1 INFO-VAX	Fri, 27 Sep 2002	Volume 2002 : Issue 534       Contents: Re: "inview" Article0 Re: "Underqualified tipster backed HP merger"...0 Re: "Underqualified tipster backed HP merger"... Re: %SYSTEM-F-DEVREQERR  Re: %SYSTEM-F-DEVREQERR  Re: A query   alpha hpicide come home to roost, Re: Alphaserver ES80 Technical Specification% ANN: Data Plotting Library DISLIN 8.0  Apache error_log flushing  Re: Apache error_log flushing  Re: Apache error_log flushing C Re: Builder Xcessory 6.1 Available - GUI Builder for VMS (Announce) C Re: Builder Xcessory 6.1 Available - GUI Builder for VMS (Announce) ; Re: Celera Throws Compaq Servers In Trash - Forbes Magazine ; Re: Celera Throws Compaq Servers In Trash - Forbes Magazine ; Re: Celera Throws Compaq Servers In Trash - Forbes Magazine ! Re: Changing Strings in EXE files  China clones Alpha ?% Re: Copy of OpenVMS hobbyist CD in UK  Re: DSM Docs On Line?  Errors running Oracle7 on VMS ! Re: Errors running Oracle7 on VMS  Exit(status) strategy  Re: Exit(status) strategy  Re: Exit(status) strategy  Re: Exit(status) strategy  Re: Exit(status) strategy  Re: Exit(status) strategy  Re: External Authentication  Re: External Authentication @ Infinite loop bug in TCPIP 5.3 NFS client with structure level 5D Re: Infinite loop bug in TCPIP 5.3 NFS client with structure level 5 Re: Memory for DS25  Newcomers to VMS OpenVMS is strategic Re: OpenVMS is strategic Re: OpenVMS is strategic Re: OpenVMS is strategic Re: OpenVMS is strategic PL/I for OpenVMS( Re: Port of a recent Ncurses lib for AXP  Re: Q: Running OpenVMS diskless?  Re: Q: Running OpenVMS diskless?  Re: Q: Running OpenVMS diskless?& Security problem with CSWS T1.3 Beta ?* Re: Security problem with CSWS T1.3 Beta ? Shadow set members Re: Shadow set members Re: Shadow set members Re: Shadow set members Re: Shadow set members The perils of reusable code  Re: Tomcat and CSWS (Apache) Re: UCX security advisory  Re: UCX security advisory  Re: UCX security advisory  Re: UCX security advisory  Re: UCX security advisory A Re: Unexpected <CR> and <FF> after a Setup module of a queue FORM A Re: Unexpected <CR> and <FF> after a Setup module of a queue FORM  Re: Version control on OpenVMS Re: Version control on OpenVMS Re: Version control on OpenVMS0 Re: VMS future (oh not not another one of these) Re: VMS, C++ and flush Re: warning during copy  Re: warning during copy  Re: warning during copy 2 Re: What IO performance expected on GS140 PCI bus?A Re: [mike@akitanet.co.uk: OpenVMS POP server local vulnerability] A Re: [mike@akitanet.co.uk: OpenVMS POP server local vulnerability] A Re: [mike@akitanet.co.uk: OpenVMS POP server local vulnerability] A Re: [mike@akitanet.co.uk: OpenVMS POP server local vulnerability]   F ----------------------------------------------------------------------  % Date: Fri, 27 Sep 2002 17:17:18 +0100 ' From: Andrew Harrison SUNUK Consultancy  Subject: Re: "inview" Article . Message-ID: <3D94848E.4020803@nospamn.sun.com>   Arne Vajh=F8j wrote: > Let us try and recapitulate !  >=20 > You wrote: >=204 >  >>>>> Case in point would be your suggestion that5 >  >>>>> Linux/Windows makes a cost effective web and  >  >>>>> apps server platform. >  >>>>>3 >  >>>>> There is only one general J2EE apps server 3 >  >>>>> benchmark that attempts to model this kind 6 >  >>>>> of app on a price/performance basis and thats >  >>>>> eCperf. >  >>>>>; >  >>>>> And this may be counter intuitive, eCperf actually 9 >  >>>>> shows that a typical Linux/Windows2000 with UNIX ; >  >>>>> DBMS infrastructure is more costly per transaction > >  >>>>> than say a Sun Apps and Sun DBMS server configuration7 >  >>>>> with the Sun solution delivering nearly double < >  >>>>> the throughput per CPU and using many fewer servers >=201 > Which is clear FUD when we look at the results:  >=20
 > Linux-Unix:  >   NONE >=20 > Windows-Unix:  >   #17  >=20  9 The boxes IBM used as apps servers for their Windows-UNIX 6 and Linux-Windows eCperf tests were the same. When you; look at the test results with the same version of WebSphere < Linux and Windows2000 delivered almost identical performance< with Win2K being slightly faster on a per system and per CPU basis.  > The boxes cost the same, so how do you think that substituting; Linux for Win2K with a UNIX server would change the overall ; price performance. If anything given that Win2k is actually 6 slightly faster a Linux-UNIX combination would provide/ marginally worse performance at the same price.   7 When you have something that actually proves your point 4 rather than a set of vague suppositions on your part6 then please feal free to respond. In the mean time the; argument seems rather pointless so I guess we can call it a  day.  : Incedentally I did however rather like your stretching the6 allready loose definition of FUD used by this group to3 encompass anything that you don't agree with :-):-)      regards  Andrew Harrison    ------------------------------  % Date: Fri, 27 Sep 2002 12:16:41 -0400 ; From: "Brian Tillman" <tillman_brian@notnoone.notnohow.com> 9 Subject: Re: "Underqualified tipster backed HP merger"... $ Message-ID: <3d9484c1$1@news.si.com>  K >Apparently you are movie challenged.  I was recounting the basic plot to a 0 >silly movie, of which I can't recall the title.  L You're correct, I am.  I rarely go to movies.  I never saw "The Secret to my	 Success".  --A Brian Tillman                   Internet: tillman_brian at si.com A Smiths Aerospace                          tillman at swdev.si.com = 3290 Patterson Ave. SE, MS      Addresses modified to prevent < Grand Rapids, MI 49512-1991     SPAM.  Replace "at" with "@"8        This opinion doesn't represent that of my company   ------------------------------  % Date: Sat, 28 Sep 2002 18:59:39 +0200 @ From: Arne =?ISO-8859-1?Q?Vajh=F8j?= <arne.vajhoej@mail.tele.dk>9 Subject: Re: "Underqualified tipster backed HP merger"... + Message-ID: <3D95DFFB.7000708@mail.tele.dk>    JF Mezei wrote:    > Fred Kleinsorge wrote:I >>My response is "So he lied on his resume.  So what?".  The fact that he J >>authored the opinion doesn't mean A) it was wrong, or B) that he had theH >>ability to issue such an opinion without the advice and consent of his >>employer   > M > However, that employee's credibility became nil when this was discovered so P > his recommendations lost value. It doesn't mean they were wrong, but just that > they weren't credible.    1 He may have lost credibility in the eyes of many.    Many other may not care.   That depends on on the viewer !    Arne   ------------------------------    Date: 27 Sep 2002 22:48:07 +0800, From: Paul Repacholi <prep@prep.synonet.com>  Subject: Re: %SYSTEM-F-DEVREQERR- Message-ID: <87ptuzebfc.fsf@prep.synonet.com>   , smahler@icbme01.ionics.com (smahler) writes:  E > By DECnet terminals, I meant LAT (sorry for the ambiguity), however C > I have come to learn that we are also having problems with TELNET > > connections.  The previous posts by Hoff and David have beenD > helpful.  If you are also able to offer something that may help meD > better understand the possible causes of this issue I'd appreciate > it.   D Oh, your network is farucked. Note, LAT is not the cause, it is justE reporting the errors, somethng alien to near all IP stacks. If any of C your network `experts' claim LAT or DECnet are the cause, terminate B them with extreme prejudice. Some depths of stupidity are just not worth plumbing.   E You are going to have to go over the network and find all the missing A terminators, crunched coax, screwed UTP, autoterminating Mac on T + pieces, rabid NFS servers and all the rest.   @ The NCP line counters are a big help for finding the area of theD damage. Oh, and do ALL the network. If one section has got that bad," you can bet the rest is no better.    --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------  % Date: Fri, 27 Sep 2002 12:21:15 -0400 ; From: "Brian Tillman" <tillman_brian@notnoone.notnohow.com>   Subject: Re: %SYSTEM-F-DEVREQERR$ Message-ID: <3d9485d3$1@news.si.com>  ! >By DECnet terminals, I meant LAT   F And, of course, DECnet and LAT have nothing whatsoever to do with each. other.  They're completely separate protocols. --A Brian Tillman                   Internet: tillman_brian at si.com A Smiths Aerospace                          tillman at swdev.si.com = 3290 Patterson Ave. SE, MS      Addresses modified to prevent < Grand Rapids, MI 49512-1991     SPAM.  Replace "at" with "@"8        This opinion doesn't represent that of my company   ------------------------------  % Date: Fri, 27 Sep 2002 15:58:10 +0100 ( From: Nic Clews <sendspamhere@127.0.0.1> Subject: Re: A query) Message-ID: <3D947202.BE3F515D@127.0.0.1>    Alfred Falk wrote: > ; > >>Quickly, cos I'm off to see Patrick Moore this evening,  > > : > > Helping him with his Galaxy configuration I assume :-)# > > "My god. It's full of starlets"   = Performing, if that's the right word, at a theatre in Bolton.   J > Wow!  The guy is still going!  First book I ever bought was Moore's "TheH > Boys Book of Astronomy".  I still have it.  It's almost 10 years older! > than my Digital Logic Handbook.   E So is his car, which is a Ford Prefect and _has_ done enough miles to 2 cover the distance between the Earth and the moon.  E He was talking about life on other worlds and (for the benefit of the E SETI on VMS users), why our messages are mathematical / targeted, and H why we're looking for the same, rather than something else in the searchE for intelligent life. Basic sort of stuff, but presented in Patrick's D own way. My first book by him was the "Observers book of Astronomy".   --  ? Regards, Nic Clews a.k.a. Mr. CP Charges, CSC Computer Sciences  nclews at csc dot com    ------------------------------    Date: 27 Sep 2002 22:54:05 +0800, From: Paul Repacholi <prep@prep.synonet.com>) Subject: alpha hpicide come home to roost - Message-ID: <87lm5neb5e.fsf@prep.synonet.com>   " Drop over to the register and read  2 http://www.theregister.co.uk/content/53/27309.html  - More of the positive merger synergy no doubt.    --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------  # Date: Fri, 27 Sep 2002 08:47:25 GMT . From: peter@langstoeger.at (Peter LANGSTOEGER)5 Subject: Re: Alphaserver ES80 Technical Specification 3 Message-ID: <xWUk9.33304$142.551003@news.chello.at>   R In article <3d92bf41$1@194.70.94.92>, "Paul Sharrock" <psharrock@dabs.com> writes:E >Does anyone know the ES80 specification and its approx release date.  > K >I know its 8 CPU but does it use the same 4 CPU QBB architecture of the GS % >series or is it a true 8 CPU system.   E No. It is a EV7 (MARVEL) system and is therefore completely different J to Turbolasers (8200/8400 aka GS60/GS140) and Wildfires (GS80/GS160/GS320)J but there is a Marvel which also gets the GS name (probably GS640/GS1280).  M EV7 systems are still "not out" though some friendly customers might run them J for a couple of months now (surely a NDA thingie). I expect the MARVELs to* pop up in the next couple of weeks/months.  N The ES80 is a rack made of 4 biprocessor boxes with 4 bi-IO boxes, a 16channel. hub, possibly some storage and power supplies.  F The new DS (maybe DS25) is a single one of this biprocessor boxes in a pedestal cover.   M The new GS (maybe GS640/GS1280) is made of some (8 or 16) octoprocessor boxes ) and is most likely up to 5 racks in size.   M NOTE: My info is based on COMP.OS.VMS and a OpenVMS technical day many months ? ago. Details, Dates and problems may have changed since then...    --   Peter "EPLAN" LANGSTOEGER % Network and OpenVMS system specialist  E-mail  peter@langstoeger.atP A-1030 VIENNA  AUSTRIA              I'm looking for (a) Network _and_ VMS Job(s)   ------------------------------  % Date: Fri, 27 Sep 2002 10:05:34 +0200 , From: Helmut Michels <michels@linmpi.mpg.de>. Subject: ANN: Data Plotting Library DISLIN 8.0- Message-ID: <3D94114E.8305D814@linmpi.mpg.de>   B I am pleased to announce version 8.0 of the data plotting software DISLIN.   ; DISLIN is a high-level and easy to use plotting library for C displaying data as curves, bar graphs, pie charts, 3D-colour plots, A surfaces, contours and maps. Several output formats are supported E such as X11, VGA, PostScript, PDF, CGM, WMF, HPGL, TIFF, PNG and BMP.   B The software is available for several C, Fortran 77 and Fortran 90C compilers. Plotting extensions for the interpreting languages Perl, B Python and Java are also supported for the most operating systems.  G DISLIN distributions and manuals in PDF, PostScript and HTML format are # available from the DISLIN Home Page         http://www.dislin.de    and via FTP from the server   (      ftp://ftp.gwdg.de/pub/grafik/dislin    -------------------   Helmut Michels=   Max-Planck-Institut fuer Aeronomie  Phone: +49 5556 979-334 =   Max-Planck-Str. 2                   Fax  : +49 5556 979-240 B   D-37191 Katlenburg-Lindau           Mail : michels@linmpi.mpg.de   ------------------------------    Date: 27 Sep 2002 02:20:12 -0700  From: theop@itex.je (Theo Platt)" Subject: Apache error_log flushing; Message-ID: <31ff7ddd.0209270120.f9e515@posting.google.com>   > Is there anyway you can see the current contents of the ApacheB ERROR_LOG without having to stop and restart the server ? I'm sure3 there must be some way - I'm just being a bit dumb.    Cheers   Theo   ------------------------------  % Date: Fri, 27 Sep 2002 12:23:43 +0200 : From: Karl Rohwedder <extern.karl.rohwedder@volkswagen.de>& Subject: Re: Apache error_log flushing/ Message-ID: <an1bec$lj61@doiweb4.volkswagen.de>    Theo Platt wrote: @ > Is there anyway you can see the current contents of the ApacheD > ERROR_LOG without having to stop and restart the server ? I'm sure5 > there must be some way - I'm just being a bit dumb.  >  > Cheers >  > TheoB Version 1.3 (currently on beta) has a APACHE$CONFIG command FLUSH.   --    - mit freundlichen Gruessen | with best regards    Karl Rohwedder                   iT-Ingenieurteam       Ellernbruch 11 D-38112 Braunschweig    . E.Mail:  karl.rohwedder(at)it-ingenieurteam.de   ------------------------------  # Date: Fri, 27 Sep 2002 12:30:08 GMT ) From: Jack Fortune <jcfortune3@fedex.com> & Subject: Re: Apache error_log flushing- Message-ID: <1103_1033129808@news.cis.dfn.de>   _ On Fri, 27 Sep 2002 12:23:43 +0200, Karl Rohwedder <extern.karl.rohwedder@volkswagen.de> wrote:  > Theo Platt wrote: B > > Is there anyway you can see the current contents of the ApacheF > > ERROR_LOG without having to stop and restart the server ? I'm sure7 > > there must be some way - I'm just being a bit dumb.  > > 
 > > Cheers > >  > > TheoD > Version 1.3 (currently on beta) has a APACHE$CONFIG command FLUSH. >   # 	$ @sys$manager:apache$config flush    ------------------------------  # Date: Fri, 27 Sep 2002 09:44:13 GMT . From: peter@langstoeger.at (Peter LANGSTOEGER)L Subject: Re: Builder Xcessory 6.1 Available - GUI Builder for VMS (Announce)3 Message-ID: <NLVk9.33653$142.212516@news.chello.at>   K In article <3D936248.165F92A7@ics.com>, Mark Hatch <mhatch@ics.com> writes: A >Integrated Computer Solutions, Inc. (ICS), announces the general @ >availability of Builder Xcessory PRO 6.1. This upgrade providesF >developers with the choice of using either Motif 1.2 or Motif 2.1 forI >building user interfaces for their applications. Evaluation copies of BX 7 >PRO are available free at http://www.ics.com/getbxpro.  > F >BX PRO consists of Builder Xcessory(tm), the industry's most advancedH >graphical interface builder for Motif and Java; ViewKit(tm), a flexibleG >and mature C++ reuse framework that can be tailored to accommodate the C >specific needs of individual businesses; and EnhancementPak(tm), a @ >library of 30 reusable components including business graphs andI >sophisticated user interface controls. More detailed product information % >is available at http://www.ics.com/.   D Am I blind or why do you post this info to comp.os.vms, when you are" not offering BX Pro for OpenVMS ?    Title:; Tools and services for professional UNIX & Linux developers    BX Pro Platform Availability: 
 No OpenVMS   Download, GUI Builders: 
