1 INFO-VAX	Mon, 16 Oct 2000	Volume 2000 : Issue 578       Contents: Re: 500A to 500AU C Announce: FRONTPORT 1.0  - Yet Another UN*X to VMS Porting Library! B Re: Anybody else having bugchecks after installing recent patches?B Re: Anybody else having bugchecks after installing recent patches? Re: cluster questions ( Re: Compaq Sollutions Alliance - OpenVMS3 Re: Encompass Bylaw Ballot: THE ENVELOPE, PLEASE... 3 Re: Encompass Bylaw Ballot: THE ENVELOPE, PLEASE... 3 Re: Encompass Bylaw Ballot: THE ENVELOPE, PLEASE... " Free Express Mail Server Software!4 Re: Getting status value's text message (lib$signal) Re: Halon dump - a data point  Re: Halon dump - a data point  Re: Halon dump - a data point  Re: Halon dump - a data point  Re: Halon dump - a data point  Re: html on vms  Image backup & dns$cache files Re: Mail forwarding $ Re: malloc: SYSTEM-F-ROPRAND failure$ Re: malloc: SYSTEM-F-ROPRAND failure$ Re: malloc: SYSTEM-F-ROPRAND failure Map a shared image  Re: Migrating the last VAX users Re: MV3100/80 boot problem.  Newbie DEBUGger question Re: Newbie DEBUGger question Re: nonfree Freeware Re: nonfree Freeware pthread and 7.3 / Re: Thinking of switching from Multinet to UCX. / Re: Thinking of switching from Multinet to UCX. . Re: time to consolidate the TCP/IP work on VMS. Re: time to consolidate the TCP/IP work on VMS9 Re: WAY OT: Ballistic (was Re: Halon dump - a data point)  Re: Welcome to Encompass! $ Re: Why has Compaq retired DECnet ??  F ----------------------------------------------------------------------  % Date: Sun, 15 Oct 2000 17:32:06 -0700 5 From: "cstranslations" <cstranslations@email.msn.com>  Subject: Re: 500A to 500AU) Message-ID: <eCPMhhwNAHA.323@cpmsnbbsa09>   < Malcolm Dunnett <nothome@spammers.are.scum> wrote in message& news:JZu5tUJ3a2bk@malvm1.mala.bc.ca...+ > In article <eNgPqmkNAHA.367@cpmsnbbsa07>, < >    "cstranslations" <cstranslations@email.msn.com> writes: > J > >> and found out (through much hard work) that (supposedly) the 500A can NOT D > >> run OpenVMS.   (note: I'm not very happy with myself right now) > >  > > Been there, done that. > > H > > Supposedly if you replace a few pieces of hardware (controller among them) F > > you can get OpenVMS running on it. When I went through this little	 adventure I > > (about 10 months back) I seem to remember that I was going to have to  forkL > > out somewhere between $600 - $800 and there was no 100% guarantee (thereL > > were some emails back and forth with Hoff that were initiated by someone in > > the LUG I'm a member in).  >  > If youK > already have a generic SCSI disk drive chances are it will work. Probably F > around another $100 for a SCSI CD drive ( this may not be absolutely necessary - H > if you have another VMS system available you could boot as a satellite node; > and us backup to make an image copy of its system disk ).   . Mmmm. Must admit that idea never came to mind.  K I'm getting rid of the dec 3000 that I'm running (actually the eBay auction L just ended however I just tried to check the results and eBay is down). I'veL been setting up a 200/233 to dual boot OpenVMS and Tru64 this weekend. AfterK most of yesterday and today I'm almost there. Still have a few minor things H set up (queue manager and tcp/ip under OpenVMS, the printer under both).H I'll have to give it a try sometime in the not to distant future (at the. moment I a bit tired of looking at computers).   Joe    ------------------------------  % Date: Sun, 15 Oct 2000 17:48:22 -0500 ) From: "John E. Malmberg" <wb8tyw@qsl.net> L Subject: Announce: FRONTPORT 1.0  - Yet Another UN*X to VMS Porting Library!/ Message-ID: <sukcuk8t4rvh06@corp.supernews.com>   D Another library for porting UN*X or LINUX programs to OpenVMS is now
 available.  E This is a result of my work in getting SAMBA 2.0.6 to run on OpenVMS.    It can be downloaded from:  , http://eisner.decus.org/~malmberg/frontport/  J It will be required for my upcoming release of SAMBA-VMS (2.0.6) that I am working on packaging up.   This porting library handles:   K Unix filenames that would not be legal on VMS ODS-2 even after translation.    fcntl() locking.  - crypt() that will return a VMS password hash.   5 mapping of "/" and "/dev/mumble to VMS logical names.   - mapping of UN*X UID 0 to a specified VMS UIC.   , Other routines not present in the DEC C RTL.  & A start at LINUX readline() emulation.  H And a standard disclaimer: This is released under the GPL, no support or warrentee offered.  K Send bug reports and other comments to wb8tyw@qsl.network (make the obvious  fix)   -John  wb8tyw@qsl.network   ------------------------------  # Date: Sun, 15 Oct 2000 22:25:27 GMT ( From: "Peter Weaver" <weaverp1@home.com>K Subject: Re: Anybody else having bugchecks after installing recent patches? < Message-ID: <r9qG5.138240$i5.5585536@news1.rdc2.on.home.com>  : "Bruce Vinson" <r4887c@email.sps.mot.com> wrote in message+ news:39E7201E.80E84AAF@email.sps.mot.com... J > I can't help with the problem, but we are running a cluster of ES40's onF > VMS 7.2-1 with those ECOs installed, and have not seen any problems, > bugchecks or otherwise.   H Try doing a SHOW SYSTEM/FULL and see if every uic shows as [0,0]. If youE remove DEC AXPVMS VMS721_SYS V6.0 then the UIC will be correct again.    ------------------------------  % Date: Sun, 15 Oct 2000 18:45:36 -0400 2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>K Subject: Re: Anybody else having bugchecks after installing recent patches? 7 Message-ID: <200010151845_MC2-B701-94C2@compuserve.com>   J         According to the comments in SYLOGICALS.COM, the OPERATOR.LOG fil= e ; is not created for workstations in a cluster.  Try defining  OPC$LOGFILE_ENABLE as "TRUE".   . Message text written by "Jean-Franois Marchal"A >As I read this, I checked I had only installed VMS721_SYS-v0500, B but I'm surprised not to found the perator.log file on the system.  ? The only OPC logical I define during boot is OPC$OPA0_ENABLE to ? "FALSE" to have the console of the XP900 quiet, so boot is much  faster.   B Not sure of the behaviour of this logical, I deassigned it, killed8 the OPCOM process and issued $ sys$system:STARTUP OPCOM.  , There is still no OPCOM file in sys$manager.  = As it is a security issue, I really need to fix this problem.  <    ------------------------------  % Date: Sun, 15 Oct 2000 20:29:52 -0500 ) From: "John E. Malmberg" <wb8tyw@qsl.net>  Subject: Re: cluster questions/ Message-ID: <sukmdnrteipg42@corp.supernews.com>   7 "Phillip Helbig" <helbig@astro.rug.nl> wrote in message ( news:8scimo$apb$1@info.service.rug.nl...G > Prior to configuring some clusters, I've been thinking about the best B > way to do things.  Yes, I have RTFM, but if the answers to these& > questions were there, I missed them. > I > Leaving aside the possibilities of volume shadowing (requires twice the I > number of disks) and dual-ported disks (doesn't work with SCSI disks on F > VAXen), it seems to me that common disks which are used all the timeJ > from various nodes of a cluster, DISK$USER and so on, should actually beJ > served by a SATELLITE.  If such a disk is served by a boot server, sinceI > that boot server has by definition a system disk, it will be out of the E > cluster whenever the system disk is being upgraded---OS update, new F > layered products installed, in some cases when and after patches areG > applied.  On the other hand, if such a disk is served by a satellite, G > the satellite can boot off any boot server which is available, and of J > course one can always make sure that at least one (probably at least two > or three) are available.  