1 INFO-VAX	Tue, 01 Jul 2003	Volume 2003 : Issue 359       Contents:4 Re: $brkthruw from a symbiont returns SS$_DEVOFFLINE& Re: CNXMGRERR with V7.3-1 DSSI cluster: Re: Compaq Solutions Aliance - "Does" it still existing  ?: Re: Compaq Solutions Aliance - "Does" it still existing  ?6 Re: Compaq Solutions Aliance - Do it still existing  ? Re: cxx performance  Re: cxx performance 3 Re: Dallas Security Seminar presentations available 3 Re: Dallas Security Seminar presentations available 3 Re: Dallas Security Seminar presentations available 1 Re: DS25 dual power supplies - are they redundant % Re: DSSI disks and allocation classes ( Re: HP Powers More TOP500 Supercomputers( Re: HP Powers More TOP500 Supercomputers Re: HP to drop hpux?I Re: HP Webcast this morning on Next-Generation Intel Itanium 2 processors I Re: HP Webcast this morning on Next-Generation Intel Itanium 2 processors I Re: HP Webcast this morning on Next-Generation Intel Itanium 2 processors I Re: HP Webcast this morning on Next-Generation Intel Itanium 2 processors I Re: HP Webcast this morning on Next-Generation Intel Itanium 2 processors I Re: HP Webcast this morning on Next-Generation Intel Itanium 2 processors I Re: HP Webcast this morning on Next-Generation Intel Itanium 2 processors I Re: HP Webcast this morning on Next-Generation Intel Itanium 2 processors I Re: HP Webcast this morning on Next-Generation Intel Itanium 2 processors 8 ISV enthusiasm grows for OpenVMS on HP Integrity servers New OpenVMS-to-Itanium FAQ Re: New OpenVMS-to-Itanium FAQJ Re: OpenVMS Pearl - OpenVMS V8.0, first release on Itanium, ships today!!!P Re: OpenVMS Pearl - OpenVMS V8.0, first release on Itanium, ships today!!! today) OpenVMS Success Story: Aurora Health Care  Re: Rethinking  V.M.S  Re: Running VMS off CD Selling: Microvax 4000-108H Re: Subject: Reading long files in BASIC that appears to be "recordless" Re: system disk on ODS-5F Tri-architecture cluster demonstrated at DECUS Ottawa Technical UpdateJ Re: Tri-architecture cluster demonstrated at DECUS Ottawa Technical UpdateJ Re: Tri-architecture cluster demonstrated at DECUS Ottawa Technical Update+ Re: Virtual Memory? (was: Rethinking V.M.S)  Re: VMS backup to NFS shares Re: VMS backup to NFS shares
 vms-pstree.el  XFC, IO Size and Backup   F ----------------------------------------------------------------------  # Date: Mon, 30 Jun 2003 23:45:28 GMT # From: hoff@hp.nospam (Hoff Hoffman) = Subject: Re: $brkthruw from a symbiont returns SS$_DEVOFFLINE 1 Message-ID: <sY3Ma.3629$hh.2299@news.cpqcorp.net>   d In article <bdc65a53.0306230757.21da7390@posting.google.com>, gspamtackett@yahoo.com (Galen) writes:  ?   There's not enough source code included here to be certain...   F :If executed from a symbiont, though, the I/O status block gets an I/O :status of SS$_DEVOFFLINE.  E   There might be a terminal involved somewhere that is XOFF'd or that E   has broadcasts disabled or such, or the username is not uppercased.   D :Can someone tell me what's happening and how I might get around it?  D   The broadcast didn't get everywhere, and you can check the IOSB toC   see how many places it did get to and how many it did not get to.   F   Visit the AskQ site (see the FAQ) for the following support article:  9     Example-C SYS$BRKTHRUW Using BRK$C_USERNAME Send Type +     STARS Unique Id: KS2-VMS000000-00150-01   N  ---------------------------- #include <rtfaq.h> -----------------------------K     For additional, please see the OpenVMS FAQ -- www.hp.com/go/openvms/faq N  --------------------------- pure personal opinion ---------------------------E         Hoff (Stephen) Hoffman   OpenVMS Engineering   hoff[at]hp.com    ------------------------------    Date: 30 Jun 2003 11:53:16 -07001 From: keithparris_NOSPAM@yahoo.com (Keith Parris) / Subject: Re: CNXMGRERR with V7.3-1 DSSI cluster = Message-ID: <cf15391e.0306301053.1388acab@posting.google.com>   x "Martin Vorlaender" <martin.vorlaender@pdv-systeme.de> wrote in message news:<bdp9nb$ve9qu$1@ID-56200.news.dfncis.de>...C > Environment is a two-node cluster of 2-CPU AlphaServer 4000s both G > running a fully patched VMS 7.3-1. Each node has its own system disk. F > All disks are connected via KFPSAs to two HSDs (including the quorum? > disk). Ethernet is all DE602-AA with the first port providing E > connectivity, and the second one connecting the two machines with a D > crossover cable (as a means of failover cluster interconnect). All" > ports are set to auto-negotiate. > E > As long as both machines boot at the same time, all is well. When I B > boot either machine alone, it comes up okay. But: when the otherD > machine boots, the second machine bugchecks as soon as the clusterE > is to be formed (most times with CNXMGRERR in SYS$CLUSTER, but I've D > seen INVEXCEPTN in SYS$SCS as well). What bothers me is that it isD > reliably only the second machine that crashes, i.e. boot A - okay;> > boot B - B crashes. Boot B - okay; boot A - B crashes again.  E You might learn more from having someone analyze the crash dumps, but D in general, a CNXMGRERR -- Connection Manager-detected error (or theD similar LCKMGRERR -- Lock Manager-detected error) bugcheck indicates@ that the Connection Manager (or Lock Manager, respectively) codeF detected bad data in an incoming SCS packet as it tried to decode it. E This might be caused by data corruption of the packet in-flight (such @ that the ECC didn't catch it, or else the packet would have been@ dropped), problems in memory on either end, code stomping on the( packet as it resides in memory, or such.  : I'd use SHOW CLUSTER/CONTINUOUS to identify which hardware= interconnect the VMS$VAXcluster SYSAP connection is over, and D scrutinize that interconnect.  If it helps during debugging, you canD use SCACP to shift cluster traffic to another interconnect while you do troubleshooting.    ------------------------------  # Date: Mon, 30 Jun 2003 18:34:07 GMT 9 From: "Fred Kleinsorge" <my-last-name@stardotzko.dec.com> C Subject: Re: Compaq Solutions Aliance - "Does" it still existing  ? 2 Message-ID: <zo%La.3587$LL7.2152@news.cpqcorp.net>  # "Andrew Harrison SUNUK Consultancy"  > F > I am using a SunRay to post this mail, graphics performance is good,$ > MPEG videos work fine for example.  L I would expect things such as image output to be fine.  The real testers areH things where a small number of things are changed over a large area in aK single operation - like a 500 pixel long XOR line.  You can optimize this a L number of ways with tradeoffs on where you want to burn CPU, throughput, and@ latency.  This isn't sophisticated graphics we're talking about.  K None-the-less, I said "it's probably OK for simple stuff".  Images by their K nature are going to consume bandwidth regardless of where they end up.  The I only ugly part is the fact that a browser (for example) will take a JPEG, L decompress it and display it, only to have the server software recompress itL and transmit it to the SunRay to be uncompressed and displayed.  But as longH as you aren't trying to send 300 x 900x900 x 24 bit images a second... I) doubt most web browser users will notice.    > It isn't a 3D device but then B > the vast majority of desktops don't need high performance 3D and< > X11 to a remote display was never that great at it anyway.  I I *deliberately* did not say anything about 3D.  Since I know of no "thin 7 client" that will do high performance 3D graphics well.    > @ > CPU cost is low, we run our desktop environment in this office> > which is ~50 SunRays using a 6 CPU E4500 (400 Mhz) with 6 GB? > of RAM. Performance is good, shared libs and shared code keep B > the memory footprint down and we have competely silent desktops. >   L I guess it depends on what you are doing.  Shared libs are not the resourcesH I'm talking about.  Each display needs (Z/8)*x*y bytes of memory for theJ display, plus other state like colormaps.  Then each display is driving anH IP connection.  Plus it is doing the X11 operations in software and thenG doing compression to send partial image updates to a LAN.  So it really G depends on just how much you guys really beat on things - and what your  expectations are.  > > > The SunRays do have a separate network, however this isn't a= > problem, we just have a network in the machine room for the : > other servers and the SunRay server and then two desktop8 > LAN's. 1 small one for non SunRay clients and a larger > one for SunRays. >   D It may not be a problem for you, but it might be for some customers.    J > SunRays in this environment have the great advantage of no having to runF > and X-Terminal emulation, that and having a native Tarantella clientA > means that they perform rather well for this kind of mixed mode  > environment. >   L And it's disadvantage is that it is a complete emulation of everything.  TheD SunRay doesn't run X11, it displays the results of X11 being done inE software and the results being send to it.  I've used VT525 emulators K running from my PC over Telnet/LAT/DECnet/Serial, and they are every bit as I fast as a DECterm.  The Multia implemented a "native" X11 server - direct K access to the HW (IIRC) so the performance was no worse than any remote X11 L terminal (network bandwidth limited).  But these days, any random PC is fastK enough to be able to emulate X11 so quickly that the network will always be 
 the limit.  I > > Yeah, most of these are Windows CE or XP based under the covers.  But  hey,I > > most of the worlds desktops *are* Windows.  And most customers have a  needK > > to run Windows applications even if they continue to use VMS or UNIX or  > > LINUX on their servers.  > > F > > A VT320 didn't run an OS, it was just an appliance.  A thin client	 should be L > > viewed as just a desktop appliance like the terminal.  Just make sure it has L > > all the right connections, to get to all your server systems.  The thinkL > > clients above have no real moving parts.  Just turn them on.  They don'tI > > require virtual servers on a host to operate, or a private LAN.  They  are G > > flexible, and OS neutral.   They are *not* the perfect solution for  someone ) > > who really wants a VMS "workstation".  > >  > D > Th SunRay doesn't run an OS, and like your VT320's which had theirE > dedicated Serial connection to the host each SunRay uses a 100BaseT 0 > connection into a generally dedicated network. >   E Bzzzzt.  A SunRay cannot exist without a Sun backend server.   Unplug J everything except a serial port on it - can you turn it on and use it as aI VT550?  Nope - because all the intellegence exists as software on the Sun H server.  It is just the "glass tube" front end for the Sun server in theF backroom.  You *can* take an EVO (for example) and ignore it's networkD capabilities and use it as a "terminal".  You *don't* have to have aI specific *dedicated* system someplace that serves it.  It could be one or 