 No OpenVMS   Download, Libraries:2 OpenVMS is selectable, but then comes the message:O "We regret that the evaluation for your platform is not available via download"    --   Peter "EPLAN" LANGSTOEGER % Network and OpenVMS system specialist  E-mail  peter@langstoeger.atP A-1030 VIENNA  AUSTRIA              I'm looking for (a) Network _and_ VMS Job(s)   ------------------------------   Date: 27 Sep 2002 16:01:14 GMT! From: Mark Hatch <mhatch@ics.com> L Subject: Re: Builder Xcessory 6.1 Available - GUI Builder for VMS (Announce)' Message-ID: <3D9480C3.E7E58EAE@ics.com>   T No, you're not blind. We screwed up our webpage in the last revision and dropped theR link to OpenVMS in the product description pages. Thanks for spotting it! I'll get it corrected ASAP.   Mark   ---    Peter LANGSTOEGER wrote:  M > In article <3D936248.165F92A7@ics.com>, Mark Hatch <mhatch@ics.com> writes: C > >Integrated Computer Solutions, Inc. (ICS), announces the general B > >availability of Builder Xcessory PRO 6.1. This upgrade providesH > >developers with the choice of using either Motif 1.2 or Motif 2.1 forK > >building user interfaces for their applications. Evaluation copies of BX 9 > >PRO are available free at http://www.ics.com/getbxpro.  > > H > >BX PRO consists of Builder Xcessory(tm), the industry's most advancedJ > >graphical interface builder for Motif and Java; ViewKit(tm), a flexibleI > >and mature C++ reuse framework that can be tailored to accommodate the E > >specific needs of individual businesses; and EnhancementPak(tm), a B > >library of 30 reusable components including business graphs andK > >sophisticated user interface controls. More detailed product information ' > >is available at http://www.ics.com/.  > F > Am I blind or why do you post this info to comp.os.vms, when you are# > not offering BX Pro for OpenVMS ?  >  > Title:= > Tools and services for professional UNIX & Linux developers  >  > BX Pro Platform Availability:  > No OpenVMS >  > Download, GUI Builders:  > No OpenVMS >  > Download, Libraries:4 > OpenVMS is selectable, but then comes the message:Q > "We regret that the evaluation for your platform is not available via download"  >  > -- > Peter "EPLAN" LANGSTOEGER ' > Network and OpenVMS system specialist  > E-mail  peter@langstoeger.atR > A-1030 VIENNA  AUSTRIA              I'm looking for (a) Network _and_ VMS Job(s)   ------------------------------  % Date: Fri, 27 Sep 2002 07:05:15 -0400 2 From: rdeininger@mindspring.com (Robert Deininger)D Subject: Re: Celera Throws Compaq Servers In Trash - Forbes MagazineJ Message-ID: <rdeininger-2709020705150001@1cust36.tnt2.nashua.nh.da.uu.net>  5 In article <3D93EF70.8D8BD808@videotron.ca>, JF Mezei % <jfmezei.spamnot@videotron.ca> wrote:     K >One option would have been to sticj with Alpha for a few more years before  >making a decision. However: > N >Consider that EV7 is incompatible with existing Wildfires. So Celera would beI >stuck with EV68s  in the "slow" Wildfires. Will HP commit to continue to L >produce  CPU modules for wildfires ? What if adding additional modules justO >doesn't give Celera the performance they need ? What if EV7 is really the only K >way to up the computing capacity significantly for the next 2-3 years ? It B >would mean a large investment to buy marvels and ditch wildfires.  E I don't know that I ever saw it spelled out, but I assumed the Celera F systems were based on ES40, not on Wildfire.  For a massively parallelJ computing problem, Wildfires are much less attractive systems.  There wereH 1 or 2 speedups in the ES40 line, but I don't expect any more since ES45G is the upgrade path. The more recent alpha supercomputing installations J have been based on ES45 bulding blocks, but they were not available yet in the Celera time frame.  O >If you know that you will eventually need to move away from Tru64, better stop J >investing in that platform and start looking for a permanent replacement.  ) Yes, SC must be a hard sell for HP now.     I >> On the other hand, it may be partly due to money and bragging rights.   > M >Yep, and I bet IBM will have no problems bragging about Celera and its super O >computers because they will want to attract as many customers to that platform B >as possible, something which Compaq didn't want to do with Alpha.  F The nasty rumor about all these big SC wins is that the companies haveF little or no profit margin.  The whole reason for the sale is braggingI rights.  HP likely can't do that (yet, if ever) with Itanium, and there's H really no reason anymore to push alpha boxes into SC sales if there's noF profit.  OTOH, Power4 is at a point in its lifecycle where it might be< worthwhile to lose money selling high-profile SC "clusters".   ------------------------------  % Date: Fri, 27 Sep 2002 14:27:09 -0000 - From: wspencer@ap.nospam.org (Warren Spencer) D Subject: Re: Celera Throws Compaq Servers In Trash - Forbes Magazine5 Message-ID: <92966D1D8warrenspencer1977@216.168.3.30>   5 Here's the corresponding article from The Register...    Warren Spencer --    0 IBM, EMC win big Applera HPC server storage deal By ComputerWire  Posted: 09/27/2002 at 06:26 EST        D IBM Corp and EMC Corp were yesterday awarded contracts with Applera G Corporation, the parent company of Celera Genomics Group which is best  H known as the cracker of the human genetic code, Timothy Prickett Morgan H writes. Celera Genomics has been a long-time user of AlphaServer server K clusters from the former Compaq Computer Corp, now part of Hewlett Packard  + Co, and Symmetrix storage arrays from EMC.    J IBM is understandably ebullient about the 2 teraflops supercomputer based E on clusters of its pSeries 690 Regatta-H Power4-based servers, which  J Applera will be installing to support Celera's computing needs as well as K those of two other units of the company. The IBM servers will be replacing  H the installed base of AlphaServers at Applera's units, but the machines J fall far short of the 100 teraflops monster supercomputer that Compaq and K Celera Genomics were going to build with the help of Sandia National Labs,  8 which is administered by the U.S. Department of Energy.   J Part of the reason that Applera can get by with a smaller supercomputer - J if one can call a 2 teraflops supercomputer small - is that it has in the I past year been transforming itself from being a genome-cracker to a drug  F developer, with the drugs created based on its knowledge of the human H genome and on its substantial understanding of how to use computers for  molecular modeling.   D Earlier this summer, Celera quietly backed out of the Compaq-Sandia K partnership, and according to a Celera spokesperson, this was done because  K of changes in Celera's business model and because of changes in the Compaq  D product lines following the merger of Compaq and HP. Celera did not L elaborate further, but it seems clear that Celera felt that if HP was going J to push AlphaServer-Tru64 customers to HP-UX in the long term, given that J its business model was changing anyway, it was a good time to open up the  bidding for a new HPC cluster.    H At about the same time this was going on, Sandia inked a $90m deal with K Cray, Inc, the Seattle-based supercomputer maker under its "Red Storm" 100  G teraflops supercomputer contract, which is being funded by the DOE and  H which had originally placed their order with Compaq in association with I Celera back in January 2001. In July 2002, when Cray nosed in on the Red  K Storm contract from Sandia, we were given the impression that this did not  D necessarily spell doom for HP in pursuit of the Red Storm contract.   J Sources at Sandia say that while the final negotiations for the Red Storm L contract between Sandia and Cray are not done, an announcement awarding the K deal to Cray is impending within the next few weeks and HP is not going to  L be awarded any portion of the Red Storm contract. Neither are IBM Corp, Sun J Microsystems, SGI, by the way. Unless something goes terribly wrong, this  will be an all Cray deal.   I HP has the changes at Celera, the changes forced upon Compaq to kill off  C the Alpha processors in favor of Itanium chips, and the increasing  J aggressiveness of Cray to thank for its being bumped out of the Red Storm C contract. This is a big deal, both in terms of money and prestige.    L Red Storm calls for the 100 teraflops machine to be operational by 2004 and I could account for hundreds of millions of dollars in hardware, software,  E storage, and services over the long haul. That said, HP is still the  L dominant supplier of HPC equipment, because of its strength in the midrange L HPC market, even if other vendors may get the glory - however fleeting - of K building the world's fastest supercomputer. HP takes down plenty of deals.    L For instance, HP just closed an order for a 9.2 teraflops supercomputer for L the DOE's Pacific Northwest national Laboratory that will be based on 1,400 H future Itanium processors, a machine that is worth $24.5m. HP also this I week announced a $22m deal at the Wellcome Trust Sanger Institute in the  L U.K. for a $22m AlphaServer cluster for genomics research that is comprised K of 38 ES45s and four ES40 four-way AlphaServers, two DS20 two-way servers,  J and a single 32-way GS320 server, including three times the disk capacity  of the original configuration.    E IBM, of course, is no slouch when it comes to HPC. Machines like the  J Power4-based Regattas are designed specifically to take market share away J from HP-Compaq in the HPC market, and it is not surprising that IBM could F win the Applera deal and boot AlphaServer equipment from the account, E regardless of the historical use of Compaq gear at Applera's main IT  $ resource consumer, Celera Genomics.   I IBM has shown a willingness to compete aggressively on price to get into  I new and high-profile accounts, and it undoubtedly did so at Applera. IBM  F says that the dozen 32-way pSeries 690 servers alone (which have been K configured with only 64GB of memory apiece) have a list price of $25m, and  J the 150TB of disk capacity that Applera is acquiring from EMC is probably L worth somewhere in the neighborhood of between $15m and $20m at list price. H Considering the size of the deal, the poor economic conditions, and the H desire to retain account control for EMC (you can bet IBM bid its Shark J arrays as part of the deal) and to get into the account for IBM, it seems L reasonable that the core hardware was acquired at around 50% of list price.   K Applera did not say what software it was installing on the machine, but it  H is paying IBM for migration services and a multi-year services contract D that will bring big bucks to IBM. The Regatta machines will run the J forthcoming AIX 5.2 operating system (due to be announced in a few weeks) K and will be clustered using IBM's latest "Colony" SP switching technology.  H This Regatta cluster will provide data processing support for Applera's K Celera Genomics, Applied Biosystems (which sells the human genome data and  A other products and raked in $1.6bn in sales in 2001), and Celera  L Diagnostics (which makes medical equipment) units. The IBM machines will be 1 located in Celera's Rockville, Maryland offices.    J IBM also said that Applera and Big Blue are in preliminary discussions to I develop life sciences products, although Peter Ungara, vice president of  J HPC at IBM, was not able to say what these products would be or who would = sell them because discussions are only now getting underway.      ComputerWire     ------------------------------  + Date: Fri, 27 Sep 2002 05:15:28 -0700 (PDT) . From: Fabio Cardoso <fabiopenvms@yahoo.com.br>D Subject: Re: Celera Throws Compaq Servers In Trash - Forbes Magazine? Message-ID: <20020927121528.9383.qmail@web20208.mail.yahoo.com>    What shame !!!  , It is related to the Alpha and Tru64 ending,, So obvious ! Nobody want "legacy" equipment.   Regards    FC=20 & --- John Smith <a@nonymous.com> wrote:/ > http://www.forbes.com/2002/09/26/0926cra.html  >=20' > Celera Throws Compaq Servers In Trash & > Matthew Herper, 09.26.02, 3:50 PM ET >=20, > When scientists at Celera Genomics and the  > government-funded Human Genome4 > Project created their maps of all human genes, one > indisputable winner was * > Compaq Computer. Both groups used Compaq > supercomputers. Today, Celera 6 > announced that it would be parting with those Compaq > servers in favor of new  > ones built by IBM and EMC. >=201 > IBM will provide Celera and its sister company,  > Applied Biosystems, with a. > supercomputing cluster capable of 2 trillion > calculations per second. EMC4 > will supply automated data storage for the system. > The financial value of3 > the contract was not revealed, and Celera did not  > name the firms that bid # > against IBM and EMC for the deal.  >=203 > The contract is an obvious victory for IBM's life  > sciences division, which5 > has spent the past two years establishing itself as  > a force in4 > supercomputing for drug and biotech companies. But > Celera is not the jewel 6 > it once was, and its stated reasons for choosing IBM > reveal the difficulty , > of its new mission: to use its status as a > gene-sequencing powerhouse to 5 > boost itself into a bona fide manufacturer of drugs  > and medical diagnostics. >=204 > Making that switch will be difficult. As investors > have realized just how2 > difficult, Celera and such competitors as Incyte > Genomics and Curagen have 3 > seen their stocks plummet by more than 75% in the  > last 12 months. Celera, 6 > with $908 million in cash and its scientific cachet, > may have the best 5 > chance of transforming itself, but it has been hurt  > by the departure of 0 > founder Craig Venter, who has started up a new > gene-sequencing shop of his  > own. >=203 > "I frankly think that Celera is still the company  > that mapped the human 6 > genome," says Caroline Kovac, general manager of IBM > Life Sciences. "It is a 3 > company that is extremely well-positioned, having  > done that, to incorporate , > genomics into the drug discovery process." >=201 > Still, the way Celera chose IBM illustrates the  > Rockville, Md.-based6 > genomics firm's growing pains. When Celera signed up > Compaq in 1998, the / > company put a great deal of importance on the  > ability of various5 > supercomputing systems to run a single gene-mapping  > program. Now, says John 4 > Reynders, vice president of informatics at Celera, > it has much more diverse4 > computing needs. "The constellation of things that > we're doing is very  > rich," he says.  >=206 > Among the factors that made IBM and EMC Celera's top > choice is IBM's 3 > experience migrating data from old systems to new  > ones. Another factor is 3 > that IBM will have to integrate two firms. Celera  > and Applied Biosystems are6 > both tracking stocks for a larger corporation called > Applera. Lately,4 > Applied has taken over Celera's original business, > selling gene-sequence / > subscriptions to drug companies, while Celera  > focuses on developing its own  > products.  >=20. > An issue that also may have played a role in > Celera's decision: Now that 4 > Compaq has been acquired by Hewlett-Packard, it is > in flux. To add more, > questions, Compaq will at some point start# > outsourcing chip design to Intel. 6 > "I think there's no question that there is no Compaq > anymore, and Compaq has . > issues with the chip, HP has issues with the > platform," says Kovac. "But 3 > it's not as if Celera didn't have other choices."  >=201 > The question for Celera, however, is not merely  > whether it made the right 0 > choice in picking IBM. Even with its rich cash > position, the company needs a 1 > real line of revenue-producing products. And its > needs to develop them as > quickly as possible. >=20 > ----- 5 > There are a number of other articles from differente > sources that place( > somewhat different spin on this event. >=202 > Suffice it to say that with Tru64 dead/dying and > Alpha dying, CeleraT2 > certainly had ample reason to look ahead and see > what was viable *for them*5 > over the course of the porting work they have aheadr > of themselves to movel  > from Tru64/Alpha to AIX/Power. >=206 > On the other hand, it may be partly due to money and > bragging rights. IBM- > lost many battles like this to Alpha in the " > pharmaceutical industry over the. > past several years. Now with their only real! > competition (Alpha) dead/slowlys5 > dying, perhaps it's simply a case of IBM willing toX > step up to the plate and& > make offers that are hard to refuse. >=20* > Personally, I don't expect many high-end  > pharmaceutical wins for Itanic0 > anytime in the next 3-4 years. So long to that > business stream HP.s >=20 >=20 >=20 >=20 >=20     =3D=3D=3D=3D=3DxL =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D  F=E1bio dos Santos Cardoso OpenVMS System Manager Rio de Janeiro - Brazilc fabiopenvms@yahoo.com.brL =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3Dc  2 __________________________________________________ Do you Yahoo!?) New DSL Internet Access from SBC & Yahoo!  