J With out multiple network paths to the existing system disk, the satelliteL will require a reboot to change it load host.  Not much of an improvement in reliability.  G > My impression is, however, that this is not a very common setup.  Why @ > not?  (Perhaps sites this concerned with availability will use0 > dual-ported disks or volume shadowing anyway.)   It would be very confusing.   J With out at least a dual path to the common storage, you have at least one single point of failure.  L Your highest fault tolerance is from each system having it's own system diskJ (shadowed), and a vote.  The multiple system disks allow a rolling update.L Of course any satellite should be shutdown while the system disk it is using is being upgraded.  I > I've only set up relatively simple clusters to date.  What happens when ? > the boot server a satellite booted off of leaves the cluster?   K Sooner or later, the satellite discovers that it's system disk is gone, and J the system will likely go into a mount verification state.  If the rest ofK the cluster survived, the satellite will likely hang.  Other bad things can 8 then result from the hang, and it may eventually reboot.   >  Does the J > satellite automatically reboot from another boot server (assume that allH > boot servers in the cluster will let all satellites boot off of them)?  J Once it starts the reboot sequence, it really does not care who feeds it a8 load image.  You must configure the MOP or LANCP for it.  I > Does this happen immediately or only when the satellite needs access to 3 > the system disk and notices that it is not there?   I If there are other voting members, it just goes into a mount verification  wait.   J > Is this a good idea at all, to let a satellite boot off of more than one > boot server?  J I can see a situation where the satellite is of a very simple application.9 I would not want anything critical to be dependant on it.   A > These days, when upgrading system disks, it seems that the most A > efficient way is to do it once on a "master disk" and then do a > > disk-to-disk image backup to other system disks, followed byI > @CHANGE_NODE_NAME.COM (has anyone written this yet?).  This assumes, of H > course, that the system disks should be as identical as possible (i.e.F > identical in everything except node name, IP address etc).  However,I > there might be some slight differences, such as queue names, user names G > etc which might be different on different disks.  It seems logical to I > keep these things which are different on a different disk, so that they G > survive the disk-to-disk image backup used to update the system disk.   : I would expect that a rolling upgrade would be much safer.  J The problem with using a CHANGE_NODE_NAME.COM is that it really depends onF what products that you have installed, other local customizations, andJ dealing with any new dependencies that a VMS upgrade may introduce.  ThereE is a section on this in either the OpenVMS FAQ or the Ask the Wizard.   I Alternatively you can maintain different system roots on the system disk, # and let the boot flags sort it out.   J > Unless each system has a disk for this purpose, one might should put allA > of these things on another disk in the cluster, for the reasons G > mentioned above perhaps on a disk served by a satellite.  Assume that E > just one boot server and the satellite are available.  Is it always F > possible to boot the satellite soon enough after the boot server hasH > started booting that the information regarding differences between theD > system disks of the various boot servers is available before it is! > needed in the startup sequence?   @ I would not count on it.  I would expect the results to be poor.  ' > Similar arguments and questions apply I > for the case of files such as SYSUAF.DAT and so on, which should be the I > same for all members of the cluster and thus logicals should be used to I > point to the common file.  Again, does it make sense to put these files I > on a disk served by a satellite and can the satellite, even if only one C > boot server is available, always be booted before these files are 	 > needed?   E I think you might have a problem if and when the imortal power supply I keeping the whole thing running quits and you have to get one of the boot  servers running.  J > Unless some additional redundancy is needed, is there any point in using0 > dual-ported disks if one has volume shadowing?  K Remember that dual ported is not the same as dual or multiple pathed.  With K a dual port disk such as an RAxxx type disk, only one port is active.  When K the dual ports are connected to two controllers then you are guarded from a J controller failure.  When the controllers are on different hosts, you will3 still have a path to the disk after a host failure.   L Since only one port is active at a time, I think that only dual ported disksK served off an HSC controller can be used as a cluster common system disk by  boot nodes.   K Dual or multiple pathed disks are available to multiple hosts concurrently. F These are your DSSI disks, HSC served disks, or for ALPHA only - SCSI.9 These can be used as a cluster common disk by boot nodes.   L Volume shadowing is mainly to guard against media failure.  You can also useI it to guard against host or controller failures.  Of course if you do not L have a redundant path to all members of a shadow set from the hosts that areE serving them, then you can have interesting problems if the path gets D broken.  This places some practical restrictions on using shadowing.   -John  wb8tyw@qsl.network   ------------------------------   Date: 15 Oct 2000 15:15:58 PDT* From: abuse@flying-disk.com (Alan Frisbie)1 Subject: Re: Compaq Sollutions Alliance - OpenVMS ! Message-ID: <WfxSn49j3$gi@flying>   M In article <OF9C316E84.B9AB35D8-ON83256977.0049A8F5@ep-bc.petrobras.com.br>,  + fabio_compaq@ep-bc.petrobras.com.br writes:   I > An important number: there are 719 companies under the Compaq Solutions & > Alliance with products for  OpenVMS.  6 That would be 720, but for a major screw-up by Compaq:  = I was a member of ASAP for many years, until they tripled the : price and reduced the benefits.   This, on top of spending= several years convincing me to port my software to Windows NT # on Alpha, then saying "never mind".   9 At the recent CETS2000 show, I wandered into the CSA room ; to see what the current program offered.   Two ladies there 9 did a pretty good sales job, but were unable to answer my 7 question about what the current discount structure was.   7 They offered to take me over to the trade show room and 5 introduce me to someone who could answer my question. 8 We walked over, they introduced me to a young, slender, 7 good-looking black woman (I don't recall her name), and  left me with her.   ; I explained that I had dropped out of the ASAP program, but ; was considering re-joining the CSA program.   I just wanted ; to know what the current disount structure was for members. 5 She refused to tell me, saying that I had to be a CSA  member before I could be told!  9 I protested that the reason I had been brought to her was 8 to convince me that it was worthwhile to join CSA.   She9 replied that it was up to *me* to decide that I wanted to 4 partner with Compaq, NOT for them to recruit me.   I; suggested that this was the wrong attitude, and she replied ; that *I* had the wrong attitude.   