 many systems.   H This really is a matter of where to you do the distribution.  X11 was anH attempt to do it as low as possible.  SunRay takes that even lower (in aL proprietary way).  You can do the same thing in an "open" way with VNC - andL use any client.  What SunRay *doesn* let you do is to use the right level ofK distribution for the job at hand.  You can't download a Java applet to it - F the Java applet runs on the Sun server - not the SunRay - for example.  F > We never have to upgrade the SunRay's and they leverage something we" > have plenty of which is servers. >   ' Yup.  SunRays exist to consume servers.    ------------------------------  % Date: Mon, 30 Jun 2003 22:13:10 -0500 , From: "Dave Gudewicz" <k9jdk@NOSPAMarrl.net>C Subject: Re: Compaq Solutions Aliance - "Does" it still existing  ? / Message-ID: <vg1v25iud99q09@corp.supernews.com>   L There are some who say that CDE and DECwindows are a bit outdated.  Not much going on in those 2 camps.  E There are some who say that other X based GUI's have more development L activity going on.  Take a look at KDE  and GNONE, 2 other popular GUI's and# compare them w/ CDE and DECwindows.   H Read recently that hp was leaning towards GNOME but reconsidered.  Don't know the details.   H Also read that Red Hat was the way then last week something about SUSE??   Hard to keep up.   Dave...   5 "Michael Rice" <marice@whiteice.com> wrote in message ) news:vfuno3ehp08m6a@corp.supernews.com... , > On 6/29/2003 4:17 PM, Fabio Cardoso wrote:
 > > People > > > > > This is because I think OVMS should have an better graphicA > > interface. You cannot run away from th GUI. Of course command > > > line still important but we dont have a graphical terminal; > > development in the past 10 years. May be the future for F > > OVMS is a Citrix Server like product. Or may be porting some linux6 > > graphic interface. Or may be Amiga Intuition ! :-) > > < > > What will be th next VT terminal ?  The OVMS engineeringB > > should think about the clients too... an OS is not just kernel  > > and communication protocols. > >  > > Regards  > >  > > FC > [...]  > , > What, you don't like DECwindows or CDE? ;) > F >  From a GUI toolkit perspective, there is a small effort underway toD > port the Trolltech Qt toolkit (www.trolltech.com) to OpenVMS.  I'mG > hoping to  contribute to that effort in the near future.  It would be G > fantastic if HP folks could help out - much like they do in providing ) > the port/builds of Mozilla for OpenVMS.  > A > If your not familiar with Qt, it is the best (IMHO) application J > framework available.  It provides for developing cross-platform apps (weE > use it for IRIX, Solaris, Linux, Windows, and soon OS X).  It has a J > great set of classes that provide for GUIs, I/O (files, generic streams,I > network, etc.), collections (arrays, lists, etc.), multi-threading, and 
 > a lot more.  >    ------------------------------  % Date: Mon, 30 Jun 2003 21:40:07 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net> ? Subject: Re: Compaq Solutions Aliance - Do it still existing  ? ' Message-ID: <3F00F487.CD015618@fsi.net>    David Froble wrote:  >  > David J. Dachtera wrote: >  > > Tom Linden wrote:  > > G > >>If you think you can avoid using windows (or mac perhaps) in todays J > >>world you are swimming against a verey strong current.  My windows boxN > >>has at least one window open (using Putty) to each cluster member, one forH > >>for a router, one for the browser, and one for Outlook pop client to > >>TCPIP5.1 smtp. > >> > >>What could easier? > >> > > I > > Believe it or else, there are folks here who find WhineBloze rather a K > > puzzlement, since they are hard-core VMS hackers. Forcing them to learn C > > a new platform distracts them from their bread-and-butter work.  > M > There's nothing wrong with learning multiple systems, and a few things very  > right with doing so. [snip]   G I'd have to do to you what my management does to me: what value does it B add? ...and does that value outweigh the loss of productive (read: billable) time while learning?   --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------    Date: 30 Jun 2003 11:27:02 -07001 From: usenet_vms@lehrerfamily.com (Joshua Lehrer)  Subject: Re: cxx performance= Message-ID: <477e0934.0306301027.6c11e569@posting.google.com>   Y Nic Clews <sendspamhere@[127.0.0.1]> wrote in message news:<bdp5lo$d20$1@lore.csc.com>... H > From your description, I would place a priority on increasing the ACP_H > caches to bring the hit % as noted in AUTOGEN to or above 95% hit. I'd  B I have no idea what that means.  I'm not a VMS engineer.  How do I3 check the value of this parameter and our hit rate?   A > also try some other approaches, DFU has a directory compression C > function, use that before a compilation (this speeds up directory    DFU?  J > lookups), have generous CHANNELCNT and corresponding FILLM (not sure howH > the compilers and linkers for C fare here YMMV). Have generous working > sets.    we do.   > G > You could consider hardware mirrored disks, or even volume shadowing,  > for performance of reads.  > J > I also mentioned BYPASS privilege in a previous message. The reason hereA > is that file protection lookups are skipped when it is granted.  >   F My testing does not indicate this.  Using "set watch file/class=major"F I see no difference in file access between when I have BYPASS and when I don't.  A > If you are pre 7.3 I'd also consider a third party disk cacher.  >    NA, we are post 7.3   C > This list is not exhaustive, there are more things which could be I > relevant, and most of the stuff above targets the I/O side of life, and J > if you're running high on CPU usage, it's not going to buy you much more$ > time unless we're saving overhead. > # > BTW, Any relation to Tom Lehrer ?    Nope, sorry.  
 joshua lehrer  factset research systems NYSE:FDS   ------------------------------  % Date: Mon, 30 Jun 2003 21:33:20 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net>  Subject: Re: cxx performance' Message-ID: <3F00F2F0.618F9703@fsi.net>    Joshua Lehrer wrote: > [ > Nic Clews <sendspamhere@[127.0.0.1]> wrote in message news:<bdp5lo$d20$1@lore.csc.com>... J > > From your description, I would place a priority on increasing the ACP_J > > caches to bring the hit % as noted in AUTOGEN to or above 95% hit. I'd > D > I have no idea what that means.  I'm not a VMS engineer.  How do I5 > check the value of this parameter and our hit rate?   E AUTOGEN tells you this in the AGEN$PARAMS.REPORT it produces when you / run it. See the system manager's documentation.   C > > also try some other approaches, DFU has a directory compression E > > function, use that before a compilation (this speeds up directory  >  > DFU?  B A freeware program from Digital / Compaq( / HP). Does lots of neatB stuff: can sometimes undelete files, can defragment files (but notA freespace), can delete entire directory trees fairly quickly, can  optimize directories, ...   L > > lookups), have generous CHANNELCNT and corresponding FILLM (not sure howJ > > the compilers and linkers for C fare here YMMV). Have generous working	 > > sets.  >  > we do.  G How 'bout page file quota? (It's a user process parameter in AUTHORIZE: < pgflquota. See also the system parameters pql_dpgflquota and pql_mpgflquota.)   --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------    Date: 30 Jun 2003 18:38:19 -00004 From: Doc.Cypher <Use-Author-Address-Header@[127.1]>< Subject: Re: Dallas Security Seminar presentations available5 Message-ID: <20030630183819.9826.qmail@nym.alias.net>   D On 30 Jun 2003, susan_skonetski---hotmail.com (Sue Skonetski) wrote:  G >Last week there was a security seminar in Dallas.  Those presentations ; >are now available on http://vmsone.com/web_sec/index.html.   7 I guess everyone is trying to download this, it's slow.   ; Can you let John know the 2002 presentation links are 404d.   + >From: Wisniewski, John (OpenVMS Services)  $ >Sent: Sunday, June 29, 2003 9:45 PMM >To: cynthia.nichols---usdoj.gov; Kevin.Nixon---cw.com; bill.hancock---cw.com H >Cc: Howard, Dale; Demers, Leo; Novak, Craig R; tshannon3---comcast.net;  H Please remember to munge or remove email addresses to keep people's spam levels down.     Doc. --  K OpenVMS.         Eight out of ten hackers prefer *other* operating systems.    ------------------------------  # Date: Mon, 30 Jun 2003 18:41:50 GMT " From:   VAXman-  @SendSpamHere.ORG< Subject: Re: Dallas Security Seminar presentations available0 Message-ID: <00A2229E.342304C5@SendSpamHere.ORG>  q In article <857e9e41.0306300741.4f905c43@posting.google.com>, susan_skonetski@hotmail.com (Sue Skonetski) writes:  >Dear Folks, > G >Last week there was a security seminar in Dallas.  Those presentations ; >are now available on http://vmsone.com/web_sec/index.html.  >  >Speakers were: @ >Linda C. Groves,  Assistant U.S. Attorney Deputy Criminal Chief >Cybercrime/Terrorism Section ? >Kevin M. Nixon Sr. Director, Security Business Strategy  Chief ) >Security Officer's Staff Cable &Wireless 0 >Leo Demers, HP OpenVMS Security Product Manager: >John Wisniewski, HP Security Services Solutions Architect >  >Warm Regards, >Sue  B I'd very much appreciate if these Mickey$chlock format files wouldB be converted to something that I might be able to view from VMS.  < If PDF is the alternative, so be it.  I'd prefer PostScript.   --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COM              5   "Well my son, life is like a beanstalk, isn't it?"     ------------------------------    Date: 30 Jun 2003 19:05:43 -0000= From: Doc.Cypher <Use-Author-Supplied-Address-Header@[127.1]> < Subject: Re: Dallas Security Seminar presentations available6 Message-ID: <20030630190543.22468.qmail@nym.alias.net>  E On 30 Jun 2003, Doc.Cypher <Use-Author-Address-Header@[127.1]> wrote: E >On 30 Jun 2003, susan_skonetski---hotmail.com (Sue Skonetski) wrote:  > H >>Last week there was a security seminar in Dallas.  