Re: DirMaint CA's VM:Backup Software Installation
Create a file with the directory entry VMIRMAINT shows you called VMBACKUP DIRECT. Then use DIRM ADD VMBACKUP to add it. DIRMAINT does the rest. Once DIRMAINT is active, you should NEVER update USER DIRECT. All that is behind you - use the DIRMAINT tools to do directory management. From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of Dave Keeton Sent: Wednesday, October 15, 2008 6:39 PM To: IBMVM@LISTSERV.UARK.EDU Subject: DirMaint CA's VM:Backup Software Installation I am attempting to install VM:Backup from CA and it wants me to update the directory with the user info for VMBACKUP. Trouble is, I'm using DirMaint and I can't seem to figure out how to manually add the user as the installation program has spelled out the entry for me. Can I update the USER DIRECT file and then get DirMaint to re-read it? Can I tell DirMaint to manually accept the user info? What's the best way to approach this? The VMIMAINT installation program lets me pick the volume on which to create the minidisks, but DirMaint would be using the EXTENT CONTROL definition instead of specific cylinders on a volume. Not sure how to proceed Thanks in advance, Dave
Re: DF/DSS backup of z/VM volumes
What is wrong with using CPFMTXA to format and allocate your VM volumes. Most Modern DASD subsystems are doing the basic formatting of the DASD volumes and all that is needed on VM is to format cylinder 0 and Allocate the volume as all PERM. Then format each individual minidisk as needed. If you need CP OWNED areas, Make sure you format all these areas using CPFMTXA. You actually answered your own question. If you don't make a habit of doing the CPVOL FORMAT on every volume, at some point in the future, you'll allocate a disk with non-PERM areas, forget that you haven't done the CPVOL FORMAT, and Much Sadness Will Ensue. Better to just make a habit of doing the format as part of putting the volume online to VM. One less bullet to shoot yourself with. -- db
OpenSolaris for System z finally available for download...
We've finally received all the clearances necessary, and the system images for build 95 of OpenSolaris for System z are finally available for download from http://distribution.sinenomine.net/opensolaris. 3 packaging choices are available: VMARC, AWSTAPE and a DVD image. All three are identical code, and can be used with the installation instructions available in the package and on the www page mentioned above. Docs are provided in the VMARC files in plain text, PDF and Bookmaster/READ format. You need a z/VM 5.3 or 5.4 system with VM64466 applied, and you must be running on a z9 BC, EC, or z10 to use these images. They will *NOT* run in an LPAR or on any form of 9672, z800, z890, z900, or z990. Support and education are available - please contact me offlist for more details. It's been two years in the making, but we're pretty proud of it. Please send questions, comments or bug reports to the OpenSolaris for System z discussion list at [EMAIL PROTECTED] We encourage anyone working with this code to subscribe (by sending mail to [EMAIL PROTECTED] with the words SUBSCRIBE SOL-390 firstname lastname in the body of the message) and help us make OpenSolaris even better. Happy downloading, -- db David Boyes Sine Nomine Associates
Re: z/vm 5.4 DVD installation
How this will finally be delivered is still TBD. As I indicated, this is a known requirement that we continue to work to deliver a solution. Even though it sounds simple to burn DVDs, there are other complications here. Just a suggestion: HP solved this problem by encrypting the existing software packaging with a key that (combined with the CPUID) would allow you to extract the package to local disk and then install it normally. I get 20 DVDs every 6 months that have EVERYTHING (every single HP operating system, service pack, applications, the works) on them; I can only decrypt and use the stuff I've paid for. I can get keys from a web site and trade them in if I change CPUids. Scales well in that HP gets to make a really BIG production run of the DVDs, and they're worthless without the keys, so HP really doesn't care if we copy them for other sites or give them to people -- in fact they encourage it because it saves them money on DVD production and they get green company points for saving energy and reducing waste. No product owners had to do anything at all to accommodate the method, and the approach of having EVERYTHING on local media with 0 delivery time is a real plus. Makes DR a lot easier when I know I've got install media for everything we own that has an HP logo in a form that fits in a lunch-box. Just something to think about. -- db
Re: z/OpenBSD -- wibni?
On Thu, 9 Oct 2008, David Boyes wrote: ... AIX/370 was a perverse evil mutant thing from hell that didn't die a second too soon. Modern AIX (ie post-AIX 4) is actually a pretty nice OS, and a lot different from AIX/370 or AIX/RT (which spawned it). ... Umm... To clarify what the Doctor is saying, it's true that AIX/370 was strange and kind of half-hearted. AIX/370 perverse? Perhaps, but not exactly evil. As I put it, half-hearted. It was obvious, when one drove it from the console, that the developers learned as little about the mainframe as they had to in order to make the port. And they were working from AIX/RT as a base, which was weird by design. AIX/370 mutant? YES! It had, among other things, something called hidden directories, which amounted to a weird attempt at multi-platform support. Borrowed from AFS. There was also still a lot of offload thinking to the extent that you were EXPECTED to couple AIX/370 with AIX/PS2, which many shops were simply not interested in doing. (AIX/PS2 would be the front-end handling all that blasted byte-at-a-time silliness. Turned out that even as stupid as byte-at-a-time is, even the 24-bit S/370 kept up.) The problem was not that you had to have the PC, but that the PC had to be micro-channel, and be directly attached to a 370 block-mode channel. The channel card cost more than the PC (which had to be a mod 60 or higher, mod 80 recommended).
Re: z/OpenBSD -- wibni?
The solution for customers with lots of small iSeries/System i boxes today running IBM i (formerly i5/OS formerly OS/400) is to Good grief -- it's like going to IKEA with these folks. Name change just Because We Can. put in a Power blade (JS12 or JS22). You save a lot of space and get to share a BladeCenter with x86 blades. Yes, that's a possibility, but it doesn't really address the base problem, though, which is to reduce the number of platforms to 1 or 2 at the most. What we observe is the customers stratifying down to Intel + Z (mostly due to the embedded nature of z/OS into core business), and any of the specialty architectures are having a lot of rough sailing to stay in play. Certainly SPARC and PA-RISC are past it, and losing Apple put a really big nail in the Power coffin. While I really like the Power chipset design and the pSystem management tools, it's a very difficult sell right now unless you have specialized problem sets that play to its strengths (large-vector operations like those common in bioscience or image processing). Writing an emulator for the iSeries on Intel would pretty much kill its advantages for non-NIC completely; the cost per performance numbers are just too good for raw MIPS on Intel/AMD compared to the cost of support for exotica.
Re: YAST
If you're trying to use the starter system install server to install the Virtualization Cookbook golden image, then use the directory entry from the cookbook, log in to the client you're installing, #CP LINK NOVSTART 19F 19F RR, and run SLES. Then use the directions in the cookbook on selecting software, etc. When asked, specify the starter system server as the install source (via HTTP, I'd suggest, but any of the supported methods provided in the starter system will work). There's no right or wrong way to do this; the two methods were designed by different people, and each is equally valid.
Re: Monitoring the backup OSA on the vswitch
How are folks monitoring their backup osa's on the vswitch? How do you know they haven't lost connectivity and won't be there when you need it? (asking because, well, yeah, that can happen :) You can also use any SNMP-capable monitor and look at the SNMP physical link state variable in the OSA hardware MIB and the switch MIB for that interface. If that drops, then the interface isn't usable.
Re: Process or tool to verify linux/oracle environment under vm
On 10/11/08 8:13 AM, Barton Robinson [EMAIL PROTECTED] wrote: Nagios and Hobbit are a naive solution for very small operations. They can be a very small part of what operations want, but any installation wanting to manage multiple servers has more complete z/VM and Linux requirements. Tools based on simple availability and CP Indicates don't meet any knowledgable set of requirements. If this is your understanding of the capabilities of Nagios and Hobbit, then you need to take another look, minus the attitude and blinders. Both offer the ability to conduct active testing of an application outside of the application host without an agent installed on the application host, as well as a library of pre-built and user-contributed application test modules.
