Re: OT: If you use linkedin.com...
On Thu, Aug 11, 2011 at 6:06 AM, David Boyes dbo...@sinenomine.net wrote: I don't know about you, but if they want to use my name and face, I expect them to rent it by the hour. I don't know about you, but I would not give up my sysprog job ... ;-)
Re: How many IFLs on my box?
On Tue, Aug 9, 2011 at 11:41 PM, Marcy Cortes marcy.d.cor...@wellsfargo.com wrote: Is there away from VM to tell how many IFLs are installed on a z box? Not just defined to my LPAR but on the entire box? (as that is how IBM licenses sw!) (yes, more ILMT fun - want to automate ini file build). Hi Marcy, The ESAHDR report (on ESAMAP) should give you the numbers. I think you'd look at CP's Configured as the number of CPs and In Physical Partition as the total (CP + IFL) If that's what you're looking for, I should be able to come up with an ESAMON extract as well. Rob
Re: Anyone remember this?
On Mon, Aug 8, 2011 at 3:40 PM, Jeff Gribbin jeff.grib...@gmail.com wrote: ... or a very garbled version of the story that the original IEFBR14 didn't clear R15 and so, 'completed' with a non-zero return-code ... I'd say the reference to VM is misplaced as well, since we would not do IEF macros unless forced to... but what do I know, I'm just a newbie...
Re: Question about Linux shutdown
On Fri, Aug 5, 2011 at 1:18 PM, Michael MacIsaac mike...@us.ibm.com wrote: I would normally set the default to 600 seconds We just moved from 5 minutes (SET SIGNAL SHUTDOWN 300) to 10 minutes (SET SIGNAL SHUTDOWN 600) as some of our systems were not shutting down within 5 minutes. Systems with high memory overcommit benefit from lowering LDUBUF. This makes the penguins leave in a more orderly fashion when they remain seated on the eligible list until there are sufficient resources to complete their exit. At one installation we actually did the evacuations by groups of servers rather than rush them all to the exit at the same time. Based on the monitor data you could even come up with the proper pacing. Another interesting approach is be to trigger drop_caches as part of the shutdown process and then inflate and deflate the CMM balloon according to the resources that it frees. I take donations from those with copious spare time ;-) Rob van der Heij Velocity Software http://www.velocitysoftware.com/
Re: Time off running z/VM 5.4 1101 on z196 for first time
On Tue, Aug 2, 2011 at 8:54 PM, David Boyes dbo...@sinenomine.net wrote: How come Z hardware doesn't come with one of these? 8-) Because the machines are mostly installed in places where you don't have sufficiently open sky to receive GPS... Similar concerns prevented usage of the radio beacons that feed consumer grade automatic clocks. And you probably experienced once in a while that your navigation system was off by far... you would need to harden the signal seriously. It's amazing how complicated it gets when you want to make it reliable enough to hook up to the mainframe. PS Found an imitation railroad clock using the radio beacon that deliberately is slow so you can see it adjust at the full minute :-)
Re: Time off running z/VM 5.4 1101 on z196 for first time
On Mon, Aug 1, 2011 at 8:25 PM, Alan Altmark alan_altm...@us.ibm.com wrote: o The SE BOC will be steered the match the CEC TOD instead of making large jumps, avoiding a Paradox that could destroy the universe. I would not be suprised to find the steering to be more conceptually speaking. The common approach with consumer grade battery clocks is that the regular synch compares the two and records the apparent drift of the battery clock. Whenever you need to use the battery clock (at POR) you have the time of last synch, the current time according to the battery clock and the recorded drift. Those 3 let you make a pretty good guess about the current true time. Time will tell ;-)
Re: Modifying the System Clock on a Running VM System
On Sat, Jul 30, 2011 at 4:48 AM, Stephen Powell zlinux...@wowway.com wrote: Theoretically, it should be possible to do this with an assembler program which uses the SCK (SET CLOCK) instruction, in conjunction with the CP EXIT facility. However, I have no idea what effect this will have on CP. The design of CP assumes that the TOD clock does not change between IPL and SHUTDOWN; and if it does change, strange things may happen. Those who know the internals of CP may choose to elaborate on what things might happen as a result. A virtual machine that issues the SCK instruction will cause CP to update the TOD offset in the VMDBK for that virtual machine, provided it is allowed to do so with the VTOD directory option. The offset is applied by SIE on each STCK instruction that the virtual machine issues afterwards. As far as I know, if you were CP and issued a SCK instruction (so not under SIE) it will be picked up by PR/SM and kept as the offset for this LPAR to be applied on each STCK that CP issues in that LPAR. This is exactly what happens when the operator decides to use his $5 Mickey Mouse watch during the z/VM IPL process to adjust the clock of a $5M machine... The real hardware TOD should run UTC. That clock increments only. The S/390 hardware clocks I've seen all run a little bit slower than the true time, ETR is used to speed up the TOD to catch up with true time. Rob
Re: Modifying the System Clock on a Running VM System
On Fri, Jul 29, 2011 at 2:44 PM, Alan Altmark alan_altm...@us.ibm.com wrote: The SET VTOD command is available if you want to influence the TOD seen by a guest. (It has no effect on CMS since CMS uses CP time.) Uh, I'd say it has not desirable effect. CMS uses CP for the seconds, and TOD for the smaller bits... | Rob
Re: Ficon CTC's between LPAR's in same Box
On Wed, Jul 20, 2011 at 8:29 AM, Crispin Hugo crispin.h...@macro4.com wrote: Am I right to assume, if I am clever(?) with the configuration, I could connect all 15 LPARS to each other with CTC’s ? If you also plan to run ISLINK connections between them, you need to be careful. We had lots of problems in the past when multiple paths exist between a pair of systems. You can't avoid through connections to build, so while you're bringing up system various indirect connections already are active. When that intermediate system goes AWOL, you have lost your conversation anyway. The approach we used was to have several sets, each with a few systems. Each set has one linking pin through which the traffic between the sets goes. You pick your linking pin based on service levels etc. You could have spare links just to start when one system has a longer outage. And only join systems that are in the same administrative domain. That is, user management by a single organisation to enforce one of the following models - a single user population such that userid X on each system is owned by the same warm body - entirely disjunct user groups, userids defined on only one system - very small group of consenting adults and no desire to protect resources (works best for N = 1) Because of the intimate nature of cross-system IUCV, other options take a very good understanding of the applications, a steady hand and a large badge. And you might still run into technical limitations. Rob
Re: TPF and PAX numbers
On Wed, Jul 13, 2011 at 9:34 AM, Gregg reed.gr...@gmail.com wrote: entries which would generate 'cute' 6 character responses.. I recall someone telling at SHARE that the list of forbidden 6-char PNRs was extended on a regular basis, and that these changes made good conversation at lunch to determine who would find them offensive for what reason... ;-) In NL there were similar concerns when our car license plate scheme went to 00-XXX-0 (for the curious foreigners, some English 4-letter words take just 3 in Dutch). For starters, we dropped the vowels. Then dropped the offensive words that look like someone just dropped the vowels out of it. And dropped the abbreviations for offensive sayings, and company names, etc... No surprise we run out of this name space pretty quick. http://en.wikipedia.org/wiki/Vehicle_registration_plates_of_the_Netherlands Rob
Re: Two simple TCPIP / FTPSERVE questions.
On Wed, Jul 6, 2011 at 4:35 PM, David Boyes dbo...@sinenomine.net wrote: That approach tests not only whether the server is logged in, but whether it’s actually functioning. Works well for lots of things, and is low-cost (no cost if you run Debian or Fedora for Z). Depending on how the FTP server fails, you might also see it in your performance monitor... With anything that monitors a service by probing, the issue is only noticed after some time. Once you increase the polling to detect it quick enough that you still have time to fix it within SLA, the process uses a lot of resources (in a former life we tested e-mail delivery like that until 50% of our traffic was probes and 90% of the outages were caused by them...) Using SCIF on the FTPSERVE userid, PROP can watch the console and notice the outage immediately. CMS gives you easy tools to arrange the restart of the server. Running a Linux virtual machine with agents just to poll the FTP server is only no cost when you have too much resources and spare time. | Rob
Re: Extending DASD format?
On Mon, Jun 27, 2011 at 11:35 PM, Leland Lucius lluc...@homerow.net wrote: A few years ago, I modified dasdfmt to allow specification of the start and end track for formatting. I was using LDL formatted volumes do I didn't have to worry about the VTOC. A few more years ago, dasdfmt allowed the user to format only a range of cylinders. I believe that support was removed because they got too many support calls from customers who incorrectly only formatted part of the volume. ;-) IIRC that broke our process using flashcopy to format a new mini disk using another (very large) empty disk and run dasdfmt to initialize the part that depends on the size of the disk... | Rob
Re: Last Workshop Re: VM Workshop - Session Grid Now Available
On Sat, Jun 25, 2011 at 11:03 PM, Mike Walter mike.wal...@aonhewitt.com wrote: I checked the reviewa of the Varsity Inn South. Reports of Bed Bugs there removed it from my list. Debugging and Dump Reading Class ?