Those presentations< >>are now available on http://vmsone.com/web_sec/index.html. > 8 >I guess everyone is trying to download this, it's slow.  N Finally got Leo's presentation.  Nice, can I have a copy of WarFighter please? :)     Doc. --  K OpenVMS.         Eight out of ten hackers prefer *other* operating systems.    ------------------------------  % Date: Mon, 30 Jun 2003 21:38:34 -0400 2 From: rdeininger@mindspring.com (Robert Deininger): Subject: Re: DS25 dual power supplies - are they redundantL Message-ID: <rdeininger-3006032138340001@user-uinj5ar.dialup.mindspring.com>  = In article <58ba0101.0306300513.55035e39@posting.google.com>, 7 andrew.rycroft@intrinsitech.com (Andrew Rycroft) wrote:    >Hi, > A >I have a DS25 with a single processor, and 1GB of memory running C >OpenVMS. By default the systems come with 2 power supplies, and an # >optional third for N+1 redundancy.  > G >Will the dual power supplies I have currently provide redundancy, i.e. G >do I really need two power supplies to drive 1CPU, and 1GB of memory ? G >The DS25 technical summary is very confusing on this, and the response  >I have received is confusing.  I Did you read the DS25 quickspecs?  That summary seems pretty clear to me.   H If you have the older, low-power supplies, you need 2.  The third supplyJ is redundant.  Unless you have more than 8 GB of memory, in which case the third supply is mandatory.  J If you have the newer, high-power supplies, 2 is the minimum and the third; is always redundant, no matter how much memory you install.   = Who gave you the confusing response?  That ought to be fixed.    ------------------------------  # Date: Mon, 30 Jun 2003 23:30:10 GMT # From: hoff@hp.nospam (Hoff Hoffman) . Subject: Re: DSSI disks and allocation classes1 Message-ID: <6K3Ma.3627$hh.3424@news.cpqcorp.net>   w In article <bd416s$j9c$1@online.de>, helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) writes:  : H :Based on this and other responses, I think I can change the DSSI stuff.D :What is still unclear is the relationship of this parameter to the 5 :normal allocation class parameter used in a cluster.   C   The following sections in the OpenVMS FAQ will undoubtedly be of     some interest here.   +     How do I rename a DSSI disk (or tape?)? 5     Explain disk (or tape) allocation class settings? <     How to configure allocation classes and Multi-Path SCSI?0     Tell me about SET HOST/DUP and SET HOST/HSC @     Does volume shadowing require a non-zero allocation classes?A     What is the correct value for EXPECTED_VOTES in a VMScluster?   (   There are likely a few other sections.  B   The just-assigned HP corporate "jumpstation" URL for the OpenVMS	   FAQ is:   %     http://www.hp.com/go/openvms/faq/   .   Yes, this new URL will be in the FAQ update.    N  ---------------------------- #include <rtfaq.h> -----------------------------K     For additional, please see the OpenVMS FAQ -- www.hp.com/go/openvms/faq N  --------------------------- pure personal opinion ---------------------------E         Hoff (Stephen) Hoffman   OpenVMS Engineering   hoff[at]hp.com    ------------------------------    Date: 30 Jun 2003 14:01:01 -07001 From: keithparris_NOSPAM@yahoo.com (Keith Parris) 1 Subject: Re: HP Powers More TOP500 Supercomputers = Message-ID: <cf15391e.0306301301.6ed41fbc@posting.google.com>   v keithparris_NOSPAM@yahoo.com (Keith Parris) wrote in message news:<cf15391e.0306241428.2185fab5@posting.google.com>...C > With 159 entries, representing more than 30 percent of the posted D > sites, HP has more installations on the TOP500 list than any other > technology supplier.   It's also interesting to note:  C HP has held the top spot consecutively for the past three published  lists.  ; Sun plummeted from 88 entries six months ago to only 9 now.   E The representation of industry standard Intel processors (both 32-bit   and 64-bit) grew from 56 to 119.   ------------------------------  + Date: Mon, 30 Jun 2003 17:44:57 -0400 (EDT) + From: Lord Isildur <isildur@andrew.cmu.edu> 1 Subject: Re: HP Powers More TOP500 Supercomputers J Message-ID: <Pine.LNX.4.55L-032.0306301741450.27619@unix44.andrew.cmu.edu>  D just because it is from intel does not make it an industry standard. isildur     ( On Mon, 30 Jun 2003, Keith Parris wrote: > G > The representation of industry standard Intel processors (both 32-bit " > and 64-bit) grew from 56 to 119. >    ------------------------------    Date: 30 Jun 2003 15:14:53 -07001 From: keithparris_NOSPAM@yahoo.com (Keith Parris)  Subject: Re: HP to drop hpux? = Message-ID: <cf15391e.0306301414.6aa99f7d@posting.google.com>    Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> wrote in message news:<bdp1ab$pkd$1@new-usenet.uk.sun.com>... A > You clearly don't want to mention the departure of Bruce Perens   E The single example I gave, that of Jon "Maddog" Hall, was also of one B who has departed, but yes, Bruce Perens contributed to Linux whileB employed by HP, and continues to contribute, last I heard.  MartinA Fink is another good example -- his recent book on "Linux and the D Economics of Open Source" shows excellent understanding of Linux andA its implications.  Somebody at Sun should get a copy and read it.   C > Getting Linux ported onto your platform so that you can sell that B > platform hardly counts as major donation to OpenSource, think of9 > it as the same as IBM getting Linux ported to Z series.   C The work that HP did in leading the development of Linux on Itanium D will benefit all vendors of Itanium servers, not just HP: IBM, Dell,B Gateway, SGI, NEC, Fujitsu Siemens, etc.  If Sun ever wised up and/ dumped SPARC for IA-64, Sun would benefit, too.   < > So what of note have you provided that delivers benefit to > the OpenSource community ????   = Sadly, I haven't provided anything of note to the Open Source  community. :-)  E There are others who could better provide a list of the contributions E to Linux from the HP family, which includes DEC, Tandem, and Compaq.  B I do remember big ones like DEC's gift of a DS20 to Linus Torvalds	 early on.   - > > HP's Linux sales now exceed $2B annually.  >  > So that's a donation ?????  E HP's position as the number-one vendor of Linux-based solutions shows > the high level of support HP has for Linux, the high degree of< importance Linux has to HP, as well as the support the LinuxB marketplace has shown for HP as a vendor of Linux-related servers,  storage, software, and services.  D (And it's quite possible that if things continue to go well, somedayE HP's Linux sales might catch up with its OpenVMS sales revenues.  :-)  )   > > all you have come up with so far is HP is committed to Linux$ > because it sells tin and services.   Plus software, and storage.   F With your viewpoint, IBM doesn't support Linux, either, because theirs is a services-centric approach.   E Gifts of code aren't the only contributions to Linux that matter.  It F also matters how a vendor treats Linux, in its product line and in itsA interactions with customers.  HP actually recommends Linux to its C customers in areas where it is an appropriate solution, rather than 9 reluctantly providing it if we can't talk them out of it.    ------------------------------  # Date: Mon, 30 Jun 2003 18:16:44 GMT 4 From: brad@.gateway.2wire.net (Bradford J. Hamilton)R Subject: Re: HP Webcast this morning on Next-Generation Intel Itanium 2 processors. Message-ID: <g8%La.67341$Ab2.140995@sccrnsc01>  q In article <cf15391e.0306300745.7b697c07@posting.google.com>, keithparris_NOSPAM@yahoo.com (Keith Parris) writes: A >HP solutions based on next generation Intel Itanium 2 processors  > ; >Today, businesses are demanding more -- more agility, more F >accountability, and a better return on their IT investment. Learn howE >HP continues to deliver innovative business and technology solutions C >with the entry of a new era in enterprise and technical computing.  > D >Find out how your company can benefit from world-leading server andE >workstation performance, best-in-class deployment, lasting value and C >investment protection with HP's latest Itanium 2-based solutions.  >  >Keynote presentations by: > B >Peter Blackmore - HP Executive Vice President, Enterprise Systems >GroupE >Scott Stallard V Senior VP & GM. Enterprise Storage & Servers Group A >Mark Hudson V Vice President V Marketing, Enterprise Storage &  >Servers Group > 3 >Follow the link from the main HP website, or go to  >http://www.hp.com/cgi-bin/large/exit.cgi?promo=wwmain_b2/Itaniumwebcast&url=http://www.hpbroadband.com/program.cfm?key=itanium2&?mtxs=home-ent&mtxb=B2&mtxl=L1   K I reviewed the PPP slide presentation, and caught a little of the webcast.  J There were several VMS mentions in the slides, and I notice that two press* releases sent by hp today mention OpenVMS.  * I guess someone in hp got the message.	:-)  F Congratulations to the Engineering group for getting 8.0 out the door!  A _________________________________________________________________ 0 Bradford J. Hamilton			"All opinions are my own"= bMradAhamiPltSon-at-atMtAbi.cPoSm	"Lose the MAPS, and replace  					'-at-' with @"    ------------------------------  # Date: Mon, 30 Jun 2003 18:56:27 GMT 9 From: "Fred Kleinsorge" <my-last-name@stardotzko.dec.com> R Subject: Re: HP Webcast this morning on Next-Generation Intel Itanium 2 processors2 Message-ID: <vJ%La.3590$CI7.2373@news.cpqcorp.net>  J "Bradford J. Hamilton" <brad@.gateway.2wire.net> wrote in message news:g8% > L > I reviewed the PPP slide presentation, and caught a little of the webcast.L > There were several VMS mentions in the slides, and I notice that two press, > releases sent by hp today mention OpenVMS. > , > I guess someone in hp got the message. :-) > H > Congratulations to the Engineering group for getting 8.0 out the door! >   I Yup.  And on that rotten corpse of an architecure IA64 - you know the one I that we just announced systems from 1 CPU to 64 CPU's.  The one with 1318 H SpecInt2000 and 2106 SpecFP2000.  Yup, the one that pretty much everyone2 except Sun (and Apple ;-) are selling systems for.  9 http://www.hp.com/hpinfo/newsroom/press/2003/030630b.html   L http://www.hp.com/products1/evolution/alpha_retaintrust/openvms/v8release.ht ml   ------------------------------  % Date: Mon, 30 Jun 2003 15:16:53 -0400 $ From: "rob kas" <news@paychoice.com>R Subject: Re: HP Webcast this morning on Next-Generation Intel Itanium 2 processors/ Message-ID: <vg13n644svdv64@corp.supernews.com>   K > Yup.  