I turned on my heel and  < walked out.   If this is the way CSA recruits ISV's, I don't want any part of it.   --  B --  Alan E. Frisbie             Frisbie "AT" Flying-Disk "DOT" Com@ --  Flying Disk Systems, Inc.   Abuse "AT" Flying-Disk "DOT" Com   ------------------------------  % Date: Sun, 15 Oct 2000 13:01:29 -0500 7 From: "David J. Dachtera" <djesys.nospam@earthlink.net> < Subject: Re: Encompass Bylaw Ballot: THE ENVELOPE, PLEASE...- Message-ID: <39E9F0F9.C15337A3@earthlink.net>    Jeff Killeen wrote:  > J >          <<< EISNER::DRA1:[NOTES$LIBRARY]CHAPTERCRACKERBARREL.NOTE;1 >>>7 >                            -< ChapterCrackerBarrel >- N > ============================================================================ > ====H > Note 140.0                   DECUS's legal standing                    [snip]   FWIW...   F I've been "doing business as" DJE Systems for some years now. On thoseF occasions when I've received contracts, checks, etc. with the name DJEG Systems instead of David J. Dachtera, I have asked that the document be   re-issued with the correct name.  E The entity undertaking the aspects of legally-enforceable transaction 9 must, itself, be a legally empowered/recognizable entity.   B What's in a name? That's up to you. If voting for the by-laws willH create a "legally empowered/recognizable entity", does it matter whether4 they call it "DECUS", "EnCompAss", or anything else?   IMO...  G "DECUS" is now more a paradigm of how the symposia should be shaped andcE conducted than the name of a user society. "D.E.C." no longer exists,OE except in our memories. So, the name itself has certain associations,pH but not much meaning beyond that. The fact that I personally cling to itE is today more a statement of personal convictions that have yet to be * reconciled with Compaq than anything else.  D Even "Enterprise Computing Association" falls somewhat short, but isD nonetheless the name chosen to replace "DECUS". I don't like it, butH there is little I can do about it other than state my protests. The termC "Enterprise", while not of itself exclusive since it (by dictionaryMD definition) includes businesses of all sizes, has acquired a meaning> that refers only to "big business" - an exclusive connotation.  : Still, the user society need to achieve a certain, legally9 recognized/recognizable status. That entity needs a name..  C That which we call a rose, by another name, would smell as sweet. IwD think that's how it goes - I'm a not student of the "Bard" (don't go there, please).e   -- s David J. Dachteraw dba DJE SystemsA http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/   F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.t   ------------------------------  % Date: Sun, 15 Oct 2000 13:15:50 -0500l7 From: "David J. Dachtera" <djesys.nospam@earthlink.net> < Subject: Re: Encompass Bylaw Ballot: THE ENVELOPE, PLEASE...- Message-ID: <39E9F456.4008C5E6@earthlink.net>    "Karl S. Erbland" wrote: > [major snippage]J > I think this is why you have the dubious distinction of being "known" as > Mr. Spin.a  F Actually, I believe that distinction is widely accorded to Andrew (the	 Sun guy).h   --   David J. DachteraM dba DJE SystemsW http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/   F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.s   ------------------------------  % Date: Sun, 15 Oct 2000 19:49:57 -0400h% From: Karl S. Erbland <karl@ksme.net> < Subject: Re: Encompass Bylaw Ballot: THE ENVELOPE, PLEASE...5 Message-ID: <MPG.14540ca37768a0869896cb@news.alt.net>   D In article <8scg7r$1s4$1@nntp9.atl.mindspring.net>, Jeff@IDM-IO.com  says...e   > L > Finally anyone who suggests we can just form a corporation without a bylawG > change is ignoring the not-for-profit tax issue at the federal level.t > 2 > Jeff Killeen - www.Killeen.cc (All contact info)    F I believe I pointed out in my response that tax status was definitely 2 affected by entity status, that is, incorporation.  G So, you point out that this IS a MAJOR change for the organization and n not something trivial.   Hey, we agree.     Karl   --   Karl Erbland KSME/Business Groups Tiffin, OH 44883  ) I want to be your Encompass board member! ) http://www.superorg.com/goodbyedecus.htmll   ------------------------------  % Date: Sun, 15 Oct 2000 17:49:24 +0001  From: emc823@lycos.com+ Subject: Free Express Mail Server Software! - Message-ID: <0G2I00CK5489A3@mx.west.saic.com>p  A   *******New List 10-14-00!!********                             M      6  Visit the site for more details! http://virtue.nu/bcemc  B The key to success in marketing online is reaching the people who " are really interested in your ad!   9 You need targeted e-mails of business opportunity seekersy< who are ACTIVELY marketing online and trying to expand their business TODAY!a   -->   Highly Deliverable Lists!rB   Our lists are cleaned and updated on a daily basis! Which means B we have the lowest undeliverable rate out there!  We've also made = sure to filter out all those unwanted e-mail address such as . mil, .org, .org, .gov, etc.(                           ! -->   What YOUR Reponse Could Be!eA   If you order the list of 200,000 and achieve a reponse rate of dA only 1%, you could have over 2,000 reponses to your ad, and that J< is just in ONE mailing.  Note:  it usually takes someone an < average of 7 times to be exposed to a message before he/she  responds to it.     Visit the site for more details! http://virtue.nu/bcemc  0  10,000 opportunity seekers e-mails for only $15 **New List 10-14-00**o0  25,000 opportunity seekers e-mails for only $250  50,000 opportunity seekers e-mails for only $350 100,000 opportunity seekers e-mails for only $500 200,000 opportunity seekers e-mails for only $75  
 - Promotions!n  A **FREE with EVERY order, demo of ListMan e-mail manager software  @ to manage your e-mails list and Credit Helper E-book with Links  to Guaranteed Visa's and MC's!  B **Order 50,000 or more e-mails and receive Express Mail Server to  send your e-mails FREE!  1; -Send your e-mails safely bypassing your ISP's mail server!N= -This is not a demo but a permanent license for the software!t  , **Order 100,000 or more e-mails and receive:= - CheckMAN software to accept checks online, by phone, or faxo+ - InfoDisk  with 1000+ Money Making Reports. - URL cloaking softwaret) Combined value of over $250!  Yours FREE!n? _______________________________________________________________JB I received your e-mail as someone interested in Internet Business A Opportunities. If I received your e-mail in error, or you are no gB longer interested, you can be removed at no cost to you simply by 8 pressing "Reply" and place "Remove" in the Subject line.A _________________________________________________________________   t     i  ,  d  s  e  t  i  d  h  P   ------------------------------  # Date: Sun, 15 Oct 2000 19:52:04 GMTP From: richard_maher@my-deja.com = Subject: Re: Getting status value's text message (lib$signal) ) Message-ID: <8sd1t3$u1j$1@nnrp1.deja.com>,   Hi,   G You may have mislaid my previous DECdtm examples that IMHO contain manyRC good programming snipits, one of which is using $putmsg to send the E results of a remote constraint violation over the network back to them requesting client.  C If you are using Deja then search for "out_msg." and select the oneo. with the title "DECdtm DECnet Client example".   Hope it helps.   Regards Richard Maher.  + In article <fmhTw5GtTgMA@eisner.decus.org>,d<   Kilgallen@eisner.decus.org.nospam (Larry Kilgallen) wrote:7 > In article <39E835A9.26A7EF95@videotron.ca>, JF Mezeit& <jfmezei.spamnot@videotron.ca> writes:B > > How can I, inside of a C program, obtain the string value of a status codee > > from a system service etc? > >l > > ? > > I know I can use LIB$SIGNAL to cause stuff to be written toe SYS$OUTPOUT, asbH > > well as a small stack dump, but I just want to have the message textG > > associated with the error code in a variable I can deal with insidee the program. >lA > As an alternative to $GETMSG you can use $PUTMSG with an actionh routineiF > to get the FAO arguments evaluated.  The boolean you return from the5 > action routine controls whether anything is sent toe SYS$OUTPUT/SYS$ERROR.h >a    & Sent via Deja.com http://www.deja.com/ Before you buy..   ------------------------------  # Date: Sun, 15 Oct 2000 22:01:09 GMTl$ From: plugh@NO.SPAM.PLEASE (Caveman)& Subject: Re: Halon dump - a data point/ Message-ID: <FOpG5.196$K94.75227@nnrp3.sbc.net>,  8 In article <d7uisssni1r3vf59a6qb08ospokhoapfuc@4ax.com>,+ Richard  <dr_vmsREMV@techieREMV.com> wrote:l >eD >We had the power shunt to our entire data center a few months ago. ? >The tenants above us have a shower, which someone left on.  ItgF >overflowed and water dripped straight into a ceiling smoke detector. D >It instantly sent the fire system into a stage II alarm and shuntedB >the power.  An on-duty operator was alert enough to hit the halon, >abort switch about 3 seconds before a dump. >eE >16 OpenVMS systems, numerous NT systems, 8 Tru64 systems, and 8tb oft) >storage...only 4 hours to deadstart.  :)s  C This is typical of the inability of most IT/IS managers and CIOs to ' comprehend intelligent risk management.e  F It also is a strong argument against the present demand for "darkroom"H data centers with no on-duty operations staff.  It pisses me off to haveD to come into work myself at 3AM just to let in a Field Circus person+ when the whole building is empty of humans.h           --  D      "A great deal of intelligence can be invested in ignorance when%       the need for illusion is deep."t           -- Saul Bellow   ------------------------------  # Date: Sun, 15 Oct 2000 22:17:59 GMT $ From: plugh@NO.SPAM.PLEASE (Caveman)& Subject: Re: Halon dump - a data point0 Message-ID: <r2qG5.1016$Gz3.91864@nnrp1.sbc.net>  K This thread seems ironic to me, because of the totally preventable disasteraJ of the loss of the C90 at Suitland, MD simply because some fuckwit managerK didn't have the sense to file a response plan with the local FD.  There was H a fire in one of the power supplies, and the FD responded with what theyH normally would for an electrical fire, dry chemical extinguishers.  TheyK sprayed this extremely corrosive substance all over the inside of the Cray,fH essentially destroying the entire unit beyond repair.  This could EASILY= have been avoided by a simple and inexpensive countermeasure._  K While the soot from the fire would likely have resulted in a serious repairtG problem in itself, it is likely that the unit could have been salvaged.n  F On the other hand, the corrosiveness of the dry chemical extinguishingG agents used is such that there was no possibility of salvaging anythingi it touched.e  F The accuracy of longrange civilian weather and climate forecasting wasF impacted severely by this loss, not to mention the cost to replace the capital asset.  H This is just another example of blindly throwing "solutions" at problemsB one hasn't done the work to define, then blindly trusting in their1 reliability until the inevitable disaster occurs.u  G As an amusement, when I worked at a hospital data center, the engineersnE who "tested" the fire alarm about 30% of the time managed to induce aiD Halon dump, requiring evacuation and shutdown of the hospital's dataB center, a large S/390 complex running SMS (Shared Medical Systems)	 software.s  C All it takes is three morons in parallel to block one side of a six B lane freeway, affecting hundreds or thousands of people's commute.     -- hD      "A great deal of intelligence can be invested in ignorance when%       the need for illusion is deep."a           -- Saul Bellow   ------------------------------  # Date: Sun, 15 Oct 2000 22:31:51 GMTt$ From: plugh@NO.SPAM.PLEASE (Caveman)& Subject: Re: Halon dump - a data point/ Message-ID: <rfqG5.628$7v5.79148@nnrp2.sbc.net>t  8 In article <9eghus4cm1o26jeproaan924uba4kk1mu3@4ax.com>,5 Bruce Bergman  <blCHURRObergman@earthlink.net> wrote:tE >  Remember there are new regulations.  Even if you never use it, andoC >you have it at home where you aren't in a commercial or industrial @ >area where inspections & refills are required by the local fireB >department, it still needs periodic service.  New Rule: Every sixF >years the extinguisher needs to be recovered into a big cylinder, allE >the O-ring seals replaced, and refilled.  This is in addition to thef >12-year cylinder hydrotest. > F >  Otherwise the O-ring seals rot, the Halon leaks out, and either youG >have a fire and the extinguisher just goes "Pff...." (followed by evenoF >louder cursing than before...) or you take it to be serviced and find4 >out just how $%#@*&!! much the stuff costs a pound. >sG >  Assuming you can even get Halon extinguishers or refill agent at ALLrE >in a few more years...  (HINT:  If you've been putting off getting aoF >few for by your computers, go soon.  And buy the good commercial-dutyD >all-metal-head units from Amerex or other comparable units, not the >junk^W inexpensive Kidde.)s  D I agree strongly with your points about maintenance.  It is possibleH to install thermal monitoring equipment on the tanks which will indicateG the liquid level without actually weighing them although simply settingkC the tank on a freight scale isn't a bad solution either.  A digitalmE freight scale doesn't cost that much, and can be integrated into yourk) system monitoring software pretty easily.u  D By the way, this also applies to home fire extinguishers.  Check andC maintain yours regularly, or you will have a useless chunk of metali when you need it.l  G On the other hand I should point out that there are modern replacementssH for Halon which are equally effective but do not possess these egregiousM environmental hazard regulations.  I don't recall details, but something liketL DuPont FE-36 (TM) 1,1,1,3,3,3-hexafluoropropane is much less environmentally
