Re: 64-bit Virtual Questions
This discussion got me thinking about my systems, z/OS R4 on twin zSeries processors. Specifically, what would I 'see' from TMON/MVS if a Job/Task/TSU were to use above-the-bar storage? So I copied the assembler code that Todd Burch posted, added a IARV64 GETSTOR/DETACH requests that I copied from an example from Section: 4.9.1 Operation in the manual: z/OS MVS Programming: Assembler Services Reference, Volume 2 (IARR2V-XCTLX), Manual Number: SA22-7607-04. As my luck goes ( which is usually not great ), the coding of the example from the manual is not correct. From code from the manual: IARV64 REQUEST=GETSTOR,SEGMENTS=NUMSEG, ORIGIN=O, RETCODE=LRETCODE,RSNCODE=LRSNCODE, MF=(E,V64L) The NUMSEG label was coded as below: NUMSEG DCD'2' The correct coding of the value for the NUMSEG label should be: NUMSEG DCAD(2) I checked, the z/OS R7 manual still has the value coded incorrectly. The IBM'er I spoke to ( via an ETR ) indicated he would submit a Reader's Comment Form to correct the coding error. Glenn Miller PS: If you execute the code from the example with the incorrect coding of the NUMSEG label, you should receive the following abend: System Abend Code: DC2, and REASON CODE=4A001520 --- This message (including any attachments) is confidential and may be privileged. If you have received it by mistake please notify the sender by return e-mail and delete this message from your system. Any unauthorised use or dissemination of this message in whole or in part is strictly prohibited. Please note that e-mails are susceptible to change. ABN AMRO Bank N.V. (including its group companies) shall not be liable for the improper or incomplete transmission of the information contained in this communication nor for any delay in its receipt or damage to your system. ABN AMRO Bank N.V. (or its group companies) does not guarantee that the integrity of this communication has been maintained nor that this communication is free of viruses, interceptions or interference. --- -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: 64-bit Virtual Questions
Glenn Miller wrote: As my luck goes ( which is usually not great ), the coding of the example from the manual is not correct. From code from the manual: IARV64 REQUEST=GETSTOR,SEGMENTS=NUMSEG, ORIGIN=O, RETCODE=LRETCODE,RSNCODE=LRSNCODE, MF=(E,V64L) The NUMSEG label was coded as below: NUMSEG DCD'2' The correct coding of the value for the NUMSEG label should be: NUMSEG DCAD(2) When I first started writing 64-bit programs back in 2002, I found the IARV64 macro would generate D'n' instead of AD(n) if you coded SEGMENTS=n where 'n' was an integer value! Naturally, the request would always fail with a DC2 abend. I opened a PMR with IBM that resulted in APAR OW54887. (How embarrassing for the RSM folks!) It never occurred to me to check to see if documentation samples were wrong as well. They were probably provided by the same genius that coded the macro. :-D -- - | Edward E. Jaffe|| | Mgr, Research Development| [EMAIL PROTECTED]| | Phoenix Software International | Tel: (310) 338-0400 x318 | | 5200 W Century Blvd, Suite 800 | Fax: (310) 338-0801| | Los Angeles, CA 90045 | http://www.phoenixsoftware.com | - -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: 64-bit Virtual Questions
In [EMAIL PROTECTED], on 09/02/2005 at 03:06 PM, Michael Cleary [EMAIL PROTECTED] said: What I am trying to test before having a crisis or outage is how TMON and RMF can be used to easily identify 64-bit virtual users who are impacting the system. Impacting the system how? You need enough local paging to cover all of your address and data spaces, whether they are a few big ones or lots of smaller ones. Similarly, you need enough real storage to hold your working set, regardless of how it is distributed. Certainly a large space will chew up more in page tables than a small one will, but why would you expect a 16 GiB space to chew up more than 16 1 GiB spaces? -- Shmuel (Seymour J.) Metz, SysProg and JOAT ISO position; see http://patriot.net/~shmuel/resume/brief.html We don't care. We don't have to care, we're Congress. (S877: The Shut up and Eat Your spam act of 2003) -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: 64-bit Virtual Questions