Re: Moving on
On Wed, Jun 22, 2011 at 6:51 PM, Schuh, Richard rsc...@visa.com wrote: After 48 years in the industry, involved with VM for the last 38 of them, I will be retiring early next month. I don't think it is possible to find a better group of people than the VM List. The professionalism, the willingness, even eagerness, to help others is outstanding. You have made my job easier. I wish you all the best. It has been nice, sometimes even fun, to know and work with such an exemplary group of people. Thanks for hanging out here all those years. I'm going to miss a familiar name on the list... enjoy your retirement. Rob
Re: IPLing z/VM 2nd level with many lines and columns
On Wed, Jun 15, 2011 at 7:30 AM, Michael MacIsaac mike...@us.ibm.com wrote: We ran into a problem IPLing a z/VM 2nd level using a 3270 emulator (PComm) with a setting of 62x160 (LINESxCOLS). Setting it back to 43x80 worked around the problem. Is this a known issue? Thanks. You're being vague. What is on the business end of your connection? Connecting to first level TCPIP and then dial into the guest works for me with large 3270 screens. If you connect to the TCPIP inside the guest, then maybe you forgot to raise the databufferpoolsize there? That would cause problems when the screen fills. Rob
Re: IPLing z/VM 2nd level with many lines and columns
On Wed, Jun 15, 2011 at 10:33 AM, Michael MacIsaac mike...@us.ibm.com wrote: Rob, You're being vague. What is on the business end of your connection? A SAPL screen. So we do a: == IPL 200 CLEAR And z/VM starts to IPL. Press F10. With an emulator that has the large screen size, the session gets disconnected. Buffer size of the 1st level TCPIP maybe? Have you been able to do very large writes already there? I believe XEDIT of a wide file with real content would show. I just tried it myself with Tom Brennan's Vista TN3270. SAPL forces a 24x80 formatted screen, and after PF10 the system console then is an unformatted screen in default size. When OPERATOR is there, my XEDIT session sees the 60x150 logical screen as well. Rob
Re: PROBLEMS WITH DEDICATE DASD IN z/VM 6.1
CCW translation is to get real address for the guest real address. The extent check is verified by the CU. Adding the offset of the mdisk is trivial (and happens even for dedicated, iirc). A significant diff may be MDC when the hit ratio is too small to be effective. You can disable that and still have benefit of disk management. On Jun 3, 2011 5:28 PM, gclo...@br.ibm.com wrote:
Re: zvm directions
On Thu, May 26, 2011 at 5:06 AM, Philip Tully tull...@optonline.net wrote: With all do respect: Contacting our IBM rep under NDA does not fit public road map I think the customers are letting IBM know, that they are not ready to relinquish control of this asset. It may not be the story IBM mgmt wants to hear but it is the one that is being told. I may no longer go onsite to customers on a regular basis, but when I was, I often needed access to the HMC and it was pretty consistent that there was significant access control for the HMC. Neither may be parts of IBM. At least two installations told me that IBM requires that the original HMC user/pw combinations remain in place for the (different) IBM support person to be able to support them. I suppose that when the customer was more persuasive they could convince their support person of something else. Some Large shops have a separate LAN for delicate stuff and implement access control with RSA gear. That includes a process to expire access when people change roles, etc. This is where you find their HMC as well the local consoles for the LPARs. You can't seriously tell them to move some of that back into the public LAN and do local password management again. Rob
Re: Sort IP addresses
Have a look at the builtin ip2socka stage. After that you can sort. After that you convert back with the reverse or strip the added key off. On May 24, 2011 5:29 PM, Dave Jones d...@vsoft-software.com wrote:
Re: password_on_cmds feature statement in CONFIG
On Fri, Apr 22, 2011 at 9:24 PM, Scott Rohling scott.rohl...@gmail.com wrote: Hi Terry - If RACF is installed and controlling links - then the password on the link statement is likely being ignored and is there from pre-RACF days.. The password is ignored now because access control is done by RACF. But my recollection is that when he changes CP to not allow password_on_cmds the LINK statement with a password would be rejected despite the fact that RACF does not use the password. If so, then the change in the configuration file might break things that work now... I believe this was what justified a local mod for one of our systems. We could not go through the code to check for statements with inline password, but did not want to allow people to type their password on logon in plain text either. Rob
Re: password_on_cmds feature statement in CONFIG
On Fri, Apr 22, 2011 at 10:24 PM, Scott Rohling scott.rohl...@gmail.com wrote: Terry - one more thing .. to make sure that ALL of your Linux guests are authorized to link (not just the one you tested on) you might want to check the RACF definitions: RAC RLIST VMMDISK LNXADMIN.191 AUTH I don't think *that* should be his concern since RACF would have denied access before despite the useless password supplied. This is the kind of analysis that you have to do when you introduce RACF to a system that ran without RACF before... This should give you a list of groups/users that are authorized.. You're okay if: - The Universal Access (UACC) for LNXADMIN 191 is READ or - The authorization list contains all the Linux guests with READ access or - The authorization list contains a group with READ access to which all Linux guests belong (and LIST OF GROUPS ACCESS CHECKING is active in RACF) Scott Rohling Let me take the opportunity to emphasize that you really should avoid putting users in the access list, only groups. One of the reasons is to avoid the extra work of checking access for all users when dealing with shared resources and the risk of missing some when you add new users to the system. The other part is that you can't remove a user if he's still in one of the access lists (and you can find which one). It also greatly simplifies the work of the auditors (as well as the sysprogs answering their questions). PS Thanks for confirming the 118E msg... | Rob
Re: password_on_cmds feature statement in CONFIG
On Fri, Apr 22, 2011 at 11:00 PM, Martin, Terry R. (CMS/CTR) (CTR) terry.mar...@cms.hhs.gov wrote: So if I remove the password from the LINK statements and set LINK to ‘NO’ than I will need to make sure that access is defined in RACF is this correct? If things are working now, then access is already granted by RACF. The password on LINK does nothing. When you set the option to no in the configuration file, the password on LINK will break it. And when you go make these changes, consider putting the LINK statements in the CP directory instead (in an include profile even, if applicable). I think there was a PTF to make sure that as long as you try to link to the same disk that was already in the directory entry (and granted by RACF) you would get away with the inline password (but maybe Scott and check that - must be 15 years ago that I did such things) Also if I run across any AUTOLOGs in EXECS how are these handled if I set the AUTOLOG to ‘NO’? If the AUTOLOG now works, then RACF authorisation is already ok. And like with LINK, the inline password will break it when you disallow that. Rob
Re: password_on_cmds feature statement in CONFIG
On Fri, Apr 22, 2011 at 11:28 PM, Scott Rohling scott.rohl...@gmail.com wrote: Nope -- the inline password still fails with HCP118E even if there is a LINK in my directory for the disk. I believe the PASSWORD_ON_CMDS LINK NO has no exceptions .. once you set it NO - you can't use an inline password - you must be prompted for it (which won't happen if you have RACF controlling LINK). Too bad. The scenario I recall was the password prompt on a non-RACF system. I believe CP was changed (back) to avoid prompting the user when the disk was already linked. I don't see a problem if CP would allow the inline password even when disallowed... Rob
Re: 391 391 RR?RE: password_on_cmds feature statement in CONFIG
On Fri, Apr 22, 2011 at 11:24 PM, Schuh, Richard rsc...@visa.com wrote: Might I ask a dumb question? If it is in the PROFILE EXEC, why not put it in the directory entry? If it is there and is detached, it can be reacquired by the command, CP LINK * 391 391 RR without having to enter a password. And without ever having to know that it is a disk defined in the MAINT entry. You're correct. I tried to avoid changes to the PROFILE EXEC because I fear it will be on each of the servers and they may have the 191 R/W (but we already discussed options to make the guest detach that disk). One reason you find this kind of things in programs is when the userid or mini disk address is determined by some programmed logic. Mini disk passwords are a pain in that situation and don't really provide any security because the link is issued in the user virtual machine. One of my peers was very proud of his program that scrambled the password in the code (so you could not see it when you browse the module). He was pretty shocked when I did a trace on the DIAG 08 to see the password... :-) Rob
Re: Tapeless segment tranfser between LPARs?
On Thu, Apr 21, 2011 at 4:53 PM, Daniel Bewley daniel.bew...@gmail.com wrote: Is there another freely available utility out there for moving segments from spool to minidisk and back again? Or is there some combination of CP commands to retrieve and load non-CP/CMS data into a segment? There's DCSSBKUP and DCSSRSAV on the MAINT 193 disk... Rob
Re: Detaching A disk from z/Linux guest dynamically
On Wed, Apr 20, 2011 at 10:13 PM, Martin, Terry R. (CMS/CTR) (CTR) terry.mar...@cms.hhs.gov wrote: HCPFOR070E – Basically it is an authorization issue with the FOR command. We are running RACF here so do you know what the profile would be and what RACF class would need to be activated to define and permit this resource? Can I also do this via the SECUSER command? The SECUSER can issue these #CP SEND CP LINUX007 DET 191 #CP SEND CP LINUX007 LINK * 191 191 MR But in general I consider it abuse of power to use a privileged userid (with FOR authorisation or as SECUSER) when the user could have done it himself. You could also do it from Linux (use modprobe vmcp if that does not happen at boot time already) vmcp det 191 vmcp link * 191 191 mr Rob
Re: Detaching A disk from z/Linux guest dynamically
On Wed, Apr 20, 2011 at 11:22 PM, Tom Huegel tehue...@gmail.com wrote: If that is the rule, then we don't need SECUSER or FOR, or when would you use them?? Just courious. Don't know whether it is *the* rule, but it's mine. When you do your job with the least amount of special privileges, there's less to justify. In many cases you need to audit the use of these special privileges, so it helps to avoid the need for special privileges. And there's the risk of mistakes. Even though you can, you shouldn't use MAINT for all kind of trivial work that could be done also on a normal userid. There's a trade-off between convenience and security, and only you can do that yourself. YMMV depending on how long ago you shut down the production system or forced off the wrong virtual machine. The purpose of SECUSER is to perform programmed interaction with a virtual machine that runs a program that does not offer that amount of control. When you SET SECUSER to steal control in such a situation, you risk integrity because not all console output is trapped anymore. The FOR command avoids stealing secuser, but does risk the guest being in console function wait when the legitimate controller wants to issue a command. And you might confusing things by producing unexpected console output. One of my other rules is not to hide evidence of what you've done, and leave a clear trail in case you get lost and people come looking for you. To see in the operator logging who logged on to the virtual machine may save you hours when searching for clues. It often outweighs the few seconds you gain by sneaking a command under the covers with FOR. Rob
Re: Detaching A disk from z/Linux guest dynamically
On Wed, Apr 20, 2011 at 11:54 PM, Scott Rohling scott.rohl...@gmail.com wrote: There are certainly valid uses for SECUSER/SEND/FOR commands -- I don't agree that their use is an abuse of power. But of course they could be depending on what you do with them. I suppose you could consider a SHUTDOWN command an abuse of power as well... ;-) Your response shows that I failed to make my point. What I mean with abuse of power is doing the job with tools that are sharper or heavier than needed. I have no reason to walk around all day with a userid capable to issue a SHUTDOWN. Most serious installations that I know have rearranged their privilege classes such that 1) don't use a CLASS A userid unless you have things that require it 2) rearrange popular commands (like LOCK and UNLOCK) to avoid 1) as long as you can 3) move the SHUTDOWN out of CLASS A to avoid mistakes by those that need CLASS A Rob
Re: Something wrong with my USERID
On Wed, Apr 6, 2011 at 8:52 PM, Martin, Terry R. (CMS/CTR) (CTR) terry.mar...@cms.hhs.gov wrote: Yes that was it. What normally happens to get this In this mode? Happens when the System Operator userid was logged off and next person to log on with a CLASS A wins the prize. You should start OPERATOR again and run whatever you use there (eg PROP). The SET SYSOPER lets you give it back to the proper userid. Rob
Re: PIPEDDR and attached DASD
On Tue, Apr 5, 2011 at 10:07 PM, Brian Nielsen bniel...@sco.idaho.gov wrote: PIPTCQ1015E ERRNO 54: ECONNRESET. PIPMSG004I ... Issued from stage 3 of pipeline 3 name iprestore. PIPMSG001I ... Running tcpdata. PIPUPK072E Last record not complete. PIPMSG003I ... Issued from stage 2 of pipeline 1. PIPMSG001I ... Running unpack. Data restore failed. Ready(01015); T=0.01/0.02 13:39:17 Sounds like PIPEDDR is not properly handling the termination of the TCP/IP connection (like the sender going AWOL while the last piece of data is still in transit). If the pipe leaks, subtle timing changes may get your feet wet. I never looked at what PIPEDDR does for flow control, but I do recall that I had to master similar things when I did mine... | Rob
Re: OT: The weather in Endicott
On Thu, Mar 24, 2011 at 3:29 PM, Shimon Lebowitz shim...@iname.com wrote: On Thu, Mar 17, 2011 at 6:26 PM, Pamela Christina in sunny warm Endicott NY chris...@gdlvm7.vnet.ibm.com wrote: On Wed, Mar 23, 2011 at 9:40 PM, Pamela Christina in springtime snowy Endicott chris...@gdlvm7.vnet.ibm.com wrote: Don't you love seeing Pam's cheery weather reports? I really do! :-) Tee hee. They sure have their ups and downs. Global warming seems to be mostly marketing. http://www.wunderground.com/history/airportfrompws/KBGM/2011/3/24/MonthlyHistory.html
Re: Simple RACF question re: LOGONBY NOPASSWORD
On Tue, Mar 22, 2011 at 1:44 PM, Colin Allinson cgallin...@amadeus.com wrote: Seems I am getting old and had forgotten that I had asked the self same question before. You're probably going to ask that every 60 days until it's resolved :-)
Re: Social Security Confronts IT Obsolescence
On Wed, Mar 16, 2011 at 4:07 AM, David Boyes dbo...@sinenomine.net wrote: Its interesting that the z10s that actually hold the SSA data are the least demanding on that front of any of the systems in that facility. .. even when data center planners charge double for it because of the dual power feeds... ;-)
Re: Old codger question: Can SFS be networked across systems?