 hazardous.             -- tD      "A great deal of intelligence can be invested in ignorance when%       the need for illusion is deep."n           -- Saul Bellow   ------------------------------  % Date: Sun, 15 Oct 2000 21:12:44 -0400.+ From: Tim Shoppa <shoppa@trailing-edge.com>o& Subject: Re: Halon dump - a data point1 Message-ID: <39EA1DCC.3C8B456D@trailing-edge.com>u   Caveman wrote:I > On the other hand I should point out that there are modern replacements J > for Halon which are equally effective but do not possess these egregiousO > environmental hazard regulations.  I don't recall details, but something likenN > DuPont FE-36 (TM) 1,1,1,3,3,3-hexafluoropropane is much less environmentally > hazardous.  @ Hmm, I'll have to check that out.  In a previous career I workedB at places that used Sulfur Hexafluoride in the machines, and I wasC quite impressed at how well it did its job of preventing dielectrici breakdown in the "air" gaps.   Tim.   ------------------------------  # Date: Mon, 16 Oct 2000 02:31:49 GMT 3 From: Bruce Bergman <blCHURRObergman@earthlink.net>o& Subject: Re: Halon dump - a data point8 Message-ID: <u83kuscmc37btcmgg3bua6iek4d9qie7ka@4ax.com>  7 On Sun, 15 Oct 2000 12:05:07 -0500, "David J. Dachtera"P$ <djesys.nospam@earthlink.net> wrote:   >Bruce Bergman wrote:   G >>   Remember there are new regulations.  Even if you never use it, and E >> you have it at home where you aren't in a commercial or industrial B >> area where inspections & refills are required by the local fireD >> department, it still needs periodic service.  New Rule: Every sixH >> years the extinguisher needs to be recovered into a big cylinder, allG >> the O-ring seals replaced, and refilled.  This is in addition to thet >> 12-year cylinder hydrotest. >> dH >>   Otherwise the O-ring seals rot, the Halon leaks out, and either youI >> have a fire and the extinguisher just goes "Pff...." (followed by evennH >> louder cursing than before...) or you take it to be serviced and find6 >> out just how $%#@*&!! much the stuff costs a pound. >-D >Here's what you can do to help ensure that you're not losing agent:I >weigh the cylinder periodically. If the weight starts going down, you'vea
 >got a leak. : >EI >Remember: this agent, in the cylinder, is in a saturated condition; thatfH >is, both vapor and liquid are present. The pressure won't change unlessH >the temperature of the cylinder changes, or the liquid is totally spent >and only vapor remains.  D   True for some of them - I forget which is which between Halon 1204F and Halon 1301.  As you note, one of them is "Self Dispensing" and theB pressure will not drop, but the other is very low pressure at roomF temperature, and has dry nitrogen in the cylinder as a propellant  You> will see the pressure drop, but if it's a "liquid" leak at theF discharge O-ring, it will be rather slight.  If it's from the cylinderE to head seal or the gauge seal, you could get lucky and not lose much / agent, even if all the nitrogen has leaked out.   I >>   Assuming you can even get Halon extinguishers or refill agent at ALLlG >> in a few more years...  (HINT:  If you've been putting off getting arH >> few for by your computers, go soon.  And buy the good commercial-dutyF >> all-metal-head units from Amerex or other comparable units, not the >> junk^W inexpensive Kidde.)o >tI >Then again, if the cheap units are all you can get, GET THEM! ...just be 3 >vigilant about monitoring the weight of each unit.n  C   But the cheap "disposable" ones are not refillable, and IIRC theynD have long been discontinued.  The best you can do if you have one ofE these (often sold for Kitchen and household use...) is have the agento@ recovered by an extinguisher service shop and get a couple bucks credit toward a real one...r         --<< Bruce >>--t -- tK Bruce L. Bergman blCHURRObergman@ NOearthSPAMlink.netEVER  Remove the caps.s9 Troubleshooter - Electrician, Phones, HVAC, Plumbing,... e: 'Current'ly with Westend Electric, Agoura, CA 818/889-9545  < WARNING:  No Unsolicited Commercial E-mail is EVER accepted.   ------------------------------  % Date: Sun, 15 Oct 2000 16:29:22 -0400h- From: JF Mezei <jfmezei.spamnot@videotron.ca>a Subject: Re: html on vms, Message-ID: <39EA139B.255078A3@videotron.ca>   John Nixon wrote:tF > However, the release notes and documentation are all in HTML format. > ' > I reckon I need to install a browser.n  M search your system dick for "mosaic". You may already have it on your system.mW (It came on the VAX at 7.2) You can also get it from other places (mor recent version).    ------------------------------  % Date: Sun, 15 Oct 2000 18:45:34 -0400s2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>' Subject: Image backup & dns$cache fileso7 Message-ID: <200010151845_MC2-B701-94C1@compuserve.com>1  G         The files are temporary files that existed while INDEXF.SYS wasiB being scanned to create the list of files to be backed up and were subsequently deleted.5  J         You could shut down DNS during the backup.  You could try startin= ghH your backup at a different time.  The most practical course, IMHO, is to learn to live with it.  J         What you *should* be worrying about is the fact the version numbe= raJ of disk:[]dns$cache.version is too close to the limit of 32767 for comfor= t.E  Things are likely to get interesting if the version actually reachesoB 32767.   The software *may* be smart enough to cope but I've neverA encountered any that didn't need professional help at this point.f  6 Message text written by INTERNET:blasthand@my-deja.com' >Can anyone through some light on this.a  0 Image backup that runs every night on the system/ disk. The backup completes (checked the saveseth. listing) but for the last couple of nights the! log file contains the following:-E   %backup-e-openin, error openingG& disk:[]dns$cache.0000029487;1 as input" -system-w-nosuchfile, no such file   %backup-e-openin, error opening  disk:[]dns$cache.version;29466 -system-w-nosuchfile  ( These files are definately not present:-  $ $ dir disk:[sys%.sysexe]dns$cache*.*   Directory disk:[sys0.sysexe]  / dns$cache.0000008023;1   dns$cache.0000029513;1) dns$cache.version;29491    Directory disk:[sys2.sysexe]  / dns$cache.0000008014;1   dns$cache.0000029502;1  dns$cache.version;29481   - My feeling is that between backup reading the - file details from the indexf.sys and actuallye1 backing them up the dns clerk has already deletedB/ the versions that the indexf.sys pointed to andy- replaced it with a new versions of the files.r  ! Is there a way I can avoid this?<-   ------------------------------  % Date: Mon, 16 Oct 2000 07:19:35 +020030 From: Didier Morandi <Didier.Morandi@Easynet.fr> Subject: Re: Mail forwarding* Message-ID: <39EA8FE8.4D1DDDD7@Easynet.fr>   Maulis Adam wrote: > $ > MAIL> set forward _NEO,xxx@xxx.comN > %CLI-W-NOLIST, list of parameter values not allowed - check use of comma (,) > /_NEO,xxx@xxx.com/  F Sorry. Bad memory. I didn't have access to VMS when I posted this, and+ of course I forgot about the Set COPY_SELF.    D.   ------------------------------  % Date: Sun, 15 Oct 2000 19:00:45 -0500 / From: Chris Scheers <chris@applied-synergy.com> - Subject: Re: malloc: SYSTEM-F-ROPRAND failurenO Message-ID: <2238E0D12CACE4E6.1BC0BB0EC77B7429.C3C0CEBAA63FD527@lp.airnews.net>i   JF Mezei wrote:u > T > A C program that has worked for a lon time has recently developped an indigestion. >  > VAX-VMS 7.2, DEC-C V 6.0 > E > The program runs fine for the first x number of messages. But then,nK > inexplicably, VMS forces it to barf the following against its will, afterO# > which the image is killed by VMS:s > E > %SYSTEM-ROPRAND, reserved operand fault at PC=00010C9D PSL=03C00004e > (followed by the traceback). > J > I have a feeling that one part of my program fucks up some code or otherE > structure. Can the above information yield any clue on what it is ?a    D Usually, when I see something like this, it means that the stack hasE become corrupt.  When trying to return through the corrupt portion of " the stack, you may see this error.  F Since the language is C, the most likely cause of this is going beyond& the limits of an auto allocated array.  G Another common cause which bites me is to allocate an IOSB on the stackoD and then do an asynchronous $QIO.  If the routine returns before theB $QIO completes, the $QIO will corrupt the stack when it writes theE IOSB.  This particular problem can be highly timing dependent and may*D work for a long time and then manifest itself after a hardware or OS upgrade.  