Re: z/OpenBSD -- wibni?
z/OpenBSD -- wibni? Yeah, it would. I just don't think anyone's got the spare time to spend on something that has even less commercial support than Linux and OpenSolaris. The justification for doing both Linux and OpenSolaris had to do with mindshare and commercial access, with Linux attaining a growing mindshare, and OpenSolaris acting as a skills bridge, as well as bringing some of the unique Solaris extensions that Linux hasn't acquired yet. OpenBSD would be fun, but you can run everything that it does on Linux, so it'd be a hard sell. If we do another native OS port, I'd be more interested in bringing i5OS, AIX or OpenVMS over. AIX has a few things that neither Linux or Solaris has developed, and has an application following already, while OpenVMS would be even more attractive from the canned system front Still, it'd be fun to have a true BSD derivative on 390. AOS was nice on the RTs, and there still are a bunch of semantic things that I really hate about SVID-based systems. -- db
Re: Process or tool to verify linux/oracle environment under vm
Nagios or Hobbit are quite nice for that sort of thing, or just use the same tools you use for the distributed systems. If you don't have such tools, that's a nice thing to offer the rest of the organization as (another) reason for the Linux on Z distribution. From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of Robert J McCarthy Sent: Thursday, October 09, 2008 2:55 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Process or tool to verify linux/oracle environment under vm We are looking to implement a tool or develop a process for use by the Operations staff to use to check the availablilty of our linux guests running under vm. When we IPL vm, the vm guests all start, however sometimes, linux and/or Oracle database running under linux donot correctly complete initialization. We need an operator friendly tool or process to verify linux and oracle availablility at IPL time, whenever a linux guest is recycled, or any other time operations deems necessary. We are trying to find out what others are using. Thanks,Bob
Re: z/OpenBSD -- wibni?
While I admit that time is a determining factor for an individual, you know how much (time) a port costs and ... Painfully well. Especially in having to spend time with lawyers to CMA. The problem is that it's not the oh it boots and gets to single user and..., multiuser, networking, drivers,... but then it's QA, keep it running, bundle it, adopt apps, mgmt apps, write docs, guides, ..., support it. Teams, more people, different temas, PYs PYs PYs. Exactly. Lot's of consulting hours to sell to migrate Sparc people to z aeh linux and solaris to BSD on z;) More like lots of non-billable hours to spend to convince people to consider doing projects in the first place. Harder than it looks. OpenBSD would be fun, ... On the other hand another problem I see for going BSD(*) (even for fun) is the lack of (public) documentation - I don't mean principle of operations or the like but OSA*/HS/crypto/... for example. I guess if getting there, that would not be a problem anymore - at least with 3-letters-paper. Having recently done this with OpenSolaris, I can say that, modulo crypto, this is near to being possible without the gory details of how the hardware works. You need to look very closely at APAR VM64466 for the final set of information to get a server-oriented system going; the rest of what you need is in the CP Programming Services manual, which is publically available information. (You can thank the OpenSolaris project later for getting the bit of work in VM64466 made public). Hopefully IBM will keep their promise on publishing documentation for the services therein on schedule. I always worry what would be next if one gets there? To keep this relevant to this list - Would the answer be 'find someone and go for VM'? See above. You can do it much more easily, if you start with the assumption of z/VM being present. After that, it's just learning more than you ever wanted to know about cross-compilation...8-)
Re: z/OpenBSD -- wibni?
If we do another native OS port, I'd be more interested in bringing i5OS Already virtualized on its own platform, sounds rare and kinky to duplicate that on the 390. Not really. There are a lot of large shops that have significant numbers of small to medium iSeries boxen that they'd give their grandmother's eyeteeth to move to either Intel or Z virtual machines so that they could completely eliminate an entire hardware platform (and an enormous amount of floor space). AIX Died a slow and natural death on the mainframe in the 1980's. Not really a fair comparison. AIX/370 was a perverse evil mutant thing from hell that didn't die a second too soon. Modern AIX (ie post-AIX 4) is actually a pretty nice OS, and a lot different from AIX/370 or AIX/RT (which spawned it). As pseudo-realtime OSes go, modern AIX has a lot to be said for it (it actually would be a much superior replacement for Comm Server for Linux guests; the Linux/SNA glue is really unnatural and weird, while the SNA bits on AIX are much better integrated). OpenVMS would be even more attractive from the canned system front VMS was a lot of fun, but is anyone still using it outside maintenance mode? Been in an machine tool factory or IKEA recently? We're actually seeing an uptick in OpenVMS now that it runs on reasonably inexpensive Intel hardware. It's a tight, clean, secure little OS that is fabulously good at canned application systems. OpenVMS's largest historical problem was the expensive DEC/Compaq/HP custom hardware vs commodity gear.
Re: VM operator console?
Trouble with a 'real' console is the need to be a locally attached 3270 device. I understand that to mean a 3174 or similar device. We don't have any. All our MVS consoles use the ICC OSA interface. If you've got an ICC, then define one of the logical devices on it to the VM LPAR and have it be the console in SYSTEM CONFIG. It's a waste of a perfectly good terminal, mostly - with no CMS users, there's very little for an operator to do on VM - but it makes all things equal from an operator POV. If you have some way to set up a NJE link to z/OS, you can route things with PROP to z/OS and handle them in Netview there. Either RSCS or the NJE Bridge does a nice job of that.
Re: VM operator console?
For an MVS shop, that all things equal helps with operators. I think that's a universal...8-) we have OPERATOR logged on that Mast er Console running the Performance Toolkit program PERFKIT. With an minimum FCONX $PROFILE, you get a nice scrolling console with some (not great) logging of console activity to disk. Yes, good suggestion. Perfkit also has a much easier to use and more sophisticated message filtering and management capability than PROP. It's also a lot easier to implement message forwarding to remote destinations, and multiple users viewing the console as well.
Re: Mainframe education/training
What's the good and bad news about today's mainframe education? I think the good parts are that there is more of it occurring. The negative side is that there continues to be a great deal of confusion (most of it generated by IBM) that mainframe = z/OS. Much, if not all, of the advances in training have covered exclusively z/OS education, not System z education. That may change soon, but I can't say the state of real System z education has improved much at all. What's missing? z/VM education Linux on z TPF VSE Large systems engineering Batch processing education (this seems to have completely disappeared) What works? Classroom instruction? Online training? Self-study courses? Books? User groups? Combinations of the above. Online only is of limited value, especially if you don't have a lab system you can break. Do employers pay for it? Is it a necessity or a golden perk? Is it safe in the budget or a first thing to get whacked? If what I needed existed, I'd pay for it. User groups are generally as effective, especially if well-organized. What's happening in local mainframe user groups? Well, Hillgang does pretty well.
Re: Mainframe education/training
I tend to agree with 99% of what Mr. Boyes said. However, the educational responsibility needs to be placed more upon academia. For them to teach UNIX / Linux is a given ability. The lack of educators with mainframe skills and knowledge along with their respective teaching tools limit in today's schools. I think another problem with the current approach is that IBM and others approach the problem with a prepared curriculum that they want to have taught. This approach works pretty well with technical and vocational schools, but I think it misses the mark with the 4 year schools because it's not easy to get new curriculum added to a true university program, and most of those organizations have a difficult time accepting classes they have to rely on remote resources to teach. There's a contractual obligation there that IBM has partially addressed with donating hardware to some schools, but expecting pre-prepared materials to go into place is a hard sell when competing with other areas. I think another interesting observation has been the fleeing of students from CS of any kind over the last few years. Many of them were going into finance or business - I wonder if that will continue? If not, then there may be some interesting opportunities to capture those students as they try to change course midstream.