On Wed, Mar 16, 2011 at 5:25 PM, Les Koehler vmr...@tampabay.rr.com wrote: A friend and I were (dis)cussing SFS and he thinks it can be networked cross-country. Possible? I would guess that it would be an authentication nightmare at the user level. Thoughts? If you share via ISFC that means both sides must be in the same administrative domain. But given the length of the FICON connection, it will only work for small countries ;-) IPGATE run over TCP/IP connections and does long distance. Latency can make it less fun for regular use, but IMHO you really should not have critical data on a remote system. You define user access and mapping in the IPGATE configuration files. Provided there is some understanding between userid management on both sides, you could even use dummy userids as the target of the mapping (eg map MAINT at SYSA to user MAINT-A at SYSB). Rob
Re: Tape drives : MVS VM
On Tue, Mar 15, 2011 at 8:21 AM, Shimon Lebowitz shim...@iname.com wrote: We actually wanted the opposite - tapes shared between machines, all attaches under control of one central program, and since detaching an assigned drive forces a RUN, we wanted no ASSIGN at all. I implemented this via CP Exit FFB, which altered the ATTACH to include NOASSIGN on the command, unless ASSIGN was explicitly specified. These days, the GIVE command might be your friend if you write a tape manager. I would assume the unit remains assigned during the operation, since you don't want another system to take it away after you validated the right tape is mounted. We used to have problems with VSE machines doing that kind of tricks. But tape management on z/VM remains a challenge since a virtual machine can unload the tape and get another volume mounted outside control of the tape manager. I wanted to make CP trap that and involve the tape manager, but I might very well have found additional holes. Rob
Re: zVM User Definitions
On Wed, Mar 9, 2011 at 7:30 PM, Kris Buelens kris.buel...@gmail.com wrote: At my former customer, we created several RACF groups. To name a few: LBSYST to control LOGONBY to various users by system programmers LBOPER for the operators' group SYSALL to permit the system programmers to link to most MDISKs Right. Those with exposure to RACF in a real life have learned that you grant access to groups rather than users. Somehow our requirements are not as unique as we may think, and using groups cuts down the administrative effort. A good reason is that we don't have an easy way to list the profiles where the user is on the access list. You do need to enable the GRPLIST option (which isn't by default, iirc) You should also look into RACFVARS to combine related service virtual machines and use a single LOGONBY profile for them: RDEF RACFVARS LNX ADDMEM(LINUX01, LINUX02, LINUX03) RDEF SURR LOGONBY.LNX PE LOGONBY.LNX CL(SURR) ID(ADMINS SYSPROGS) ACCESS(READ) Now when you define a new Linux guest, you only have to add it to the LNX profile. Rob
Re: cpuplugd Daemon
On Fri, Mar 11, 2011 at 1:50 PM, Michael MacIsaac mike...@us.ibm.com wrote: Hello list, We are activating cpuplugd process for dynamic CPU and memory management for Linux guests running in z/VM. We have found a reference in Virtualization cook book for SLES11 SP1 how to make necessary configuration for CMM modules within Linux.But couldn't find any reference about the configuration to be done within z/VM for CMM. Is any configuration required or z/VM comes with CMM enabled by default. I've been waiting for some performance person to chime in on what I perceive as a disparity, but it has not happened. So I will ask the question - Aren't we talking about two different technologies here - cpuplugd and CMM/VMRM? Yes, the new Virtualization Cookbooks describe cpuplugd in the section Utilizing the cpuplugd service. This was based on presentations from Hans-Joachim Picht, et al, but it does not involve CMM and VMRM. In previous versions of the book there were sections on those, but it was agreed that they should be removed - probably the references to loading the cmm module at Linux boot time should also be removed (perhaps this was the source of the confusion). Yes, two different things. I've been telling people that you need to combine VM and Linux metrics to get it right. VMRM tries with just the VM data, cpuplugd tries with just the Linux data. IMHO the results confirm my claim... So I believe the answer to the original question is - you don't need to configure CMM and VMRM for the cpuplugd service to work. That section should pretty much stand on it's own. You also need to distinguish the two different tuning parts (memory and CPU) which leaves these to consider: * VMRM for memory: uses just the VM memory metrics to tell CMM in Linux to use less memory when there is less available. While recent enhancements support setting a minimum to avoid killing Linux, it remains a close your eyes and cross your fingers approach. Even the believers suggest you don't use it for serious business workload. * cpuplugd for memory: apart from the examples being wrong (and maybe never tried) it is just too simplistic. The controls are either too aggressive or too soft. Typically cpuplugd is too slow in giving back resources. * cpuplugd for cpu: it's a bit like a solution looking for a problem, and to some extent it creates its own problem (it confuses the z/VM scheduler about the priority of the workload). The cases where it would make sense would be addressed better by proper configuration, and in several other scenarios it makes things worse. It would be more relevant to Linux in LPAR. And please, could we have a bit less FUD without really digging into the question? Thanks. I'm not sure what your request is... the devil is in the detail, as always. If it didn't even sound like a great idea to the casual observer, then we would not even have this discussion. I'm perfectly happy to explain, but can't do without going into the details. Rob -- Rob van der Heij Velocity Software http://www.velocitysoftware.com/
Re: cpuplugd Daemon
On Fri, Mar 11, 2011 at 3:31 PM, Michael MacIsaac mike...@us.ibm.com wrote: Memory sizes can also be set by the cpuplugd service. However, unlike processors, there is no good generic default value. The following example is in the Device Drivers book: MEMPLUG = swaprate freemem+10 freemem+10 apcr MEMUNPLUG = swaprate freemem + 1 However, this is just a starting point. You should test any setting that you want to implement against a representative workload that your Linux systems will be running. Details are beyond the scope of this section. So how is it just plain wrong if we say the setting is dependent on your system's workload? Thank you. I rarely get accused of spreading FUD. Maybe my attempt to be friendly and not insult you in public did not come through very well ;-) I was obviously also confused by your request to comment but not go into detail... Your examples seem the same as in the Device Driver book, so my comments apply to yours too. The MEMUNPLUG says as much as when swapping more than 2,500 pg/s, remove some memory - this qualifies as plain wrong in my perception. When the system is thrashing, you don't want to reduce available memory. To someone like me with physics background, the formula can't be right anyway because it adds metrics with different units. Such a formula tends to work only for very small range of values where one of the factors is constant or can be ignored. The example would qualify as FUD since it suggests a lot of smart thinking behind it that just isn't there. Adding that it depends on your workload makes the complex formula even more silly. I discussed this in my presentation on Memory Management as well - http://www.rvdheij.nl/Presentations/zLX45.pdf (pg 34-36) It seems the developer meant something like this: MEMPLUG = swaprate 0 MEMUNPLUG = swaprate 10 This is like when we're swapping, add some memory. when we're not really swapping, take some memory out and probably easier to understand. The 0 vs 10 is to introduce some hysteresis, but it is hard to avoid that this thing will be tinkering with CMM all the time (and thus cause overhead). When you make the increment small, it takes way too long to inflate and deflate the balloon. But when you make it larger, it causes problems by taking large chunks of memory. Rob -- Rob van der Heij Velocity Software http://www.velocitysoftware.com/
Re: EDEVICE Overhead
On Fri, Mar 4, 2011 at 1:10 PM, John Hanley jhan...@courts.state.va.us wrote: I'm looking to use the new XIV box to for DB2 space in a DB2/WebSphere environment. The I/O and channel utilization currently is not a problem, but CPU can be. I thought someone told me XiV would be doing ECKD as well, but I don't see that at all. Maybe the representative in the booth was just tired or defaulted to yes like a good sales person :-) Given the XiV, your option would be than native FC or EDEV, right? WebSphere should not be an issue, since it normally does not really do I/O. DB2 might be sensitive to I/O, depending on the type of workload. You should collect performance data and review that. If you want a very rough rules of thumb: I ran a pure I/O workload (with dd) and EDEV had a T/V ratio 1.5 and ECKD was around 1.1 This sounds bad, but it does not tell you anything unless you know how much CPU your application would spend on I/O. Assume your DB2 server is busy 20% of the time, and during that time is 20% in I/O wait. Add a bit for disk writes (your performance data shows you how much you really write). So it would be doing I/O 5% of the time. Some of that is CPU and some is waiting for the disk. Imagine that is 1:5, this would mean 1% of a CPU is spent to issue the I/O. If CPU is your concern already, it means 90% of CPU is used for non-I/O related things. So even when CP adds 0.5% on top, it does not really make a lot of difference. The difference obviously does show in synthetic lab experiments, but how many of us run those as their normal workload. It's a poorly kept secret that I can often be talked into reviewing performance data... Rob -- Rob van der Heij Velocity Software http://www.velocitysoftware.com/
Re: Query Proc
On Fri, Mar 4, 2011 at 10:04 PM, Mike Walter mike.wal...@aonhewitt.com wrote: A quick Google search didn't turn up anything obvious. Our z/OS folks never heard of Reserved processors in z/OS (they see processors by the command: D M=CPU). Perhaps them are using the term Reserved for processors Dedicated to an LPAR? It says: A logical CPU is in the reserved state when it is in the level-2 configuration, is not available to be used to execute programs, and cannot be made available by issuing instructions to place it in the configured state as opposed to a standby processor that *can* be made available, and configured processors that *are* available to execute programs. The LPAR configuration has fields for initial and reserved number of CPUs. My understanding is that initial is how many logical CPUs come online when you activate the LPAR, and that you can add logical CPUs upto the reserved without an outage of the LPAR. | Rob