 Good luck!  G -----------------------------------------------------------------------s$ Chris Scheers, Applied Synergy, Inc.  C Voice: 817-237-3360            Internet: chris@applied-synergy.com _   Fax: 817-237-3074y   ------------------------------  % Date: Sun, 15 Oct 2000 21:42:36 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca> - Subject: Re: malloc: SYSTEM-F-ROPRAND failuret, Message-ID: <39EA5D05.31E7DA85@videotron.ca>   Chris Scheers wrote:F > Usually, when I see something like this, it means that the stack hasG > become corrupt.  When trying to return through the corrupt portion of-$ > the stack, you may see this error.  A Obviously, something is getting corrupt. But it is quite strange.sE I have about 90 email messages waiting on the POP server. The programpK consistently has been bombing on the 31st for about 2 days now, on a REMQHI ; instruction in LIB RTL called by DECC-RTL called by malloc.   M So, I manually read the 1st message from the POP server and deleted it. I waswL hoping the program would bomb on the 31s message again. NO ! It now bombs onK the 67th message, on a simple IF statement. The pointer to a structure gets.' zapped with either a 0 or the value 44.e  N This is a program which, just a few weeks ago was able to process 700 messages without problem...  I > Another common cause which bites me is to allocate an IOSB on the stacki# > and then do an asynchronous $QIO.o  H Tisk Tisk... Bad Boy ! All of my QIOs are in a different module and have everything declared static.?   > Good luck!  D I don't want luck ! I need patience  :-( :-( :-( :-( :-( :-( :-( :-(  7 OK, gotta go, it is about to get to that 67th message !o   ------------------------------   Date: 16 Oct 2000 03:32:58 GMT1 From: haimo@pclinuw01.cern.ch (Haimo G. Zobernig)s- Subject: Re: malloc: SYSTEM-F-ROPRAND failurew* Message-ID: <8sdsta$h4a$1@sunnews.cern.ch>  K Just a guess, but often when I got this error it was due to a function call M where caller and callee didn't agree on the number of arguments.  The actual  N error usually happened at return and it was not always easy to locate the bad F code. But with some luck, it may not be too far from your malloc call.    Good Luck, Haimo   G Haimo G. Zobernig                                  tel   +41 22 7677361tG Univ. of Wisconsin                                 fax   +41 22 7678370t c/o CERN Bld. 32 R-A02G 1211 Geneve 23, Switzerland                email Haimo.Zobernig@cern.ch   e   ------------------------------  % Date: Sun, 15 Oct 2000 23:30:44 +0200nB From: "Torsten.Robitzki" <torstenDELETE.robitzkiTHIS@arcormail.de> Subject: Map a shared image , Message-ID: <39EA2204.D957910A@arcormail.de>  
 Hello to all,w> I'm trying to use shared memory with c++ objects. This objects; should be shared by several processes. The problem is, that ; an object with a virtual methode will have a pointer to thes8 virtual methodetable (vtab). The pointer to vtab becomesH initialized while object contstructing. So the value of the vtab-pointer  D and the content of the vtab depends on the address where the code isB maped into virtual addresspace of the object constructing process.@ The try to call one of the virtual methodes of the shared object? from an other process ends up in jump to nirvana (assuming that ' the code is maped to an other address).b? I can see no chance to change the construction of the vtable or.1 to change the code for calls to virtual methodes.-> So now I'm trying to build a shared imaged with the objectcode< and than map that code to the same virtual address for every@ process. To install the image /resident/share=address_data seemsA to do the job. But every shared image my image is linking to must B be install the same way. And this is the problem because I can not install all images this way.C So the question is: how can i bring my objectcode to a fix location ) in virtual addresspace for every process? C Please forgive me my worse english :-) and thanks for every help inx advance, Torsten    ------------------------------   Date: 16 Oct 2000 06:24:14 GMT- From: djweath@attglobal.net (Dave Weatherall)t) Subject: Re: Migrating the last VAX userse5 Message-ID: <DTiotGxQ0bj6-pn2-GdMzJdsRDqEJ@localhost>   C On Thu, 12 Oct 2000 19:10:04, mathog@seqaxp.bio.caltech.edu (David r Mathog) wrote:  i > In article <hQdHaBz6umHY@eisner.decus.org>, Kilgallen@eisner.decus.org.nospam (Larry Kilgallen) writes:h  ? >> But if you think Compaq is going to keep VMS/VAX up to date   forever, well,I > it just doesn't seem likely.  Draw a graph showing the fraction of new hM > features in VMS releases that are making it into VMS/VAX (which is falling tM > with each release) and you'll be able to estimate as well as any of us whenl' > the de facto de-support will occur.  a  D It suddenly struck me,. If Larry prefers debugging on VAX, a view I F share even if its only because I can read the M32 code and thus don't F usually need the /NOOPT qualifier, maybe VMS engineers find it easier  to do some things on VAX too?    Cheers - Dave.   ------------------------------  # Date: Mon, 16 Oct 2000 00:28:50 GMT / From: "Tom Simpson" <simpsont@xxx.mediaone.net> $ Subject: Re: MV3100/80 boot problem.F Message-ID: <6ZrG5.24086$h%1.117983@typhoon.jacksonville.mediaone.net>  B Does it pass the power-on self tests?   My guess would be a memory problem...but it'sJ just a guess.  The MV3100 has built-in tests for the SCSI.  Enter either T SCSI or T 10 to run the RZ tests.  J I'm in Jacksonville also, and I have an external RRD42 (I think. I have toE check and see it's still around) that you could borrow if it comes toe that...  How do you do backups?r   -Tom  - <champ_clark666@my-deja.com> wrote in messagek# news:8s8269$dq4$1@nnrp1.deja.com...RH > In article <ROKF5.20126$h%1.104262@typhoon.jacksonville.mediaone.net>,4 >   "Tom Simpson" <simpsont@xxx.mediaone.net> wrote:2 > > Even with the STARTUP stuff bypassed (again)?? > >r > > -Tom > * > No no.. Let me explain a little better.. >rB > - I got a MV3100,  and decided to put it on the Internet. Booted9 > the system (successfully) and modified somefiles in the,! > twg$tcp:[netdist.etc] diectory.n >e> > - Rebooted the system,  and it would not bring me to a login; > prompt on reboot (The system was technically booted,  butt> > there was a error in one of the twg$tcp:[netdist.etc] file).< > It kept saying something about a invalid queue name.  Over& > and over again this message came up. >j< > - Reset the system and at the ">>>" prompt (this system is9 > not currently setup to autoboot"),  did a "B/1",  whichL7 > brought me to the SYSBOOT> prompt.  Did the nessesaryn& > commands to "bypass" the STARTUP.COM >s< > - That worked,  and I got to a DCL prompt.  I modified the7 > "broken files",  and SHUTDOWN the system.  This time,n= > I was going to let it boot like normal. (without the "B/1")b > @ > - I got to the ">>>" prompt, and type "b" to boot (as normal).3 > This time,  the system HALT'ed with the followingl > code:a >a