David Bond did an excellent session at SHARE in Washington DC which is in the proceedings 8158 - 64-bit z/OS Assembler Coding http://ew.share.org/client_files/callpapers/attach/Washington_DC/s8158a.pdf 64-bit support is now in z/OS but there are few examples of why and how to write 64-bit code. Come to learn about the hardware and software support for 64-bit registers, addressing mode and storage and how to exploit them. The demo program is embedded in the foils. I will drop a note to David and maybe he can stick his sample program on his web site http://tachyonsoft.com where you can find his excellent cross assembler (free for MVS 3.8 use) or on CBT or both. Best Regards, Sam Knutson, GEICO Performance and Availability Management mailto:[EMAIL PROTECTED] (office) 301.986.3574 Anyone who has never made a mistake has never tried anything new. - Albert Einstein -Original Message- 2) Has anyone built a simple assembler program using IARV64 to get 64-bit virtual storage, that we can run on a test LPAR to check reality and see what TMON and RMF see of the 64-bit virtual workload, if anything. If so, can you share the program source? Cheers, Michael This email/fax message is for the sole use of the intended recipient(s) and may contain confidential and privileged information. Any unauthorized review, use, disclosure or distribution of this email/fax is prohibited. If you are not the intended recipient, please destroy all paper and electronic copies of the original message. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
64-bit Virtual Questions
Greetings, We are poised to migrate to z/Architecture mode on z/OS 1.4. All of the software vendors have been contacted and all of them state that we are on levels of their products to function under z/Architecture mode. None of them have claimed that their products exploit 64-bit virtual. I am thinking about globally setting SMFPRMxx MEMLIMIT to something very low, as a starting point; maybe 2GB or 4GB. The unknown variable is what actual software at execution time will eploit 64-bit virtual under z/Architecture mode. Possibly, none. We are on DB2 V6 (ok, stop laughing) and are migrating to DB2 V7 (ok, I said stop laughing). From what I can see, DB2 V6 and DB2 V7 exploit 64-bit real, but not 64-bit virtual as DB2 V8 does. We run DFSORT 14.0. So, a few of questions: 1) Has anyone identified any z/OS 1.4 components, IBM software or ISV software that will exploit 64-bit virtual? If so, can you share the details? 2) Has anyone built a simple assembler program using IARV64 to get 64-bit virtual storage, that we can run on a test LPAR to check reality and see what TMON and RMF see of the 64-bit virtual workload, if anything. If so, can you share the program source? 3) The official z/Architecture (64-bit) migration checklist has a generic warning about the LRA instruction and ABEND 0D3. I have applied all of the IBM maintenance for ABEND 0D3. I have searched all user assembler code and freeware, and the only occurance of LRA is in ShowMVS, which Roland says is not a problem. Has anyone experienced any ABEND 0D3 in ISV code? If so, can you share the details? Cheers, Michael Start your day with Yahoo! - make it your home page http://www.yahoo.com/r/hs -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: 64-bit Virtual Questions
I am thinking about globally setting SMFPRMxx MEMLIMIT to something very low, as a starting point; maybe 2GB or 4GB. Not a bad plan but (a) if you have any 64-bit exploiters, make sure your paging configuration is up to the job and (b) you may have a few 64-bit exploiters you were not planning for, e.g. the JVM requires some storage above the bar when running on a 64-bit system. I don't recall how much exactly, but if your memlimit is too low the JVM simply craps out and won't run at all. Most embarrassing if you happen to be running any JAVA work. 1) Has anyone identified any z/OS 1.4 components, IBM software or ISV software that will exploit 64-bit virtual? If so, can you share the details? See above. I am not aware of any specific list of exploiters. The system itself and related middleware can be exploiters. Prior to V8 DB2 doesn't do any exploitation I am aware of, but other components may and a good many ISV products do too. You should contact the vendors of your products. 