Re: VM Total time in $ACCOUNT files
On Wed, Feb 23, 2011 at 12:34 PM, Gregg reed.gr...@gmail.com wrote: I wish I had SAS and I see now where I had an error in my rexx program re: seconds/ms calc. much closer. my bad sorry. What's BC mode? z900 or earlier basic mode or 1 LPar/dedicated lCPs? I believe someone is confused here due to early re-use of acronyms ;-) Long ago, PR/SM was optional, and hardware could be configured to run in basic mode versus LPAR mode but with current hardware (since z9) it's not optional anymore (unrelated to the z9-BC vs z9-EC difference). This means that on supported hardware, z/VM (or any OS) runs in LPAR, even when that is the only LPAR on the machine. More recently, IBM introduced IFL-only models that have no CPs installed and thus don't run The Operating System but only z/VM and Linux. Depending on one's context and background, this without a z/OS LPAR may be observed as without an LPAR - which obviously is not the correct interpretation. PR/SM is the only component to know how much CPU cycles went into LPAR overhead and LPAR management overhead. The z/VM monitor (like RMF) get that information from the hypervisor and pass it along to your favority performance monitor. Any LPAR on that machine would get the same global utilization info from PR/SM (even though MVS does some hide and seek to keep the IFLs out of the total license charges). The z/VM account records do not account for that overhead (no pun intended). The overhead is normally small enough that you could not even tell when you drive all CPUs in a loop and see how far you're away from N*100%. The account records do provide the total for z/VM management overhead (which is different from the LPAR overhead). So if you care to know the LPAR overhead, you need to collect performance data from any one of the partitions on the machine. With ESALPS we feed the data into MXG and MICS, which works well for most installations that do capacity planning. From a license point of view it surely beats the alternative of getting a CP on the machine and run z/OS there just to collect LPAR usage statistics. Rob -- Rob van der Heij Velocity Software http://www.velocitysoftware.com/
Re: Xedit question
On Mon, Feb 21, 2011 at 9:16 PM, Martin, Terry R. (CMS/CTR) (CTR) terry.mar...@cms.hhs.gov wrote: I want to insert the following command Q DA in front of each entry. What is the best way to do this using xedit? This sound like how can a drill a hole with... type of challenge. I probably would not use XEDIT to do it but rather :0 pipe xedit | spec ,Q DA, 1 w2 nw | cp | cons (assuming you want to issue the commands, not just have a file with the commands in it...) Surprised none of the more experienced XEDIT people suggested ARBCHAR ? set arbchar on $ :0 ch /$ /Q DA / * Rob
Re: VM Total time in $ACCOUNT files
On Fri, Feb 18, 2011 at 11:50 PM, Ackerman, Derek derek.acker...@infocrossing.com wrote: I am simply summing the total CPU times for each CP, for my report I keep them separate, here is an example. I believe you're confusing virtual CPUs with logical CPUs. The CPU number in your account record is for the virtual CPU of the guest (with exception of the SYSTEM record). A virtual CPU is not tied to a logical CPU, so your UNIPROD's CPU 00 gets dispatched on logical CPU 00, 01 or 02 as z/VM likes. It is even possible for a single virtual CPU to consume more cycles than the average of any of your logical CPUs (when it hopped from one to the other). Rob -- Rob van der Heij Velocity Software http://www.velocitysoftware.com/
Re: VM Total time in $ACCOUNT files
On Sat, Feb 19, 2011 at 12:17 AM, Les Koehler vmr...@tampabay.rr.com wrote: Is a logical cpu equivalent to a physical cpu? I realize that the h/w now has multiple physical cpus, but I was brought up on System/360 and never have quite grokked how it is all controlled and accounted for! z/VM itself runs in an LPAR under control of PR/SM. And just like z/VM does for virtual machines, PR/SM will dispatch the LPAR's logical CPUs on any physical CPU as it sees fit. Even when you have a dedicated CPU, it's PR/SM who decides which physical CPU is dedicated to that logical CPU (for almost all the time). And we're not even talking about hot-sparing of CPUs You're in a maze and all CPUs look the same ;-) Rob -- Rob van der Heij Velocity Software http://www.velocitysoftware.com/
Re: z/VM Monitor Records
On Wed, Feb 16, 2011 at 3:21 PM, Billy Bingham billy.bingham...@suddenlink.net wrote: Hello all, Does anyone have a procedure that they use to collect and process z/VM monitor records that they would be willing to share? Thanks in advance, Billy If you're interested in a commercial solution, make sure to get in touch. Along with our performance monitor, you also get people to help you to make sense of the numbers (which happens to be the hardest part of it). PS If your time is so cheap that you can afford to write your own monitor, we might be able to take advantage of that ;-) Rob -- Rob van der Heij Velocity Software http://www.velocitysoftware.com/
Re: z/VM 5.4; 6.1
On Wed, Feb 9, 2011 at 6:56 PM, Alan Altmark alan_altm...@us.ibm.com wrote: I recommend that when you're done with installation, go back and reallocate those volumes so that they are entirely SPOL or PAGE. after you're done formatting those remaining cylinders, I assume... (and if you get ICKDSF wrong, good news is that you probably still remember how the install goes) | Rob
Re: SET SHARE ABSOLUTE/RELATIVE
On Mon, Feb 7, 2011 at 10:26 PM, Tom Duerbusch duerbus...@stlouiscity.com wrote: We have sufficient enough resources so set share isn't a big deal, however, I do have production set at set share relative 150 and the test systems as set share rel 50. When they are competing, the production machines end up with 75% of the processor and the test machines with about 25%. That seems to be cool with me. Having plenty of resources will let you get away with a lot of settings, but it is very rewarding to be able to improve performance even in that situation. Your objectives of 3:1 are more realistic than what I sometimes hear. When management tells you to have a 100:1 ratio, you should not be surprised to get only 1% of your CPU for test... Also, your situation on CPs is a bit easier since the scheduler still works for those. It's the IFLs that have the problem... You don't want certain machines to be cpu starved. We need CICS users, both test and production to be serviced and all IP stuff to be serviced. That includes DB2 when accessed via DRDA. So, in my minds eye (since we don't have any good performance monitors), I use the relative share to pretty much allow the test systems about 12% of the processor. This should take care of any CICS and IP traffic. A long time ago, I tried dropping the test machines in the basement, priority wise. CICS started abending AICA, IP would fail on timeouts and DB2 DRDA would faile on timeouts. Many times it was quicker to reipl the test system then to determine what failed and restart the task. Since then, test systems always get some cpu. If your workload is pretty constant, you can compute relative into absolute and the other way around. The challenge is to come up with settings that handle the case where workload does not behave as usual. The value of relative share goes down when there is more competition. This is what you normally want for the regular workload. Absolute share guarantees a baseline service, useful for critical services like RACFM, TCPIP and the performance monitor. PS Ending up with a total absolute share of more than 100% is a configuration error and results are not pretty. It sometimes happens when the MVS people are made to do VM configuration without reading the book. -- Rob van der Heij Velocity Software http://www.velocitysoftware.com/
Re: Assembler Question
If you change these MVC MAQUINA+0(1),=XL1'E5' by MVI MAQUINA+0(1),X'E5' then things don't look obvious anymore. Alternatively, pass the userid as a parameter and let someone else worry :-) | Rob
Copyfile with PACK on MVS ?
Friends, When a RECFM V file is transferred through download/upload with a PC, we need to protect the record layout. And when it's non-text you can't stick CRLF between the lines. On VM we normally have people use the PACK option of COPYFILE, which puts enough info in the file to restore the structure. What do people do on MVS? I thought a simple IEBGENER to convert to VB or whatever would do, but that does not seem to be common practice... does TERSE maybe put stuff in to retain the record structure? And does our DETERSE recover that? Thanks, Rob
Re: General CMS minidisks and SFS on PAV DASD?
On Mon, Dec 20, 2010 at 5:59 PM, George Henke/NYLIC george_he...@newyorklife.com wrote: The IO Supervisor has not kept up with the hardware. It still thinks of a disk device as a spinning platter when in fact it is a rank of RAID devices striped over numerous HDs and cached in a disk controller from where it is actually being read thereby permitting multiple IOs to the same device number.. The architecture guarantees that the I/O's to the device are serialized, that is the 2nd queued I/O only starts when the first one completes. This architecture is exploited by OS and applications to ensure that data on disk is in a consistent state. Ever heard of shops where 5000 PROFS users had to go through fsck on their CMS disk after a power failure? ;-) Sometimes that guarantee is not needed. Often when two completely unrelated I/O's go for different data that happens to reside on the the same real volume, you couldn't care less. Or when the OS does not provide such a guarantee anyway (aka lazy write) you can't exploit the hardware guarantee. When the channel program guarantees that they are really unrelated (so not writing or reading the same data) we can leave it to the I/O subsystem to change the order if it makes sense. Whether it makes sense is not easy to tell. It make a lot of sense for your hardware vendor. It makes sense when you do single-threaded lab benchmarks that need to saturate the I/O subsystem. There's a lot of cases where it does not make sense (like when SFS does its own smart things to spread I/O). If someone has relevant data, I'm always interested to see whether it makes sense... Rob
Re: Copyfile with PACK on MVS ?