http://sbc.yahoo.com   ------------------------------  # Date: Fri, 27 Sep 2002 16:56:58 GMTn4 From: Tim Llewellyn <tim.llewellyn@blueyonder.co.uk>* Subject: Re: Changing Strings in EXE files0 Message-ID: <3D948BCD.FD9D9805@blueyonder.co.uk>   David Froble wrote:a >  > Tim Llewellyn wrote: >  > >t > R > > Given the current propensity for companies to change ownership/name, shouldn'tM > > best practice be to make the company name a parameter for the application F > > (say a logical name translated at runtime)? Or would that throw up > > problems of its own? > P > Been doing that for 25 years!  Codes file look-up.  Many, many parameters.  Is > there any other way?  S Well, I guess rebuilding the application from code or patching the executables, ando then retesting :-).    >  > Dave   -- e tim.llewellyn@blueyonder.co.uk a  H * PLEASE NOTE tim.llewellyn@cableinet.co.uk address is NO LONGER VALID *   ------------------------------  % Date: Fri, 27 Sep 2002 17:22:23 -0000e- From: wspencer@ap.nospam.org (Warren Spencer)k Subject: China clones Alpha ?e5 Message-ID: <92968AC57warrenspencer1977@216.168.3.30>r  ) I guess Bob C was right - Alpha lives on?.  L "Other sources tell us that China has also successfully built clones of the % DEC, now Intel Alpha microprocessor."-  8 Full story at:  http://www.theinquirer.net/?article=5609   ws   -- e   Warren Spencer' Senior Software Engineer (not a writer)U The Associated Press  < ** Time flies like an arrow.  Fruit flies like a bananna. **   ------------------------------  % Date: Fri, 27 Sep 2002 09:45:12 +0200 0 From: Alex Leibovici <aleibovici@compuserve.com>. Subject: Re: Copy of OpenVMS hobbyist CD in UK8 Message-ID: <0s28puk12upgv2fprvcrr9ovfo272fhl56@4ax.com>  C On 2 Sep 2002 05:58:57 -0700, issinoho@slayme.com (issinoho) wrote:   S >Have you broadband? If so, I can put up an image for you to grab on my FTP server.n  D Please, can I have tale a copy from your FTP server too ? Just got a? 3-month trial licence for Charon-VAX from an exhibition here ina
 Switzerland. 2   Thank in advance !   Alex   ------------------------------    Date: 27 Sep 2002 11:28:51 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen)> Subject: Re: DSM Docs On Line?3 Message-ID: <+ZZy4OyX7984@eisner.encompasserve.org>n  W In article <up7if2t7oc4m03@news.supernews.com>, "John Vottero" <John@mvpsi.com> writes: ? > I'm no MUMPS expert but I think you can find what you need at  > http://www.intersystems.com/ > - > I believe that DSM has morphed into Cache'.e  = I was under the impression that Cache' is a different producthI and that InterSystems still supports DSM (but does not add new features).c   ------------------------------  % Date: Fri, 27 Sep 2002 09:46:25 -0700U7 From: "Robert Cervantez" <rob@DONOTSPAM.symcas-tsg.com>t& Subject: Errors running Oracle7 on VMS% Message-ID: <3d948590_1@news.vic.com>   J Not sure if this makes much sense to anyone, but I'm wondering if this log helps..N  0 This is an alphaserver 2100A running VMS 6.2-1H3  L Apparently, users reported the server to run very slowly whenever one of theK Oracle tables wouldn't open.  The cause for the tables not opening may havelA something to do with the following logs (given to me by he OraclelH consultant).  I was unable to come up with the cause of the error on the( system, given the VMS Error codes below:   *** SESSION ID:(3.1)K Could not a write message to alert file, probably too busy. (VMS error was:eJ 98954): The message will be written to this trace file instead: Checkpoint not complete  J 19-SEP-2002 07:40:42.50: Could not a write message to alert file, probablyK too busy. (VMS error was: 98954): The message will be written to this trace$
 file instead: L   Current log# 2 seq# 107998 mem# 0: DISK$USER:[ORACLE7.DB_CRIS]ORA_LOG2.RDO   19-SEP-2002 07:41:29.99:K Could not a write message to alert file, probably too busy. (VMS error was:vH 98954): The message will be written to this trace file instead: Thread 1 advanced to log sequence 107999    19-SEP-2002 07:42:12.21:K Could not a write message to alert file, probably too busy. (VMS error was:r? 98954): The message will be written to this trace file instead: L   Current log# 1 seq# 107999 mem# 0: DISK$USER:[ORACLE7.DB_CRIS]ORA_LOG1.RDO   19-SEP-2002 07:42:55.83:K Could not a write message to alert file, probably too busy. (VMS error was:?H 98954): The message will be written to this trace file instead: Thread 1( cannot allocate new log, sequence 108000      Any help is greatly appreciated!   ------------------------------    Date: 27 Sep 2002 10:50:22 -07001 From: nothome@spammers.are.scum (Malcolm Dunnett)h* Subject: Re: Errors running Oracle7 on VMS- Message-ID: <wppl$7+J0YB0@malvm7.mala.bc.ca.>c  % In article <3d948590_1@news.vic.com>,n<    "Robert Cervantez" <rob@DONOTSPAM.symcas-tsg.com> writes:  L > Not sure if this makes much sense to anyone, but I'm wondering if this log	 > helps..m > 2 > This is an alphaserver 2100A running VMS 6.2-1H3 > N > Apparently, users reported the server to run very slowly whenever one of theM > Oracle tables wouldn't open.  The cause for the tables not opening may have-C > something to do with the following logs (given to me by he OracleeJ > consultant).  I was unable to come up with the cause of the error on the* > system, given the VMS Error codes below: >  > *** SESSION ID:(3.1)M > Could not a write message to alert file, probably too busy. (VMS error was:e
 > 98954):   &     $write sys$output f$message(98954)5     %RMS-E-FLK, file currently locked by another userD  ?   Some other process has the alert log file open. This is not a B normal situation. This file should be in the [.trace] subdirectoryG under the ORA_DB: directory. It should be called <host>_<sid>_alert.lognH ( where <host> is your host name and <sid> is the SID of the instance ).  H   Are you trying to run more than one instance of Oracle? If so, you mayF have it misconfigured ( eg same SID on both ) so that they are sharing files they shouldn't.m  I   Try doing a $SHO DEV/FILES on the relevant disk and look to see who hasn that file open.c   ------------------------------  % Date: Fri, 27 Sep 2002 04:29:02 -0400v- From: JF Mezei <jfmezei.spamnot@videotron.ca>  Subject: Exit(status) strategy, Message-ID: <3D9416B8.2E3DCB16@videotron.ca>   Ok, I have a C program.a@ I want it to set $STATUS to indicate status, or various errors.    If I   exit(2008) ;   U then the program ends , but DCL then says "%SYSTEM-W-NOMSG - Message number 000007D8"   L Is there a way to exit nicely, set $STATUS to a value but not get DCL to try to say anything ?   M I don't really want to go through the trouble of defining messages (I display G the error message on screen, but want to make it possible for a command F procedure to check the status in cases where this would be automated).  J Also, is there a "safe" range of exit status I can use without interfering with DCL etc ?  F This is for a foreign command/utility, so I want it to be as simple as. possible (no SET COMMAND, no SET MESSAGE etc).   ------------------------------  % Date: Fri, 27 Sep 2002 12:22:00 +0200 - From: "Bert de Geus" <bert.de.geus@xs4all.nl>l" Subject: Re: Exit(status) strategy* Message-ID: <an1bp1$t70$1@news1.xs4all.nl>  : "JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in message& news:3D9416B8.2E3DCB16@videotron.ca... > Ok, I have a C program.-A > I want it to set $STATUS to indicate status, or various errors.0 >  > If I   exit(2008) ;a >rF > then the program ends , but DCL then says "%SYSTEM-W-NOMSG - Message number 000007D8" > J > Is there a way to exit nicely, set $STATUS to a value but not get DCL to try  > to say anything ?s >lG > I don't really want to go through the trouble of defining messages (Ie display I > the error message on screen, but want to make it possible for a commandiH > procedure to check the status in cases where this would be automated). >yL > Also, is there a "safe" range of exit status I can use without interfering > with DCL etc ? >aH > This is for a foreign command/utility, so I want it to be as simple as0 > possible (no SET COMMAND, no SET MESSAGE etc).  : Add bit 31 to the exit value. This suppresses the message.C Please remember that bits 0-2 represent the severity level. The DCL ! procedure reacts on these values:a    0    Warningn     1    Success     2    Error     3    Information     4    Severe (Fatal) errorc  I This behaviour of DCL can be influenced by the SET ON, SET NOON, ON ERROR ! ..., ON WARNING ... etc commands.   
 Greetings,   Bert   ------------------------------  % Date: Fri, 27 Sep 2002 06:28:05 -0400a) From: "Neil Rieck" <n.rieck@sympatico.ca>R" Subject: Re: Exit(status) strategy: Message-ID: <dpWk9.9673$l44.1841257@news20.bellglobal.com>  L First off, the lowest 3 bits control the DCL error response (odd numbers are cool) like so:  
 0 -w- warning0
 1 -s- successo 2 -e- errore 3 -i- informationalu 4 -f- fatal  5 -?-p 6 -?-  7 -?-   L In progams that I write to communicate with DCL, I use the following scheme:   main(){a     ... &     exit_code = 2;    // 1=ok, 2=error5     exit_reason = 9; // can be any user defined valuehL     exit((exit_reason * 32768) + exit_code);    // this is placed in $STATUS }.  # Then in DCL we need something like:a   $say := write sys$output $run my_proegam.exea $my_status = f$integer($STATUS)   $if ((my_status .and. 1) .ne. 1) $thenp  $    my_reason = my_status/32768< $    say "program exited with error for reason: ", my_reason $    if (my_reason .eq. 1)	 $    thend $!        do something
 $    endif $endif  
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.8 http://www3.sympatico.ca/n.rieck/links/cool_vax_vms.html  : "JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in message& news:3D9416B8.2E3DCB16@videotron.ca... > Ok, I have a C program.aA > I want it to set $STATUS to indicate status, or various errors.u >h > If I   exit(2008) ;l >gF > then the program ends , but DCL then says "%SYSTEM-W-NOMSG - Message number 000007D8" >dJ > Is there a way to exit nicely, set $STATUS to a value but not get DCL to tryI > to say anything ?o >oG > I don't really want to go through the trouble of defining messages (Ib displayrI > the error message on screen, but want to make it possible for a commandrH > procedure to check the status in cases where this would be automated). >eL > Also, is there a "safe" range of exit status I can use without interfering > with DCL etc ? > H > This is for a foreign command/utility, so I want it to be as simple as0 > possible (no SET COMMAND, no SET MESSAGE etc).   ------------------------------    Date: 27 Sep 2002 07:55:35 -0600 From: briggs@encompasserve.org" Subject: Re: Exit(status) strategy3 Message-ID: <l7Iq++ziQwc4@eisner.encompasserve.org>E  \ In article <3D9416B8.2E3DCB16@videotron.ca>, JF Mezei <jfmezei.spamnot@videotron.ca> writes: > Ok, I have a C program.eB > I want it to set $STATUS to indicate status, or various errors.  >  > If I   exit(2008) ;  > W > then the program ends , but DCL then says "%SYSTEM-W-NOMSG - Message number 000007D8"n > N > Is there a way to exit nicely, set $STATUS to a value but not get DCL to try > to say anything ?   ? Set the STS$V_INHIBMSG bit in the status code.  This is the low-A order bit of the high order nibble of the 32 bit code.  Bit 28 if  you count from zero.   For example:  
 $ exit %xc< %SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual : address=000000000000000C, PC=000000007AF82950, PS=00000000 $ exit %x1000000ct $   F Then you can just mask that bit out when examining the resulting code.  E If you want to play by the rules, you should define your own facilityrB number in the appropriate range and set bits 16-27 accordingly and: then decide if your codes are facility-specific or not and set bit 15 accordingly.b  D And you could even build a message file for your custom status codes5 and let the system print some error messages for you.-   	John Briggs   ------------------------------  % Date: Fri, 27 Sep 2002 08:34:14 -0400 2 From: Atlant Schmidt <atlantnospam@mindspring.com>" Subject: Re: Exit(status) strategy. Message-ID: <3D945046.F368B73C@mindspring.com>   JF Mezei wrote:r   > Ok, I have a C program. A > I want it to set $STATUS to indicate status, or various errors.  >  > If I   exit(2008) ;t >sW > then the program ends , but DCL then says "%SYSTEM-W-NOMSG - Message number 000007D8"  >tN > Is there a way to exit nicely, set $STATUS to a value but not get DCL to try > to say anything ?e  , Well, you could "succeed" instead of "fail".  * Remember that VMS is *VERY* different from/ Unix in its use of $STATUS values. On Unix, "0"-5 is usually success and all othe rvalues are failures,w- but on VMS, the low three bits of the $STATUSn- value represent the "severity" where any evena- value represents some sort of failure. Here's & Neil Rieck's table from another reply:   > 0 -w- warning  > 1 -s- successh
 > 2 -e- error  > 3 -i- informational 
 > 4 -f- fatal  > 5 -?-L > 6 -?-- > 7 -?-   , So if "2008" is meant to imply success, then+ you'll have some conflicts with VMS. "2009"e" would be a better "success" value.   Atlant   ------------------------------  % Date: Fri, 27 Sep 2002 12:39:42 -0400-; From: "Brian Tillman" <tillman_brian@notnoone.notnohow.com>k" Subject: Re: Exit(status) strategy$ Message-ID: <3d948a26$1@news.si.com>   >If I   exit(2008) ; >8L >then the program ends , but DCL then says "%SYSTEM-W-NOMSG - Message number	 000007D8"w >eI >Is there a way to exit nicely, set $STATUS to a value but not get DCL to  tryh >to say anything ?  ? Of course.  Set the "no display" bit.  Your value would then bes* exit(268437464) (i.e., %x100007D8).  Look:   $ x= %x000007d8 	 $ exit 'x ( %SYSTEM-W-NOMSG, Message number 000007D8 $ x= %x100007d8o	 $ exit 'x  $P --A Brian Tillman                   Internet: tillman_brian at si.com A Smiths Aerospace                          tillman at swdev.si.com = 3290 Patterson Ave. SE, MS      Addresses modified to preventR< Grand Rapids, MI 49512-1991     SPAM.  Replace "at" with "@"8        This opinion doesn't represent that of my company   ------------------------------  % Date: Fri, 27 Sep 2002 10:02:33 +0200t, From: "Rainer Giese" <waste.not@welcome.net>$ Subject: Re: External Authentication5 Message-ID: <an13ag$9snle$1@ID-138444.news.dfncis.de>   > "Shael Richmond" <ksrich@bellsouth.net> schrieb im Newsbeitrag' news:3D9397C5.D679EA67@bellsouth.net... ? > I saw that you can load the external authentication componenttG > of Advanced Server but it still seems to require full Advanced Servert; > on one node of the cluster.  I really didn't want to loadr > Advanced Server.  0 No experience, but from AS installation manual :  I "A cluster member that is not running the complete Advanced Server can be F configured to authenticate logon requests from network users if it hasE access to external authentication software on a shared cluster disk."s  @ In my memory I remember an installation option "install external authentication only".o   Rainer Giese   ------------------------------  # Date: Fri, 27 Sep 2002 12:02:30 GMT + From: "Rick Barry" <barry@star.zko.dec.com>o$ Subject: Re: External Authentication0 Message-ID: <qNXk9.6$XW5.56376@news.cpqcorp.net>  L External authentication requires an Advanced Server installation on at leastL one node of the cluster, running the "logon" service. Advanced Server is theJ communication channel into the Microsoft user domain database and servicesL the actual authentication requests or passes them off to a Domain Controller elsewhere on the network.   