And on that rotten corpse of an architecure IA64 - you know the one K > that we just announced systems from 1 CPU to 64 CPU's.  The one with 1318 J > SpecInt2000 and 2106 SpecFP2000.  Yup, the one that pretty much everyone4 > except Sun (and Apple ;-) are selling systems for. >  >         Sadly no one is buying them   ------------------------------  # Date: Mon, 30 Jun 2003 19:59:34 GMT 9 From: "Fred Kleinsorge" <my-last-name@stardotzko.dec.com> R Subject: Re: HP Webcast this morning on Next-Generation Intel Itanium 2 processors2 Message-ID: <GE0Ma.3599$6V7.2912@news.cpqcorp.net>  / "rob kas" <news@paychoice.com> wrote in message ) news:vg13n644svdv64@corp.supernews.com... I > > Yup.  And on that rotten corpse of an architecure IA64 - you know the  one H > > that we just announced systems from 1 CPU to 64 CPU's.  The one with 1318L > > SpecInt2000 and 2106 SpecFP2000.  Yup, the one that pretty much everyone6 > > except Sun (and Apple ;-) are selling systems for. > >S > >u >r >e  >    Sadly no one is buying them >t  H Hmmm.  Let's see - they were announced today.  So why don't we give it aF couple days.  The availability of both HW platforms and SW is about at critical mass.   ------------------------------  % Date: Mon, 30 Jun 2003 15:35:41 -0400a$ From: "rob kas" <news@paychoice.com>R Subject: Re: HP Webcast this morning on Next-Generation Intel Itanium 2 processors/ Message-ID: <vg14qcnt5a4s4e@corp.supernews.com>d  # "> >    Sadly no one is buying theme > >0 >uJ > Hmmm.  Let's see - they were announced today.  So why don't we give it aH > couple days.  The availability of both HW platforms and SW is about at > critical mass. >i  H                   mmmm Let's see   in 3 years how many IA-64's have been
 sold......   ------------------------------    Date: 30 Jun 2003 20:19:40 -07001 From: keithparris_NOSPAM@yahoo.com (Keith Parris) R Subject: Re: HP Webcast this morning on Next-Generation Intel Itanium 2 processors= Message-ID: <cf15391e.0306301919.37c5a7f1@posting.google.com>r  j brad@.gateway.2wire.net (Bradford J. Hamilton) wrote in message news:<g8%La.67341$Ab2.140995@sccrnsc01>...L > There were several VMS mentions in the slides, and I notice that two press, > releases sent by hp today mention OpenVMS.  @ I particularly enjoyed Scott Stallard's statements about OpenVMS? during the webcast today, especially when one realizes how much C farther up the management chain Scott is today compared with a year  ago.  B In discussing HP's multi-OS strategy for the HP Integrity Servers:  9 Scott: "OpenVMS, as you well know, has a very loyal, very A mission-critical following, in a lot of key industries around theaE world.  We intend to continue to, in fact, grow this business -- withs5 investment protection, [and] increased ISV coverage."C  C And a bit later, talking about Independent Software Vendors (ISVs):t  C Moderator: "Now, Scott, what about OpenVMS?  Where are we as far ase? that? And of course that's coming next year, but I would assumehB [we're] already working with the key ISVs, to make sure they're in9 place when it comes to market later this, er, next year."-  C Scott: "Well, absolutely.  And we share a lot of these ISVs, as you E know, with the other parts of the HP family, as well as some that areuF very specific and unique to the very world-class application areas forD OpenVMS.  We've been greatly aided by the fact that it's on a commonF set of platforms, so now we have the development environments in termsF of production systems we can put out there, as well as the developmentC environments, so these ISVs are coming on board very, very rapidly,IB and so we expect to have, as you mentioned, next year, a complete,A robust offering that will complement our Alpha line-up, again, tosE allow our customers to start to migrate to OpenVMS on the new Itanium E 2 family, the Integrity family, when it's right for them -- for theire
 business."   ------------------------------  % Date: Mon, 30 Jun 2003 22:19:54 -0500a, From: "Dave Gudewicz" <k9jdk@NOSPAMarrl.net>R Subject: Re: HP Webcast this morning on Next-Generation Intel Itanium 2 processors/ Message-ID: <vg1vepf5sd0fa4@corp.supernews.com>:  K Heard the Madison stuff and saw the slides.  I think VMS got fair attention- this time around.-  K Although also saw that the rx5670 (entry level 4-way) will not be supported/J by VMS.  Wonder why?  Hope this isn't the start of another white-box AlphaE campaign.  Remember the white Alpha boxes?  Most would rather forget.t   Dave...   > "Keith Parris" <keithparris_NOSPAM@yahoo.com> wrote in message7 news:cf15391e.0306300745.7b697c07@posting.google.com...jB > HP solutions based on next generation Intel Itanium 2 processors >6< > Today, businesses are demanding more -- more agility, moreG > accountability, and a better return on their IT investment. Learn how-F > HP continues to deliver innovative business and technology solutionsD > with the entry of a new era in enterprise and technical computing. >sE > Find out how your company can benefit from world-leading server andsF > workstation performance, best-in-class deployment, lasting value andD > investment protection with HP's latest Itanium 2-based solutions. >a > Keynote presentations by:a > C > Peter Blackmore - HP Executive Vice President, Enterprise Systemso > GroupoF > Scott Stallard V Senior VP & GM. Enterprise Storage & Servers GroupB > Mark Hudson V Vice President V Marketing, Enterprise Storage & > Servers Grouph > 4 > Follow the link from the main HP website, or go to >r http://www.hp.com/cgi-bin/large/exit.cgi?promo=wwmain_b2/Itaniumwebcast&url=http://www.hpbroadband.com/program.cfm?key=itanium2&?mtxs=home-ent&mtxb=B2&mtxl=L1   ------------------------------    Date: 30 Jun 2003 20:41:57 -07001 From: keithparris_NOSPAM@yahoo.com (Keith Parris)oR Subject: Re: HP Webcast this morning on Next-Generation Intel Itanium 2 processors< Message-ID: <cf15391e.0306301941.d64122e@posting.google.com>  [ "rob kas" <news@paychoice.com> wrote in message news:<vg13n644svdv64@corp.supernews.com>...   >    Sadly no one is buying them  6 Not true anymore -- the floodgates seem to be opening.  F Here's a small list of customers for HP Itanium 2 systems.  (These areA only the ones who happen to have testimonials on HP's web pages.)h   123Multimedia  
   Airbus UK o   Bank Sarasin    Bioinformatic Institute   BPr
   BMW Group y&   California Institute of Technology &   Center of Piezoelectrics by Design    ChevronTexaco l   Chrysler Group    Delaware North Companies    Electra Sweden AB t   EPAQ International '   ESIEE (Engineering School of Paris) f   Exact Holding NV 
   Fiat Auto     INRIA   JetBlue Airways -   Johns Hopkins University    JT International #   Landspitali University Hospital m   Lower Saxony Police c,   MIT and National University of Singapore 
   Multiyork l   NTT COMWARE Corporation z   Ohio Supercomputing Center 0   Pacific Northwest National Laboratory (PNNL)    Queen's University Belfast    Raymond James Financial     realTech AG v   Rice University x,   Samsung Advanced Institute of Technology    SHARCNET    University of Houston     VTG-LEHNKERING AG m   Wells Fargo    C See http://www.hp.com/products1/itanium/testimonials/index.html forg2 details on each one.  A couple of recent examples: --- 7 HP showcases new customer, investments, and programs in ! high-performance Linux clusteringg  E Showcasing its work in the hardware clustering market, HP announced aoD new customer, the expansion of the HP Scalable Cluster Center, and a? new program for developing Linux clustering systems built on HP = ProLiant and Itanium 2-based servers. The National Center forsE Supercomputing Applications (NCSA) has selected HP's high-performance E Linux cluster systems. The NCSA is using a 64-node cluster of Itaniumg9 2-based HP rx2600 servers running Linux to test software,s= interconnects, and compute-intensive scientific applications.  ---nA JetBlue has implemented an Intel Itanium 2-based solution for itsd frequent flier program  ? JetBlue is a low-fare, low-cost passenger airline that provides E high-quality customer service. Since launching operations in February F 2000, the airline has carried more than 11 million passengers. JetBlueC operates a fleet of 40 new Airbus A320 aircraft and is scheduled tosE place into service another 13 A320s by the end of 2003. JetBlue has aaD longstanding relationship with HP, and an IT infrastructure based onE HP servers, desktop, and laptop computers. JetBlue has implemented anoF Intel Itanium 2-based HP rx5670 server for its frequent flier program.A Based on the positive results from deploying its first HP Itanium,E 2-based server, JetBlue is looking forward to deploying more high-endc4 servers such as HP Integrity Superdome systems.  See$ http://www.hp.com/products1/itanium/ ---a   ------------------------------  % Date: Mon, 30 Jun 2003 23:57:34 -0400g* From: "Bill Todd" <billtodd@metrocast.net>R Subject: Re: HP Webcast this morning on Next-Generation Intel Itanium 2 processors2 Message-ID: <yYqcnUVMlqcrm5yiXTWJjg@metrocast.net>  D "Fred Kleinsorge" <my-last-name@stardotzko.dec.com> wrote in message, news:vJ%La.3590$CI7.2373@news.cpqcorp.net...   ...   ? on that rotten corpse of an architecure IA64 - you know the oneb8 > that we just announced systems from 1 CPU to 64 CPU's.  J But you've been pretty shy about announcing benchmark results for anythingE beyond 4 CPUs so far.  And IBM just beat out your 64-CPU TPC-C resulteL (again) - again using only 32 CPUs (on only 16 chips, each about the size ofJ one Madison chip...) and offering that performance at a bit lower cost perJ tpmC as well.  Kind of sounds like the GS80/160/320 all over again - but IH guess that's what those jolly server guys who helped knife Alpha wanted.  L Any place where I could find out what SuperDome memory latencies (both localB and remote) are?  I did find some information about both local andG off-module peak bandwidth (not all that impressive, but at least you'vem published them).     The one with 1318e" > SpecInt2000 and 2106 SpecFP2000.  L Which mostly proves (again) that if you're willing to throw enough power andH chip area (especially cache) at the problem you can indeed make the barnI door fly, albeit inelegantly.  Looks as if the compiler people managed tosJ eke about another 10% of SPECint performance out of EPIC during the courseF of the past year:  at that rate, the long-awaited dramatic benefits ofG compile-time optimization look as if they'll pretty much continue to be G awaited until that pig of a core gets replaced by whatever the EV8 teamt# comes up with in 3 - 4 years' time.,  I Of course another question is whether that added SPECint performance willlJ actually translate into anything worthwhile in real-world applications.  