 > HTL INST > PC = 00000B15 PSL = 041F0004 > = >   I'm almost starting to thinks it a hardware (SCSI) error.m> > Its just wierd.. It was booting fine,  then in a five minuteB > period, it stop and HALT'ed with the above!  I haven't kick this< > box (Hell,  the machine is in Tallahassee, Fl.. and I'm inA > Jacksonville).  I was playing around at the ">>>" console,  andh> > I hope I just screwed something up there and it'll be a easyA > fix.  My gut tells me the HD died.. Which sucks because I don't < > have any sort of install media (sigh).   Any help would be@ > great (because I don't want to drive to Tallahassee)... Thanks >M7 > Oh yeah.. Even the "b/1" to get me to the SYSBOOT.EXEl> > doesn't work!  It never gets that far,  and just returns the > same "HLT" message.  >o >c > >n( > > <champ@vistech.net> wrote in message% > news:8s7q09$6qv$1@nnrp1.deja.com...- > > > This is killing me...u > > >cH > > > I've read tons of FAQ's,  but I can't seem to find any informationF > > > regardinging my problem (sigh).  So,  I've decided to post here. > > >eD > > > I recently got a MicroVAX 3100/80 with VMS 5.5-2 loaded.   The > systemF > > > was working fine and dandy.  I decided to reconfigure TWG TCP/IPE > > > settings.  In the process,  I mangled my STARTINET.COM.  No bigr
 > deal.  II > > > booted the VAX from the ">>>" with the "B/1" to get to the SYSBOOT>sH > > > prompt and bypassed the STARTUP.COM stuff... Anyways.. I fixed the4 > > > problem,  then rebooted... And now I get this! > > >r > > > HLT INST! > > > PC = 00000B15 BSL =041F0004r > > >tJ > > > Could the Harddrive had _died_ in that brief amount of time?   I wasJ > > > messing around in the little boot enviroment (type "halt", etc), butH > > > don't think I would have messed up anything terribly serious.  For; > > > example,  I didnt reset the boot devices or anything.g > > >k> > > > Anyone have any idea what this error message might mean? > > >h5 > > >                                   - Champ ClarkB > > >w > > > , > > > Sent via Deja.com http://www.deja.com/ > > > Before you buy.  > >  > >u >v >t( > Sent via Deja.com http://www.deja.com/ > Before you buy.a   ------------------------------  % Date: Sun, 15 Oct 2000 22:10:59 -0400n- From: JF Mezei <jfmezei.spamnot@videotron.ca>E! Subject: Newbie DEBUGger question , Message-ID: <39EA63AB.7260166B@videotron.ca>  9 Can't seem to do what I would expect is a very easy task.i   Character cell mode debugger.   K Once one has SET MODE SCREEN, but prior to starting execution, how does one N display the source code of another module so that one can set a break point on a line in that module ?   N eg: I want execution to be stopped when it reaches a certain line of code in aK routine which is in a different compilation unit than the "main" routine's, H and I want to set that break point prior to execution beginning. Is this
 possible ?   I have tried dif   ------------------------------  % Date: Sun, 15 Oct 2000 22:18:00 -0400o+ From: Tim Shoppa <shoppa@trailing-edge.com>e% Subject: Re: Newbie DEBUGger questione1 Message-ID: <39EA2D18.15C1F5A2@trailing-edge.com>e   JF Mezei wrote:s > ; > Can't seem to do what I would expect is a very easy task.g >  > Character cell mode debugger.  > M > Once one has SET MODE SCREEN, but prior to starting execution, how does onelP > display the source code of another module so that one can set a break point on > a line in that module ?a  > There are hundreds of ways to do this, but the two most commonB ways that I use (depending on what exactly I'm trying to watch) if. I'm trying to debug gode in ANOTHERMODULE are:  H 1. Do SET SCOPE ANOTHERMODULE.  Then you can do a TYPE nn:mm to find the< line you want to debug, and then a SET BREAK %LINE whatever.  ? 2. You can, if you want, do a TYPE ANOTHERMODULE\nn:mm from any = module to look at lines nn through mm, and then set the breaku7 at line kk there with SET BREAK ANOTHERMODULE\%LINE kk.m  / HELP from inside the debugger *is* your friend!t   Tim.   ------------------------------  % Date: Sun, 15 Oct 2000 15:05:40 -0400 + From: Tim Shoppa <shoppa@trailing-edge.com>l Subject: Re: nonfree Freeware 1 Message-ID: <39E9C7C4.3CE20146@trailing-edge.com>    David J. Dachtera wrote: >  > Hoff Hoffman wrote:o > >e > >   If y'all want, [snip]u > 9 > Been talking to the dee-eff-dubbya guys, have ya, Hoff?.  = Speaking of the DFW-guys, I have a stupid question: why can'tnH www.montagar.com sell the Freeware CD anymore?  Did Compaq ban them fromD selling it, or did the montagar people decide that there were better projects to do?   % ISTR it was just $6.95 from Montagar.C   Tim.   ------------------------------  % Date: Sun, 15 Oct 2000 18:45:38 -0400N2 From: "Richard B. Gilbert" <DRAGON@compuserve.com> Subject: Re: nonfree Freeware-7 Message-ID: <200010151845_MC2-B701-94C3@compuserve.com>F           Please don't!i  C         We have a few very noisy advocates of "everything should bev; cheap/free" here but they are, I think, among the minority.t  D         If they think $25 is too much to pay for the CD, they can doJ without by way of protest.  Perhaps they can find someone who will buy it=  J for $25 and make them copies for free!  I would like to be able to purcha= seD the disk.  What is the part number again?  And where can I order it?  J         Some of us understand that you get what you pay for.  