BatchPipeWorks. BatchPipes Version 2 is still a marketed IBM product. It was SmartBatch that I was thinking of. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +966 556730876 --- On Fri, 20/3/09, Nemo plumbersar...@gmail.com wrote: From: Nemo plumbersar...@gmail.com Subject: Re: Any Batch Pipes experience? To: IBM-MAIN@bama.ua.edu Date: Friday, 20 March, 2009, 5:43 PM On Fri, 20 Mar 2009 17:32:31 +, Terry Draper wrote: I seem to remember that there was a JOB restructuring function available with Batch Pipes. Not aware that it was used much. Or if its still around. The problem with that product was that it relied on history. As soon as you change a JOB, for the next N (not sure how many) runs it does no optimization. This is while it relearns the new JOB structure. I think any JCL change to the JOB would stop the optimization. Thus the batch window will run longer for this period of relearning. Also there were several things it could not detect. You have confused IBM's Batch Pipes (BatchPipes/MVS) with BMC's MAINVIEW Batch Optimizer. BMC's BO relies on the job history, much as you said. BatchPipes relies on a one-time manual conversion and it is all smooth sailing from there. (No history, no downstream relearning.) And, as the poster from Brazil stated - it works great. Me, I would go with a manual process to fix the big gains. Me, too. One and done. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: SMF reporting question
We use RMF Spreadsheet reporter. The customer does not have SAS or MXG. It has lots of canned graphs for z/OS things. It comes complete with RMF. We include a lot of these graphs in our monthly technical report to management. The tool is not user friendly but it does the job. It will not address the who is using or CICS things. I have taken the 30 record and built a cut down flat file (no repeat sections) which I can report on easily. I used assembler - you get the mappings with z/OS - so easier to use. Certainly look at using RMF Spreadsheet Reporter. Then look elsewhere for the others. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +966 556730876 --- On Thu, 26/2/09, David Betten bet...@us.ibm.com wrote: From: David Betten bet...@us.ibm.com Subject: Re: SMF reporting question To: IBM-MAIN@bama.ua.edu Date: Thursday, 26 February, 2009, 7:17 PM I'm not that familiar with it but the RMF Spreadsheet Reporter might povide you with some z/OS performance reporting. I'm pretty sure it comes with RMF so it probably meets your requirement of being something you already have. Have a nice day, Dave Betten DFSORT Development, Performance Lead IBM Corporation email: bet...@us.ibm.com DFSORT/MVSontheweb at http://www.ibm.com/storage/dfsort/ IBM Mainframe Discussion List IBM-MAIN@bama.ua.edu wrote on 02/26/2009 11:07:43 AM: I realize that most people likely use SAS for SMF reporting. We did this in the past. Management has declared that SAS on the z is simply too expensive. They also declared that installing the Windows version on our main user's PC is also too expensive. End of discussion on that point. So, what else could be used? It must either be something that we already have, such as COBOL, EasyTrieve Plus (without the SMF add on), REXX, HLASM, ... . We cannot spend any hard money on this. Oh, it would be nice if if were very CPU efficient because we just downgraded our z9BC from a V02 to a T02 in order to save on software costs. And it must be such that doing ad hoc requests can be responded to quickly. Yes, I know, give me the world, but don't spend any money. I am actually looking at downloading the raw SMF data to a Linux box (my desktop) using BINary and SITE RDW. I did this for last week's SMF and had about 14Gib of data. I know how to read this with Java. This may actually be what I end up looking at doing. But I am the only person in my group who is even mildly Java literate. The main performance person is not. And he doesn't have access to my PC anyway. Of course, that is one reason that I'm looking at Java. I have written Java in the past (minor application) which truly was run anywhere. At least it ran, as compiled on the Linux box, on 32 bit Linux/Intel, 64 bit Linux/Intel, Mac OS/X, 32 bit Windows, and on the z. I just transferred the jar file and ran it. Any thoughts or commiserations appreciated. -- John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Z11 - Water cooling?
Peter, You are correct that some of the 9672's had water cooling. It started at the top of the range (cannot remember which model). If I remember correctly it was used for all models on later ranges. I could find out if anyone wants to know. I put water cooled in quotes because it depends what you mean. There is the case of a water enclosed system that chills the air before it is passed over the CPU s. This can be called water cooled, but no water passes to the CPU packaging. This is what appeared in the 9672's. This compares to the old water cooled processors, like 3090's, where the water actually flows through the CP module. I am not aware of anything like this being used on the latest processors. It depends if you would really call the first a water cooled system. The rumour needs to be clarified on what exactly they are implying. If I have time I will research and give the full picture. Terry Draper zSeries Performance Consultant w...@btopenworld.com mobile: +966 556730876 --- On Mon, 2/2/09, Hunkeler Peter (KIUK 3) peter.hunke...@credit-suisse.com wrote: From: Hunkeler Peter (KIUK 3) peter.hunke...@credit-suisse.com Subject: Re: Z11 - Water cooling? To: IBM-MAIN@bama.ua.edu Date: Monday, 2 February, 2009, 8:17 AM With the z9 EC, IBM used what they called hybrid cooling. The first z/Architecture machines (zSeries 900) already had them. Didn't remember if the latest 9672 also had MCU's but a quick search showed that at least some of the 9672 G5 and G6 models also had a closed loop liquid cooling system. -- Peter Hunkeler Credit Suisse -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html