2) Has anyone built a simple assembler program using IARV64 to get 64-bit virtual storage, that we can run on a test LPAR to check reality and see what TMON and RMF see of the 64-bit virtual workload, if anything. If so, can you share the program source? What's to test? IARV64 is relatively simple and it works exactly like the books say it does. The workload has nothing to do with it. The system is trimodal. There is an unpredictable mix of code switching between 24, 31 and 64 bit modes constantly according to the state requirements of the running program. What would you expect TMON or RMF to see? 3) The official z/Architecture (64-bit) migration checklist has a generic warning about the LRA instruction and ABEND 0D3. I have applied all of the IBM maintenance for ABEND 0D3. I have searched all user assembler code and freeware, and the only occurance of LRA is in ShowMVS, which Roland says is not a problem. Has anyone experienced any ABEND 0D3 in ISV code? If so, can you share the details? Have I seen them? Sure. Everyone in the business probably has. 64-bit real support has been around long enough now that most of those have probably been flushed out, assuming you're reasonably current on product maintenance. After all, 64-bit real showed up with OS/390 2.10 and z/OS 1.4 is now the oldest supported release. CC -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: 64-bit Virtual Questions
-Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Craddock, Chris Sent: Friday 02 September 2005 11:26 To: IBM-MAIN@BAMA.UA.EDU Subject: Re: 64-bit Virtual Questions Michael Joseph Cleary wrote... 1) Has anyone identified any z/OS 1.4 components, IBM software or ISV software that will exploit 64-bit virtual? If so, can you share the details? See above. I am not aware of any specific list of exploiters. The system itself and related middleware can be exploiters. Prior to V8 DB2 doesn't do any exploitation I am aware of, but other components may and a good many ISV products do too. You should contact the vendors of your products. I agree with Crash. Vendors know best. Some of my current employer's products are running with AMODE 64 because of other products now exploiting 64 bit. But I wouldn't necessarily call that exploitation. A prior employer of mine, Software AG, announced ADABAS 64-bit exploitation on the day of IBM's announcement. We had access to 64-bit capable (under NDA) a long time earlier, and we had beta customers testing this in the field. I would expect that data base products were amongst the first to first exploit 64 bit. Best regards, Ray -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: 64-bit Virtual Questions
In a message dated 9/2/2005 2:19:47 P.M. Central Standard Time, [EMAIL PROTECTED] writes: would expect that data base products were amongst the first to first exploit 64 bit. If you're on a supported level and a z/Box! -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: 64-bit Virtual Questions
Thanks Chris, your comments are always valuable. We do not have JVM or DB2 V8. What I am trying to test before having a crisis or outage is how TMON and RMF can be used to easily identify 64-bit virtual users who are impacting the system. I have seen 64-bit virtual war stories posted here, and just wanted to get a feel for what this would like like to our monitoring products so that the 64-bit virtual users can be managed. The monitors have panels for diagnosing CSA, SQA, etc issues, I just wanted to get a feel for the capability of the monitors to asssist in identifying large 64- bit virtual users, other than just by the size of auxilary storage that they are using. We have a pretty robust and over sized local paging system. Have a great holiday weekend. Cheers, Michael -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: 64-bit Virtual Questions