IF notice that your 4-processor SPECweb99_SSL result has not yet improvedF (despite being run on an HP-UX base and hence presumably being able toC benefit from the same compiler that one might suspect generated theuI SPECint_base score you refer to above:  SGI only managed to get 1077 withoF presumably the best compiler they had available to them) - and has nowJ dropped to third place behind not only the 4-processor 1.8 GHz Opteron butJ now the 4-processor 1.7 GHz POWER4+ as well.  And the dual-Madison results2 remain only a smidge ahead of both Opteron and P4.  (   Yup, the one that pretty much everyone4 > except Sun (and Apple ;-) are selling systems for.  D Only if they can find people to buy them.  And unless "Just wait forI Madison!" turns out to be a lot better a prediction than first "Just waittL for Merced!" and then "Just wait for McKinley!" did over the past many yearsD that Itanic enthusiasts have been chanting (and then changing) theseL mantras, I'd say it's time to start looking around seriously for a lifeboat.   - bill   ------------------------------    Date: 30 Jun 2003 13:49:59 -07001 From: keithparris_NOSPAM@yahoo.com (Keith Parris)aA Subject: ISV enthusiasm grows for OpenVMS on HP Integrity serversi= Message-ID: <cf15391e.0306301249.35d1fe25@posting.google.com>   C "We look forward to the price/performance improvement and continuedlA scalability, reliability, security, and predictability offered byeD OpenVMS on Itanium architecture," says Graham Beare, senior platformD architect at Schlumberger.  With over 180 systems worldwide and overE 150 technical specialists, Schlumberger is one of the world's leadingeC providers of telecommunications infrastructure solutions, including F messaging, intelligent applications, fraud solutions, SS7 over IP, andD development platforms.  "We are very excited about moving to ItaniumF with our OpenVMS product suite," concurs Tim Sleath, product manager. D "This change will offer more cost saving and improved performance to our customers."r  N See http://www.hp.com/products1/evolution/alpha_retaintrust/openvms/index.html   ------------------------------    Date: 30 Jun 2003 13:43:43 -07001 From: keithparris_NOSPAM@yahoo.com (Keith Parris)h# Subject: New OpenVMS-to-Itanium FAQ = Message-ID: <cf15391e.0306301243.77f9536c@posting.google.com>s   New OpenVMS to Itanium FAQss  E With the recently announced evaluation release of OpenVMS version 8.0e= for Integrity Servers, we have achieved yet another strategicgC milestone in the porting of OpenVMS to the Itanium architecture. InmC response to the bevy of questions, we have created a new OpenVMS to E Itanium FAQ section on the Alpha RetainTrust website.  Included amongi> the many questions that are addressed are:  What are the plansF (roadmap) for OpenVMS from here on?  What kind of help will HP provideE to assist with the transition? Will all of the features on which I'vetF come to rely on in the OpenVMS operating system on AlphaServer systems> be available on the OpenVMS Industry Standard 64 (OpenVMS I64)! versions of the operating system?)  B For answers to these and many other questions that you might have,O visit http://www.hp.com/products1/evolution/alpha_retaintrust/openvms/faqs.htmlo   ------------------------------  # Date: Tue, 01 Jul 2003 04:10:35 GMTa6 From: "Craig A. Berry" <craigberry@mac.com.spamfooler>' Subject: Re: New OpenVMS-to-Itanium FAQa@ Message-ID: <e60a2f4cf474a0ecb144f15dabded474@free.teranews.com>  = In article <cf15391e.0306301243.77f9536c@posting.google.com>,e3  keithparris_NOSPAM@yahoo.com (Keith Parris) wrote:l   > New OpenVMS to Itanium FAQs    snip   K > http://www.hp.com/products1/evolution/alpha_retaintrust/openvms/faqs.html,  C I'd suggest a bit of proof-reading and fact checking.  There are a  F number of small but distracting typos (e.g. in item 9), and also some ( misleading and incomplete information.    H Item 19 occurs twice with somewhat different text, but both times seems A to imply that the only option for PL/1 users is to migrate their tH applications to another language.  Kednos has previously stated in this E forum their intention to port to Itanium by mid-2004 and has also -- cE clearly in vain -- pleaded with Compaq/HP personnel to stop implying  G that the PL/1 compiler (and any porting decision) belongs to HP rather tE than Kednos.  The claim that HP is retiring PL/1 is also made on the a7 openvms_plans page as well as being implied in the FAQ.>  F Also in the language porting category (item 18), why is COBOL missing C when it is listed on the separate products page as something being e ported?e  H Item 21.  The discussion of IEEE floating point issues should obviously 8 contain a pointer to the white paper on that very topic.   ------------------------------  % Date: Mon, 30 Jun 2003 20:46:28 +0200g+ From: "Hans Vlems" <hvlems.nieuw@zonnet.nl> S Subject: Re: OpenVMS Pearl - OpenVMS V8.0, first release on Itanium, ships today!!!l5 Message-ID: <bdq0ia$vlosu$1@ID-143435.news.dfncis.de>?  > Very nice indeed! Can I have an evaluation system at home? :-)  @ "Sue Skonetski" <susan_skonetski@hotmail.com> schreef in bericht7 news:857e9e41.0306300930.3b056994@posting.google.com...  > From: Gorham, Mark% > Sent: Monday, June 30, 2003 1:26 PM $ > To: Skonetski, Susan; Schimel, LinA > Subject: FW: VMS V8.0, first release on Itanium, ships today!!!c > C > I am very pleased to be able to announce HP OpenVMS for IntegrityoF > Systems V8.0 has been shipped to 10 evaluation customers today, JuneF > 30, 2003.  This is the result of 24 months of work by a creative andH > dedicated team of engineers.   Today you probably also heard about theF > great new Itanium-based systems from HP, with a new brand name of HPH > Integrity Servers.  These include for the first time Superdome systems0 > - up to 64 CPUs per server shipping in August! >pB > For our first release, we intended to ship a lightweight OS, twoC > network stacks, and some cross compilers and development tools to E > allow our key horizontal ISVs to start moving their applications tooG > Itanium.  We have greatly exceeded this in 8.0, adding clustering (weoB > even have mixed architecture Alpha/Integrity clustering!!!), SMPH > (symmetric multi processing), DFS, XFC (extended file cache), Ramdisk,E > and ZIP/UNZIP.  Stability is much better than expected, and we have.G > already established with one customer's application a proof point forbE > recompile/relink application portability between Alpha and Itanium.d >sH > Work with key ISVs starts immediately.  Our FastTrack program will useG > our lab and central engineering resources to accelerate ISV adoption.a > D > We are on track to ship V8.1 in December of this year, and to shipB > V8.2, our first production-quality release of OpenVMS on Itanium > systems, in H2 2004. >gH > If you have any questions on what we're doing with VMS, please send meB > mail or check out our web site at  <http://h71000.www7.hp.com/>. >- >-	 > Regardsa >  > Mark > G > Mark Gorham   Vice President   OpenVMS Systems Group  Hewlett-Packard 
 > Corporationx0 > email: mark.gorham@hp.com  phone: 603-884-01183 > admin: Susan Christie phone: 603-884-0100  email:a > susan.christie@hp.comi >e >g > VMS V8.0 includesl >. >.6 >  Base O/S (includes RMS, XQP, Image Activator, ...)H >  Utilities (BACKUP, BAD, COPY, CLI, DCL, DIR, MAIL, UTIL32, EDT, ...) >  (i2000), rx2600 supportd? >  LAN support (Intel 8255x 10/100, Broadcom 5701 10/100/1000)  >  SCSI Storage (U160, U320)t >  FDDI support > * ODS-5 supportU >  IPCl >  Monitor  >  RMS Convert and FDLSHR > * Dual CPU support >  INSTAL/PRIVmD > * Clusters (Initial support including mixed architecture clusters)) >  DECnet Phase IV (SET Host, Copy, ...) E >  TCP/IP (FTP client and server, Telnet client and server, sockets,  > QIO interface, NET, ...) >  LATp > * DFS. > Development Tools: >  Cross Linker >  Cross Librariand >  Cross Message Compiler/ >  Cross and Native Command Definition Utilityy >  CHECKSUM >  LIBRTL, CRTL, MATHRTL, CVT
 >  XDELTA6 >  Cross and Native SDA (System crash, process crash) >  Bugcheck$ >  DECset (DTM, CMS, MMS, TPU, LSE) >  DEBUG (Initial support)i
 > * ZIP/UNZIP ) > Cross Compilers and Related Components:( >  BLISS 32/64s >  Ce
 >  IMACRO >  Fortranu? >  C++ less template library (Initially for internal use only)q >  Cross Assembler 
 >  LIBOTS
 > * SORT32 > * LIBOTS2/
 > * Cobol RTL/ > Security:l3 >  Base security (Except ACME and Security Server)/ >  Security System Services >  LOGINOUT
 >  Authorizen >  SET/SHOW Securityr" >  Audit Server and Analyze/Audit >  Encryption for OpenVMS >  Accounting > OTHERp >  DECthreads >  DDTM >  DECwindows Client1 > * XFCa- > * Ramdisk (available for internal use only)n >  PCSI