 Rick Barry OpenVMS System Software Group- Hewlett-Packard Companya
 Nashua, NH  8 "Shael Richmond" <ksrich@bellsouth.net> wrote in message' news:3D9397C5.D679EA67@bellsouth.net...U= > Is there anyway besides loading Advanced Server to have VMSs9 > user names/passwords use the NT LAN username/passwords?n >e? > I saw that you can load the external authentication componentsG > of Advanced Server but it still seems to require full Advanced Servera; > on one node of the cluster.  I really didn't want to load2 > Advanced Server. >a< > I am running VMS 7.3-1, but most of my sites are at 7.2-1. >o > 	 > Thanks,  >o > Shael  >e   ------------------------------  % Date: Fri, 27 Sep 2002 14:16:02 +0100 % From: Alan Greig <a.greig@virgin.net>tI Subject: Infinite loop bug in TCPIP 5.3 NFS client with structure level 5c8 Message-ID: <qpl8pusog3fnpu0ogm3qlo0egcetl45mht@4ax.com>  D Is there an ECO out for TCPIP 5.3 yet? If so any chance it fixes the+ following problem with /structure-5 mounts.     tcpip sho ver  7   Compaq TCP/IP Services for OpenVMS Alpha Version V5.3a4   on a AlphaStation 400 4/233 running OpenVMS V7.3-1   /structure=2  / $ dir/siz/date     dnfs1:[weealf_manman_backup]n  & Directory DNFS1:[WEEALF_MANMAN_BACKUP]  5 ALAN.BCK;3            216468  26-SEP-2002 13:05:15.40t5 ALAN.BCK;2            110250  26-SEP-2002 11:16:49.69n5 ALAN.BCK;1                 0  26-SEP-2002 11:16:14.39 5 MANMAN.BCK;1         2419137  25-SEP-2002 14:51:56.20v5 MANMAN.DIR;1               1  25-SEP-2002 12:11:46.64 5 SADMAN.BCK;2            4788  25-SEP-2002 13:45:08.33X5 SADMAN.BCK;1           45612  25-SEP-2002 13:42:18.19y5 ZLAN.BCK;3           1054976  27-SEP-2002 13:52:11.40O5 ZLAN.BCK;2           1005102  27-SEP-2002 13:44:13.66a5 ZLAN.BCK;1            178983  27-SEP-2002 10:46:06.06   " Total of 10 files, 5035317 blocks.  4 $ dir/siz/date  DNFS1:[weealf_manman_backup]alan.bck  & Directory DNFS1:[WEEALF_MANMAN_BACKUP]  5 ALAN.BCK;3            216468  26-SEP-2002 13:05:15.40A5 ALAN.BCK;2            110250  26-SEP-2002 11:16:49.69M5 ALAN.BCK;1                 0  26-SEP-2002 11:16:14.39,    Total of 3 files, 326718 blocks. $i    + But now the same thing mounted /structure=5:    & Directory DNFS1:[weealf_manman_backup]  5 alan.bck;3            216468  26-SEP-2002 13:05:15.4015 alan.bck;2            110250  26-SEP-2002 11:16:49.69F5 alan.bck;1                 0  26-SEP-2002 11:16:14.39v5 manman.bck;1         2419137  25-SEP-2002 14:51:56.20o5 manman.DIR;1               1  25-SEP-2002 12:11:46.64 5 sadman.bck;2            4788  25-SEP-2002 13:45:08.33t5 sadman.bck;1           45612  25-SEP-2002 13:42:18.19k5 zlan.bck;3           1054976  27-SEP-2002 13:52:11.40 5 zlan.bck;2           1005102  27-SEP-2002 13:44:13.66t5 zlan.bck;1            178983  27-SEP-2002 10:46:06.06a  " Total of 10 files, 5035317 blocks.   But nowy  4 $ dir/siz/date  DNFS1:[weealf_manman_backup]alan.bck  & Directory DNFS1:[weealf_manman_backup]  5 alan.bck;3            216468  26-SEP-2002 13:05:15.40.5 alan.bck;3            216468  26-SEP-2002 13:05:15.40a5 alan.bck;3            216468  26-SEP-2002 13:05:15.40W5 alan.bck;3            216468  26-SEP-2002 13:05:15.40W5 alan.bck;3            216468  26-SEP-2002 13:05:15.40 5 alan.bck;3            216468  26-SEP-2002 13:05:15.40y5 alan.bck;3            216468  26-SEP-2002 13:05:15.40a5 alan.bck;3            216468  26-SEP-2002 13:05:15.40E5 alan.bck;3            216468  26-SEP-2002 13:05:15.40    and so on for ever -- Alan   ------------------------------  % Date: Fri, 27 Sep 2002 11:05:51 -0400 5 From: David Beatty <David.Beatty@qwertysasasdfgh.com>mM Subject: Re: Infinite loop bug in TCPIP 5.3 NFS client with structure level 5t2 Message-ID: <F3OUPWQKc06bKAVLLTkyoJ4UV3Nm@4ax.com>  % Yes, ECO 1 is available and there areh quite a few NFS client fixes.C   David R> Beatty   C On Fri, 27 Sep 2002 14:16:02 +0100, Alan Greig <a.greig@virgin.net>t wrote:  E >Is there an ECO out for TCPIP 5.3 yet? If so any chance it fixes the , >following problem with /structure-5 mounts. >a > tcpip sho vera > 8 >  Compaq TCP/IP Services for OpenVMS Alpha Version V5.35 >  on a AlphaStation 400 4/233 running OpenVMS V7.3-1  >r
 >/structure=2  > 0 >$ dir/siz/date     dnfs1:[weealf_manman_backup] >s' >Directory DNFS1:[WEEALF_MANMAN_BACKUP]O >P6 >ALAN.BCK;3            216468  26-SEP-2002 13:05:15.406 >ALAN.BCK;2            110250  26-SEP-2002 11:16:49.696 >ALAN.BCK;1                 0  26-SEP-2002 11:16:14.396 >MANMAN.BCK;1         2419137  25-SEP-2002 14:51:56.206 >MANMAN.DIR;1               1  25-SEP-2002 12:11:46.646 >SADMAN.BCK;2            4788  25-SEP-2002 13:45:08.336 >SADMAN.BCK;1           45612  25-SEP-2002 13:42:18.196 >ZLAN.BCK;3           1054976  27-SEP-2002 13:52:11.406 >ZLAN.BCK;2           1005102  27-SEP-2002 13:44:13.666 >ZLAN.BCK;1            178983  27-SEP-2002 10:46:06.06 >B# >Total of 10 files, 5035317 blocks.a >d5 >$ dir/siz/date  DNFS1:[weealf_manman_backup]alan.bckn >f' >Directory DNFS1:[WEEALF_MANMAN_BACKUP]n > 6 >ALAN.BCK;3            216468  26-SEP-2002 13:05:15.406 >ALAN.BCK;2            110250  26-SEP-2002 11:16:49.696 >ALAN.BCK;1                 0  26-SEP-2002 11:16:14.39 >a! >Total of 3 files, 326718 blocks.t >$ >d > , >But now the same thing mounted /structure=5 >B >s' >Directory DNFS1:[weealf_manman_backup]a >p6 >alan.bck;3            216468  26-SEP-2002 13:05:15.406 >alan.bck;2            110250  26-SEP-2002 11:16:49.696 >alan.bck;1                 0  26-SEP-2002 11:16:14.396 >manman.bck;1         2419137  25-SEP-2002 14:51:56.206 >manman.DIR;1               1  25-SEP-2002 12:11:46.646 >sadman.bck;2            4788  25-SEP-2002 13:45:08.336 >sadman.bck;1           45612  25-SEP-2002 13:42:18.196 >zlan.bck;3           1054976  27-SEP-2002 13:52:11.406 >zlan.bck;2           1005102  27-SEP-2002 13:44:13.666 >zlan.bck;1            178983  27-SEP-2002 10:46:06.06 >p# >Total of 10 files, 5035317 blocks.c >u >But now >e5 >$ dir/siz/date  DNFS1:[weealf_manman_backup]alan.bckt >s' >Directory DNFS1:[weealf_manman_backup]e >f6 >alan.bck;3            216468  26-SEP-2002 13:05:15.406 >alan.bck;3            216468  26-SEP-2002 13:05:15.406 >alan.bck;3            216468  26-SEP-2002 13:05:15.406 >alan.bck;3            216468  26-SEP-2002 13:05:15.406 >alan.bck;3            216468  26-SEP-2002 13:05:15.406 >alan.bck;3            216468  26-SEP-2002 13:05:15.406 >alan.bck;3            216468  26-SEP-2002 13:05:15.406 >alan.bck;3            216468  26-SEP-2002 13:05:15.406 >alan.bck;3            216468  26-SEP-2002 13:05:15.40 >l >and so on for everi   ------------------------------    Date: 27 Sep 2002 05:31:39 -0700( From: rrb35146@yahoo.com (Robbie Benton) Subject: Re: Memory for DS25= Message-ID: <dba64bc2.0209270431.72353fa9@posting.google.com>   \ carl@gerg.tamu.edu (Carl Perkins) wrote in message news:<26SEP200220041500@gerg.tamu.edu>...  E > Suggestion: Don't look for "200 pin DIMMs", look for "DS25 memory"."  ? That was kind of my point.  The hp literature keeps referencingf@ "industry-standard memory" for this product.  If the modules are> unique to DS25s only, then it is not industry standard, is it?   > @ > Crucial doesn't list the DS25 on their web site, but they haveF > memory for the DS20 - if it is the same, you can use it (it probablyB > isn't the same - I think it is clocked faster: it is part number+ > MS350-xx whereas the DS20's is MS340-xx).    It is different.  D > The systems may be too new for the 3rd party memory people to have9 > made memory available for them, if they are being slow.   D I did have some contact offline from various vendors who are sellingF the memory.  It is just not on their web sites yet.  So I will be able to save a bundle of money.  
 > --- Carl   Thanks for the reply.c   Robbie   ------------------------------    Date: 27 Sep 2002 08:37:08 -0600B From: clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley) Subject: Newcomers to VMSe3 Message-ID: <vj0VyjCHEyfx@eisner.encompasserve.org>   C Although due to real world time pressures I have only been skimming B comp.os.vms for the last couple of months, I have still managed toE see a number of new names appearing asking basic beginners questions.g   First of all, welcome to VMS.l  C Second, I am interested in knowing why there have been a upsurge ofaC newcomers to VMS. Was it anything to do with the Slashdot article ?r  D Third, how have you found VMS to work with ? What are your comments, both good and bad ?t   Simon.   -- oB Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP       + Microsoft: The Lada of the computing world.t   ------------------------------  % Date: Fri, 27 Sep 2002 12:31:22 +0100e% From: Alan Greig <a.greig@virgin.net>H Subject: OpenVMS is strategic$8 Message-ID: <k2f8pusq88umvuklce9vqagc7nd7v7cobp@4ax.com>  @ Just received a new set of brochures from HP in the UK for AlphaD related products. I have to admit that someone has gone out of theirE way to ensure that it, at least, tries to send a strong message aboute VMS.  F "and remember, HP OpenVMS is now a *strategic platform* (underlined inE original doc)" and "*OpenVMS is strategic* (large font page headline) E as an enterprise server platform for the new HP" No mention of VMS on 7 low end Itanium systems (iMultia, Mutanium?...) though.   D The roadmaps are described as "cast in stone" which surely indicatesC that someone in marketing has at last started to figure out that we & are a bit wary of roadmaps these days.  C Lots of offers including 33% trade in against any Sun or IBM systemn  for any new AlphaServer ordered.  B Not a word about helping anyone move from VMS to HP-UX or NT thankE goodness. You'd almost think someone is paying  attention to feedback6 in comp.os.vms :-)  @ 8/10 for the glossies.  Now let's see how well they can execute. -- Alan   ------------------------------    Date: 27 Sep 2002 07:51:25 -0600B From: clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley)! Subject: Re: OpenVMS is strategics3 Message-ID: <LsmglpxNjfCq@eisner.encompasserve.org>(  ` In article <k2f8pusq88umvuklce9vqagc7nd7v7cobp@4ax.com>, Alan Greig <a.greig@virgin.net> writes: > B > Just received a new set of brochures from HP in the UK for AlphaF > related products. I have to admit that someone has gone out of theirG > way to ensure that it, at least, tries to send a strong message abouta > VMS. >   H Assuming that this was a standard mailing, then my copy hasn't caught upJ with me at my new location yet, so I have a question: Is this a general HPH brochure or was it produced by the UK Alphaserver ["Top Dog" :-)] unit ?  F For those outside of the UK, this unit has managed to produce stronglyI pro-VMS literature even in the bad old days, so I'm interested in knowingaG if this is HP originated material or the Alphaserver unit promoting VMSa again.   Simon.   -- kB Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP       + Microsoft: The Lada of the computing world.    ------------------------------  % Date: Fri, 27 Sep 2002 08:35:55 -0400e2 From: Atlant Schmidt <atlantnospam@mindspring.com>! Subject: Re: OpenVMS is strategic . Message-ID: <3D9450AB.9E55C6CF@mindspring.com>   Alan Greig wrote:   / > The roadmaps are described as "cast in stone"   ! Soapstone, sandstone, or granite?-   Atlant    1 ("I bring you these fifteen, (thump/crash!), err,  *TEN* commandments...")o   ------------------------------  % Date: Fri, 27 Sep 2002 15:17:28 +0100F% From: Alan Greig <a.greig@virgin.net>1! Subject: Re: OpenVMS is strategicc8 Message-ID: <hhp8pugphnr2l137uonnhb8vdrk9jkg993@4ax.com>   On 27 Sep 2002 07:51:25 -0600,C clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley) wrote:y  a >In article <k2f8pusq88umvuklce9vqagc7nd7v7cobp@4ax.com>, Alan Greig <a.greig@virgin.net> writes:e >>  C >> Just received a new set of brochures from HP in the UK for AlphatG >> related products. I have to admit that someone has gone out of their H >> way to ensure that it, at least, tries to send a strong message about >> VMS.m >> . >tI >Assuming that this was a standard mailing, then my copy hasn't caught up K >with me at my new location yet, so I have a question: Is this a general HPbI >brochure or was it produced by the UK Alphaserver ["Top Dog" :-)] unit ?   D The brochures say "Printed in UK, September 2002", However the largeE fold-out one appears to not be UK specific. and ends with a paragrapheE (and photo) by Scott Stallard The smaller of the two brochures has UKaA email and web addresses. The covering letter is signed by Richard % George. All carry the HP invent logo.t  @ Hey, maybe they've let the top-dog team design the international$ brochures. That would be a big plus!  G >For those outside of the UK, this unit has managed to produce stronglyoJ >pro-VMS literature even in the bad old days, so I'm interested in knowingH >if this is HP originated material or the Alphaserver unit promoting VMS >again.a >  >Simon.e   -- Alan   ------------------------------  # Date: Fri, 27 Sep 2002 17:12:05 GMT 4 From: Tim Llewellyn <tim.llewellyn@blueyonder.co.uk>! Subject: Re: OpenVMS is strategic 0 Message-ID: <3D948F59.6344B251@blueyonder.co.uk>   Alan Greig wrote:y >   e > E > Lots of offers including 33% trade in against any Sun or IBM systemh" > for any new AlphaServer ordered.  $B Alan, you didn't mention the VMS unlimited user licence offer with any new server!n  D However, offers cannot be combined and end on 13th October. How many6 people can push an upgrade order through that quickly?  8 Anyway, I like the overall message and the presentation.   A So, people, start buying new VMS alpha's and hiring people to runo7 them. Especially if you are in South West England. :-).i     -- o tim.llewellyn@blueyonder.co.uk t  H * PLEASE NOTE tim.llewellyn@cableinet.co.uk address is NO LONGER VALID *   ------------------------------  % Date: Fri, 27 Sep 2002 06:54:18 -0700 # From: "Tom Linden" <tom@kednos.com>' Subject: PL/I for OpenVMSa9 Message-ID: <CIEJLCMNHNNDLLOOGNJIEEAJFNAA.tom@kednos.com>e  ; Kits and documentation available at http://freja.kednos.comg  @ I guess we have put off long enough having a (terse) web site:-) Tomn --- & Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).@ Version: 6.0.391 / Virus Database: 222 - Release Date: 9/19/2002   ------------------------------   Date: 27 Sep 2002 11:28:05 GMT/ From: Thomas Dickey <dickey@saltmine.radix.net>o1 Subject: Re: Port of a recent Ncurses lib for AXPh* Message-ID: <an1fc5$n9r$1@news1.radix.net>  8 Bernard Giroud <bernard.giroud@creditlyonnais.fr> wrote:; > We envisage to use curses for an application on OVMS/AXP. ; > Some features are absent from the standard DEC C RTL, butq > present in ncurses.   > > Before starting the port ourselves, we would be delighted to; > just download patches for sources and recompile the wholer > thing.  C I'm not aware of a port.  It's probably doable, but noting that VMSkD doesn't have POSIX terminal I/O, there's a fair amount of work to beF done (changing terminal modes, etc.).  For a first cut - just to focus on this - I'd do a: 	configure --disable-database --with-fallbacks=vt100,vt220 and 