[no subject]
Most common reason is a COLD start. Remember, segments live in spool. Second reason, spool volume offline. VMFBUILD ( ALL should fix it. From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of Dean, David (I/S) Sent: Monday, October 06, 2008 11:08 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Help. This is a second level test system that used to work. Any ideas what I did? DMSDCS1083E Saved segment CMSPIPES does not exist DMSDCS1083E Saved segment CMSPIPES does not exist DMSDCS1083E Saved segment CMSVMLIB does not exist David Dean Information Systems *bcbstauthorized* Please see the following link for the BlueCross BlueShield of Tennessee E-mail disclaimer: http://www.bcbst.com/email_disclaimer.shtm
Re: DMSDCS1083E Saved segment CMSPIPES does not exist
Log in as MAINT on 2nd level. VMFSETUP ESA CP (or whatever your level of z/VM expects) VMFBUILD ( ALL See the VM Service Guide manual. It'll explain all the goodness and mercy contained in VMF utilities to you. From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of Dean, David (I/S) Sent: Monday, October 06, 2008 11:21 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: DMSDCS1083E Saved segment CMSPIPES does not exist Last question (ahh, probably not) but where does vmfbuild live? Drive?
Re: DMSDCS1083E Saved segment CMSPIPES does not exist
2nd. You want the current level from the 2nd level system built in 2nd level spool. From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of Dean, David (I/S) Sent: Monday, October 06, 2008 11:16 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: DMSDCS1083E Saved segment CMSPIPES does not exist On 2nd level? Or first? David Dean Information Systems *bcbstauthorized* From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of David Boyes Sent: Monday, October 06, 2008 11:13 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Most common reason is a COLD start. Remember, segments live in spool. Second reason, spool volume offline. VMFBUILD ( ALL should fix it. From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of Dean, David (I/S) Sent: Monday, October 06, 2008 11:08 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Help. This is a second level test system that used to work. Any ideas what I did? DMSDCS1083E Saved segment CMSPIPES does not exist DMSDCS1083E Saved segment CMSPIPES does not exist DMSDCS1083E Saved segment CMSVMLIB does not exist David Dean Information Systems *bcbstauthorized* Please see the following link for the BlueCross BlueShield of Tennessee E-mail disclaimer: http://www.bcbst.com/email_disclaimer.shtm Please see the following link for the BlueCross BlueShield of Tennessee E-mail disclaimer: http://www.bcbst.com/email_disclaimer.shtm
Re: SLES10 Client
I've added an annotated console log for both the starter system install and a client system install as appendices in the document, assuming that Novell does another release. -- db
Re: SLES10 Client
*** login using 'ssh -X [EMAIL PROTECTED]' *** *** run 'yast' to start the installation *** Am I suppose to disconnect from client and then ssh to it and enter YAST; Yes.
Re: SLES10 Client
While implementing the SLES exec for the SLES10 full-service server, it asks for a MAC address. (I thought the MAC address was automatically assigned. Why is it asking for a static address?) We are using OSA GB-EXP. You should respond no to the layer 2 prompt here. Nothing in the setup process needs layer 2 access. In general, it's asking so that it correctly identifies the interface you want to use. CP generates a random MAC for the last 3 bytes of any virtual interface (the first 3 bytes are set by MACPREFIX in SYSTEM CONFIG), so if you are going to use the layer 2 function, it matches the virtual MAC to identify which interface to use.
Early draft of architecture and porting guide for OpenSolaris on Z available
An early draft of the architecture and porting guide for OpenSolaris for Z is available from distribution.sinenomine.net. It covers the release 95 build. This is a draft, so there will be a few changes yet, but comments and corrections are always welcome. File is at http://distribution.sinenomine.net/opensolaris Happy reading, -- db David Boyes Sine Nomine Associates
Re: VM RSCS as RJP to JES3, replace current process and last 3745.
Doesn't JES3 speak TCP/IP these days? If so, use the TCP/IP NJE driver to connect RSCS to JES3. I think it does (or it's trivial to use our NJE Bridge to provide an impedance matcher), but it sounds like the original setup has some kind of modified RJE exit on the JES3 side that sets the default for output to the JES system, not the originator of the job. In that case, you'd have to apply those mods to the NJE driver in JES3, which is not trivial. I think they are going to have to make some changes in the job stream to tell JES to dump the output locally, or make the mod to the NJE driver if they want the same functionality. It's not clear to me whether the specialized commands he's talking about are on the JES side or on the RSCS side, but in either case, they're looking at some refit.
Re: VM RSCS as RJP to JES3, replace current process and last 3745.
That would probably work for output, but if they're used to job progress messages on VM from JES via NJE msgs, that would break message routing, I think. Worth a try, though. Rob's idea is intriguing; the approach of letting a service machine mung the JCL into IEBGENR and submitting it via the internal rdr is clever (and gets the text processing work out of assembler and into REXX and Pipes, which would be double-plus good in terms of easy maintainability as well as getting rid of a JES exit or mod). If the //*MAIN ORG idea works, that'd be the way to implement it. Whatever they do, though, I think those custom commands are probably toast. From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of Smith, Ann (ISD, IT) Sent: Friday, September 26, 2008 1:33 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: VM RSCS as RJP to JES3, replace current process and last 3745. I'm not sure what the 'JL' command is but it sounds like you want your job sysout, jesmsg, etc. to stay over on the JES3 side. You might try editing the jcl on the VM side to include a //*MAIN ORG statement to override the origin similar to //*MAIN ORG=jes3node.LOCAL where jes3node is your zOS BDT node name. You can use OUTPUT or FORMAT statements as well but //*MAIN ORG might be best. I don't know what 'JL' is but we send jobs and data from zVM to zOS sendimng output to various vendor softcopy and output storage products. Some need special output classes. But if your current jcl works from VM with RJE then NJE may work fine if you try options with //*MAIN ORG. Always glad to see someone else using JES3.
Re: VM RSCS as RJP to JES3, replace current process and last 3745.
Since we have VM/RSCS appearing as a RJP workstation to JES3 using EP and a 3745, can it do the same using the SNA protocol to look like a 3770 device, but minus the 3745? If I'm reading the RSCS docs correctly, no. The RJE drivers really do expect a physical line, as far as I can tell. The exit customization guide also seems to point that way.
Re: Spool
Where could a 'newbie' find a comprehensive explanation of Spool? ie What it is intended for, how its used by VM and potentially other products/tools/applications. Ask here. 8-) Originally, on CMS-intensive systems, spool areas were for two purposes: 1) a place for virtual machines to submit unit record input and output for processing by CP on real devices (eg, print files that CP would drive on a real printer, incoming card decks from real card readers that needed to be directed to a virtual machine, etc). 2) a place for virtual machines to put data that needed to be processed by other virtual machines, eg console logs, files, etc that are produced by one virtual machine but another virtual machine has a need for processing. The PRINT and PUNCH commands produce examples of #1. SENDFILE, DISKLOAD and friends are examples of #2. Around VM/XA, spool acquired two new purposes: 3) Store DCSS (discontinuous shared segments) which can represent shared data mapped into multiple virtual machines. 4) Store NSS (named saved systems), which are IPLable blocks of code that can be IPLed by name, eg CMS, GCS, etc. DCSS and NSSes were around before this, but they were stored in special areas that had to be managed carefully, and you often had to fuss over a lot of things to keep them updated; turning them into files in spool made that process a LOT easier (and gave you an easier way to back up and restore them via SPTAPE and SPXTAPE). Today, most VM systems that support primarily Linux guests use spool only to capture console logs and present them to other virtual machines for archiving via tools like PROP, or to communicate with z/OS via RSCS. CP can also allow paging to overflow into spool if things are really overtaxed, although this should NOT be a normal thing on your system. Other products often use spool to communicate with each other by sending small transactions as files in spool from one virtual machine to another, kind of an implied transaction queuing system. RSCS (or other NJE implementations) can move files between spool areas on different systems, so servers can be separated. There's lots of other useful stuff you can do with spool. It's hard to describe it all, because it's so varied. I like to capture console logs and send them to a single virtual machine for analysis and archiving. I could do it with a central syslog server, but what if the network breaks? Spool doesn't easily break, and I don't lose messages due to losing UDP packets over a busy LAN. I can also take advantage of PROP automation and lots of other free VM goodies to help manage the system. Does that help?