 >  VMSINSTALS   ------------------------------  % Date: Mon, 30 Jun 2003 21:54:49 -0500s1 From: "David J. Dachtera" <djesys.nospam@fsi.net>nY Subject: Re: OpenVMS Pearl - OpenVMS V8.0, first release on Itanium, ships today!!! todaya' Message-ID: <3F00F7F9.3B43DF13@fsi.net>    Hans Vlems wrote:F > @ > Very nice indeed! Can I have an evaluation system at home? :-)  9 Agreed! Any chance this might be available for hobbyists?w   Low-cost (less than $50US) CDs?H   A downloadable ODS image?a   -- g David J. Dachterac dba DJE Systemse http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/n   ------------------------------    Date: 30 Jun 2003 13:23:38 -07001 From: keithparris_NOSPAM@yahoo.com (Keith Parris)32 Subject: OpenVMS Success Story: Aurora Health Care= Message-ID: <cf15391e.0306301223.39a3d33c@posting.google.com>a  B New success story highlights Aurora Health Care's OpenVMS database prescription  C With the need to run 750 gigabytes of live data on a single server,SB Aurora Health Care turned to a high-performance HP Next GenerationB AlphaServer GS1280 system running the OpenVMS operating system.  AF newly published success story tells how the Milwaukee-based consortiumC eliminated a major bottleneck in its Cerner Millennium environment,r? reducing its 90+ percent database system utilization rate to 25tE percent or less.  As technical services manager Lee Oettinger states,,> "Now our users are happy, and we can sleep at night."  QuotingC extensively from Aurora executives, the story describes how many HPyD groups --from the account team to OpenVMS engineering to services --F worked together to deliver a solution that would support growth as theF organization adds more hospitals and clinics.  According to the story,C the new system's quicker response time has encouraged more staff to F use the application, as Aurora plans to expand Millennium applications? to other facilities.  Moreover, Aurora has pioneered innovativemD OpenVMS Millennium capabilities, such as read-only Web access to theB Oracle database.  With confidence in HP's product roadmaps, seniorC systems manager David Harrold says, "I'm sure that in the future we<) will move to the Itanium-based platform."   A See http://h71000.www7.hp.com/openvms/brochures/aurora/index.html G (HTML) or http://h71000.www7.hp.com/openvms/brochures/aurora/aurora.pdft (PDF).   ------------------------------   Date: 30 Jun 2003 18:23 CDTI' From: carl@gerg.tamu.edu (Carl Perkins)s Subject: Re: Rethinking  V.M.S- Message-ID: <30JUN200318230593@gerg.tamu.edu>F  2 Fabio Cardoso <fabiopenvms@yahoo.com.br> writes...4 }The VMS was developed 25 years ago when memory was ( }expensive and Virtual Memory an option.3 }But do we need a Virtual Memory System nowadays ? g2 }May be we should think in another architecture as8 }the memories are becoming cheaper. May be the VM System9 }can be improved or rethinked... What are the performance.; }issues related in having the VM (paging, swapping, relatedl4 }to disk I/O delay, etc ...) What do u think about ? }  }Regards }  }FC   H You probably still need it. You can tell because the pagefiles are used.H VMS doesn't normally page stuff out unless it needs to (unlike Windows).  8 The overhead is not very large when you aren't using it.  A The page (and swap) files are very nice to have when you need it.oA If you ahve no virtual memory, what do you do when you run out ofaA space in your real memory? You fail to do whatever it is you wereeC trying to do and you won't be able to do it untill you upgrade your E system's memory. With virtual memory it is much easier - when you run,D out of real memory it just uses virtual memory, and when it runs outF of that you just add another pagefile. Even with todays incredibly lowC memory prices, disk space is still much much cheaper. Do you reallydG want to buy an extra 8GB (or whatever) of memory just in case you mighteE someday need some extra memory during the peak load? You have to have-0 enough to cover the peaks, not just the average.   --- Carl   ------------------------------  % Date: Mon, 30 Jun 2003 21:49:08 -0500a1 From: "David J. Dachtera" <djesys.nospam@fsi.net>n Subject: Re: Running VMS off CDo' Message-ID: <3F00F6A4.EAD1133B@fsi.net>e   "Doc.Cypher" wrote:v > N > For those that have noticed I have an interest in remailers (duh!). The goodM > news is that the new Type-III remailer (Mixminion) is written in Python andm# > will therefor be runnable on VMS.s > L > Any system such as that should be as secure as possible, and I'm wondering6 > about the possibility of booting and running off CD. > N > Has anyone done this?  If so, is XFC enough of a performance booster to makeM > this a viable proposition?  What would be a reasonable amount of memory forl
 > the system?O >  > Thanks for any advice.  F I'm sure it can be done, provided you are prepared to hack up your ownG bootable image. I've done it. It's not easy, and the time investment ise substantial.  F I'd have to question the reason why, however. What does a write-locked@ system disk do for you? ...and are you prepared to deal with the ramifications of that?   -- e David J. Dachtera- dba DJE Systems- http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/S   ------------------------------  % Date: Mon, 30 Jun 2003 16:53:22 -0400n, From: "Island" <dbturner@nospamislandco.com># Subject: Selling: Microvax 4000-108g/ Message-ID: <vg18the09gbf5e@news.supernews.com>-   We have four incomingn Enquire within!7     --   David B Turner Island Computers US Corporationj 2700 Gregory St., Suite 1800 Savannah GA 31404e Tel: 912 447 6622. Fax: 912 201 0402@ Email: dbturner@hpaq.net http://www.hpaq.neto   ------------------------------   Date: 30 Jun 03 11:13:32 PST From: mckinneyj@cpva.saic.com Q Subject: Re: Subject: Reading long files in BASIC that appears to be "recordless" ( Message-ID: <K+f0xasQDfMv@cpva.saic.com>  = In article <6f29699e.0306300816.450c0ec6@posting.google.com>,t*  dennisb@wvhmhc.org (Dennis Baker) writes:D > Thanks for your timely input, Randy and Dave, but unfortunately it > didn't help. >   C I think that you'll need to adopt a scheme where you use sys$assignsE to open a channel to the disk and $qiow's to access the file and read-@ the data. You'll want to use the IO$_READVBLK function code with; the $qiow reads to step through the blocks within the file.s  C Don't have any BASIC examples. Perhaps google can help? If not, and1B you'd like an example in C or Macro send a request to mckinneyj at$ saic.com and I'll see what I can do.   --   - Jimn   ------------------------------  # Date: Mon, 30 Jun 2003 23:47:58 GMTt# From: hoff@hp.nospam (Hoff Hoffman)-! Subject: Re: system disk on ODS-5 1 Message-ID: <O_3Ma.3630$hh.3515@news.cpqcorp.net>   f In article <a7234bb1.0306241159.3968ae2@posting.google.com>, timasmith@hotmail.com (Tim Smith) writes:R :I hear this is not recommended - what is the current status/issues for vms 7.3-1?  H   Donno.  I've been running ODS-5 on all of my V7.3-1 (and later) boxes.H   Works OK here.  You may or may not have local applications or criticalG   LPs that will work (or that will fail) in the environment, of course.t   (I don't.)  = :How big a deal is it to convert a system disk back to ODS-2?r  F   The sequence is documented in the OpenVMS manuals, of course -- withG   the details and your knowledge of your local system environment, you  "   can best decide this difficulty.    N  ---------------------------- #include <rtfaq.h> -----------------------------K     For additional, please see the OpenVMS FAQ -- www.hp.com/go/openvms/faqfN  --------------------------- pure personal opinion ---------------------------E         Hoff (Stephen) Hoffman   OpenVMS Engineering   hoff[at]hp.come   ------------------------------    Date: 30 Jun 2003 11:19:38 -07001 From: keithparris_NOSPAM@yahoo.com (Keith Parris)rO Subject: Tri-architecture cluster demonstrated at DECUS Ottawa Technical Update = Message-ID: <cf15391e.0306301019.1fb14925@posting.google.com>   F A VMS cluster with VAX, Alpha, and Itanium systems was demonstrated at" the DECUS Ottawa Technical Update.  ; View of Cluster from system ID 58693  node: CTHX03         . 23-JUN-2003 21:18:32@ +----------------------------------------------------+---------+@ |                       SYSTEMS                      | MEMBERS |@ +--------+--------------------------------+----------+---------+@ |  NODE  |             HW_TYPE            | SOFTWARE |  STATUS |@ +--------+--------------------------------+----------+---------+@ | CTHX03 | AlphaServer ES40               | VMS V7.3 | MEMBER  |@ | CTHOPS | VAXstation 4000-60             | VMS V7.3 | MEMBER  |@ | I64CDN | HP rx2600                      | VMS X9TM | MEMBER  |@ +--------+--------------------------------+----------+---------+   ------------------------------  # Date: Mon, 30 Jun 2003 20:45:00 GMT / From: brooks@cuebid.zko.dec.nospam (Rob Brooks)sS Subject: Re: Tri-architecture cluster demonstrated at DECUS Ottawa Technical UpdateL- Message-ID: <Tp17TEWDh2yx@cuebid.zko.dec.com>l  7 David Beatty <David.Beatty@qwertysasasdfgh.com> writes:n > ? >     Hmmm ... I wonder if that means HP will decide to supporta6 > VAX and Itanium in the same cluster, since it was my5 > understanding that they had not planned to do that.s  C It is *highly* unlikely that we'll _support_ VAXes and IA64 systemsPL together in a cluster.  We have said, however, that we're not planning to doJ anything that would prevent a VAX-IA64 cluster.  For us, there is a ratherO large committment to qualification and validation for supported configurations;M5 that time and money is likely better spent elsewhere.)  D As always, if someone has a strong business need  to see support for7 VAX-IA64 clusters, please contact andy.schneider at hp.0   -- n  M Rob Brooks    VMS Engineering -- I/O Exec Group     brooks!cuebid.zko.dec.comh   ------------------------------  % Date: Mon, 30 Jun 2003 15:30:55 -0400l5 From: David Beatty <David.Beatty@qwertysasasdfgh.com>nS Subject: Re: Tri-architecture cluster demonstrated at DECUS Ottawa Technical Updatev2 Message-ID: <qo8AP2WLKLUgDuM3PuLSdEwRjGz6@4ax.com>  =     Hmmm ... I wonder if that means HP will decide to supporte4 VAX and Itanium in the same cluster, since it was my3 understanding that they had not planned to do that.e   David R. Beattyi  B On 30 Jun 2003 11:19:38 -0700, keithparris_NOSPAM@yahoo.com (Keith Parris) wrote:  G >A VMS cluster with VAX, Alpha, and Itanium systems was demonstrated at:# >the DECUS Ottawa Technical Update.  >e< >View of Cluster from system ID 58693  node: CTHX03          >23-JUN-2003 21:18:32 A >+----------------------------------------------------+---------+nA >|                       SYSTEMS                      | MEMBERS | A >+--------+--------------------------------+----------+---------+iA >|  NODE  |             HW_TYPE            | SOFTWARE |  STATUS |/A >+--------+--------------------------------+----------+---------+wA >| CTHX03 | AlphaServer ES40               | VMS V7.3 | MEMBER  | A >| CTHOPS | VAXstation 4000-60             | VMS V7.3 | MEMBER  |tA >| I64CDN | HP rx2600                      | VMS X9TM | MEMBER  |eA >+--------+--------------------------------+----------+---------+g   ------------------------------  # Date: Tue, 01 Jul 2003 00:23:41 GMT_# From: hoff@hp.nospam (Hoff Hoffman)d4 Subject: Re: Virtual Memory? (was: Rethinking V.M.S)1 Message-ID: <hw4Ma.3633$hh.1701@news.cpqcorp.net>c  0 Fabio Cardoso <fabiopenvms@yahoo.com.br> writes: :-4 :The VMS was developed 25 years ago when memory was ( :expensive and Virtual Memory an option.  ?   This shows more in the VAX instruction set design than in thee@   virtual memory design -- VAX instructions traded off speed for?   reduced storage.  Alpha and Itanium take different approachese   here than VAX, of course.   3 :But do we need a Virtual Memory System nowadays ? e  G   Yes.  I will assume that this is not a troll, and that you are truelytG   rather unfamiliar with these architectural considerations, of course.   