 	make sourcesi& before copying the source tree to VMS.   -- n= Thomas E. Dickey <dickey@radix.net> <dickey@herndon4.his.com>i http://dickey.his.com  ftp://dickey.his.com   ------------------------------  # Date: Fri, 27 Sep 2002 14:50:47 GMT 4 From: Tim Llewellyn <tim.llewellyn@blueyonder.co.uk>) Subject: Re: Q: Running OpenVMS diskless?d0 Message-ID: <3D946E3B.633BFD5C@blueyonder.co.uk>   Brian 'Jarai' Chase wrote: > : > In article <00A14872.CB412C56@SSRL04.SLAC.STANFORD.EDU>,* >  <winston@SSRL.SLAC.STANFORD.EDU> wrote: > N > > You can do a local area VAXcluster connected by Ethernet.  This would meanM > > that you'd have a boot node that owned the system disk (which is designedcP > > for this - SYS$COMMON and SYS$SPECIFIC trees mean that the shared bits don'tS > > have to appear multiple places on the disk) and MOP-booted the satellite nodes, 9 > > which sounds pretty much like what you're asking for.a >  > Yes, excellent.c  W It will work. I remember using 3100M30's like this. 8MB RAM, no local disk. I preferred M the one with mono not colour graphics:-). Suffice to say, we used to type our R password, then go and have a LONG coffee break waiting for DECWIndows to start:-).   > M > > However, it's a really good idea to have some local disk on the satelliteeK > > nodes for paging and swapping, because performance tends to suck if you L > > page and swap across 10megabit Ethernet, especially if you have a lot of; > > satellites paging to the same disk across the Ethernet.  > >mN > > However, if you're willing to take the performance hit, or if you know forN > > sure you're not going to be paging or swapping, it's perfectly legal to go$ > > diskless on the satellite nodes. > G > I don't think paging will be an issue for me, at least not initially.oG > The main satellite I'm concerned with is a VAX 6000.  I'll have about E > 512MB of RAM in it, but it lacks disks and a disk controller at the:K > moment.  Nor do I have the means to make use of the CI controller; I lacku< > a coupler and another /small/ CI capable system with disk.  R Does 6000 series support remote boot? Maybethe ROMS were upgraded, but I rememeberT being told by support that our efforts to boot a 6210 as a satellite were pointless.   Regards,   >   > Now... to get the Hobbyist CD. > 	 > -brian.  > --H > --- Brian Chase | bdc@world.std.com | http://world.std.com/~bdc/ -----J > This counter is [6,177,399,753] times as pointless as a real one.  -- K.   -- i tim.llewellyn@blueyonder.co.uk a  H * PLEASE NOTE tim.llewellyn@cableinet.co.uk address is NO LONGER VALID *   ------------------------------  # Date: Fri, 27 Sep 2002 15:19:07 GMTo+ From: Ryan Moore <rmoore@rmoore.dyndns.org>-) Subject: Re: Q: Running OpenVMS diskless?0< Message-ID: <Pine.LNX.4.31.0209270812120.5691-100000@jaipur>  ) On Fri, 27 Sep 2002, Tim Llewellyn wrote:-T > Does 6000 series support remote boot? Maybethe ROMS were upgraded, but I rememeberV > being told by support that our efforts to boot a 6210 as a satellite were pointless.  G I remember booting a VAX 6000 off an Infoserver to do VMS upgrades.  WeuH didn't have a CD-ROM on our VAX 6420, so we'd use the Inforserver.  Sure@ it took a while, but it worked.   I don't see how booting off anD Infoserver would be different than booting sattelite into a cluster.I Either way, it's MOP.  We only had a TK70 and TU81 on the VAX, and I knowiH for sure we didn't use either to load VMS.  We would have had to use our Infoserver.2  I What was that file name?  ISL_LVAX_060 or something like that for the VMSi 6.0 upgrade.   -Ryanh   ------------------------------  % Date: Fri, 27 Sep 2002 12:56:13 -0400g; From: "Brian Tillman" <tillman_brian@notnoone.notnohow.com> ) Subject: Re: Q: Running OpenVMS diskless?r$ Message-ID: <3d948e05$1@news.si.com>  H >Do most people live with satellites being dependent on one boot server,I >or do folks allow more than one boot server per satellite (with the sidepG >effect that there is more than one SYS$SPECIFIC for such a satellite)?   K I sure don't understand this, because all twelve of our satellites have twoeJ different boot servers (in one cluster, of course), and none has more than one SYS$SPECIFIC.  --A Brian Tillman                   Internet: tillman_brian at si.com A Smiths Aerospace                          tillman at swdev.si.coma= 3290 Patterson Ave. SE, MS      Addresses modified to preventu< Grand Rapids, MI 49512-1991     SPAM.  Replace "at" with "@"8        This opinion doesn't represent that of my company   ------------------------------  % Date: Fri, 27 Sep 2002 10:04:40 +0200  From: Dirk Munk <munk@home.nl>/ Subject: Security problem with CSWS T1.3 Beta ?s5 Message-ID: <kiUk9.135718$H6.10941514@zwoll1.home.nl>l  O I've installed the Compaq Secure Web Server T 1.3, and now it seems there is a r" problem in the SSL part somewhere.  O I don't have secure web pages, but occasionaly I see multiple connections from tN one or more remote addresses to port 443. When these connections are visible, Q there is always one of those apache processes looping. I suppose the connections yI are malicious attempts to hack my system, but luckily I see no direct of   buffered IO taking place.,  Q After killing the looping process all port 443 connections are gone. So the only cM    'damage' they inflict is using up connections and making the system a bit a7 sluggish. However this should not happen of course ....r   ------------------------------  # Date: Fri, 27 Sep 2002 12:26:37 GMT + From: "Rick Barry" <barry@star.zko.dec.com>S3 Subject: Re: Security problem with CSWS T1.3 Beta ?:1 Message-ID: <18Yk9.8$u_5.151669@news.cpqcorp.net>o  I Have you enabled MOD_SSL? Otherwise the server should not be listening on 	 port 443.i  L The CSWS T1.3 version is using a pre-patched OpenSSL 0.9.6b library which isJ susceptible to certain buffer-overflows. These may cause access violationsH or unpredictable behavior. CSWS V1.3 (should be announced next month) isB built with a patched version of OpenSSL 0.9.6b which addresses the buffer-overflow problems.8  J I don't know for sure if this is related to the problem you're seeing. AreC there entries in your error log that show SSL handshake errors? I'mhI investigating a similar problem where I've found these entries in the logpD which lead me to suspect OpenSSL buffer-overflows as the root cause.  J (For anyone who doesn't already know, a patch available for CSWS V1.2 thatI addresses the OpenSSL buffer-overflow problem. You can download the patchc fromL http://www.openvms.compaq.com/openvms/products/ips/apache/csws_patches.html. )-  
 Rick Barry Compaq Secure Web Server Teamm OpenVMS System Software Groupn Hewlett-Packard Companyp
 Nashua, NH  + "Dirk Munk" <munk@home.nl> wrote in message3/ news:kiUk9.135718$H6.10941514@zwoll1.home.nl...rK > I've installed the Compaq Secure Web Server T 1.3, and now it seems there  is a$ > problem in the SSL part somewhere. >nK > I don't have secure web pages, but occasionaly I see multiple connections  fromF > one or more remote addresses to port 443. When these connections are visible,F > there is always one of those apache processes looping. I suppose the connectionspJ > are malicious attempts to hack my system, but luckily I see no direct of > buffered IO taking place.  >eI > After killing the looping process all port 443 connections are gone. Sog the onlyJ >    'damage' they inflict is using up connections and making the system a bito9 > sluggish. However this should not happen of course ...._ >d   ------------------------------  % Date: Fri, 27 Sep 2002 16:53:24 +0530i5 From: "Sandeep Yelwatkar" <Sandeep_Yelwatkar@bmc.com>t Subject: Shadow set membersy/ Message-ID: <up8ffu26opnpf7@corp.supernews.com>o   Hi  L In an OpenVMS cluster system, how can one identify all the disks in a shadowI set. Also which member of the shadow set is a master and which is member.-   Thanks in advance for any help	 - Sandeep9   ------------------------------  # Date: Fri, 27 Sep 2002 12:08:46 GMT(- From: "labadie" <labadie_g.tocardsa@decus.fr>6 Subject: Re: Shadow set membersp1 Message-ID: <iTXk9.7$o46.231566@news.cpqcorp.net>u  @ "Sandeep Yelwatkar" <Sandeep_Yelwatkar@bmc.com> wrote in message) news:up8ffu26opnpf7@corp.supernews.com...M > Hi >mG > In an OpenVMS cluster system, how can one identify all the disks in au shadowK > set. Also which member of the shadow set is a master and which is member.u >   > Thanks in advance for any help > - Sandeeps >   dor $ help lex f$getdvi argp and look at the shdwxxxr shdw_member, shdw_master...n   regardse   grard   ------------------------------    Date: 27 Sep 2002 08:24:25 -0600+ From: young_r@encompasserve.org (Rob Young)h Subject: Re: Shadow set memberst3 Message-ID: <bC+6y5ntIp5J@eisner.encompasserve.org>C  a In article <iTXk9.7$o46.231566@news.cpqcorp.net>, "labadie" <labadie_g.tocardsa@decus.fr> writes:0 > B > "Sandeep Yelwatkar" <Sandeep_Yelwatkar@bmc.com> wrote in message+ > news:up8ffu26opnpf7@corp.supernews.com...s >> Hil >>H >> In an OpenVMS cluster system, how can one identify all the disks in a > shadowL >> set. Also which member of the shadow set is a master and which is member. >>! >> Thanks in advance for any help  >> - Sandeep >> >  do  > $ help lex f$getdvi arg  > and look at the shdwxxxM > shdw_member, shdw_master...0 > 	 > regardst >   ? 	And you have to whip through members.  Here is a hack that wasr 	laying around:h     $ item    = "dsa220"/ $ mbr1    = f$getdvi(item,"shdw_next_mbr_name")s/ $ mbr2    = f$getdvi(mbr1,"shdw_next_mbr_name"),0 $! mbr3    = f$getdvi(mbr2,"shdw_next_mbr_name")0 $! mbr4    = f$getdvi(mbr3,"shdw_next_mbr_name")6 $ master_name_mbr1 = f$getdvi(mbr1,"shdw_master_name")6 $ master_name_mbr2 = f$getdvi(mbr2,"shdw_master_name")7 $! master_name_mbr3 = f$getdvi(mbr3,"shdw_master_name")  $ !i $ show symbol mbr1 $ show symbol mbr2 $! show symbol mbr3A $! show symbol mbr41 $ show symbol master_name_mbr1 $ show symbol master_name_mbr2 $! show symbol master_name_mbr3e $ !l    > 	Generalize that and you have a DCL routine to do it.  Another 	quick and dirty is to:    	$ analyze/systemf 	SDA> show dev dsa220:  ! 	Page 2 you see things like this:   B                 --- Shadowing Descriptor Block (SHAD) 80FC4E80 ---    + Virtual Unit SCB Status:        0001 normal-  P Total Devices          3    VU_UCB          80C5A8C0                            P Source Members         3    SCB LBN         0086DE60                            P Act Copy Target        0    Generation      00A14950                            P Act Merge Target       0     Number         C67AFB47                            P Last Read Index        1    VU Site Value   00000000                            P Master Mbr Index       1    VU Time Out       262800                               Device $1$DGA8&   Index 0 Status    000000A0 src,valid8   Read Cost 0000002A    Site 00000000   SM Timeout 180004                         UCB  80DFC700   VCB 80EDB580N Device $1$DGA4104       ...  Master Member         !!<<<<<<<<<================&   Index 1 Status    000000A0 src,valid:   Read Cost 0000002A    Site 00000000   SM Time Out 2592004                         UCB  80DD9080   VCB 80FB5640 Device $1$DGA5&   Index 2 Status    000000A0 src,valid     				Robm   ------------------------------  # Date: Fri, 27 Sep 2002 13:46:21 GMT - From: "labadie" <labadie_g.tocardsa@decus.fr>n Subject: Re: Shadow set memberst2 Message-ID: <NiZk9.10$ac6.342104@news.cpqcorp.net>  7 "Rob Young" <young_r@encompasserve.org> wrote   Another1 > quick and dirty is to: >  > $ analyze/system > SDA> show dev dsa220:t >T-  > Device $1$DGA4104       ...  Master Memberh !!<<<<<<<<<================    Good idea, another hacku  F pipe write sys$output  "show dev dsaxxx:" | ana/system | search/nohigh sys$input "Master Member"t  , Some more work and we will get a one-liner !   Regardsh   Grard   ------------------------------    Date: 27 Sep 2002 12:59:15 -0600+ From: young_r@encompasserve.org (Rob Young), Subject: Re: Shadow set members 3 Message-ID: <0fseDMjZ20TI@eisner.encompasserve.org>w  b In article <NiZk9.10$ac6.342104@news.cpqcorp.net>, "labadie" <labadie_g.tocardsa@decus.fr> writes: > 9 > "Rob Young" <young_r@encompasserve.org> wrote   Anothern >> quick and dirty is to:b >> >> $ analyze/systeme >> SDA> show dev dsa220: >>/ >  > Device $1$DGA4104       ...  Master Membere > !!<<<<<<<<<================r >  > Good idea, another hacku > H > pipe write sys$output  "show dev dsaxxx:" | ana/system | search/nohigh > sys$input "Master Member"6 > . > Some more work and we will get a one-liner ! >   6 	That is slick and indeed works.  Here is how you find 	all your master members:e  5 $ pipe write sys$output  "show dev ds" | ana/syst | -  	search/nohigh sys$input master   = 	Make sure you are properly privved, or it won't work and you & 	won't see the error message!  (Grin).   				Robu   ------------------------------  % Date: Fri, 27 Sep 2002 09:27:24 -0400t; From: "Webb, William W Raleigh, NC" <wwebb1@email.usps.gov> $ Subject: The perils of reusable codeK Message-ID: <BA52530E3149734A9BAABDBBFA808E49028BEB0D@rlghncst964.usps.gov>   : The reuse of some object-oriented code has caused tactical: headaches for Australia's armed forces. As virtual reality= simulators assume larger roles in helicopter combat training,r6 programmers have gone to great lengths to increase the> realism of their scenarios, including detailed landscapes and,: in the case of the Northern Territory's Operation Phoenix,; herds of kangaroos (since disturbed animals might well giveb away a helicopter's position).  ; The head of the Defense Science & Technology Organization'si9 Land Operations/Simulation division reportedly instructed-7 developers to model the local marsupials' movements and  reactions to helicopters.t  ; Being efficient programmers, they just re-appropriated somee; code originally used to model infantry detachment reactionsi6 under the same stimuli, changed the mapped icon from a: soldier to a kangaroo, and increased the figures' speed of	 movement.0  : Eager to demonstrate their flying skills for some visiting9 American pilots, the hotshot Aussies "buzzed" the virtualy: kangaroos in low flight during a simulation. The kangaroos: scattered, as predicted, and the visiting Americans nodded9 appreciatively... then did a double-take as the kangaroos 7 reappeared from behind a hill and launched a barrage ofl; Stinger missiles at the hapless helicopter. (Apparently theh4 programmers had forgotten to remove that part of the infantry coding.)y  8 The lesson? Objects are defined with certain attributes,: and any new object defined in terms of an old one inherits; all the attributes. The embarrassed programmers had learned 8 to be careful when reusing object-oriented code, and the; Yanks left with a newfound respect for Australian wildlife.M  8 Simulator supervisors report that pilots from that point9 onward have strictly avoided kangaroos, just as they were:	 meant to.-  > From June 15, 1999 Defense Science and Technology Organization7 Lecture Series, Melbourne, Australia, and staff reports@   ------------------------------  # Date: Fri, 27 Sep 2002 11:52:36 GMTL+ From: "Rick Barry" <barry@star.zko.dec.com> % Subject: Re: Tomcat and CSWS (Apache) 1 Message-ID: <8EXk9.4$816.201864@news.cpqcorp.net>   I mod_jserv uses AJPV12 protocol to speak to the servlet engine, either the I older JSERV subprocess of Apache or the newer Tomcat detached process. Asd9 you said, mod_jk uses AJPV13, and has better performance.   E The pre-compiled versions of mod_jserv and mod_jk can be found in then CSWS_JAVA kits.   