On Tue, Dec 21, 2010 at 12:20 PM, Kris Buelens kris.buel...@gmail.com wrote: On VM at least TERSE creates a F 1024 file, wouldn't it do that on MVS too? Well, you can also talk IND$FILE into making F 1024 file, but that does not help... ;-) But since I think the service process uses tersed VMFPLC images, I got more optimistic that the DETERSE should be able to rebuild the record layout. I thought we only got DETERSE with z/VM, but that may have changed. | Rob
Re: Copyfile with PACK on MVS ?
Thanks for all the on-line and off-line enhancements of my rusty MVS skills. I should have some options now. Two platforms separated by a common code page ;-)
Re: Vswitch Grant as a CMD in User's Directory?
On Wed, Dec 8, 2010 at 7:38 PM, Alan Altmark alan_altm...@us.ibm.com wrote: I've been saying for several years, You need an ESM. More and more z/VM security management will be focused on ESMs, not native CP. If your fave ESM doesn't simplify things for you, gripe to the vendor. That's self-fulfilling prophecy, Sir. You also created the mind boggling approach where the VM Sysprog needs to change hats and perform both steps of the ritual. But I stopped years ago saying that one word of the VM sysprog should be enough for things he controls. So when it already requires magical powers to get a NICDEF statement into the directory, there is no problem in having that imply the GRANT as well. Different when the class G command is used to define the NIC. Yes, this is different from a LINK in the directory because we assume that the owner of the resource manages access to it. In that case it is appropriate that the owner decides whether the LINK can actually work (and can revoke access). | Rob
Re: Jumbo Frames
On Thu, Nov 18, 2010 at 7:20 AM, Alan Altmark alan_altm...@us.ibm.com wrote: Also, a guest's ability to use an OSA is limited by the guest's access to the CPU. Very much. In many scenarios that I worked on, disk I/O throughput or CPU was the limiting factor. I still recall one case where the customer did a Hipersockets benchmark showing only a few MB/s. They were using scp to transfer from a guest in one z/VM LPAR to the other. Both LPARs shared the same (single) CPU and he assured me CPU was not the bottleneck because both LPARs used only 50% of the capacity... Obviously they really measured symmetrical encryption and decryption in openssh rather than Hipersockets bandwidth. Having real data from the experiment helps a lot, though it's still not trivial to explain all factors. Rob -- Rob van der Heij Velocity Software http://www.velocitysoftware.com/
Re: Jumbo Frames
On Wed, Nov 17, 2010 at 7:41 PM, Tom Duerbusch duerbus...@stlouiscity.com wrote: If you have a single engine, I wonder if you are hitting the same class of problem that Websphere has on a single engine? I wouldn't know which class that is. Sure Java can absorb a lot of CPU, and the more power the better. But I encourage anyone to show me evidence that two half CPUs is better than one full CPU ;-) That is, you have two active tasks in competition for CPU time. i.e. two stacks, one engine. stack A sends a packet The cpu is taken from Stack A to service Stack B. Stack B gets the cpu and sends back the ack. The cpu is taken from Stack B to service Stack A. It's not that bad with QDIO because the hardware (or CP, in case of VSWITCH) can handle the buffers without dispatching the virtual machine. So it allows for some asynchronous work. If you do the math, you can see that the default amount of buffers is normally too small for large bandwith. Each time when the entire set of buffers is processed by the hardware, the virtual machine needs to be dispatched again to pull the packets out of the buffers. Depending on the number of virtual machines, that can take a while (and longer when people defined all virtual-MP guests). With virtualization, most of these things average out well. You can increase efficiency when you allow for some latency. But latency is the price you pay for sharing a resource. Just have to realize that you can't estimate the total bandwidth by measuring a single task. | Rob -- Rob van der Heij Velocity Software http://www.velocitysoftware.com/
Re: A how to ?...
On Wed, Nov 17, 2010 at 9:32 PM, Brian France b...@psu.edu wrote: Folks, Our storage folks have come across a dasd volume that apparently by it's name was one of our VM volumes. Now, none of my 3 vm's have this volume on line and I would like to just have a look see as to what was on the volume. Is there a utility to do this? We have Vmdirect but I don't believe by adding this volume into the pool a map is going to give me anything since no users have it. IF this was z/OS, I would simply use ISPF 3.4 to list the contents of the volume and this is what I was wondering about, it something like that existed in z/VM. Maybe this is something that can't be done due to the nature of it all. Thanx in advance... This is VM, anything can be done ;-) With DDR (or trackread in CMS Pipelines) you can see the virtual cylinder number. If you would scan the entire pack cylinder by cylinder, it would nicely show you where the mini disks were. If it's a CMS mini disk, the VOL1 record has the number of cylinders that CMS had, and you can normally skip the rest of that mini disk in your scan. Alternatively, you can DDR real cylinder 1 to a small T-disk and try ACCESS to have CMS tell you what was there. That information gives you the amount to increment your cursor and try the next disk. I've also seen people use DEF MDISK but I like it less because it goes R/W... | Rob
Re: A how to ?...
On Wed, Nov 17, 2010 at 9:43 PM, Rob van der Heij rvdh...@gmail.com wrote: With DDR (or trackread in CMS Pipelines) you can see the virtual cylinder number. If you would scan the entire pack cylinder by cylinder, it would nicely show you where the mini disks were. If it's a CMS mini disk, the VOL1 record has the number of cylinders that CMS had, and you can normally skip the rest of that mini disk in your scan. Forgot to mention... to scan the entire volume, you start at the last cylinder N to determine the start of that mini disk. If the two don't match, you check cylinder N-1 and try there, etc. When they do match, you've identified one mini disk and scan down from there... | Rob
Re: FTP Question.
On Thu, Nov 4, 2010 at 10:19 PM, Schuh, Richard rsc...@visa.com wrote: The entire process is done via VMFTP. It takes a long time to transfer all of the files. I read somewhere that a new data path is created for each file. This appears to be where the time is being spent. Is there any was to cut down on or eliminate this time? It would seem to me to be much more efficient if all of the data were transferred over the same path without all of the reconnecting. Could it be that the reverse lookup for the IP address of your VM system is leading to dark zones of the network? When DNS zones are not properly configured, you may end up waiting for a time-out. Either ftpd or tcpwrappers could be trying the reverse lookup. Running tcpdump on Linux for UDP data might reveal whether that's the case. If Linux could offer the data via HTTP, you might be able to retrieve multiple files via a persistent connection... | Rob
Re: Seinfeld's Contribution to the The Principles of Operation
On Fri, Oct 8, 2010 at 7:33 PM, Bill Holder hold...@us.ibm.com wrote: I believe the reason for allowing some leeway has to do with hardware design concepts for things like pipelining, reach ahead and out-of-order execution - with the hardware in essence speculatively performing certain operations ahead of time or in parallel (to maximize utilization and throughput), the potential exists for the processor thinking something is going to be referenced, and performing reference recording on the fetch from storage, only to find out later that that path wasn't really taken. To correct such apparent overindication of reference would mean tracking de-referencing and having to worry about (and serializing against) reference recording actions from other processors. You could well ask why not defer reference recording until the path decision is made?, but I would answer that doing so would break the pipeline and constrain the design of the processor / storage interface in ways that would defeat the whole purpose. I would even say it's better from a paging perspective. If the code makes a U-turn just at the end of the page and causes the CPU to decode the next instruction already, it would be annoying if the OS would take that 2nd page away because nobody needs it anyway... assuming that such a case would also cause a page fault? Just don't bring your reference bits into court as evidence... ;-) | Rob
Re: simplest little pipe
On Wed, Oct 6, 2010 at 10:50 PM, Hughes, Jim jim.hug...@doit.nh.gov wrote: Thanks Mark. I've never issued HELP PIPE let alone HELP PIPE CP. I use PIPE HELP or PIPE AHELP when assistance is required. You learn new stuff everyday. One also learns bad habits every day ;-) PIPE AHELP is what you want (and a recent PIPELINE HELPLIB for it). The help files shipped with CMS are normally less accurate and not always correct. You're right that the description of the CP stage does not state that you can't have an extra CP command as the argument when using the secondary output (the reason is probably that we'd lose the one-to-one correspondence between processing input records and producing something on the secondary output). | Rob
Re: SFS - misunderstanding
On Tue, Oct 5, 2010 at 2:10 PM, Mark Pace pacemainl...@gmail.com wrote: What am I missing that allows me to edit a file on a R/O accessed SFS? 'SET RORESPECT ON' It protects against accidentally editing a file when you access it R/O (and it is kind of strange that the default is R/O for any filespace that is not yours). | Rob
Re: z/VM ISFC links
On Mon, Oct 4, 2010 at 3:24 PM, John P. Hartmann jphartm...@gmail.com wrote: It has to be a real CTC where one end is attached to the virtual machine and the other end is activated to CP. Right. Experience shows that a lot of experts will insist that this is invalid in the IOCP configuration. It isn't, but rather shows their MVS sysplex background where you don't do this (and maybe their IOGEN / HCD flags them as suspicious). But it does not make you go blind nor does it grow hair on your hands... ;-) When you have them between systems / LPARs then it's a moot point because you invested in the two CHPIDs already. But for shops with just one z/VM image and a guest, it's a pity CP does not connect to virtual CTCs. | Rob
Re: z/VM ISFC links
On Mon, Oct 4, 2010 at 4:02 PM, Mark Pace pacemainl...@gmail.com wrote: I wonder how much I lose, or overhead gained, by having the other LPAR being a gateway for the 2nd level to the first level resources. Most obvious one is in availability. When the other LPAR has an outage, your ISFC link to 2nd level goes too. I've never measured the cost of transfer, but might be relevant with very high data volumes. I would expect the 2nd level guest being the biggest cost factor. | Rob
Re: A confused CCW question.