Re: Spool
YES IMMENSELY. Linux guests are our sole customer on /VM, so that is where my 'need to know' lies. Good. I have a personal interest in making sure BCBS gets this right. 8-) One other observation on DCSS: these are being exploited for Linux as containers to use for seldom-changed filesystems that are mounted on lots of systems. Mapping the contents of the filesystem into a memory block, and then mapping that memory block into multiple virtual machine's address spaces is really really clever, if a somewhat of a bear to manage. Great for /usr on mostly-identical systems, though. I'm sure this wheel has already been discovered: Is there a 'doc' on using spool and a central 'syslog server' to capture Linux guest console logs? I think I did a presentation on it a few years back, which I can't put my hands on right now. I'll root around and see if I can find a copy. Basically the idea is threefold: 1) spool the virtual console of your Linux guest to another machine that acts as a collector for the files and processor and have it automatically close the console file every 750-1000 lines (or however much risk you're willing to tolerate). 2) configure /etc/syslog/syslogd.conf on Linux to log everything to /dev/console 3) Run PROP plus a homebrew exec in the collector machine that wakes up when a file arrives in it's virtual reader, reads it and sends each record to your central syslog server via UDP (or however you want them archived). Repeat forever. If you get an error in logging, save the incoming file and reorder it for later processing. Very easy, and quick to do. If I find the presentation, I'll put it up on the SNA www site and let you know where to find it. - db
Re: SLES10 Client
Do I need a secondary user defined It's not required, but it is a good idea to have somewhere central for all the console logs on your system to go by default, rather than have them scattered all over the spool. If you want to silence the message, change the CONLOG on the CONSOLE statement in the directory entry to any valid userid on your system, or remove it if you want the console logs to end up in CLIENT's rdr. do I need both CPU statements? No. Most guests seem to feel healthier and more responsive if you have more than one logical CPU, but it's not required or desirable in some cases. Do note that there are some applications that don't build properly in a UP system, so we put the extra statement in to try to catch that. As Rich said, it's an example. Tailor it as you see fit. 19F and the disks you plan to use to put the Linux system on are the important bits. The rest is as you like it.
Re: TNVT100 (was Re: LOGONBY)
OK, I'll byte. Is the most current TNVT100 located by google at: http://ukcc.uky.edu/~tools/1997 Or is a newer one located somewhere beyond google's wide vision? That's the version I have. I don't think Arty's done anything to it in a very long while. Emacs on a 3270 is a very strange experience. -- db
Re: Running os/390 as a guest
May I run OS/390 V2R8 as a guest under any of the supported releases of z/VM? I believe it is unsupported by IBM, but, does it work? OS/390 is unsupported so I don't think I'm any worse off. I'm looking at bringing in a smaller, newer, IBM mainframe and trying to save some money on my IBM monthly license fees. But, I need to support an old guest running OS/390 V2R8. I can't afford to upgrade all the third party software that this guest is using. Assuming that the release of OS/390 can deal with whatever disk and hardware you will be running it on, and you have standard CPs available, then it ought to work OK. Don't expect it to work with newer disks, OSAs, etc - you will probably have to isolate it as much as possible from the real hardware using minidisks, VSWITCH/gLAN, etc.
Re: SLES10 Client
No, you should use an different IP address from the same subnet, or from a subnet that is reachable from the starter system. You need the starter system to be up when you try to build the new system. What page in the docs are you referring to? If I goofed, I'd like to make sure I fix the docs. -- db
Re: SLES10 Client
No, the SLES exec punches the install deck to the new client rdr and IPLs from that. NOVSTART does have to be up to deliver the files to the client during installation. -Original Message- From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of Austin, Alyce (CIV) Sent: Wednesday, September 24, 2008 5:17 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SLES10 Client Does NOVSTART have to be up to IPL the client?
Re: SLES10 Client
OK, then Rich is correct. You're telling it the address of the install server (where it's going to get the files to install the new server). The new server will eventually need a unique address of it's own, though.
Re: LOGONBY
So the only thing you are buying here is that you keep TCPMAINT password secret is that the whole idea behind LOGOnBY? So then you only add certain user ids to do LOGONBY for this user id correct? Think of it more as a role: you are assuming the role of TCPMAINT, using your own login credentials to validate your claim to the role. The idea is minimum privilege; shared ids should not be directly logged into, because you lose the audit trail of who did what. You give individual ids minimum privilege (essentially with the combination of LOGINBY and PROP, there's rarely a real reason for any individual id to have more than class G), and they authenticate to the shared ID when they need to do something more powerful, or an extended string of things that require privileges or access to files w/o having to jump through a lot of maintenance-intensive hoops.
Re: LOGONBY
It's controllable by RACF, but is part of CP (finally). Is LOGONBY a RACF thing or z/VM???
Re: LOGONBY
It was a user mod for a while (pre-XA). I think VM/XA or VM/ESA 1.x was when it became official. That was a long time ago, though. I know I had it on SP5 as a usermod and remember that I was really glad to not have to maintain it any more.
Re: z/VM 5.4 Guide For Automated Installation and Service
On 9/23/08 4:58 PM, Michael Forte [EMAIL PROTECTED] wrote: I was able to pull a few strings and have the z/VM V5.4 Guide for Automated Installation and Service, as well as the two summaries, added directly to the z/VM library page. The Force is strong in this one
Re: LOGONBY
Geez. 28 years. Now I really feel officially ancient. Thanks a lot, dude. On 9/23/08 5:44 PM, Dave Wade [EMAIL PROTECTED] wrote: David Boyes wrote: A very long time... http://vm.marist.edu/~vmshare/browse?fn=DMKLOGft=MEMO http://vm.marist.edu/~vmshare/browse?fn=NOPSWDft=NOTE
Re: LOGONBY
On 9/23/08 5:58 PM, Martin, Terry R. (CMS/CTR) (CTR) [EMAIL PROTECTED] wrote: Hi One last thing on this. Am I logged on with my user id and password then from there logonby to another machine such as MAINT? Or do I just logon to MAINT using LOGONBY with my personal user id's password? The latter, although if you have RACF or a similar ESM, you can force the other behavior by limiting the pattern of logical terminals that certain IDs can be used from, and installing a copy of the super-fabulous SESSION tool available from most collections of useful VM tools. No system should be without it (or TNVT100). SESSION lets you create essentially a poor-mans session manager from a CMS session. Good for environments that force you to logon as you first.
Re: LOGONBY
On 9/23/08 6:14 PM, Mike Walter [EMAIL PROTECTED] wrote: No, you cannot logon to one userid when logged onto another. Mike! What happened to your copy of SESSION? Native CP can't do it, but you certainly can use SESSION or YVETTE or NV/AS or PVM to get multiple login sessions from a single terminal.
Re: WAIT STATE 9003
On 9/23/08 6:15 PM, Mike Walter [EMAIL PROTECTED] wrote: Just on the off chance... are there any CPUs defined to that LPAR? Check the HMC. Also make sure they didn't accidentally give you a CF engine by mistake. They used to look the same as an IFL on older HMCs, and some people haven't gotten used to the new HMC software.