 Rick Barry Compaq Secure Web Server Team  OpenVMS System Software Group  Hewlett-Packard Company 
 Nashua, NH  : "Keith A. Lewis" <lewis@spyder.mitre.org> wrote in message( news:amvnec$akj$1@newslocal.mitre.org...9 > "Rick Barry" <barry@star.zko.dec.com> writes in article-K <dREk9.20$oi5.355398@news.cpqcorp.net> dated Thu, 26 Sep 2002 14:29:29 GMT:sL > >OS.H is located in APACHE$COMMON:[SRC.OS.OPENVMS], so the compile command7 > >needs to include this directory on the command line:m > >< > >$ ccl >uL >mod_jserv/incl=(APACHE$COMMON:[SRC.INCLUDE],APACHE$COMMON:[SRC.OS.OPENVMS]) > " > Oh.  (duh!)  It's compiling now. >c. > >(Did you mean mod_jk in the above command?) > J > According to jakarta.apache.org (where I got Tomcat), mod_jk is supposed toE > replace mod_jserv.  It uses a newer protocol called Ajpv13 designediI > specificially for server-server communications.  I think mod_jserv usesA HTTP. > on Tomcat's port, pretending to be a client. >rC > http://jakarta.apache.org/tomcat/tomcat-3.3-doc/mod_jk-howto.htmlo >o; > >"Marty Kuhrt" <kuhrt@encompasserve.org> wrote in message 0 > >news:jHLzTRwNhDGj@eisner.encompasserve.org...G > >> I have the precompiled version working just fine.  I haven't tried $ > >> to compile it, though, so YMMV. >cF > A precompiled version of mod_jk (if so, where'd you get it?) or just Tomcat?-) > I want to augment CSWS, not replace it.  >e- > --Keith Lewis              klewis$mitre.org @ > The above may not (yet) represent the opinions of my employer.   ------------------------------    Date: 26 Sep 2002 23:43:22 -0700% From: Bart.Zorn@xs4all.nl (Bart Zorn)c" Subject: Re: UCX security advisory= Message-ID: <a98cd882.0209262243.58bb07c4@posting.google.com>n  ^ david20@alpha2.mdx.ac.uk (David Webb) wrote in message news:<amv9ts$iu8$1@aquila.mdx.ac.uk>...g > In article <a98cd882.0209260619.17d1d4e8@posting.google.com>, Bart.Zorn@xs4all.nl (Bart Zorn) writes:e > >Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com> wrote in message news:<01KMYH41GRFI9QWI2N@sysdev.deutsche-boerse.com>...h > >uG > >Actually, my main criticism is the same as someone else already saidtF > >in this thread. TCP/IP engineering does not care about security andG > >they do not take the time to figure out which privileges they reallynG > >need. I cannot imagine that they really NEED BYPASS. It's just a wayo' > >to avoid those pesky error messages.- > >- > >Bart Zorn > N > In this instance BYPASS probably is needed. The POP server needs not only toL > read the messages in the user's inbox (new mail folder) but also to deleteP > them when they are downloaded to the POP client. These messages will either beM > directly in the user's MAIL.MAI file if they are really small or will exist K > as separate MAIL$????????????????.MAI files (with a pointer in mail.mai).  > K > Hence to function as a pop server it needs enough privilege to be able toaN > delete files (the MAIL$????????????????.MAI files) owned by any user on the I > system. The usual protection on these files is (RW,RW,,) on my systems.  > O > They might be able to fudge it just using SYSPRV which would give them access:M > via the system part of the protection code and which will also I think give O > control access. Hence if the system part of the protection code doesn't allow D > access they could change the file's protection and then delete it. > M > However in this instance I don't really see much difference between a buggy2F > program with SYSPRV and one with BYPASS as to the security problems. > L > The problem isn't with the privileges themselves it is with the fact that N > what looks like some elementary checking ( a -logfile qualifier ??? ) wasn'tO > carried out. Missing a buffer overflow hidden deep in the code is unfortunate26 > but understandable missing this sort of thing isn't. >  > David Webb > VMS and Unix team leader > CCSS > Middlesex University  C That's possible. But I would require TCP/IP engineering to document]B this and every other situation where they really need BYPASS. ThatC would make them think hard about it and it would give me confidence 0 that the product is OK with respect to security.  	 Bart Zorn    ------------------------------  % Date: Fri, 27 Sep 2002 03:09:48 -04006- From: JF Mezei <jfmezei.spamnot@videotron.ca> " Subject: Re: UCX security advisory, Message-ID: <3D94042C.E3FDBF69@videotron.ca>  6 Another thing that should be changed in the software :  I enable privs needed to bind the port 110. Then disable those privileges. P  7 and only enable SYSPRV when accessing a user's mailbox.   K I have a problem with anyone being able to issue the extended command (XTNDNK which allows shutdown of the server and changing the logging level. The POP:J server should verify that the supplied username has at least SYSPRV and/or8 WORLD in his UAF setting before accepting XTND commands)   ------------------------------   Date: 27 Sep 2002 10:18:56 GMT1 From: JONESD@er6.eng.ohio-state.edu (David Jones) " Subject: Re: UCX security advisory: Message-ID: <an1bag$p47$1@charm.magnus.acs.ohio-state.edu>  , In message <3D937A3C.998BBE70@videotron.ca>,1   JF Mezei <jfmezei.spamnot@videotron.ca> writes:5L >the all mighty and reliable decthreads server from dave jones seems ok withO >just cmexec and sysprv to bind its socket to port 80. (definitely a well knownf@ >port), and it isn't defined in the TCPIP> SET SERVICE database.  L The decthreads server requires sysprv to bind 'listen' sockets to privilegedH ports numbers (e.g. port 80) and to lookup login directories via $GETUAIJ for /~username processing.  Note that sysprv is disabled while opening anyI files except those implicity opened by $GETUAI.  The cmexec privileged isnL required because $GETUAI calls are done via a $CMEXEC kludge concocted prior: to VMS 7.2's introduction of per-thread security profiles.  M >so it seems that even the documerntation is wrong. (unless he uses cmexec to=& >do really weird stuff to gain bypass)  E Cmexec isn't used to gain bypass, it's used to acquire the inner mode.H mutex to prevent other client threads inadvertantly gaining sysprv whileN a thread doing a /~username translation has sysprv enabled.  That's the theoryH anyway.  The code will skip the $CMEXEC hack if the process doesn't have= cmexec privilege, still making the $SETPRV and $GETUAI calls.   < David L. Jones               |      Phone:    (614) 292-6929- Ohio State University        |      Internet: L 140 W. 19th St. Rm. 231a     |               jonesd@er6s1.eng.ohio-state.edu: Columbus, OH 43210           |               vman+@osu.edu  1 Disclaimer: I'm looking for marbles all day long.    ------------------------------  + Date: Fri, 27 Sep 2002 12:07:49 +0000 (UTC)[+ From: david20@alpha1.mdx.ac.uk (David Webb) " Subject: Re: UCX security advisory+ Message-ID: <an1hml$b69$1@aquila.mdx.ac.uk>1  i In article <26SEP02.15451657@thuria.waisman.wisc>, karcher@thuria.waisman.wisc.edu (Carl Karcher) writes:$D >In a previous article, david20@alpha2.mdx.ac.uk (David Webb) wrote: >yO >->In this instance BYPASS probably is needed. The POP server needs not only to:M >->read the messages in the user's inbox (new mail folder) but also to delete Q >->them when they are downloaded to the POP client. These messages will either be2N >->directly in the user's MAIL.MAI file if they are really small or will existL >->as separate MAIL$????????????????.MAI files (with a pointer in mail.mai). >-> L >->Hence to function as a pop server it needs enough privilege to be able toO >->delete files (the MAIL$????????????????.MAI files) owned by any user on the  J >->system. The usual protection on these files is (RW,RW,,) on my systems. >wF >If the POP server uses the callable mail interface then BYPASS shouldH >not be needed. For example, Yahmail uses the mail API and only requires >SYSPRV. >5  M Not something I'd looked into but I'd suspect that the callable mail routines G are therefore using the method I described later of using the fact thatEN SYSPRV gives you control access which means it can change the files protection and then delete it.6  L But as I said I can't really see much difference from a security perspective? whether the buggy program has SYSPRV and BYPASS or just SYSPRV.a  
 David Webb VMS and Unix team leader CCSS Middlesex University   ------------------------------  + Date: Fri, 27 Sep 2002 12:59:43 +0000 (UTC)r+ From: david20@alpha1.mdx.ac.uk (David Webb)c" Subject: Re: UCX security advisory+ Message-ID: <an1knv$c62$2@aquila.mdx.ac.uk>c  l In article <4b6ec350.0209261213.5f4e3da0@posting.google.com>, JimStrehlow@data911.com (Jim Strehlow) writes:4 >FYI:  I do not have problems with TCP/IP v5.1 ECO 4 >rJ But have you actually got the DEC TCPIP Services POP server configured and enabled on the system ?fF If not then the image won't be installed with privileges and hence the attack won't work.  
 David Webb VMS and Unix team leader CCSS Middlesex University     >$show system/noprocesseA >OpenVMS V7.3  on node XXXXXX  26-SEP-2002 13:10:25.19  Uptime  3 	 >06:30:29i >i >$tcpip show version@ >  Compaq TCP/IP Services for OpenVMS Alpha Version V5.1 - ECO 4, >  on a AlphaServer XXX running OpenVMS V7.3 >f >$SHOW PROCESS/PRIVILEGESp? >26-SEP-2002 12:17:34.32   User: XXXXXX           Process ID:  l	 >00000322i? >                          Node: XXXXXX           Process name: 	 >"XXXXXX"  >Authorized privileges:w! > NETMBX       TMPMBX       WORLDo >  >Process privileges:0 > NETMBX               may create network device3 > TMPMBX               may create temporary mailbox > > WORLD                may affect other processes in the world >  >Process rights:$ > INTERACTIVE                       $ > REMOTE                             >  >System rights:  > SYS$NODE_XXXXXXs >e >Soft CPU Affinity: offn > / >$break_it :== $sys$system:tcpip$pop_server.exe ? >$break_it -logfile sys$system:I_SHOULDNT_BE_ABLE_TO_WRITE_HEREc >g; >2002-09-26 13:07:32 log file could not be opened: error 13dE >2002-09-26 13:07:32 starting TCPIP POP server V5.1-15B, OpenVMS V7.3r" >Alpha on host XXXXXX and port 110 >2002-09-26 13:07:32 r, >2002-09-26 13:07:32 POP configuration info:4 >2002-09-26 13:07:32    ignore_mail11_headers: FALSE. >2002-09-26 13:07:32    send_id_headers: FALSE* >2002-09-26 13:07:32    disuserpass: FALSE$ >2002-09-26 13:07:32    trace: FALSE/ >2002-09-26 13:07:32    leave_in_newmail: FALSEe0 >2002-09-26 13:07:32    nosysacm_userpass: FALSE, >2002-09-26 13:07:32    nosysacm_apop: FALSE. >2002-09-26 13:07:32    use_mail_folder: FALSE( >2002-09-26 13:07:32    fast_scan: FALSE, >2002-09-26 13:07:32    personal_name: FALSE, >2002-09-26 13:07:32    purge_reclaim: FALSE- >2002-09-26 13:07:32    max_messages: No max.tC >2002-09-26 13:07:32    sysacm_timeout: default provided by SYS$ACMM >system servicee7 >2002-09-26 13:07:32    master_log_level: INFORMATIONALS* >2002-09-26 13:07:32    security: FRIENDLY1 >2002-09-26 13:07:32    decnet_rewrite: TRANSFORM 8 >2002-09-26 13:07:32    quoted_decnet_rewrite: TRANSFORM. >2002-09-26 13:07:32    sndbuf: TCPIP default.' >2002-09-26 13:07:32    max_threads: 15s9 >2002-09-26 13:07:32    link_idle_timeout: NOT CONFIGUREDu >2002-09-26 13:07:32 a( >2002-09-26 13:07:32 Miscellaneous info:G >2002-09-26 13:07:32    TCPIP POP version: V5.1-15B, OpenVMS V7.3 Alphah+ >2002-09-26 13:07:32    POP server PID: 325a7 >2002-09-26 13:07:32    POP server process name: XXXXXX = >2002-09-26 13:07:32    POP server account default directory:. >SYS$SYSDEVICE:[TCPIP$POP]5 >2002-09-26 13:07:32    SYS$SCRATCH: myLoginDirectory  >2002-09-26 13:07:32 -- >2002-09-26 13:07:32 SMTP configuration info:6+ >2002-09-26 13:07:32    substitute_domain: E >2002-09-26 13:07:32 ) >bind: permission denied< >2002-09-26 13:07:32 error binding initial socket. errno=13, >vaxc$errno=36< >2002-09-26 13:07:32 could not create or bind initial socket> >%SYSTEM-F-NOPRIV, insufficient privilege or object protection
 >violation > ( >Jim Strehlow, Data911, Alameda, CA, USA   ------------------------------  % Date: Fri, 27 Sep 2002 12:03:32 +0200 - From: "Bert de Geus" <bert.de.geus@xs4all.nl> J Subject: Re: Unexpected <CR> and <FF> after a Setup module of a queue FORM* Message-ID: <an1amc$rm7$1@news1.xs4all.nl>  8 "Peter Weaver" <peter.weaver@stelco.ca> wrote in message/ news:amslc7$8rs9u$1@ID-141708.news.dfncis.de...n: > "Bert de Geus" <bert.de.geus@xs4all.nl> wrote in message& > news:amrv2o$9lj$1@news1.xs4all.nl... > >...J > > The module I extracted and inserted contained only escapes and had theL > > <ESC>P <ESC>\ around the setup code. Yet the <CR><FF> was emitted by the