On Sat, Oct 2, 2010 at 5:01 PM, Tom Huegel tehue...@gmail.com wrote: Thanks Alan, Rob. I found the timestamp in the Define Extent CCW. A little FYI for anyone who might care the PIPE stage DATECONVERT is a great tool to make the TOD clock readable. Life is even easier than that with c2t in specs. I use it a lot with monitor data and also to preface output with a time stamp: .. | spec tod c2t(*) 1 1-* nw | ... | Rob
Re: Maximum Virtual Storage
On Sat, Oct 2, 2010 at 1:11 AM, Gary M. Dennis gary.den...@mantissa.com wrote: If the volume limit for a z/VM page volumes is 240+, how does this relate to maximum defined virtual storage for all active guests under a z/VM image? The total amount of virtual storage in the universe is limited by the amount segment and page tables in PTRM address spaces that we'd need to span that virtual memory. Some of that is resident, some is paged out, and some is just not there. The amount resident is limited by your real memory, paged out is limited by the amount of paging space, etc. Paging volumes is due cpowned maximum and number of spool packs, plus size of a device, etc. The way I remember the presentation is that Bit points at the various restrictions independent of each other. So you know that once this restriction is gone, the next barrier in space would be that one. Or the other way around: don't start bugging us about this limit because the other one is hitting you first. We don't want the car manufacturers to list all cars maximum speed at 55 mph because that's the local limit in your area. But looking at the glossies, it does make sense to realize whether a listed 110 mph does you any good... | Rob
Re: A confused CCW question.
On Fri, Oct 1, 2010 at 11:51 PM, Alan Altmark alan_altm...@us.ibm.com wrote: According to documents on the web, timestamps are part of the PREFIX and DEFINE EXTENT CCWs. However, you will not find any documentation of PREFIX, nor of the extensions to DEFINE EXTENT (no pun intended). Another case of RTFC, as LINUX will use them. I think you need to move your tracing out of the realm of CCWs and consider whether Linux can trace the timestamps it is using in its I/O. Browsing a few Linux source files suggests that you should be able to recognize it by the data length in your Define Extent CCW. We used to do 16 byte, from what dasd_eckd.h tells me XRC adds another 16 (and uses some of the reserved bytes). An I/O trace with CCW option should reveal it... And since someone asked that a while ago; once RDC shows the device can, Linux appears to include the time stamps. | Rob
Re: Mixed page volume sizes
On Thu, Sep 30, 2010 at 12:53 AM, O'Brien, Dennis L dennis.l.o'br...@bankofamerica.com wrote: [ snip ] need them, so if I add another 32 GB of storage, my choice would be something like 212 3390-3's and 28 3390-9's, or 100 3390-9's. Which would be a better choice? As you indicated, z/VM spreads paging effectively to fill volumes with the same amount of data. This will fill your -3's quicker than the -9's. The risk is that at some point you filled the small ones completely and end up paging only with a subset of the subchannels. Mixing sizes is not recommended; depending on the actual numbers it may be a very bad idea (tm). Bottom line is to look at your performance monitor and determine whether users are unreasonably held back by paging and whether that troubles your users. Let's do the math: once you have filled 50% of your paging space (~ 340 GB) each will hold ~1.4 GB. So the -3's are at 60% and the -9's are at 20%. You'd still have 240 packs more or less contribute in paging. When you convert to all -9, you only have 100 packs doing the paging. So in this case, 240 is more than 100 ;-) This is assuming modern DASD, not the true 3390-9 that were rotating slower so you could stuff more bytes on a track (and take 3 times as long to read and write the data). If you plotted the dates when you added page packs, you may be able to predict when you post again... ;-) and knowing that you will run out of the maximum number of CP owned volumes real soon... suppose you would go for 28 extra 3390-27 instead. Once you fill that space for 50%, your -3's are stuffed and you only have space on the big ones, so you run with only ~10% of your paging volumes (that counts as bad idea). Would you have started with all -27's, you now had about twice the number of page packs working for you. PS I'm not making up these examples. We did see a customer convert to just a few very big page devices. They were not happy. Rob -- Rob van der Heij Velocity Software http://www.velocitysoftware.com/
Re: Mixed page volume sizes
On Thu, Sep 30, 2010 at 1:35 PM, Martin Zimelis martin.zime...@gmail.com wrote: And record one more vote that mixed sizes is a very bad idea (for the reasons Rob described). But your very bad may vary ;-) For anyone still awake, ponder on this one... Rob, we just put in 30 volumes 3390-9 for paging. After the IPL, we noticed that we failed to remove 540PAG (a 3390-3) from the CP owned list. Can't get it off because it is in use. Somebody said that mixing volume sizes is a very bad idea. Is it good enough to just drain that volume or should we plan for another IPL to take it off? | Rob
Re: z/VM ISFC links
On Thu, Sep 30, 2010 at 6:21 PM, Mark Pace pacemainl...@gmail.com wrote: I think I'll also look into IPGATE. But that does not do ISFC ...
Re: Applying Maintenance - Best Practice
On Wed, Sep 29, 2010 at 11:27 AM, Kris Buelens kris.buel...@gmail.com wrote: And, if the 2'nd level system has enough free spool space, you could let this 2'nd level run without page packs: CP will page in the spool when paging is full (or not-existing). **NOT** recommended for a production system, or a 2nd level system that you use every day. Or with a VDISK taken from 1st level. My PROFILE EXEC runs ICKDSF against it to format allocate one before IPL. | Rob
Re: Applying Maintenance - Best Practice
On Wed, Sep 29, 2010 at 4:11 PM, Kris Buelens kris.buel...@gmail.com wrote: That's then to confuse everyone 0 and O, who can tell the difference unless having a very good font, and even then. If you change, change it better, not more work Why not 54TRES, 54TW01 etc I guess my background of multi-image installation shows... we never ran production systems with the IBM sample volume names, but had the owning system name in the volser (independent of the release installed). Less risky when someone forgets to change the labels (or worse: picks the wrong volume to change the label). The kind of surprises that don't show until the weekend IPL... Like our RACF-person doing major changes and then remembered his dentist appointment on Friday afternoon (which ran late, so he did not go back to the office to finish his updates...). | Rob
Re: z/VSE dispatch of multi-CPU under VM
On Wed, Sep 29, 2010 at 7:30 PM, Frank M. Ramaekers framaek...@ailife.com wrote: This was stated on the z/VSE LISTSERV, can someone confirm (or deny) it? Here is a quick tip. When running under VM with multiple VSE's it is usually NOT a good idea to define multiple CPU's to VSE and expect turbo dispacher to handle them. Why? Because z/VM will not dispach a VSE unless it has ALL requested CPU available. Often VSE could be running but is waiting for z/VM to find a secind free CPU. As stated here, we can simply conclude and demonstrate that this claim is not true. The more interesting part is to understand which statement *is* true and how that led to this rumor ;-) In general, it's a bad idea to have more virtual CPUs than you can get from z/VM when you have workload to use them. The total number of logical CPUs in z/VM is an upper bound for what you can get, but when you run 100 Linux guests on 5 IFLs, it's unlikely you find a guest have all its 5 virtual CPUs dispatched at the same time. One of the challenges with virtualized multi-processor guests is about locking. When the virtual CPU holding the lock is not dispatched, the other virtual CPU ends up spinning waiting for the other virtual CPU to free the lock (which does not happen because you're burning a CPU spinning). To avoid that, the guest OS uses a voluntary time slice end DIAG44 to give up running and expect the other virtual CPU to get time to free the lock. Linux is even using a later version of that to tell z/VM which virtual CPU should be put in front of the queue (with more than 2 virtual CPUs other is a bit vague). I don't know how much locking is done in VSE. Another aspect is about SMP. Linux is symmetrical and does not care which virtual CPU runs what. Some Operating Systems deal with serialization by master only tasks. z/VM used to have a lot of that in ancient past, and got rid of almost all now. When the guest OS needs some work to run on one particular CPU (the master) but dispatches work on both virtual CPUs, you can't pick which one is dispatched first by z/VM. The question would be whether VSE has much master only work. Rob -- Rob van der Heij Velocity Software http://www.velocitysoftware.com/
Re: z/VSE dispatch of multi-CPU under VM
On Wed, Sep 29, 2010 at 8:54 PM, Tom Huegel tehue...@gmail.com wrote: With VSE's using only 1 CPU (non-dedicated) I carfully selected a one hour job mix. Giving VSE's 4 CPU's (non-dedicated) the same job mix ran close to 1hr 20min... This is wall clock time, which in final analysis is the only one that counts. If you think we disagree, then I was obviously not clear enough. With 7 VSE guests, the amount of resources each could get at any time will be (far) less than 400% (all 4 CPUs). If more than 4 of them working, each gets even less than one CPU worth of cycles. So one virtual CPU will do. The drawback is that with just one virtual CPU you can not get more than 100% of a CPU, not even when all the others are idle. That may or may not be a true concern. While you're right that wall clock time is what counts, performance measurements help you understand the difference between two experiments and allow you to improve performance other than through trial and error. Rob -- Rob van der Heij Velocity Software http://www.velocitysoftware.com/
Re: Central vs. expanded storage
On Fri, Sep 24, 2010 at 3:04 PM, George Henke/NYLIC george_he...@newyorklife.com wrote: Why now the need Expanded Storage in the VM world to accommodate LINUX guests when Expanded Storage in the MVS world is a thing of the past? Other Frequently asked Questions that I rarely bother to answer: Why have z/VM run Linux guests when MVS can't do it? Why does z/VM page at all? Our MVS LPAR has enough storage that we don't have to. As Alan says, it's different operating systems with different workloads. The only thing in common is that they can run on the same hardware, even next to each other. Isn't that cute? | Rob
Re: Central vs. expanded storage
On Fri, Sep 24, 2010 at 7:14 PM, Alan Altmark alan_altm...@us.ibm.com wrote: But, see, it's not *quite* that simple. MVS is optimized to run in an unconstrained environment. z/VM, on the other hand, has been optimized to run in constrained environments. Both reflect their respective value propositions and decades of customer spending habits (which drive IBM investment decisions). That seems to apply also to the systems programmers working on VM vs MVS... ;-) | Rob
Re: Central vs. expanded storage
On Thu, Sep 23, 2010 at 2:14 AM, O'Brien, Dennis L dennis.l.o'br...@bankofamerica.com wrote: I heard from a couple of performance people at SHARE that we should have 20% to 25% of the total storage in an LPAR configured as expanded storage. Naturally, that's a guideline and the proper amount varies by workload. What should I look at to determine if we have enough expanded storage? We use Velocity's ESALPS suite. The systems that I'm most concerned about have a Linux guest workload. One of them is all WAS, and the other is a mix of WAS, Oracle, and some other things. You're most welcome to send us data for review. We do that all the time for customers. Let's work off-list on that; would be a bit unfair to those on the list who don't have our products. We can tell from the data when you don't have enough expanded storage. But you need enough performance history to know when your peak requirement is. Too much expanded storage is rarely a problem, to little is bad. Expanded storage creates another layer in z/VM storage management. Typical Linux workloads effectively bypass some of the first layers, so you really want that last bastion before going to DASD. The challenge in tuning many of these workloads is that the middleware itself introduces a layer of memory management that does not interface with the hypervisor. The JVM grabs memory from Linux and treats it like real, Linux gets memory from z/VM and thinks it is real. And z/VM is trying to decide who should use the real memory and when. With no further info from the application, z/VM uses the what would happen when I take this away approach. When the page is still in expanded storage, that works pretty well. When it went to DASD, it gets really slow. So I can tell from performance data that z/VM has not enough expanded storage. But since the tuning itself changes the workload, it is very hard to predict how much you need (requires Crystal Ball Technology that we have not yet announced). Since workloads change over time, and too much expanded storage rarely hurts, I tend to stay on the safe side. I've heard that WAS isn't the best choice for System z, but that's not the focus of my concern. We have the workload that we have, and I just want to make it run as well as it can. Exactly my view! It's no secret you need more CPU and memory when you replace a well-tuned CICS application by SAP with all the bells and whistles people asked for the past 20 years. If you can even compare the function offered. But we also get orders of magnitude more resources in our mainframes now. Other aspects of the business justify such a move. Can make a lot of sense. Rob -- Rob van der Heij Velocity Software http://www.velocitysoftware.com/
Re: Central vs. expanded storage
On Thu, Sep 23, 2010 at 4:36 PM, David Boyes dbo...@sinenomine.net wrote: (requires Crystal Ball Technology that we have not yet announced). Sorry for the delays in the speed-of-light upgrades. Barring further interference from the future, you should have a prototype to play with two weeks from the 11th of last November...8-) /me going though the signed NDA's to see who leaked that! Oh, and code changes color when it gets near you...