Re: Network-SLES10
I'm am using network device QDIOETHERNET: OSA-Gigabit Ethernet with my OSA device addresses (E000-E002) dedicated to NOVSTART in my directory. When I IPL the SLES10 starter system, I get the following message: (It doesn't give me the choice to choose '3' as my network device.) Is there a statement I need to add to the profile exec on NOVSTART to implement the network connectivity? (The exec doesn't give me a choice to choose which network device I am using.) Others have probably answered this already, but the starter system is configured as an appliance, so you don't get full configurability. You have to attach the network adapter at virtual address 340 if you are not using a virtual NIC (and even if you are, you still have to use address 340). The systems you install from the starter system have full configurability. (God, I need an eighth day in this week. *sigh*) -- db
Re: z/Linux Network Monitor
What sort of things do you want to monitor? The traditional SNA monitors like Netview can provide connectivity monitoring remotely, and it would be fairly easy to develop some plugins for Nagios or Hobbit to do SNA monitoring and reporting back to the monitoring system via IP, but I doubt there is anything out there at the moment that runs ON Linux other than Omegamon XE which understands SNA connectivity (and even then, it's probably going to be a very limited understanding). That said, the BMC and IBM products running on remote systems can certainly monitor Linux SNA endpoints. There are also analysis tools from Cisco and others that can get Netflow data on SNA traffic and produce interesting reports on it. From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of [EMAIL PROTECTED] Sent: Thursday, September 18, 2008 6:45 PM To: IBMVM@LISTSERV.UARK.EDU Subject: z/Linux Network Monitor We moved some SNA applications from unix servers to z/Linux. Now are running with Communications Server for Linux, we successfuly implemented Enterprise Extended. But we are looking for some Network Monitor. We find some to monitor TCPIP, but none for SNA. Does anybody know any? Thanks. Claudio Testore
Re: Question about FTP to z/VM
Issue the ASCII command before transferring it on the sending side. That should force LF-CR/LF conversion (the FTP spec requires CR/LF on the wire). Most Unix FTP clients default to binary mode these days, which is why you get the raw LF. Question: are these fixed length records? If so, try SITE FIX nnn, and the FTP client will segment the records into nice neat fixed-length chunks properly.
Re: Bacula + tape size
We use Bacula and VTS to make backups I wonder if we can configure in z/V M tape volumes greater than 830 MB. It look like that what I can define max 830 MB. There's no limit in z/VM; the tape mount code we wrote writes until physical EOV; you may need to define a new media type in Bacula to represent the larger media.
Re: Bacula + tape size
When you say VTS, I assume you mean an IBM 3494 ... no, you can't. The virtual volumes are emulated to the extent that includes simulating the capacity of a physical 3490E cartridge. Ah, right. VTS, duh. Suggestion: define a disk pool with large volumes (5-10G) and do migration from disk to tape. Much more efficient. If you have FCP disk available, this is a great use for lower cost FCP disk pools.
Re: Starter SLES10
After doing the link novstart 191 f191 MR, the cmsddr restore appeared to work. This is what I got as a response: [snip] Yes, that's what you should expect.
Re: dial to 2nd level vm?
From OPERATOR or other priv id on the 2nd level system, do ENABLE ALL However, we are not able to DIAL to it. I put in some SPECIAL statements in definition within USER DIRECT which were of no help. I thinking there might be something on 2nd level VM system that needs to be changed or turned on. Suggestions?
Re: Partially Successful: OpenVMS on System z
I didn't know that OpenVMS could run on Intel Itaniums, but it sounds way cool. Yes. It's a relatively recent development (and about 20 years too late -- we could have been spared Windows entirely if DEC had bought a clue way back when...*sigh*), but it's actually fueling a renaissance in VMS usage. VMS is a nice, tight, secure, easy to manage system, with decent batch and tape processing capabilities, and excellent compilers and compatibility with the IBM extensions for COBOL, a decent CICS replacement, etc, etc. It's a terrific replacement for small VSE and z/OS shops who can't get an affordable z system these days. HP doesn't seem to have such a huge problem with hobbyist use, either. Now if someone would port REXX to VMS instead of DCL... hmm. Now that's a useful thought. What's next, Mac under z/VM? :-) TOPS-20, of course. If I say fiat lux, then I expect the OS to turn the lights on reliably, with a proper Emacs for editing the universe's config files. Galaxy on Z -- it's about time we got a decent batch system for Z. *grin* 8-) Actually, if this emulation approach we're using for the Alpha works out as it appears to be doing, PPC operating systems like AIX and i5OS would be a logical next step. Licensing for Mac OSX Server would probably just be a RPITA to get, even if I could get proper emulation working.
Re: Full volume minidisk FTP
Use the recently posted DDR and FTPPUT pipe stages. *Much* simpler for batchable tasks like that. From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of Gary M. Dennis Sent: Tuesday, September 09, 2008 3:58 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Full volume minidisk FTP Using VMFTP, how is the best way to transmit full volume minidisks (mixed CMS, VSE, and z/OS) to a backup site? --. .- .-. -.-- Gary Dennis
Re: Alternatives to 3490E
IBM 3590 costs and cartridge replacements (retired system?) VSSI VTAPE disk costs and lack of easy D/R. IBM VTL system Virtualized Tape systems EMC DISK LIBRARY Replacement for Carts D/R questions. Other possible solutions. Flex-CUB (if you can get a 3490-F10, you could keep existing carts or use it to copy incoming volumes to disk volumes for easier use) BusTech box (see above) Both the above would allow trivial mirroring to other sites, and the BusTech box provides multisystem attach and FICON, I think. Both use less expensive open-systems disk than VTAPE (although you could use FCP disk with VTAPE, I suppose). At some point you will need a 3590, as IBM is likely to stop supporting 3490 distribution tapes soon (3480 is already gone, I think).
Partially Successful: OpenVMS on System z
Latest interesting step: I've gotten a partial boot of OpenVMS in a virtual machine on System z. Some instruction emulation still needs work (the Alpha POP is a little unclear in a few areas), but we have system services initializing and hardware detection is fully functional. *grin* Anybody interested? -- db David Boyes Sine Nomine Associates
FW: Partially Successful: OpenVMS on System z
I realized I just responded to Scott. For the rest of you: From: David Boyes Sent: Monday, September 08, 2008 5:06 PM To: 'Scott Rohling' Subject: FW: Partially Successful: OpenVMS on System z Uh Yeah!! More info please! :-) I've enhanced the ES40 Alpha emulator to run more smoothly on Z (added some instruction optimization for the z9 instruction set, and some lookahead code to try to do a better job of predictive branching on Z (the Alpha does this strangely)). It seems to deal pretty well with OpenVMS 7.x Alpha install CDs up to the point of going multiuser - I think I have something wrong in the stack handler, but I'm making progress. What interests you about it? I'd like to be able to tell the IBM PTBs that someone is really interested. 8-) -- db
Re: MORE THAN HALF THE MAINFRAME MIPS IBM SELLS ARE LINUX?
*sigh* No, it certainly wasn't a SNA project, although a number of us did participate while part of other lives. As to the more than half number, what I said was that it was likely that the majority of the *increase* in MIPS was specialty engines of *some* type. I'm not sure the reporter got that zIIPs and zAAPs are also specialty engines, but cannot run Linux. Apparently not. It's nice to have our role in this mess featured occasionally, but not as claiming credit for a collaborative effort. Reporters. Sigh. -- db
Re: MORE THAN HALF THE MAINFRAME MIPS IBM SELLS ARE LINUX?
I've added a corrective comment to the posting as well. From: The IBM z/VM Operating System on behalf of Alan Ackerman Sent: Fri 9/5/2008 12:18 AM To: IBMVM@LISTSERV.UARK.EDU Subject: MORE THAN HALF THE MAINFRAME MIPS IBM SELLS ARE LINUX? Any truth to this? Someone at work forwarded it to me. The web site works . David Boyes of Sine Nomine Associates is quoted.
Re: VMFTP Return Code -5
Is there a cat belonging to somebody named Schrodinger somewhere in the neighborhood of your system? Maybe not, but he's certainly got a local demon named Maxwell. 8-) -- db
Re: VTAPE from Virtual Software Systems
VTAPE is a super product, but also consider a Flex-CUB (essentially a channel attached PC that can present itself as tape drives or slower disk volumes. You can attach a variety of lower-cost disk storage to the Flex-CUB, and mirror that, or use a variety of techniques to get that to higher-density offline storage. The Flex-CUB boxen also do emulation of 3xxx tape drives on selected SCSI drives, which is mighty handy in a lot of cases.