 > > symbiont.y > >... >eG > Are you 100% certain that there is no <CR> after the <ESC>\? I recallrL > working with someone with the same problem (might have been VMS 5.7 or 6.1G > at that time), he kept telling me that there was no <CR> in the setupVJ > module. After exchanging phone calls for a few weeks between myself, theA > programmer and the user near the printer I finally asked him tooH > DUMP/DECIMAL/BYTE/RECORD the module. It turned out that there were twoI > records in the file that the programmer never saw when typing it to his-J > screen or editing the file. He deleted the second (blank) record and the > problem went away. >R > -- > Peter Weaver   Here is the dump:pG Dump of file USER$ROOT:[GBG.WRK]132X66.TXT;4 on 27-SEP-2002 12:12:21.71   5 File ID (407,154,0) End of file block 1 / Allocated 3e  3 Virtual block number 1 (00000001), 512 (0200) bytesv  ! 59 83 77 86 93 27 0 46 ...]VMS; 0i  " 99 42 27 80 27 92 27 50 2.\.P.*c 8  $ 46 56 49 115 40 27 70 53 5F.(s18. 16  # 27 85 48 49 40 27 72 53 5H.(10U. 24I  $ 49 108 38 27 76 55 97 38 &a7L.&l1 32  # 92 27 82 48 97 38 27 69 E.&a0R.\ 40t   0 0 0 0 0 0 -1 -1 ........ 48s   0 0 0 0 0 0 0 0 ........ 56    0 0 0 0 0 0 0 0 ........ 64a   0 0 0 0 0 0 0 0 ........ 72        So no CR after de <ESC>\   Bert   ------------------------------  % Date: Fri, 27 Sep 2002 10:08:34 -0400c- From: "Peter Weaver" <peter.weaver@stelco.ca> J Subject: Re: Unexpected <CR> and <FF> after a Setup module of a queue FORM5 Message-ID: <an1op4$ae7q0$1@ID-141708.news.dfncis.de>   8 "Bert de Geus" <bert.de.geus@xs4all.nl> wrote in message$ news:an1amc$rm7$1@news1.xs4all.nl... >...5 > Virtual block number 1 (00000001), 512 (0200) bytese >l# > 59 83 77 86 93 27 0 46 ...]VMS; 0m >e$ > 99 42 27 80 27 92 27 50 2.\.P.*c 8 > ...r  J You have two <ESC>\ sequences in the file. It has been a long time since IK had to deal with issues like this, but IIRC you should only have the one at J the end. I would take the first <ESC>\ (after the VMS;2) and try it again.   -- Peter WeaverL Opinions are my own, and do not reflect the opinions of my employer, nor theK company that it sub-contracts to, nor the company that it sub-contracts to.m   ------------------------------  # Date: Fri, 27 Sep 2002 09:58:57 GMT-. From: peter@langstoeger.at (Peter LANGSTOEGER)' Subject: Re: Version control on OpenVMSe3 Message-ID: <BZVk9.33732$142.556209@news.chello.at>e  p In article <RBMk9.121200$U_.99583@news04.bloor.is.net.cable.rogers.com>, "Yong Liu" <fdu9774@rogers.com> writes:M >WE have codes scattered on OpenVMS, UNIX and OpneVMS. The question is how to D                                               ^^^^^^^^ what's this ?K >build a central version control repository. Which version control softwareiL >do you use to address this problem? Is the software deal with heterogeneityG >or you have some arangement to deal with it. If so, how did you do it?   P I'd start to have a look into DECset and the CMS client for Non-OpenVMS systems. Maybe it fits your needs.e  L There is also the CVS freeware on OpenVMS, but AFAIK/IIRC only the client...   --   Peter "EPLAN" LANGSTOEGERC% Network and OpenVMS system specialistc E-mail  peter@langstoeger.atP A-1030 VIENNA  AUSTRIA              I'm looking for (a) Network _and_ VMS Job(s)   ------------------------------   Date: 27 SEP 2002 15:12:02 GMT+ From: Dave Greenwood <greenwoodde@ornl.gov>t' Subject: Re: Version control on OpenVMS/2 Message-ID: <27SEP02.15120299@feda01.fed.ornl.gov>  = In a previous article, "Yong Liu" <fdu9774@rogers.com> wrote:r > Hi,m >  cN > WE have codes scattered on OpenVMS, UNIX and OpneVMS. The question is how to  # Glad to see the emphasis on VMS ;-)a  L > build a central version control repository. Which version control softwareM > do you use to address this problem? Is the software deal with heterogeneityfH > or you have some arangement to deal with it. If so, how did you do it? > Thanks  K I've used cvs in such cases.  It has the advantage of clients for VMS, unix I and Windows but you'd have to put the server on unix.  The clients handleoH line terminator issues as long as you don't try to do something like useJ a Windows system to commit a unix-formatted file back into the repository. And, of course, it's free.  I I generally prefer CMS to cvs but AFAIK there's no unix version.  And, of  course, it's not free.   Dave --------------9 Dave Greenwood                Email: Greenwoodde@ORNL.GOV@H Oak Ridge National Lab        %STD-W-DISCLAIMER, I only speak for myself   ------------------------------  % Date: Fri, 27 Sep 2002 13:00:17 -0400 ; From: "Brian Tillman" <tillman_brian@notnoone.notnohow.com>u' Subject: Re: Version control on OpenVMS9$ Message-ID: <3d948ef9$1@news.si.com>  J >WE have codes scattered on OpenVMS, UNIX and OpneVMS. The question is how toK >build a central version control repository. Which version control softwarel$ >do you use to address this problem?  G We use Compaq CMS (Code Management System) from the DECset suite.  As IiK understand it, there is also a Unix-style revision control system available  for download somewhere.t --A Brian Tillman                   Internet: tillman_brian at si.comtA Smiths Aerospace                          tillman at swdev.si.comg= 3290 Patterson Ave. SE, MS      Addresses modified to preventg< Grand Rapids, MI 49512-1991     SPAM.  Replace "at" with "@"8        This opinion doesn't represent that of my company   ------------------------------    Date: 27 Sep 2002 22:18:16 +0800, From: Paul Repacholi <prep@prep.synonet.com>9 Subject: Re: VMS future (oh not not another one of these)h- Message-ID: <87u1kbect3.fsf@prep.synonet.com>,  7 "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> writes:a  ? > JF Mezei wrote in message <3D91FC85.472DD8A1@videotron.ca>...a > >Fred Kleinsorge wrote:   E > >> The Multia was *never* designed to be ANYTHING but what it was -t > >> a Windows/X terminal.    > >And therein lies the problem.   > >WHY ????o  ? > BECAUSE THEY WERE THE TERMINALS GROUP!  They were building antB > appliance.  They were ahead of their time.  The appliance neededD > more things than they really wanted.  They were not building a lowB > cost workstation.  The WORKSTATION group was the group with thatB > responsibility.  They did not set out to build a general purposeA > workstation.  Most people felt that it was underconfigured, andl9 > underpowered for the task of a general purpose machine.r  @ So as we could have bought a Windows `terminal' or a T64 X term,( why couldn't we get a VMS secure X term?  ? Oh, yes, not invented here... Again... Anyone remember the moto > `Digital has it now'? and the refrain `And you can't have it.'  . At least hp seems to be fixing the first part.   -- c< Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda.E@                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------  # Date: Fri, 27 Sep 2002 17:59:29 GMT0 From: sasadmin <jec@nospam.net>  Subject: Re: VMS, C++ and flushn2 Message-ID: <87r8ffmhr4.fsf@Alethion.systasis.net>  ( Jan-Erik Sderholm <aaa@aaa.com> writes:   > Scenario: : > A C++ application works just OK on all systems, but one.; > The application is rather simple, reads on text file with = > 500 char blocks of data, moves the data around a little and 8 > writes a new textfil with 80 char blocks with a subset > of the data.  D If you're just moving data around, DCL SORT and a sort specification5 file is a powerful tool for this type of application.y   And, of course, TPU...   Cheers,I   jec    -- e Microsoft Free By 2003   ------------------------------    Date: 27 Sep 2002 06:06:55 -0700. From: spamsink2001@yahoo.com (Alan E. Feldman)  Subject: Re: warning during copy= Message-ID: <b096a4ee.0209270506.6f3ce6a2@posting.google.com>2  s spamsink2001@yahoo.com (Alan E. Feldman) wrote in message news:<b096a4ee.0209261807.627ccc93@posting.google.com>...s > G > It seems to me that it's the carriage-control method that causes TYPEtC > to misTYPE the file. The VFC files come with "Print file carriageoF > control" while the variable-length records files come with "Carriage > return carriage control".c > D > If you DUMP the combined file you can see print control charactersD > only in the first two records (which are from the VFC file) so theH > remaining lines (from the variable-length records file) are just typed8 > to the screen without any deliberate carriage control.    B Uh, make that the first three records. When I wrote this, I forgotD that the first line from the VL file overwrote the 3rd line from the	 VFC file.      Disclaimer: JMHO Alan E. Feldmani   ------------------------------    Date: 27 Sep 2002 07:23:06 -0700. From: spamsink2001@yahoo.com (Alan E. Feldman)  Subject: Re: warning during copy= Message-ID: <b096a4ee.0209270623.541010f5@posting.google.com>D  ` "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message news:<3D93AF7F.1A98442A@fsi.net>... > "Alan E. Feldman" wrote: > > d > > "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message news:<3D926AEF.5A71FF56@fsi.net>... > > > "Alan E. Feldman" wrote: > > > >sh > > > > "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message news:<3D8FC64A.D3CE58CD@fsi.net>...% > > > > > aniruddha patwardhan wrote:i >  [...]P > > > > > If your intention is simply to make a copy of the file in your currentR > > > > > default device/directory, the following will suffice and not produce the > > > > > mesage:y	 > > > > >  > > > > > $ COPY infile [] > > > >yI > > > > Well, I don't see why he got the message in the first place, so I-2 > > > > don't see how using [] would fix anything. > > > >3% > > > > Also, [] doesn't always work:z > > > >m > > > > DCL> SHOW DEFt! > > > >   SYS$SYSDEVICE:[FELDMAN]e  > > > > DCL> SET DEF SYS$STARTUP > > > > DCL> SHOW DEFy > > > >   SYS$STARTUP:[FELDMAN]  > > >sL > > > I'm guessing that's an invalid path specification since SYS$STARTUP isD > > > not a concealed (or a "rooted") logical. Since your default is > > A > > Your guess is incorrect. This "path" is invalid *only* if you. > > explicitly specify []. > > E > > So you're telling me that you can't set default to SYS$STARTUP? !e >  > Well, I'm not, but VMS is.  $ And just how is VMS telling us that?  H > SYS$SYSROOT:[SYS$STARTUP][FELDMAN] is in invalid path specification no> > matter how how slice it. It never will be correct. Same with' > SYS$SYSROOT:[SYS$STARTUP][DDACHTERA].   D I have never explicitly used either as a path specification. I typed& SET DEF SYS$STARTUP; VMS did the rest.  H > Now, that will cause other things to happen. Case in point further on. > G > > > technically invalid, "[]" is not the only thing I would epxect to/! > > > exhibit anomalous behavior.t > > H > > Nope. Any program or DCL command will use SYS$STARTUP as the default@ > > (and ignore whatever is in the brackets) and work just fine. > F > ...as long as you are looking up existing files in that search list.E > When you start trying to create files in or copy files to that path B > specification, things may not happen the way you want or expect.   Maybe not the way *you* expect.s  	 > > It ismI > > only if you insist that [] means your default that you might get inton. > > trouble. And it does not always mean that. > F > ...and no one ever said it does in the context of this discussion. IJ > suggested that it might solve his problem - I suggested nothing more. As7 > you are about to say, "a very important distinction".    Uh, you said, and I quote: s  ? "If your intention is simply to make a copy of the file in yournD current default device/directory, the following will suffice and not$ produce the mesage: $ COPY infile [] "-  E All I was saying is that in some cases, your advice above won't work, $ and that using * instead of [] will.  J > > The notation [] means "default directory", not your default. That is aI > > very important distinction. If SYS$DISK contains a directory, then []l > > is ignored.h >  > Sorry, don't buy that.  " $ SET DEFAULT SYS$STARTUP:[SYSEXE]
 $ SHOW DEF   SYS$STARTUP:[SYSEXE]   =   SYS$SYSROOT:[SYS$STARTUP]-   =   SYS$COMMON:[SYS$STARTUP]   =   SYS$SYSROOT:[SYSMGR]   =   SYS$COMMON:[SYSMGR]e $ COPY NL: AEF.TMPSKI F %COPY-S-COPIED, _NLA0: copied to SYS$SYSROOT:[SYS$STARTUP]AEF.TMPSKI;1 (0 records)  $S  < In the example above, SYS$DISK contains directory-specs, andE therefore, as I said, the "default directory", in this case [SYSEXE],lE is ignored. If I would instead explicitly specify []AEF.TMPSKI as theA9 output, *then* the file would go to SYS$SYSROOT:[SYSEXE]:    $ COPY NL: []AEF.TMPSKIdD %COPY-S-COPIED, _NLA0: copied to SYS$SYSROOT:[SYSEXE]AEF.TMPSKI;1 (0 records) $h  h& > > SYS$DISK takes precedence over []. >  > Don't buy that either.    Too bad. I just proved it above.   > > If SYS$DISK doesn'tmH > > contain a directory, then the default directory (whatever is between > > the brackets) is used. >  > ...maybe.r  D And why would VMS ignore the default directory when SYS$DISK doesn't contain one? Example please.  