Re: CP unresponsive on certain guests
On Thu, Sep 16, 2010 at 7:09 PM, Kris Buelens kris.buel...@gmail.com wrote: You are paging on one disk only, and that disk is filled too much. My guess is that the one installing VM6PG2 as paging device forgot format that disk. The 180 blocks suggests that he got at least the allocation wrong, using only cylinder 0 as PAGE. Maybe also forget to format, but we can't tell. Since you can't free the volume now to make CP see the rest of it once allocated, just take a new volume and format that completely, allocate 1-END as PAGE. | Rob
Re: PROP RTABLE and HCPCRC8082I
On Wed, Sep 8, 2010 at 9:02 PM, Frank M. Ramaekers framaek...@ailife.com wrote: Tried changing it from “TELL MAINT” to “TELLMAIN” (for an EXEC that invokes both TELL MAINT and MSG MAINT). Still can’t get it to work. (I already have a similar EXEC that works.) You sure it's a type 1 message and not 3 ? My last resort is to have a catch all others action routine that displays the characteristics of the message. | Rob
Re: PROP RTABLE and HCPCRC8082I
On Fri, Sep 3, 2010 at 5:32 PM, Frank M. Ramaekers framaek...@ailife.com wrote: I can’t seem to get HCPCRC8082I entry in PROP RTABLE to work. I’ve used message class 1-3 and start column of 1 and end column of 80 to try to get this to work. Anyone see anything wrong? I think your column 80 is wrong... it works for me: tell op cmd acnt all Ready; T=0.04/0.06 20:30:25 Command complete TellRob: CP CP RVDHEIJ RVDHEIJ 03 OPERATOR RVDHEIJ RSCS PROP TellRob:20:30:25 HCPCRC8083I Accounting record threshold has userid MAINT. Currently 0140 records are enqueued. TellRob: The routine table entries are these: /HCPCRC8082I/ 10 20 3 TELLROB /HCPCRC8083I/ 10 20 3 TELLROB
Re: PROP RTABLE and HCPCRC8082I
On Wed, Sep 8, 2010 at 10:59 PM, Tom Huegel tehue...@gmail.com wrote: But will this help Frank get started? I would think so. I took the effort of actually coding the RTABLE and showed that it works, and I pointed out the difference with his RTABLE (that his original post wanted the full 80 chars to consist only of that message) The discussion on whether that was obvious from the documentation is probably less interesting. | Rob
Re: BookManager format softcopy
On Fri, Sep 3, 2010 at 8:11 PM, Alan Altmark alan_altm...@us.ibm.com wrote: The Information Center is very nice in that regard. You can download and run it on your workstation if you like, or you can use the Internet version. Learning curve is nil. Don't make me laugh. I thought you had 2 more weeks before your new job... Yeah, I frequently have cases like I really need to find what the default msglimit is for the IUCV statement in the directory, but the learning curve for using my PDF Reader is so steep... I have a copy of most z/VM PDFs on my laptop and hacked a copy of the HTML index from the VM web site to navigate that. Way cool. Google Desktop Search even searches them when I need. I rarely agree with David, but with respect to the Information Center it's hard to avoid. You may throw that stuff as far as you can (though not in this direction please) Only use it for the VM books by accident when Google leads me there, but I've used it more often for IBM middleware (online, so the latency may be part of my user experience). What I really hate is getting returned a list of a few hits with no indication that it's still searching and will add more later. Similar things happen with text where you see a paragraph and conclude required detail is not there, and then more lines are still added. A lot of the books suffer from too deep nesting of sections. That makes for an attractive layout in print and PDF, but is tedious with expand/hide process of Info Centers. When I click in the ToC under IUCV Statement a section called Operands then I expect to see the operands like in the book itself, not an empty paragraph and again new level of headings. Go ahead, try search the WAS InfoCenter on memory tuning - this gives you lots of references that all look the same because the InfoCenter has the books for WAS on 7 different platforms all merged together. So you get all the duplicated sections as well. Try finding how to restrict the search to the pieces that are relevant for you. Sure, it can be done by constructing your own virtual slice of a book in InfoCenter... But only if you're more patient and desperate than me. I downloaded the relevant PDF and read applicable sections as if it were a book (not much of a learning curve there) | Rob
Re: ESAWEB message
On Mon, Sep 6, 2010 at 10:51 AM, Kris Buelens kris.buel...@gmail.com wrote: One of my customers is using ESAWEB. Recently, some of their web servers abend some times with these messages 14:34:46 VSIIIH0010E TCP/IP path severed: KILL -9 14:34:46 VSIIIH0010E TCP/IP path severed: TCPIP TCPIP DMSHND911E An IUCV sever error occurred on path 2, iprcode=1; severing of other paths continues 14:34:46 VSIIPS0137W Shutdown of host (Default_host) has started 14:34:46 VSIIPS0137W Shutdown of host (Default_host) has ended Ready; T=31.13/35.61 14:34:46 My customer has no ESAWEB messages manual, and google doesn't find VSIIIH0010E, Your customer is probably looking at the wrong support staff for help in this ;-) From the data that you show, I expect the web server initiated a shutdown because the TCP/IP stack was shutdown or otherwise severed the connection. If that's not what happened, I would suggest the customer to get in touch with us so we can help understand whether he runs current level and one way or the other help him getting things resolved. Rob -- Rob van der Heij Velocity Software http://www.velocitysoftware.com/
Re: ISFC Ficon CTC question
On Fri, Sep 3, 2010 at 9:12 PM, Marcy Cortes marcy.d.cor...@wellsfargo.com wrote: And while it says State: Down on the ones with a blank Node: , clearly it's not since the byte counts keep increasing... Back then, when Alan was still among us, he'd ask to draw a picture ;-) Are you sure the counts for the down links are still increasing, or is this from before it went down? I know from experience that it is best not to add redundant links that allow systems to communicate via different paths. So if you have 3 systems, I would make just use 2 connections. CP does not like two routes to the same node, and race conditions during link establishment do weird things. So (A) - (B) - (C) Obviously you would give the role of (B) to the system that is most popular in conversations (eg where the SFS server runs). That avoids overhead of the extra hop for most of the traffic. It appears less robust, but the duplicate links don't prevent a short outage (and loss of conversations) anyway. If you need that, get something done in your programmable operator to activate the other ISLINK when you have an outage. | Rob PS My experience is based on z/VM 5 and earlier. I can't comment on what future z/VM levels will do.
Re: ipl tcpip
On Fri, Jul 30, 2010 at 6:27 PM, Dean, David (I/S) david_d...@bcbst.com wrote: Thanks, I will take a look at auditor; the temporary TCPIP outage is not a problem. I have no way other than TCPIP to connect without physically going into our Secure Area where the HMC lives. As folks mention, a 2nd stack is the most flexible (also if you made a mistake in the configuration files). And be aware that for most configuration changes in VM TCP/IP, you really don't have to restart the stack... | Rob
Re: AUTOLOG2?
On Thu, Jul 29, 2010 at 3:15 PM, Tony Thigpen t...@vse2pdf.com wrote: I have always liked the password 'NOTHING'. When the auditors look at password information, they assume it can't be logged onto. :-) They might also be happy also with NOPASS (assuming it's the same as NOPASSWORD in RACF). Such experiences should show the responsible VM Systems Programmer he's on his own and should not expect any helpful guidance from the auditors. And maybe not even try to explain why the user profiles were missing for all NOLOG users... | Rob
Re: AUTOLOG2?