Re: SPOOL/PAGE and cylinder zero
Historically, there was a problem, and that was a required restriction in those days. Modern CP does handle it correctly. You lose a trivial amount of disk space by skipping cyl 0 consistently (180K or so per volume). Even so, IMHO, it is still good practice to avoid cyl 0 for all packs as a matter of consistency. One less thing to document or to forget about and have something happen later (like DIRMAINT allocating a minidisk in an area where you *really* don't want it...sigh). From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of [EMAIL PROTECTED] Sent: Wednesday, August 27, 2008 10:35 PM To: IBMVM@LISTSERV.UARK.EDU Subject: SPOOL/PAGE and cylinder zero I have always thought that CP handles properly SPOOL and PAGE areas starting on cylinder zero. It would be a shot on its own foot if CP would destroy the VOLID and the allocation map. However, someone is stating that SPOOL and PAGE MUST NOT be allocated on cylinder zero, I wonder if someone can positively answer this question and point to any documentation that may exist about this. Francisco A. S. Grossi
Re: IUCV - What's wrong with this picture?
1. Does IUCV infrastructure overhead specifically associated with number of connections become prohibitive at some well known point? There is a limit to the maximum number of connections (a parm on the IUCV statement in the directory entry; I think the max value for that parm is 64K, but check your docs -- I don't have manuals with me at the moment). 2. Has anyone had experience with an application having a high IUCV connection count like this? If so, what was that experience? Probably the best example I know of for an application that uses a lot of IUCV connections is the VM TCPIP stack (yes, I know there is a lot of VMCF too, but it also uses IUCV.). It seems to be fairly stable and scales well. At a past employer, we had a few applications that regularly had several thousand IUCV connections open and active simultaneously with no connection state related problems. The key problem with performance tuning for those applications was supplying sufficient virtual storage to provide suitable buffers for the connections, and the need to dispatch each endpoint every time you needed to send and respond to a message. Those two things were a lot bigger hit than the IUCV connection and session management processing. Something also to think about is the additional hit of dispatching and routing messages between VM images if you are planning on permitting distributed IUCV (a good idea). That's going to mean some additional work needed from CP to do ISFC, or asking CP to dispatch TSAF or AVS or IPGATE to get the message frame over to the other system. That can be significant.
Re: Where Do I Go From Here?
Who among those of us who do not speak Latin Bloody barbarians. They don't speak Greek either! 8-) So before you go on about the bloody Romans, don't you forget... you're one of 'em! -- Not the Messiah, He's a Very Naughty Boy!
Re: Where Do I Go From Here?
As I know nothing of the guts of Jovial I can't say if this would affect it. If it does then will it run on a really old VM system on Hercules. Critical question is what apps are you running and where do they get there data from and send it too... Dave Wade G4UGM Illegitimi Non Carborundum It would have to be a REALLY old version of VM, eg VM/370. You can't legally license and run any modern VM on Hercules for production work. Richard's suggession of CMSACCOM and friends is the only way you could legally do this.
Re: Isolating one Linux guest in the DMZ
ur z/VM environment currently sits behind a firewall. We would like to allow one Linux guest to act as an Internet server. We do not however want to expose the other Linux guests or the z/VM environment itself to the outside world. We are using VSWITCH. Good. That makes a few things possible. Is this possible? What options are there? Do we require a separate OSA for the Linux guest in the DMZ? Not necessarily. The easiest thing to do is to have your network guys engineer a new VLAN, and move the guest you want to expose onto that VLAN. That requires no physical hardware changes, and the networking guys can do all the routing that needs to happen outside the box. As long as there are no other network connections to the exposed guest, you can't get from there to any of the other guests. The assumption is that you're using VLAN-aware VSWITCHes, and that your networking guys understand how to make the magic connections to create and propagate a VLAN in your network. These days, that's a fairly safe bet. If you have spare money or excessively paranoid security weenies, you could get another OSA and dedicate it to that one guest. It's a waste of money, but technically valid. As I'm not a comm's guy, please keep it simple. Thanks in advance. You'll want to work closely with the network guys. Show them the 1st paragraph above, and they'll get it. -- db
Re: DDR'ing 3390 DASD To Remote Location
¿SPXTAPE will support dump to disk in the future? There's a requirement for supporting PIPE connections like the DDR mods that were just made available. Status is unknown at the moment other than IBM has seen it and seems to understand the desire for the function, but hasn't made any public statements of when or if they will do this. It would be nice, though, wouldn't it? (BTW, Jiri - the DDR work looks good for us. Beer is definitely on us next time.) -- db
Re: Where Do I Go From Here?
On 8/21/08 5:49 PM, Karl Severson [EMAIL PROTECTED] wrote: The compiler was originally written to run on MVS but it was tweaked for us t o run on VM in 370 mode back in the mid 1980¹s. If you truly still need 370 mode, you're going to be hard pressed to find a system that will still run true 370 mode. Most (if not all) the modern systems no longer have true 370 mode microcode. My organization is considering re-hosting Jovial on some sort of Windows platform Considering the amount of work on not just the applications but all the surrounding environment and data management, this would be a REALLY bad idea. You might be better off talking to HP about a VMS-based solution (they offered, and AFAIK, still offer a fairly decent Jovial compiler, and at least VMS understands labeled tapes and batch natively. It'd probably be less work to update the compiler you have for a modern CMS, or port it to Linux (either on Z or on some Intelish thing), and cost the Army less to run it as well. to keep this an IBM operation if at all possible. We also heavily use Unix, Linux and Windows on other platforms. To offset some of the cost of a new system (z9 at a minimum) maybe it could do double or triple duty replacing some of those other platforms. Good thought. A combination of Linux, CMS and OpenSolaris virtual machines would be a very compelling case. Of my options, which would be the most efficient? The latest zVM with a VM/ESA guest? The latest zOS with a VM/ESA guest? Some other combination? Or, heaven forbid, go with re-hosting to Windows? If you have source for the Jovial compiler and runtime, go to the latest z/VM and the latest CMS and ditch the VM/ESA system, or at least limit its use to a migration system. Z/OS is very unlikely to be cost-effective (and can't run VM/ESA as a guest anyway). Windows will be the most expensive option if you incorporate all the porting costs and the surrounding environment necessary to make this work. A Linux port would be very cost effective, especially if you can pick up some additional virtual workload in the process (and IFLs would make the new machine LOADS cheaper). Check out the OpenVMS solution as well. Since you're being systematic about it, the HP Integrity boxes are a lot of bang for the buck for VMS. -- db
Sol-390 mailing list created to discuss OpenSolaris for System z
To avoid dragging these two lists off-topic constantly, by grace of the fine people at Marist College, a mailing list called sol-390 has been created to discuss the forthcoming availability of the OpenSolaris port for System z. We're very, very close, and if you're interested in the topic, I'd strongly encourage you to subscribe. The process for signing up for the mailing list is the same as signing up for the Linux-390 mailing list; send mail to [EMAIL PROTECTED], and put the following in the BODY of the message (where your letter text normally goes): SUBSCRIBE SOL-390 firstname lastname You'll receive a confirmation request email. Do what it says to be automatically added to the mailing list. Thanks for your patience - the response at SHARE to Neale's session was really encouraging. We're trying to resolve the last legal issues with Sun and we're looking forward to getting this thing into people's hands. In the interim, everybody order and install VM64466 for your 5.3 z/VM systems. That's a necessary enabler to be able to use OpenSolaris. It'll help get the Powers That Be Holding Up The Party the message that people want this to happen. -- db
Re: Sol-390 mailing list created to discuss OpenSolaris for System z
Dave Jones say: In the interim, everybody order and install VM64466 for your 5.3 z/VM systems. The PTF to order that goes with this APAR is: UM32414 And note that it is for z/VM 5.3 only. Thanks, Dave. I could remember the APAR #, but not the PTF number. Too bloody many lawyers. First against the wall when the revolution comes. -- db
Re: DDR'ing 3390 DASD To Remote Location
A while ago there was a thread here about the ability to DDR DASD to remote location. Well there is an answer! I modified DDR so it can communicate with CMS PIPES (DDR can now be a pipe stage). The modules can be downloaded from here: http://www.vm.ibm.com/download/packages/descript.cgi?DRPC 'Bout darn time. First of all, THANK YOU. That's been needed for AGES. Have you talked to the product owner? There is an open requirement against DDR that you could be a hero by helping them close it. The FTPPUT and FTPGET PIPE stages were also included and documented at the above address. Cool. Notice that this project is still a work in progress, therefore feedback is welcomed! Q: does this version also include the CMSDDR mods? So, when do you start on SPXTAPE? 8-) -- db
Re: DDR'ing 3390 DASD To Remote Location
Yep, George has been working on this project with that exact requirement in mind. Fantastic. That's exactly what I wanted to hear. Q: does this version also include the CMSDDR mods? No, this is an otherwise unmodified DDR module. Yeah, about 3 seconds after hitting send I realized that with PIPE support CMSDDR becomes as simple as DDR FOO BAR A | OUTFILE DDR B, so I didn't need CMSDDR any longer anyway. Cool. Thanks for making it happen. Beer's on us next time I'm in Endicott. -- db
Re: DDR'ing 3390 DASD To Remote Location
LOL. George left one small thing off of his signature: z/VM I/O Development IBM Endicott Alan Altmark z/VM Development IBM Endicott Well, far be it from me that I suggest that VM Development begin to talk to themselves. You lot 're odd enough to begin with...8-) -- db
Re: Linux Commands
for i in $(find /usr/share/man/man*) -name \*gz; do gzip -dc $i | nroff | lpr done Note that this also doesn't include the GNU binutils documentation or any of the stuff that is in 'info' format. If you want PDF, substitute 'ps2pdf -o $i' for 'lpr'. No, don't actually do this. You will be sorry. I concur, especially if you have emacs installed. That one item is 892 pages. -- db
Re: Portable z/VM help?