 > > [snip]+ > > > > $ COPY <file-spec> *   ! works finec > > >tM > > > Can you post some evidence that it "works fine" (SHOW DEFAULT, followedn+ > > > by the display output from COPY/LOG)?  > > ) > > Since you're too lazy to even try it:m > H > I wasn't making the claim - it wasn't my place to produce evidence, it > was yours.  C Uh, proving that it works in all reasonable cases is a VERY BIG JOBaD that I don't have the time to do here, nor should I. You claim it isB bad. If that be the case, then it is FAR, FAR, easier for *you* toE produce a single counter-example than for me to try every reasonable,dD conceivable case. I believe most readers in the group are convinced.  E So let's see. Suppose someone claimed that Newton's Theory of Gravityb= were wrong. Newton would say, "how is it wrong". And you, Mr.h? Dacthera, would say, "Well, Mr. Newton, you need to measure the_B gravitational force between every body and every other body in theE universe at all possible separations before I'll believe it." And Mr.o2 Newton would say, ..., well, let's skip that part.  0 It is up to *you* to produce a counter-example.    > > $ SHOW DEF > >   DISK$DATA1:[FELDMAN] > > $ COPY SYS$MANAGER:A.TMP *9 > > %COPY-S-COPIED, SYS$SYSROOT:[SYSMGR]A.TMP;1 copied to + > > DISK$DATA1:[FELDMAN]A.TMP;1 (0 records)k > > $  > >  > > It works fine. > 
 > Hhmmm... >  > DJAS01::DDACHTERA$ sh def  >   DKA0:[DDACHTERA]( > DJAS01::DDACHTERA$ set def sys$startup > DJAS01::DDACHTERA$ sh def0 >   SYS$STARTUP:[DDACHTERA] ! >   =   SYS$SYSROOT:[SYS$STARTUP]C  >   =   SYS$COMMON:[SYS$STARTUP] >   =   SYS$SYSROOT:[SYSMGR] >   =   SYS$COMMON:[SYSMGR] / > DJAS01::DDACHTERA$ copy sys$login:login.com * 7 > %COPY-S-COPIED, DKA0:[DDACHTERA]LOGIN.COM;1 copied to 0 > SYS$SYSROOT:[SYS$STARTUP]LOGIN.COM;1 (1 block) > . > Oops! That isn't what I wanted ... or is it?  C I don't know what you wanted. Only you know that. But if you didn't F want the file to go to SYS$SYSROOT:[SYS$STARTUP], why the hell did you set default to it?  J > Well, we know that the first element of the path search list is invalid.I > SYS$SYSROOT:[SYS$STARTUP][DDACHTERA] is not a valid path specification.=E > So, I guess the software did pretty much what one would expect - it F > traversed the search list for a path that would lead to a successful1 > operation and it found one in the next element.   D SYS$SYSROOT:[SYS$STARTUP][DDACHTERA] is not the first element of theD search list. It is not even *an* element in the search list. Only if7 you specify [] or [DACHTERA] does it become the "path".-  
 $ SHOW DEF   SYS$STARTUP:[SYSEXE]   =   SYS$SYSROOT:[SYS$STARTUP]m   =   SYS$COMMON:[SYS$STARTUP]   =   SYS$SYSROOT:[SYSMGR]   =   SYS$COMMON:[SYSMGR]>
 $ DIREC/TOTAL0   # Directory SYS$SYSROOT:[SYS$STARTUP]a  r Total of 10 files.  =" Directory SYS$COMMON:[SYS$STARTUP]  c Total of 45 files.  t Directory SYS$SYSROOT:[SYSMGR]  m Total of 350 files.$  u Directory SYS$COMMON:[SYSMGR]   k Total of 405 files.e  n( Grand total of 4 directories, 810 files. $- $ DIREC/TOTAL []  e Directory SYS$SYSROOT:[SYSEXE]    Total of 529 files..    Directory SYS$COMMON:[SYSEXE]o  e Total of 305 files.D  0 Directory SYS$SYSROOT:[SYSEXE]  g Total of 529 files.k  0 Directory SYS$COMMON:[SYSEXE]"  a Total of 305 files.d  d) Grand total of 4 directories, 1668 files.r $c  aD So, as you can see above, the search list does not contain [sysexe],D even though it is a valid directory given the value of SYS$DISK. AndE if you explicitly specify [], you get [SYSEXE], just like I said. Now,E before you go and say "See, but why is it repeated twice?", that's anyA anamoly of combining DIR with a search string default and is even = alluded to in the documentation (User's manual, v6.2, 13.4.2,dF Imprecise File Scpfications). Each element of the SYS$DISK search list@ is matched to []. That's a whole other issue and that's how it's? supposed to work. Besides, if you want [SYSEXE], SET DEFAULT tot SYS$SYSTEM first!n  I > Of course, LOGIN.COM is not real useful anywhere in that portion of the-I > directory structure. However, suppose I wanted to "install" a file thathH > needs to be executed by all nodes in the cluster at system-startup (orJ > maybe even on demand). Clearly, SYS$SYSROOT:[SYS$STARTUP] would not be aD > good choice. For a node-specific proc., it might suffice, however.  C Is your middle name Tera? ;-) (Yeah, I know it starts with a J, I'mg just kidding here.)r  I > Still, I think the point is made. "Works fine" tends to be a subjectiveeG > judgement, where an attempted operation that yields an error message:  > 0 > DJAS01::DDACHTERA$ copy sys$login:login.com []C > %COPY-E-OPENOUT, error opening SYS$COMMON:[DDACHTERA].; as outputp! > -RMS-E-DNF, directory not found $ > -SYSTEM-W-NOSUCHFILE, no such file > G > ...might still be considered subjective; however, from the software'ssG > point of view (it's "subjective opinion"), something is amiss. So, intG > this case, I'd tend to defer my judgement to that of the software andi > the people who wrote it.  C What's your point here? I said it works fine if you don't explcitlyo$ specify []. What's the problem here?   > YMMV, as always, of course.o > L > > > Also, you might try F$PARSE() with your default set as above. In fact,' > > > just try this an post the output:i > > , > > [lengthy examples of parsing [] omitted] > > D > > F$PARSE works fine as long as you don't use []. (Try F$PARSE("")5 > > instead of F$PARSE("[]") and all will work fine.)o > 
 > Does it? > $ > DJAS01::DDACHTERA$ say F$PARSE("") > SYS$SYSROOT:[SYS$STARTUP].;l% > DJAS01::DDACHTERA$ say F$PARSE("*")a > SYS$SYSROOT:[SYS$STARTUP]*.;   Looks okay to me.t  H > Hhmmm... Produces the same result we saw with the COPY, but that stillG > does not guarantee that we gets the results we want or expect, we gettI > the results that fit the software's rules - nothing more, nothing less.j  C Huh? You claimed F$PARSE("[]") showed a problem. I claimed that yourD should try it without the []. And it works fine. So suddenly F$PARSEA doesn't mean anything? Huh? F$PARSE is documented to yield a null B string if the directory doesn't exist. It didn't, so the directory exists (and is fine).s   > [Help excerpt snipped.]t > ' > The bottom line is this default path:- >  > DJAS01::DDACHTERA$ sh def3 >   SYS$STARTUP:[DDACHTERA]a! >   =   SYS$SYSROOT:[SYS$STARTUP]:  >   =   SYS$COMMON:[SYS$STARTUP] >   =   SYS$SYSROOT:[SYSMGR] >   =   SYS$COMMON:[SYSMGR]s > C > ...may work for some operations and yield predictable results forRI > others, but I do not advocate forcing the system to deal with malformed G > path specifications by making them part of a search list knowing thatiC > the system will go on to the next element in the search list upon F > finding that the first is malformed. That technique may be useful in8 > solving certain problems, but it strikes me as sloppy.  E Oh - KAY. So how would you recommend setting default to SYS$STARTUP? .  rF > Again, YMMV considerably, even to the point of diametric opposition.   Disclaimer: JMHO Alan E. Feldmand   ------------------------------    Date: 27 Sep 2002 08:03:42 -0700. From: spamsink2001@yahoo.com (Alan E. Feldman)  Subject: Re: warning during copy= Message-ID: <b096a4ee.0209270703.6c491fee@posting.google.com>H  ` "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message news:<3D93AF7F.1A98442A@fsi.net>... > "Alan E. Feldman" wrote: > > d > > "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message news:<3D926AEF.5A71FF56@fsi.net>... > > > "Alan E. Feldman" wrote: > > > >lh > > > > "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message news:<3D8FC64A.D3CE58CD@fsi.net>...% > > > > > aniruddha patwardhan wrote:a >  [...] [...]tE > > So you're telling me that you can't set default to SYS$STARTUP? !  >  > Well, I'm not, but VMS is. > H > SYS$SYSROOT:[SYS$STARTUP][FELDMAN] is in invalid path specification no> > matter how how slice it. It never will be correct. Same with' > SYS$SYSROOT:[SYS$STARTUP][DDACHTERA].M > H > Now, that will cause other things to happen. Case in point further on.   [long remainder omitted]  F Please read the current online User's Manual (Software Version OpenVMSB Alpha Version 7.3--1 OpenVMS VAX Version 7.3) Section 11.7, and in@ particular, the last two paragraphs of 11.7.1 and all of 11.7.2.   Disclaimer: JMHO Alan E. Feldmanr   ------------------------------  # Date: Fri, 27 Sep 2002 13:52:30 GMT6& From: Bob Willard <BobwBSGS@attbi.com>; Subject: Re: What IO performance expected on GS140 PCI bus?v) Message-ID: <3D9462A9.8351DAFA@attbi.com>e   Bob Drennon wrote: > L > What kind of IO performance should one expect on a 8400 (GS140 cpus) KFTIAL > to DWLPB, OpenVMS 7.3 with latest SCSI-Fibre patch and KGPSA FCAs, that is: > how many megabytes per second will this PCI bus support?  E I don't have the specific data that I think you want, but I will notetD for starters that your question assumes that KGPSAs (and their O'VMSB drivers) are designed to maximize the 8400's PCI bandwidth.  WhileE KGPSAs may be well behaved, they were designed to do real work rather G than to artificially showcase PCI bandwidth.  Most I/O buses, including F the 8400's PCIs, exhibit bandwidth which is asymmetric w.r.t. transferD direction; also, bandwidth varies with the size, alignment, and typeB of DMA bus transactions used.  In some cases, the bandwidth of the> 8400's PCI depends on memory interleaving; bandwidth certainly= depends on the number of KGPSAs *and* their allocation to the?  different PCI segments on a PCI.  = That said, I will be pleasantly surprised if the VMS gang haseA measurements of precisely what you want, since the combination of?B new I/O and new software running on a relatively old system is not8 likely to be a high priority for bandwidth measurements.- {Maybe VB will prove me wrong on this count.}V -- h Cheers, Bobh   ------------------------------  + Date: Fri, 27 Sep 2002 12:50:26 +0000 (UTC)i+ From: david20@alpha1.mdx.ac.uk (David Webb)eJ Subject: Re: [mike@akitanet.co.uk: OpenVMS POP server local vulnerability]+ Message-ID: <an1k6i$c62$1@aquila.mdx.ac.uk>   \ In article <3D9356A3.62CEFCCA@videotron.ca>, JF Mezei <jfmezei.spamnot@videotron.ca> writes: >"Mark(un-MASK)Forsyth" wrote:_ >> The fact that I've been running it for a day can in no way be classed as exhaustive testing.tB >> It was the first quick'n dirty that I thought of to do the job. > N >Consider this: the process can be started on demand by the TCPIP stack, whichG >I assume has privileges even higher than BYPASS, SETPRV and CMKRNL :-)  > M >But the process itself would always run under TCPIP$POP account (or whatevers, >is defined in TCPIP> SHOW SERVICE POP /FULL > H >POP needs to bind to a known port. That, according to the doc requires:K >	The process must have a system user identification code (UIC) **and** the C >	SYSPRV, BYPASS, or OPER privilege to bind port numbers 1 to 1023.- >-N >Now, all of the TCPIP$* accounts are in the TCPIP$AUX group which is at 3655,O >VERY far away from the group numbers in the "system" range. Unless, of course, H >there is an error in the documentation and they really means "**or**".  >(emphasis mine).  >bK >*Perhaps* if the service is defined in the TCPIP> SET SERVICE database, it 2 >doesn't need all those privs to bind to the port. > J >And as far as access to the user's mailboxes, SYSPRV would be all that isN >needed, correct ? It needs "read all" to check the UAF, but needs "read-write, >all" to read and modify the user's mailbox. >aK >One note though: I believe that the POP server has the ability to compressaJ >user's mailboxes. If this means doing the equivalent of MAIL COMPRESS, itK >means that it might need bypass to delete the old files if they are set sodN >that even System has no delete access.  (SYSPRV only grants you access to theP >"System:RWED" class, right ?, whereas Bypass lets you do anything with a file).  J Many files (including mail files) are created with SYSTEM:RW ie without D J access. SYSPRV allows deletion of such files in a two step operation sinceN it also grants control access to the file. You can change the files protectionM and then delete it. So strictly speaking for file deletion if you have SYSPRVo you don't need BYPASS seea  G http://www.openvms.compaq.com/doc/731FINAL/6346/6346PRO.HTM#a654563101 e  F It doesn't really make that much difference whether this buggy program9 was installed with just SYSPRV or with SYSPRV and BYPASS.u  
 David Webb VMS and Unix team leader CCSS Middlesex University   ------------------------------  % Date: Fri, 27 Sep 2002 23:08:57 +1000 9 From: forMsytAhm@optusShom.com.aKu (Mark(un-MASK)Forsyth)=J Subject: Re: [mike@akitanet.co.uk: OpenVMS POP server local vulnerability]8 Message-ID: <slrnap8m37.29m.forMsytAhm@plague.bogus.com>  O On Thu, 26 Sep 2002 17:19:08 +0100, Mike Riley <mike@akita.co.uk> gushed forth:   	 [deletia]a   >pB >Sorry :)  I didn't mean to make anyone nervous here, rest assured: >that we are responsible around release of advisories like  ? Rest assured that I'm not questioning your responsibility here.   A >this - you'll notice this wasn't released until 2/3 months after @ >we'd reported it to Compaq and the ECO fixing it came out.  I'd7 >recommend keeping up to date with the patches, though.|    D You imply that you've found a number of vulnerabilities one of whichO is a combination of an errant qualifier and possibly careless use of privilege.0E What makes me nervous is what vulnerabilities of a similar nature youX _haven't_ found.  F I'll get up to date with the ECO's just as soon as I can beg or borrowI a more recent consolidated distro than the one I have. Of course my causerL would be helped somewhat _if_ I could find an OpenVMS job in Melbourne (Oz.)L There's been exactly _one_ system managers job advertised here in the last 2 years...:-(a   -- e     Oorooh	 Mark F...b  $ Another Optus Cable Traffic Monitor.3 http://www.members.optushome.com.au/forsythm/traff/    ------------------------------    Date: 27 Sep 2002 08:27:58 -0600B From: clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley)J Subject: Re: [mike@akitanet.co.uk: OpenVMS POP server local vulnerability]3 Message-ID: <weFF4oVhrfmT@eisner.encompasserve.org>l  R In article <amvbv4$bnk$1@uns-a.ucl.ac.uk>, "Mike Riley" <mike@akita.co.uk> writes: > H > "Mark(un-MASK)Forsyth" <forMsytAhm@optusShom.com.aKu> wrote in message4 > news:slrnap62q5.1ml.forMsytAhm@plague.bogus.com... >>H >> So am I. It's led me to wonder what else there is in there lurking. I > particularly >> didn't like seeing :- >>D >> "This issue was discovered as part of wider research into OpenVMSA >> security.  Many issues have been found, and further advisoriesa >> will be released shortly."c >>3 >> in the Bugtraq advisory. Makes me quite nervous.h >> > C > Sorry :)  I didn't mean to make anyone nervous here, rest assuredu; > that we are responsible around release of advisories likePB > this - you'll notice this wasn't released until 2/3 months afterA > we'd reported it to Compaq and the ECO fixing it came out.  I'd,8 > recommend keeping up to date with the patches, though. >   I Not speaking for Mark, but in my case it's the fact that you _have_ found6C additional issues that's of concern, not how you have released thatsF information [which appears to have been done in a responsible manner].  @ Now for some firmly worded criticism towards TCP/IP engineering:  F This is a very serious security issue, and for it to be slipped out ofG the door without warning system managers of this issue stinks. Not evenhH Microsoft acts in this manner; they publish their (many) security issuesN for all to see and you can get Microsoft security alerts sent to your mailbox.  M It's not enough to just issue the patch kit; you need to tell system managers F that the latest patch kit fixes a security problem as this affects the2 decision about how soon (or if [*]) to install it.  I Will patch kits be issued to address this problem in previous versions ofDG UCX/TCP/IP ? Will system managers be warned about this or will those onsG previous versions of UCX/TCP/IP who don't happen to see Mike's bulletin C continue running vulnerable systems without them been aware of it ?    Simon.  L [*] Just because a patch kit is issued does not mean that it gets installed.M If you are not experiencing any of the problems that the patch kit addresses, A you may not be able to justify the system downtime to install it.S  C PS: Mike, without giving any information that could be useful to an$H attacker, can you say if the other issues that you have found are in UCX or in base VMS ?   -- OB Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP       + Microsoft: The Lada of the computing world.I   ------------------------------    Date: 27 Sep 2002 11:51:01 -0600B From: clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley)J Subject: Re: [mike@akitanet.co.uk: OpenVMS POP server local vulnerability]3 Message-ID: <ySuZbzY0FzO2@eisner.encompasserve.org>   x In article <weFF4oVhrfmT@eisner.encompasserve.org>, clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley) writes: > O > It's not enough to just issue the patch kit; you need to tell system managers0H > that the latest patch kit fixes a security problem as this affects the4 > decision about how soon (or if [*]) to install it. > K > Will patch kits be issued to address this problem in previous versions ofsI > UCX/TCP/IP ? Will system managers be warned about this or will those on I > previous versions of UCX/TCP/IP who don't happen to see Mike's bulletinME > continue running vulnerable systems without them been aware of it ?: >   H Be aware people, that the security advisory has managed to make it's way onto the Inquirer:  ( http://www.theinquirer.net/?article=5611  K Of course, if we had been told when the ECO had been released to install it:J because it fixed a serious security issue and a plan had been put in placeL to handle older UCX versions, then we wouldn't have to worry about the usersJ getting to possibly see the exploit information before the system managers did.  K I hope that when the fixes for the further issues are released, that we getrK told that it's a security patch at the time that the patch kit is released.L   Simon.   -- 2B Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP       + Microsoft: The Lada of the computing world.E   ------------------------------   End of INFO-VAX 2002.534 ************************