On Fri, 2 Sep 2005 13:38:32 -0700, Edward E. Jaffe [EMAIL PROTECTED] wrote: I have seen 64-bit virtual war stories posted here, I must have missed those war stories. Could you please post links to the archives? I would like to read them. Ditto. We have been exploiting 64-bit virtual storage in our products since October, 2002. Because we were one of the first exploiters, we generated several APARs against RSM (both from in-house testing and via customers in the field). That's all ancient history now. The remaining issues of which I'm aware stem from system programmers not getting off their duff (to quote Greg Dyck from http://bama.ua.edu/cgi-bin/wa?A2=ind0504L=ibm-mainP=R50560) in order to specify a reasonable default MEMLIMIT value in SMFPRMxx. I had several conversations about IEFUSI and MEMLIMIT at SHARE with people. I'm not protecting / limiting it via IEFUSI (yet?), but I do have a default specified in SMFPRMxx. They asked me what a good default value was and my reply was the same as has been discussed previously on ibm-main - no one knows yet. Most of my LPARs have MEMLIMIT(10G), but some of the smaller ones have MEMLIMIT(5G). My fear is that even 10G is perhaps too small, but I haven't run into any problems yet (z/OS 1.6, WAS 5.1, DB2 V8). Mark -- Mark Zelden Sr. Software and Systems Architect - z/OS Team Lead Zurich North America and Farmers Insurance Group mailto: [EMAIL PROTECTED] Systems Programming expert at http://Search390.com/ateExperts/ Mark's MVS Utilities: http://home.flash.net/~mzelden/mvsutil.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: 64-bit Virtual Questions
From: Edward E. Jaffe I have seen 64-bit virtual war stories posted here, I must have missed those war stories. Could you please post links to the archives? I would like to read them. Selective memory Ed ???. Have a look for the problems Brian Peterson and I both had with RSM - in my case (at least) a SADUMP dump a week. Doesn't qualify as a war story ???. All this after we had tested 64-bit extensively at 2.10, and after I made a similar query to this very list, and got similar responses. m All that is in the past fortunately - at least Michael is lucky that way. As for the TMON testing, again old news. Right idea though - when I tested as Michael is suggesting, Sam (in his old role) got very narky when I mentioned (publicly) that they couldn't add up. The displays for above the bar were out to hell. I guess Michael could look at SMF for storage usage above the bar ... Sorry - bad joke ... Shane ... -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: 64-bit Virtual Questions
Narky? Perhaps just low on caffeine at the time:-) http://bama.ua.edu/cgi-bin/wa?A2=ind0306L=ibm-mainP=R89697I=1 If it's any consolation I am again a TMON customer and just got really narky with support about the slew of abends I got trying to use said displays to study use of memory objects by DB2 V8. Anyone who wants to use TMONMVS V3.2 to review use of memory objects will likely want these TD03490 S0C4 IN TMVS155 OFFSET 71E TD03747 STORAGE OVERLAY IN TMVS155 WHEN DISPLAYING MEMORY OBJECTS Thanks, Sam Knutson, GEICO Performance and Availability Management mailto:[EMAIL PROTECTED] (office) 301.986.3574 The computer is mightier than the pen, the sword, and usually, the programmer. -Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of ibm-main Sent: Friday, September 02, 2005 7:26 PM To: IBM-MAIN@BAMA.UA.EDU Subject: Re: 64-bit Virtual Questions From: Edward E. Jaffe I have seen 64-bit virtual war stories posted here, I must have missed those war stories. Could you please post links to the archives? I would like to read them. Selective memory Ed ???. Have a look for the problems Brian Peterson and I both had with RSM - in my case (at least) a SADUMP dump a week. Doesn't qualify as a war story ???. All this after we had tested 64-bit extensively at 2.10, and after I made a similar query to this very list, and got similar responses. m All that is in the past fortunately - at least Michael is lucky that way. As for the TMON testing, again old news. Right idea though - when I tested as Michael is suggesting, Sam (in his old role) got very narky when I mentioned (publicly) that they couldn't add up. The displays for above the bar were out to hell. I guess Michael could look at SMF for storage usage above the bar ... Sorry - bad joke ... Shane ... -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html This email/fax message is for the sole use of the intended recipient(s) and may contain confidential and privileged information. Any unauthorized review, use, disclosure or distribution of this email/fax is prohibited. If you are not the intended recipient, please destroy all paper and electronic copies of the original message. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html