Maybe IBM could create a Windows compiled Help file for those using a Windows desktop and INFO or MAN files for those using a Linux desktop. There is a pdf2docbook utility from Cambridge Systems. It produces ugly, ugly output (like most Docbook related things), but it's one way to handle PDFs. Pdfindex handles pretty much everything else I need it to do.
Re: Linux Commands
That's going to be complicated because every program you install becomes a new command, and some commands could be books all to themselves. There's also the complication that there are several variations to how documentation for Linux commands is prepared and maintained. All packages are *supposed* to include man pages, but that can be a bit spotty for some of the commands maintained by smaller groups or individuals. The man command will display summaries for each command and details. Some commands use the info command, and some supply HTML pages (which I personally detest). Example: 'man ls' will display a manual page for the ls command. If you're not sure of the exact command, try 'man -k' and a keyword, eg 'man -k mail' will get you all the commands that contain the keyword 'mail' somehow. The commands that use 'info' (usually things originating in the GNU project like GCC) work with similar syntax, but they bring up a full screen browser to navigate the documentation. Info tends to be used for more complex applications, like emacs. The emacs documentation is a full-scale book of it's own. Your best bet is a good Unix book like the O'reilly sysadmin guides. You can usually get copies of them at better tech bookstores (alas for Computer Literacysigh... RIP). - db On 8/14/08 5:39 PM, Alyce Austin [EMAIL PROTECTED] wrote: Does a manual exists that has all the SuSE Linux commands listed running on the Z series with really good examples? Thanks for your support, Alyce
Re: VM userid AUDITOR
It's part of CUF. It monitors virtual machines and restarts them if necessary. See http://www.vm.ibm.com/related/CUF/. -Original Message- From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of Gentry, Stephen Sent: Wednesday, August 13, 2008 1:58 PM To: IBMVM@LISTSERV.UARK.EDU Subject: VM userid AUDITOR Can someone tell me what the AUDITOR userid is used for on VM? Thanks, Steve
OT: Alan has a pony.
I would just like to point out that Alan Altmark's long-standing wish for a pony has been satisfied. A brown and white pony has been delivered, and he has no need for further ponies. 8-) -- db
Re: OSA Adapter TCP/IP stack association limit?
Does this approach get around the stack limit some way? If we try to put 800 machines on a VLAN, will this blow the limit referenced below? The layer 2 TYPE ETHERNET VSWITCH gets the OSA limits out of the line of fire entirely -- the OSA(s) (note multiple) servicing the VSWITCH just forward frames, acting like a bridge to the VLAN. You can put lots more than 800 guests on a VLAN. I know of people with well over 2000 guests of various types on a single VLAN sharing a VSWITCH. Downside is that unless you have a z10, I don't think you can have a OSA card in layer 3 and layer 2 mode simultaneously, so you can't share a card with a LPAR that requires layer 3 mode. Probably not a big deal for you, as you're likely to push the limits of a card and not want to share it anyway, but YMMV. The limit applies only to device triples directly attached to individual guests, which is a not so hot idea under VM anyway given that VSWITCH exists. It'll also simplify the ability to move virtual machines if IBM ever gets us such a thing -- not tying the adapters to actual physical devices is a Good Thing.
Re: CMS Files to XMIT compatible datastream
Hmm. If the package passes through RSCS, since RSCS doesn't support NETDATA with multiple datasets, won't the package get exploded on transmission? Does anyone have a procedure for combining multiple CMS files into a sing le XMIT data file? /Tom Kern /301-903-2211
Re: OSA Adapter TCP/IP stack association limit?
The z9, with updated OSA microcode, can also run the OSA port in shared layer 2 and layer 3 mode.. Nifty. Thanks -- one of these days I'll get my brain upgraded for more short term storage. 8-) --db
Re: CMS Files to XMIT compatible datastream
TAPPDS and TAPEMAC can reassemble unloaded PDSes on CMS. In looking around, also look for INMR123 REXX on your S disk. That seems to do something like what you might want. I suspect that the 'some data' is an IEBCOPY unloaded PDS. Now I have an assembler program that reads an IEBCOPY unloaded PDS but no program to wr ite one. I will have to break open the IEBCOPY program to see how a datastrea m is constructed. Maybe I can get a pipeline to create a compatible datastr eam.
Re: Portable z/VM help?
But, searching a PDF for a keyword is far less good than the search method in bookmanager read, both in performance and in the presentation of the search result. (Even though FoxitReader searches faster than Acrobat reader, both need to scan the whole PDF file what takes time). Only if the PDF is not indexed. If you use the full Acrobat package to generate the PDFs, there is a pdfindex tool that can be used to generate a index of a collection of PDFs. If you use that regularly, the search performance is comparable between the two.
Re: z/VM 5.4 Workloads
Key word: selected. You've been able to do this for a while -- .Net stuff with Mono, ASP stuff with various toolkits, etc. Nothing new here. *Moving selected Linux, Windows(r), and UNIX(r) workloads to a single System z server
Re: programmer biographies (was a long time ago: Immediate instructions )
Alan Altmark [EMAIL PROTECTED] wrote: -snip- I'm sure *someone* will speak up and say, Well *I* worked on a Babbage's first difference engine. Knew the man well. He was a geek's geek. Whoever speaks up, my hat's off to you. You win! It's more likely someone will have claimed to helped build Stonehenge, or aligned the pyramids, or Well, this light thing worked out OK in the labs, so we kicked it upstairs to the Big Guy for product announcement. The customer tests with the Let There Be Light campaign seemed positive, so we told him to run with it... 8-)
Re: R.I.P -- BookMaster.....
We still use BookMaster 1.4 here to build documents we distribute to clients and such. I think it's about the best way ever for doing computer related documentation of most kinds. Me too. Doc tools written by people who actually have to *use* them. Best system I've seen. Wonder if we could ask the current product owners to put these tools up on the VM download web page, for use as is, where is? I would rely like to be able to use ProcessMaster here as well. Second the motion. Anyone know who the current product owner is? It'd be a good place to start begging.
Re: SMAPI Java library?
There are already Java libraries to front-end Sun RPC, so that would be one way to go about it. You'd have to feed the RPC stubs to rpcgen to get a proper mapping for your system, but that's not unusual behavior for Sun RPC app interfaces. From: The IBM z/VM Operating System on behalf of Jack Woehr Sent: Thu 7/31/2008 5:05 PM To: IBMVM@LISTSERV.UARK.EDU Subject: SMAPI Java library? Has IBM or anyone wrappered z/VM SMAPI http://publib.boulder.ibm.com/infocenter/zvm/v5r3/topic/com.ibm.zvm.v53.dmse6/hcsl8b20.htm in Java classes akin to how the host servers on i/OS got wrappered by JTOpen? I have not found anything, in particular, no open source project.