2 :May be we should think in another architecture as8 :the memories are becoming cheaper. May be the VM System9 :can be improved or rethinked... What are the performancec; :issues related in having the VM (paging, swapping, relatedf4 :to disk I/O delay, etc ...) What do u think about ?    L   Virtual memory is a rather more powerful construct than you are apparentlyK   familiar with -- please read most any collegiate operating system text on0K   the subject, or (at a minimum) please consider private addressing and itspL   benefits, and process and system security and isolation, and tasks such asK   memory error handling, and global sections and sharing of address spaces.e  K   The virtual memory system is the basis of OpenVMS security, for instance.v  I   Application debugging centrally uses the virtual memory support, too -- K   the no-access-page handling is how tracepoints and watchpoints are built.   L   There is no question that I would incorporate virtual memory into most anyM   new architectural design I worked on -- unless something even more powerfulmM   than virtual memory is available and included in the new design, of course.-I   (This ignores architectures that are designed for and targeted for veryaG   specific tasks -- there are undoubtedly cases where the inclusion of t3   virtual memory support might not be appropriate.)G  H   Discussions of architectures beyond current-generation virtual memory G   designs (such as the combination of PALcode and OpenVMS code found onnG   within the OpenVMS Alpha memory management design, or the scheme usedtF   on Intel Itanium systems) are likely fodder for a doctoral thesis orG   two, and serious designs would likely be considered to be proprietaryoC   or confidential technologies and unsuitable for discussions here.u  E   If you are interested in the topic, I would encourage familiarizing D   yourself with the OpenVMS Alpha memory management (V7.0 or later) E   and with the Intel Itanium memory management design.  You will also-E   want to consider what areas you believe could be improved, and what D   you will inevitably have to trade (transistors and/or speed and/orD   design complexity and/or costs and/or ?) for the memory management:   changes -- as is often the case, there are trade-offs...  C   The Internals and Data Structures Manual (IDSM) will also provideSF   an idea of the various ways OpenVMS uses virtual memory -- check theC   most recent Alpha chapters, as I think some of the virtual memory E   details from circa V7.0 are included there.  Even if not, the older-9   IDSM materials will have useful memory-related details.-  N  ---------------------------- #include <rtfaq.h> -----------------------------K     For additional, please see the OpenVMS FAQ -- www.hp.com/go/openvms/faq3N  --------------------------- pure personal opinion ---------------------------E         Hoff (Stephen) Hoffman   OpenVMS Engineering   hoff[at]hp.com.   ------------------------------  # Date: Mon, 30 Jun 2003 20:44:51 GMTo) From: Andrew Balaam <abalaam@yahoo.co.uk> % Subject: Re: VMS backup to NFS sharest4 Message-ID: <20030630.20445100.484801457@imagnu.geo>   Jason,  I I have not used UCX, so I can't comment on how that package handles NFS =   I mounts. As I said, the default behaviour for Multinet is to preserve FDL=f =20eE information in these hidden files. These files are named as follows:-   5      .$fdl$name-of-file.name-of-tail[;version-number]   I for a backup file with a block size of 8192, the FDL file contains the=20b following:-t   FILE 	MAX_RECORD_NUMBER 8192i 	ORGANIZATION sequential RECORD
 	FORMAT fixedi
 	SIZE 8192  I I think most shells appreciate the $ characters being escaped when typin=V g=20 in the file names.  I There is a gotcha though - If you have more than one version of a file, =   I there is a file name associated with each of the versions (plus its FDL =e  I 'friend') with the name having the (;n...) semi-colon and version number=  .=20E The file name with no version number is hard linked to the lastest=20lF version of the file. So, if you TAR the area, your TAR file will be=204 larger than it needs to be because of the hard link!  I Making and restoring image backups of non-system disks should work fine =e -=20I I would suggest for all backups to NFS, that you pick a reasonable block=e =20wI size, like 8192, the same as is used by default for tapes. If you have a=y =20yI major problem, you could create VMS compatible tapes using your *nix box=t .   I I have had no problem with vms backups over NFS. I have in the past had =   I problems with other file types, but it was so long ago, I don't remember=n =20tI what it was (and it was on an earlier version of Multinet V3.something).=- =20-= I currently use v4.0 as we stopped maintenance some time ago.D  I Don't take my word for any of this - check yourself that it all works OK=  .=20I Data is (should be) too valuable to gamble on the posting of a stranger =r   on a news group! :-)   Let me know how you get on.s Andrew.k  6 >>>>>>>>>>>>>>>>>> Original Message <<<<<<<<<<<<<<<<<<  I On 30/06/03, 18:58:24, jason.mccormick@lexi.com (Jason McCormick) wrote =W  ' regarding Re: VMS backup to NFS shares:o    I > > Multinet stores the FDL info in a hidden file (ie, one starting with=t  a . > > ) I > > on an NFS server by default. So when the file is viewed from VMS, th=e ee  > > file attributes are correct.  B > So unlike UCX (as pointed out in this thread) there's no need toI > specify a qualified like /ADF=3D since it's automatic? (Just wanting t=i o  > make sure)  G > > If you are doing and image followed by daily and weekly incrementaliI > > backups, you will have to work out some method of restoring the imag=t eo& > > backup (if it is the system disk).  I > I'm going to use our old TZ86 to back up the system disks and then all=r  F > the user/application data is on other disks.  I can use backup/imageF > across NFS for the disks that aren't the system disk though correct?  I > > To read a saveset, I just copy one off the DVD into an area readable=V  by=20 myI > > VAX over NFS, gunzip it, create the hidden FDL file and then use bac=p kupS > > on the VAX.G  2 > And all the files restored correctly?  Terrific.  I > Thanks for the insight into Multinet.  And thanks to everyone else who=w  ) > had some ideas.  It's much appreciated.i  
 > -- Jason   ------------------------------  % Date: Mon, 30 Jun 2003 21:45:19 -0500m1 From: "David J. Dachtera" <djesys.nospam@fsi.net>-% Subject: Re: VMS backup to NFS sharesJ' Message-ID: <3F00F5BF.343FBFE5@fsi.net>_   Jason McCormick wrote: > H > I was hoping that a VMS guru could answer a question for me.  I have aH > VAXStation 4000 90A and I'm attempting to get a usable backup solutionE > for it.  The tape unit I've been using on it requires a data stream4G > speed that the VAX simply can't keep pushing out to the unit and it'sy" > casuing physical damage.  [snip]   Whoa, there!!!   Care to elucidate???  D How can BACKUP be causing physical damage to a tape, drive, etc. ???   -- o David J. Dachteraa dba DJE Systemst http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/l   ------------------------------    Date: 30 Jun 2003 15:34:39 -0400% From: Thien-Thi Nguyen <ttn@glug.org>u Subject: vms-pstree.el( Message-ID: <7gof0ffjg0.fsf@gnufans.net>  @ if you use emacs under vms, you might find vms-pstree.el useful.   thiN  : __________________________________________________________5 ;;; vms-pstree.el --- Display VMS processes as a treei  5 ;;; Copyright (C) 2003 Thien-Thi Nguyen <ttn@gnu.org>a  = ;; This file is released under the same license as GNU Emacs.A  D ;; GNU Emacs is free software; you can redistribute it and/or modifyG ;; it under the terms of the GNU General Public License as published bymF ;; the Free Software Foundation; either version 2, or (at your option) ;; any later version.l  ? ;; GNU Emacs is distributed in the hope that it will be useful,GA ;; but WITHOUT ANY WARRANTY; without even the implied warranty oft@ ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the/ ;; GNU General Public License for more details.s  D ;; You should have received a copy of the GNU General Public LicenseD ;; along with GNU Emacs; see the file COPYING.  If not, write to the? ;; Free Software Foundation, Inc., 59 Temple Place - Suite 330,e ;; Boston, MA 02111-1307, USA.   ;;; Commentary:u  G ;; VMS processes can be listed w/ the "SHOW USER/FULL" command, and now H ;; visualized in a tree form to show relationships between a process andH ;; its parent (if any), with vms-pstree.el.  This was written to help meG ;; decide whether or not the Emacs 21.x port to OpenVMS 7.3-1 should bedG ;; changed to actually implement "proper" subprocesses.  OK, I confess:eF ;; it is merely a display hack that actually has no bearing on whetherF ;; or not subprocess support will change.  Actually^2, it does help toH ;; lightly test the work-in-progress port so there is some merit besides@ ;; (actually^3) being another creative act of procrastination... ;;/ ;; In any case, the usage is relatively simple:s ;; ;;      M-x vms-pstree ;;E ;; Emacs asks for a user name to scan; leave empty to scan all users.t= ;; The output shows up in a buffer named "*vms-pstree USER*".  ;;B ;; There are two variables to modify the behavior of `vms-pstree':B ;; `vms-pstree-show-pid' and `vms-pstree-suppress-unix-init-envy'.1 ;; See their respective docstrings for more info.t ;; ;; Feedback welcome.  	 ;;; Code:i  > (require 'cl)                           ; use the source luke!   (defvar vms-pstree-show-pid nilwD   "If non-nil, `vms-pstree' also displays the PID of each process.")  . (defvar vms-pstree-suppress-unix-init-envy nilG   "If non-nil, `vms-pstree' does not fake an \"init\" parent process.")o   (defun vms-pstree-hash ()u+   (let ((ht (make-hash-table :test 'equal))X.         ;; hash data: children parent tty name&         (empty (list nil nil nil nil))         case-fold-search)3$     ;; unix envy rears its ugly head.     (unless vms-pstree-suppress-unix-init-envy1       (puthash "1" (list nil nil nil "init") ht))      (goto-char (point-min))'=     (while (re-search-forward "^ [A-Z][A-Z]+ " (point-max) t)i       (move-to-column 18)A,       (let ((p (point)) name pid tty parent)-         (setq name (progn (move-to-column 32)6@                           (buffer-substring p (setq p (point))))/               pid (buffer-substring p (+ 8 p))) /         (let ((snip (string-match " +$" name)))v:           (when snip (setq name (substring name 0 snip))))         (forward-char 10)e(         (cond ((looking-at "[A-Z0-9]+:")B                (setq tty (buffer-substring (point) (match-end 0)))9                (unless vms-pstree-suppress-unix-init-envya$                  (setq parent "1")))<               ((looking-at ".subprocess of \\([0-9A-F]+\\)")A                (setq parent (buffer-substring (match-beginning 1)t>                                               (match-end 1))));               (t (unless vms-pstree-suppress-unix-init-envyf'                    (setq parent "1"))))c         (when parent@           (puthash parent (let ((cur (gethash parent ht empty)))B                             (cons (cons pid (car cur)) (cdr cur)))                    ht))l8         (puthash pid (let ((cur (gethash pid ht empty)))8                        (list (car cur) parent tty name))                  ht)))     ht))  : (defun vms-pstree-spew (ht pid &optional prefix cols uncp)M   ;; PREFIX is calculated by the parent when recursing.  For the first child,wL   ;; it is a short connector, for subsequent children, it may contain uncle-J   ;; continuation marks (vertical bars) if UNCP is non-nil, at the columnsJ   ;; specified by COLS.  A number in COLS less than zero is used in offsetK   ;; calculations (that is, its absolute value is used) but does not rendern    ;; an uncle-continuation mark."   (unless prefix (setq prefix ""))$   (unless cols (setq cols (list 0)))    (let* ((data (gethash pid ht))          (name (cadddr data))o          (kids (car data))1          (ofs (+ (abs (car cols)) (length name) 2n(                  (if vms-pstree-show-pid'                      (+ 2 (length pid))r                    0))))(     (insert (format "%s%S" prefix name))     (when vms-pstree-show-pidh#       (insert (format "(%s)" pid)))o     (if (not kids)         (insert "\n")o1       (let ((last-kid (car (reverse (cdr kids))))s"             (nkids (length kids)))         (vms-pstree-spew htn#                          (car kids)n9                          (if (= 1 nkids) "-----" "--+--")n.                          (cons (+ 5 ofs) cols)%                          (< 1 nkids))          (mapcar '(lambda (kid)#                    (vms-pstree-spew                      ht kid$                     (concat (if uncp*                                 (mapconcat-                                  '(lambda (f)a=                                     (make-string (- f 5) 32))lK                                  (let ((ls (remove-if '(lambda (x) (> 0 x))mM                                                       (cons (+ 5 ofs) cols))) ,                                        fill)2                                    (while (cdr ls)A                                      (setq fill (cons (- (car ls)tC                                                          (cadr ls))s;                                                       fill)t8                                            ls (cdr ls)))(                                    fill))                                  "  |  ")d3                               (make-string ofs 32))p4                             (if (equal kid last-kid)'                                 "  `--"-'                               "  |--"))75                     (cons (* (if (equal kid last-kid) G                                  ;; A negative column counts for offsetfH                                  ;; calculations but disables display ofC                                  ;; uncle continuation marks there.r#                                  -1t!                                1)r'                              (+ 5 ofs))r                           cols)u0                     (not (equal kid last-kid))))                 (cdr kids))))))T   (defun vms-pstree (user))   (interactive "sUser (empty for all): ")r(   (let* ((SPC-user (if (string= "" user)                        usero(                      (concat " " user)))G          (buf (get-buffer-create (concat "*vms-pstree" SPC-user "*"))))t     (switch-to-buffer buf)     (erase-buffer)I     (subprocess-command-to-buffer (concat "SHOW USER/FULL" SPC-user) buf)s!     (let ((ht (vms-pstree-hash)))n       (erase-buffer)       (let (top)%         (maphash '(lambda (key value)h(                     (unless (cadr value)1                       (setq top (cons key top))))-                  ht)         (goto-char (point-min))g         (mapcar '(lambda (tree)a,                    (vms-pstree-spew ht tree)!                    (insert "\n"))b                 top)))))   (provide 'vms-pstree)f   ;;; vms-pstree.el ends herei   ------------------------------    Date: 30 Jun 2003 14:53:50 -0500+ From: young_r@encompasserve.org (Rob Young)   Subject: XFC, IO Size and Backup3 Message-ID: <5NMyTCKZ7HD7@eisner.encompasserve.org>u  @         Here is some interesting data.  Using XFC stats, you can@         see some things going on with IO via backup and/or third         party backups.  J         This first group of data below is XFC stats on the target (output)L         file when backing up a 4800006 block file disk-to-disk via a simple          backup command:n  4         $ backup *.* new_disk:[new_directory]*.*/log  * Total QIOs to this file:             75004* Read IOs to this file:                   0* Read Hits:                               0+ Hit Rate:                            0.00 %t* Read Aheads:                             0* Read Throughs:                           05 Read Arounds:                            0 (/NOCACHE)n> Read Arounds:                            0 (Function modifier)4 Read Arounds:                            0 (IO Size)  * Write IOs to this file:              75004* Write Throughs:                          05 Write Arounds:                       75004 (/NOCACHE) > Write Arounds:                       75004 (Function modifier)4 Write Arounds:                           0 (IO Size)  . Average Overall I/O response time to this file,  in milliseconds:                     1.6177+ Average Disk I/O response time to this files,  in milliseconds:                     1.61771 Accuracy of I/O resp time:                   78 %s  K         Things of note, 75004 IOs, which works out to 32K IO size - defaultbN         size as expected.  All IO is write IO and therefore bypasses XFC.  IO ;         response time is low as disks have writeback cache.k  I         This next set of stats is taking the original input file, backingiH         it up via ABC to a TSM server (3rd party TSM client for OpenVMS)  * Total QIOs to this file:              9766* Read IOs to this file:                9766* Read Hits:                               0+ Hit Rate:                            0.00 %n* Read Aheads:                             0* Read Throughs:                           05 Read Arounds:                         9766 (/NOCACHE).> Read Arounds:                            0 (Function modifier)4 Read Arounds:                         9766 (IO Size)  * Write IOs to this file:                  0* Write Throughs:                          05 Write Arounds:                           0 (/NOCACHE)o> Write Arounds:                           0 (Function modifier)4 Write Arounds:                           0 (IO Size)  . Average Overall I/O response time to this file,  in milliseconds:                    18.2225+ Average Disk I/O response time to this file ,  in milliseconds:                    18.22251 Accuracy of I/O resp time:                   71 %l  G         Curiously, ABC uses a 512 block IO.  So response time of coursekA         is high , but much is being pushed, actual backup time istE         very good.  Notice all IO bypasses XFC.  This makes sense as a&         VCC_MAX_IO_SIZE is 127 blocks.  A         Watch what happens when IO size is raised to 1000 blocks,b@         running a ABC/TSM backup against the copied target file:    * Total QIOs to this file:             84380* Read IOs to this file:                9376* Read Hits:                               0+ Hit Rate:                            0.00 %1* Read Aheads:                          2216* Read Throughs:                        93765 Read Arounds:                            0 (/NOCACHE)x> Read Arounds:                            0 (Function modifier)4 Read Arounds:                            0 (IO Size)  * Write IOs to this file:              75004* Write Throughs:                          05 Write Arounds:                       75004 (/NOCACHE),> Write Arounds:                       75004 (Function modifier)4 Write Arounds:                           0 (IO Size)  . Average Overall I/O response time to this file,  in milliseconds:                     4.1719+ Average Disk I/O response time to this fileu,  in milliseconds:                     4.17191 Accuracy of I/O resp time:                   78 %t  O         The same 9376 read IOs for the TSM backup, but 2216 of them satisified ,?         by read aheads.  Reads are through cache as compared to :         ABC read arounds when IO size was 127 (see above).  *         Checking a file backed up like so:  /         $ backup large_file.dat nla0:t.bck/savey  @         and looking at XFC stats,  nary a read.  Backup bypassesD         XFC on reads.  Makes sense - wouldn't want XFC to get in theE         way of backup as backup is double-buffering , pumping tons ofID         IO to disk.  One more layer of cache would be a hindrance toE         BACKUP but not to a 3rd party product that isn't doing doublefF         buffering.  That would be an assumption on my part (ABC isn't O         double-buffering).  A few of my ABC backups have sped up after raising rE         VCC_MAX_IO_SIZE to 1000 blocks.    It is a dynamic parameter:n  ! $ mcr sysgen show vcc_max_io_sizeeO Parameter Name           Current    Default     Min.      Max.     Unit Dynamic O --------------           -------    -------    -------   -------   ---- -------rN VCC_MAX_IO_SIZE              1000        127         0         -1 Blocks     D    #                                 Robs  K "Even if the biblical assertion is incorrect that where there is no vision,lH  the people perish, it is difficult to think what could be the engine orJ  stimulus for social behavior in a nihilistic system committed only to theI  certainty of the passage of time, without any energetic relationship to  /  another principle or purpose."  --Lionel Tigere   ------------------------------   End of INFO-VAX 2003.359 ************************