On Thu, Jul 29, 2010 at 4:39 PM, Alan Altmark alan_altm...@us.ibm.com wrote: On Thursday, 07/29/2010 at 09:27 EDT, Rob van der Heij rvdh...@gmail.com wrote: They might also be happy also with NOPASS (assuming it's the same as NOPASSWORD in RACF). It isn't. NOPASS in the directory means no password required. 'NOPASSWORD NOPHRASE' on RACF means that the user ID does not have an authenticator and end users cannot access it. No FTP. No logon. All you can do is XAUTOLOG it. Right, we know it isn't. But it isn't obvious without reading the book either... .. cannot... except for LOGONBY ... ESMs can deny NOPASS logins if they want. RACF doesn't. (Though I am increasingly tempted to add a RACF SETROPTS to allow you to do so - and turn it on by default.) We played with this before RACF/VM had the NOPASSWORD setting. For NOPASS users, our local modification would skip the password check to RACF (and thus avoid the risk of getting revoked). But we did not like the idea that with RACF inactive, all these important service machines would be wide open... Such experiences should show the responsible VM Systems Programmer he's on his own and should not expect any helpful guidance from the auditors. And maybe not even try to explain why the user profiles were missing for all NOLOG users... VM allows the ESM to override a NOLOG. I.e. you have a user profile with a password and directory entry of NOLOG. You can authenticate via FTP (for example) and access files, but you do not have a virtual machine to call your own. This lets you keep USER DIRECT and the ESM in sync. I think override is a bit strong here. So you can have a RACF user profile to access resources, even though you don't have a virtual machine with that name in the directory. And we have NOLOG virtual machines defined that never run on VM, so they don't request access to resources.There's a void space between them. Some special usage cases might nicely fit in as long as you know what you do. RACF and CP directory both have a partial view of the world for their own purpose. Attempts to align them only to simplify administration often leads to interesting experiences (like automated programs issue a DIRM PURGE for MNT540 because the RACF profile had not been touched in 90 days). | Rob
Re: Dasd Volser standards documented
On Wed, Jul 28, 2010 at 2:28 PM, Philip Tully tull...@optonline.net wrote: I was wondering if there was an IBM document which gave suggestions for DASD volser conventions and/or limitations. I have seen plenty of local written ones, but I can't remember it being in a manual. No wonder you can't remember... My code has this: /* Comply with C-S 3-8010-003, dated 1981-12 ;-) credits ALTMARKA */ | Rob
Re: Dasd Volser standards documented
On Wed, Jul 28, 2010 at 2:52 PM, Michael MacIsaac mike...@us.ibm.com wrote: In the Virtualization Cookbooks, some of which are Redbooks and thus IBM documents, we recommend using the last four characters of the volser as the DASD rdev. If this convention is followed it guarantees unique labels and makes it easy to know which DASD is which. But that leaves only two That's definitely not the old school tradition where we were told to avoid volser based on real device address. You'd name the volume after the data or purpose, not on where it is sitting. Today it's probably less common to find a volume restored on another HDA when you get back in the office. Since your approach probably will have exceptions too, you'll have to use the right info anyway (rather than code 'DETACH' substr(volser,2) for example - the lookup stage is your friend for that...) | Rob
Re: Running multiple PIPEDDRs at once?
On Mon, Jul 26, 2010 at 7:57 AM, Leland Lucius lluc...@homerow.net wrote: Is there anyway to do it? I've tried using MTREXX, but the threads run after each other. I've tried multiple stages, but they also run after each other. IIRC the trackread and trackwrite stages use diagnose I/O under the covers, so would block there. Depending on what happens in the rest of the pipe, it may make sense to build a more complicated topology (like with long haul IP connection with varying latency). But probably not with CMS MT as you suggest. You probably should run multiple virtual machines and break up the work like that (though if the disk is the bottleneck, you're not likely to achieve much speed-up). | Rob
Re: Running multiple PIPEDDRs at once?
On Mon, Jul 26, 2010 at 9:23 AM, Leland Lucius lluc...@homerow.net wrote: Yepper, I'm to the point of just doing the single threaded, but, even if the trackwrite used a DIAG, then I'd expect threading would get a chance between DIAG calls. I've tried adding a ThreadYield() in a REXX stage, but that just causes all sorts of problems. Just out of curiosity, I think I'll try DDR before throwing in the towel...though I'm not gonna try too hard. :-) Yes, I/O issued by DDR will also serialize. I fear I don't know what you're trying to achieve... If you expect to speed up things, my guess is that you gain more by starting the program now than to postpone that a few hours writing complicated code :-) | Rob
Re: New standard for networking help
On Thu, Jul 15, 2010 at 5:06 PM, Alan Altmark alan_altm...@us.ibm.com wrote: Bottom line, it enabled me to discover the problem in about 5 minutes - the NATIVE and default VLAN on DEFINE VSWITCH had the same value. With so many people getting excited, I feel un irresitable urge to assume my position on the peanut gallery this Friday afternoon... Well that may be true, but at what expense for the customer? From my current position, I obviously welcome any effort the customer is willing to put in to increase my efficiency and improve the quality of my response. And I do expect that most of those 17 pages was their normal documentation that they maintain for the system anyway. But one should ask how long that customer has been fighting the problem to make them think it required such extensive documentation. And if it only took you 5 minutes to browse those 17 pages (certainly not read it all) and find the cause and post to the mailing list, is it clear enough in the books to prevent the problem from happening. But in a former life as customer, I soon realized that vendors were asking for extensive documentation and experiments only to buy time (so once you had things collected, they could tell you that you have a really old level and could you try with the latest version). An automated program to generate such documentation with no effort - or worse, even before the vendor asks for it - really defeats the purpose... :-) Seriously, I doubt such a tell me all you know program will improve things. Especially since it only shows what the customer defined, not what he meant to define or should have defined. Much of what you can collect just is not needed in most cases. Like in this case, having the Rick's list of 16,000 volumes would not have made Alan's task any easier (depending on the layout of that list, he would have told us 285 pages of documentation to be the norm :-) Don't get me wrong. I do value some kind of standard form or checklist for each specific problem area. But I would focus on the 10% of the information that resolves 90% of the questions. My experience is that 3 questions is about the maximum you can do (beyond that, people seem to think it's multiple choice and they answer just one or two of them :-) | Rob
Re: New standard for networking help
On Fri, Jul 23, 2010 at 3:26 PM, Edward M Martin emar...@aultman.com wrote: Having good documentation helps everyone involved, and good change management (of some sort) helps debug issues. Haha. For those without proper change management, when they tell you nothing changed it sometimes helps to ask what do you think someone else might have changed to cause this difference :-) | Rob
Re: assembler error
On Tue, Jul 20, 2010 at 2:54 PM, Tony Thigpen t...@vse2pdf.com wrote: I normally have 32m of virtual storage and the program assembled correctly earlier today. I tried upping it to 128M, but I get the same error. Other info: (PTF UQ29646) Page 1 HLASM R3.0 2010/07/20 08.50 It's not really the very latest fashion in HLASM... it could be that it lives under the 16M completely. Check with STORMAP whether you have loaded some other stuff under the line that uses a large part (or fragments it) since last time you assembled. It's also possible that you normally run the assembler from DCSS, but something is now blocking your sight. Go back to your usual 32M and IPL CMS to make sure you did your best. Oh, and did you check that the record format of the file is correct? I've seen compilers give confusing messages when presented RECFM V files. | Rob
Re: question to mixed CP an IFL in one LPAR
On Wed, Jul 14, 2010 at 9:34 AM, Franz Josef Pohlen fjpohlen-maill...@gmx.de wrote: Hello listers, an IBMer has told me that a mixture of CPs and IFLs in one LPAR are supported on z10 not only with z/VM 6 but also with zVM 5.4. Is this correct? I thought that for those environments you must have z/VM 6. Your IBMer is correct. Be aware of the required licenses. It's probably most attractive if you already had both sides of the z10 licensed for z/VM 5.4 and want to lower your operating cost by having just one single image. | Rob
Re: Pipe question
On Fri, Jul 9, 2010 at 3:25 PM, Mark Pace mpac...@gmail.com wrote: and it does not create the file. change it to console and it works again. Am I missing something obvious? As Alan points out, you need to append rather than replace the file. But since it is Friday, you might learn a bit of mult-stream plumbing instead... Rather than calling the pipeline in a loop, you could make a single pipe do all the files: 'PIPE (end \)', '\ stem files.', '| o: fanout', '| pad 25', '| j: juxtapose', '| test data a', '\ o:', '| getfiles', '| locate anycase /'LookFor'/', '| j:' Depending on the number of files you're handling, you might even notice the speedup. But more important is that it makes it much easier to extend the process and do other things with the data. Rob
Re: Pipe question
On Fri, Jul 9, 2010 at 5:06 PM, Mark Pace mpac...@gmail.com wrote: Thanks, Rob - After years and years of ignoring Pipes, I decided this week I was going to learn how to use them. I found some excellent documentation, thanks to you and others, and have dived head first into the pool. But even with the Authors Edition, and Pipelines Visualized, I'm still struggling with multi-stream plumbing. I just need more practice and see good examples. The papers from Melinda on the Pipelines Home Page are classics. The Pluning On does multi-stream. And we should probably discuss those things on CMSPIP-L instead... Sir Rob the Plumber
Re: Second screen in a z/VM CMS session
On Thu, Jul 8, 2010 at 8:25 AM, Florian Bilek florian.bi...@gmail.com wrote: Rob, I would also be interested to get this code. Another userful thing that does not do what was asked for... I am very fond of Perry's SWAPCONS that does fullscreen on a dialed GRAF and linemode output on your normal terminal. Great if you want to trace REXX code that drives (XEDIT) full screen. Perry just reminded me the source is (since 1991!) on VMSHARE: http://vm.marist.edu/~vmshare/read?fn=SWAPCONSft=NOTEline=1 If people have trouble building and using it, we might need a package on the VM Download pages. R;
Re: Second screen in a z/VM CMS session
Another userful thing that does not do what was asked for... I am very fond of Perry's SWAPCONS that does fullscreen on a dialed GRAF and linemode output on your normal terminal. Great if you want to trace REXX code that drives (XEDIT) full screen. R;
Re: Sample REXX using XEDIT
On Wed, Jul 7, 2010 at 3:50 PM, Sergio Lima sergiovm...@hotmail.com wrote: We need, do here a REXX that will READ a lot of COBOL Programs. So, must, locate all CALL Statements, and think use XEDIT for this. Someone know, if have a place, or a documentation, for us get a sample program? I would say that for such situations, any time invested in learning CMS Pipelines will pay back tenfold (both in programming and runtime). My little pipeline refinery to convert assembler files to HTML pages is a serious productivity gain. It does links to follow branches and calls and to find fields in DSECTs. It also generates a symbol cross reference with links . You'd hang out on CMSPIP-L with questions about using CMS Pipelines... R; Sir Rob the Plumber