Re: Immediate instructions (was nonames)
My fave is LOAD HALFWORD RELATIVE LONG. Say it a few times and it's practically a poem! Chant it, and it'd make a decent mantra. 8-)
Re: Immediate instructions (was nonames)
Yes, but they still do not have DWIM. The Z doesn't, but my Xerox Dandelion does.
Re: Curiosity question: 370 Accommodation Facility Definition
I don't know about running a guest operating system designed for a 370 machine like MVS or DOS/VS. I think you would need SET MACHINE 370. Very likely. Also keep in mind that most of the 370 mode support in the processors themselves is gone, so SET MACHINE 370 might not work at all. Hercules in a Linux (or OpenSolaris) guest would be the way to go if you really need MVS 3.8 on a z10...8-)
Re: CSE and VMSERVx
I think it is reasonable to use VMSYS and VMSYSU in the same manner that IBM uses them in the 'default' RMSMASTER configuration. For applications that are LOCAL to this system, I put control type information (config files) i n VMSYS and use VMSYSU for work and logfiles. At disaster recovery, I do no t mind starting with an empty VMSYSU. Given that IBM discourages use of the VMSYSx: pools for real user data, would it make sense to propose that those local filepools be configured in a way that they reflected the node name they were running on, eg. if running on node FOO, then the default install would create FOOS: (for the IBM uses like RMSMASTER if that code can't be fixed to not require VMSYS:) and (optionally) FOOR:? Using a generic name like VMSYS: seems to be both confusing and not really all that optimal. If one wanted to provide a default pool for user data in the starter system, how about calling it node:, eg FOO: from our above example. It shouldn't be too hard to have the SFS code determine the current node name and publish that as the filepool name and/or APPC resource name.
Re: PIPELINES and Deblocking
I don't know why Unix used NL (==LF) instead. Perhaps a different har dware device? One reason (probably not the only one, but one): Bare CR on the model of DECwriter commonly shipped as PDP-11 console terminals tended to foul the ribbon, especially if you issued it from 40 character positions into the line. LF pushed the ribbon down into the holder (side effect: making it harder to foul the ribbon), and there was a switch on the DECwriter that would auto-add a CR to a bare LF (not the other way round), which saved a whole character frame on a 110-baud console terminal. BSD 2.x used LFs to try to optimize for that. Dunno if V6 or V7 inherited that weirdness from the PDP6 or not. The PDP6 at SAIL didn't have that problem (only a few DECwriters left by the time I got access to it, and NOTS didn't care). All the -11 DECwriters would jam if you ran 300 baud on the console.
Re: SUSE Linux Enterprise Server 10 SP2 Starter System for IBM System z
Samples are just that, samples, not Holy Writ from ${DIETY}. You can, of course, choose to do anything you like. Well, as the author, I rather like the idea of saying let us turn to chapter 2 of the Gospel of Installation... 8-) YMMV. The idea with the 15x separation was to have a small /boot separate from the bulk of the other parts, and encourage the use of a separate /home. The setup in the sample is the base we use for most of our installs, and it's proven to be fairly flexible for different purposes. What we're trying is to establish a set of basic conventions on how things are done that start from field-tested practice. I don't want to prevent you from doing whatever you want, but the default should be usable for reasonably large ranges of useful. Whenever possible, I prefer to have those two volumes only be used for the operating system itself. Any add-on products, such as WebSphere, etc., or applications, would be installed in separate file systems created from a different LVM volume group, using different physical volumes. This is good practice in any case. The default setup works well with this philosophy, and it's common good practice in the Unix world. (Mother's First Law in action...)
Re: Nice idea in blog: Should we toss x86 architecture
How can the z handle 3000 copies of Windows all running a graphic user interface (cpu-intensive) ? Probably the Windows Server 2008 configuration, which can have the GUI disabled. Gee. Shades of OS/2 Warp. 8-) -- db
Re: SUSE Linux Enterprise Server 10 SP2 Starter System for IBM System z
assume that the document SUSE Linux Enterprise Server Starter System Z' is an update to the SHARE handout. Is that the case? Nope. It's a completely separate document.
Re: SUSE Linux Enterprise Server 10 SP2 Starter System for IBM System z
We have a one year contract with Novell until May 2009. They gave us a unique activation code. They told me to download the CD1-CD2 disks from their site. Can I use the SUSE Linux Enterprise Server Starter System for System z to install it? I'm confused... Maybe this will help: The starter system is a prebuilt installation server that you use to create other virtual machines - it contains the disk1 and disk2 images plus a copy of SLES configured to act as a NFS/HTTP/SMB/FTP server to make that data available. It's the functional equivalent of downloading the CD media, setting up a server to provide access to the media, and configuring all the software you need to provide access to the media -- condensed into the single step of downloading the starter system and installing it. You use the installation server as the source to create new virtual machines. Machines that are created from the installation server get the activation codes, because once they're installed, they interact directly with the Novell servers to get updates, etc (unless you set up a local service mirror), and thus need to identify themselves as valid recipients. So, to answer your question -- yes, use the starter system to install a new guest, and when prompted, give it the activation code you got from Novell, and the end product is exactly the same as if you had downloaded the media, found someone to serve it up to you over the network, transferred the boot files, made a boot image, booted SLES from reader or tape, and answered the prompts. You just save a bunch of steps, you don't have to convince anyone to stick a CD or DVD in their machine for two or three hours, and get your virtual machines installed at memory-to-memory speeds instead of just Ethernet speed. All you do is create the new VM userid, LINK NOVSTART 19F 19F RR, ACCESS 19F R, and type SLES. From that point on, you're in the normal SLES installer.
Re: SUSE Linux Enterprise Server 10 SP2 Starter System for IBM System z
On 7/18/08 4:33 PM, Quay, Jonathan (IHG) [EMAIL PROTECTED] wrote: Ok, I have a license and I'm about to download and install SLES10SP2. Can I somehow, and I haven't completely read through the Starter Set documentation, download the SLES-10-SP2-DVD-s390x-GM-DVD1 and 2 .iso images and serve them using the NOVSTART virtual machine? The SP2 starter system contains the entire SP2 DVD image already set up for this purpose. If you set up a SP1 starter system, just use a different userid for the NOVSTART system -- there's nothing in there that cares about the userid it's installed in -- and use that as your new install server. That way you don't mess up anything that already works, and you can easily switch back and forth for production use w/o breaking things.
Re: PROP not running on OPERATOR after REIPL
At his point, CP, bloodlust sated, settles down to watch over his harem of virtual machines, secure in the knowledge The Operator will maintain order in the universe. (Sorry - it's VM Week on the Discovery Channel.) Alan Altmark z/VM Development IBM Endicott Lions and tigers and bears... oh, my!
Re: Data Security Erase utility for z/VM tapes
It sure would be nice if there were a requirement against z/VM to provide such a function. Everyone keeps asking everyone else, but no one ever asks *us*. (sniff) Alan Altmark z/VM Development IBM Endicott Ask and ye shall receive. WAVV requirement WRIBDB01 for the next meeting. -- db
Re: CMS REXX eMAIL
Alan, with your reply of I'm not Dave, but Daves not here man Someone name that album Cheech and Chong, 1971. Quality entertainment, that. But, let's not go there, man... -- db
Re: downloading CMS files
Assuming they don't have an answer for you, tell them you want to open a requirement. Submitted as WAVV requirement WRIBDB02.
Re: SUSE Linux Enterprise Server 10 SP2 Starter System for IBM System z
Not that anybody is likely to care, but this is a facility to install z/Linux under z/VM using only CMS. It is a nice way to get a z/Linux system up quickly. It also makes a very nice upgrade server if you're moving from SLES 9 or previous SLES 10 releases. 8-)