I don't wor= ke0 for free and I don't think I ask anyone else to.    > Message text written by INTERNET:hoffman@xdelta.zko.dec.nospamD > If y'all want, I can probably have the part numbers and the recentB   separately-orderable availability of the Freeware pulled off theE   price file -- I can probably get the OpenVMS Freeware back to its =u  H   previous availability only via the OpenVMS distributions and via the =  G   website, if that's the prefered approach.  (I'd prefer not to do thishG   and I'd hope that most customers would also prefer that this not be =$  I   done, of course, but I can ask that the media offering be "pulled" if =   ?   the current distribution mechanisms are viewed as a problem.)   H   What I can't manage is the creation and shipping of CD-ROM media for =  H   free -- if an organization wishes to donate to cover the costs, then =  F   I can likely organize providing the CD-ROM "masters" for the currentI   and the next release.  (The next OpenVMS Freeware release, BTW, has a =l  G   submissions deadline of 27-Nov-2000.)  And, of course, folks are free =   to duplicate the contents of the Freeware on CD-R or media.p  H   AFAIK, the current charge only covers the CD-R(OM) media replication =  E   and stocking, the handling, and the shipping costs.  Nothing more.<6   ------------------------------  % Date: Sun, 15 Oct 2000 18:16:16 -0700 5 From: "cstranslations" <cstranslations@email.msn.com>s Subject: pthread and 7.3) Message-ID: <eHIOM6wNAHA.319@cpmsnbbsa09>.  : We'll see how much I get slapped around for this one . . .  F Anyone know if applying (eco) vms721_update-v0100 to OpenVMS 7.3 (wellF actually to E7.3) is a good idea (or at least the pthead patch that it0 incorporates, vms721_pthread-v0100 - I believe).  I I'm setting up a Alphastation 200 2/233 and put OpenVMS 7.2-1 (DECwindowsiJ Motif 1.2-5) on it. BNU gives a traceback anytime Netscape (3.03) tries toF look at anything that's HTML. This came up (here) back in July and the0 solution (for 7.2-1) is the afore mentioned ECO.  K . . . I wanted to put COM on it but that requires a half dozen other things E and after spending pretty much all weekend on getting this set-up I'mNI getting tired and a bit sloppy. Anyway (being tired and sloppy) I figuredEE I'll upgrade it to 7.3 (well - E7.3) and maybe I'll get lucky and the - pthread fix will have been incorporated . . .S  D I didn't get lucky. Any suggestions welcome (well - if we could keepK "restore the 7.2-1 and then apply the patch, dope" to a minimum since I'm au bit tired and sloppy . . .).  K While we're on the subject of 7.3 (and not to ask another stupid question);@K however, this is my first exposure to 7.3. Does OpenVMS (starting with 7.3) G always (or by default) load the mutiprocessing images? I'm getting thisCK impression from the show system and show cpu commands (among other things).h   Joe    ------------------------------  # Date: Sun, 15 Oct 2000 21:05:21 GMTO  From: jlsue <jlsuexxxz@home.com>8 Subject: Re: Thinking of switching from Multinet to UCX.8 Message-ID: <tg6kuscc2f1cvb66c16prt1lvidt08vg2p@4ax.com>  y >In article <5OjC5.12053$TP6.180193@newsread2.prod.itd.earthlink.net>, "Mike Flaherty" <mflaherty2@earthlink.net> writes:y  J >> 3) Does UCX come with any migration tools.  For example, will I have to7 >> recreate all of my TCP/IP print queues from scratch?n >r  @ I have done this for another customer (I'm a Compaq ProfessionalE Services Consultant).  I believe that our TCP/IP Services engineering : team used these procedures to help Cerner convert as well.  F It isn't too tough to write procedures that will convert your printersC to UCX.  It takes some bit of testing to make sure that all of yourhE different printers work as expected - one big problem we hit was thatrF Unix applications often do *not* adhere to the PostScript standards ofF 120 characters per line (I think that's the standard limit).  This mayD caus problems if you try to use the "tcpip$telnetsym" method because% of line limits on RMS to the devices.n  E Also, if using the system as a print server, if you have a printer in_A the printcap that is remote (ie., has rp and rm fields) - usuallyuA these are printers used with the "tcpip$lpd_smb" processor - thene> you'll never actually see jobs placed in the execution queues.E Instead, the TCPIP$LPD_QUEUE queue will accept the incoming jobs, andaF then re-send them back out.  I didn't care for this because it made itB difficult/impossible to do accounting on which printers were used.F Also, it seems like one "stalled" printer could hang jobs to all other	 printers.t    1 Not speaking for anyone, certainly not DEC/Compaqw- (get rid of the xxxx in my address to e-mail)e   ------------------------------  % Date: Mon, 16 Oct 2000 00:24:22 -0500u1 From: Jeff Schreiber <SCHREIBER@Eisner.DECUS.org>c8 Subject: Re: Thinking of switching from Multinet to UCX.1 Message-ID: <01JVDVCU8P7600A6PF@Eisner.DECUS.org>d  D On Mon, 09 Oct 2000, Mike Flaherty <mflaherty2@earthlink.net> wrote:I >> 2.  Quality!  Both Multinet and UCX are ports of the Berkeley code buthM >> Multinet has a multi-year head start!  DECpaq, up until TCP/IP 5.0A was/isbK >> still finding and fixing the Berkeley bugs plus the ones they introducedh >> during the port >eK >If this is the case then why do I keep running into problems with MultinetBK >that can only be fixed with patches from Process Software that emulate UCXe	 >drivers.   M     When it comes to the BGDRIVER emulation, we don't know every undocumented(N     call into that driver.  Until we get examples of their use from customers,H     due to problems reported, we can't go implement those features.  TheN     problems with missing pieces have been few and far between in my time withK     Process [which starts far before the Multinet aquisition], and have allg,     been delt with on extremely short order.  K >Now I know that Multinet is a great product as far as third party products L >go.  However, it is still a third party product and and it will always haveL >to assimilate to Compaq's code which means it will always be a step behind.  I     Ok...  As far as the BGDRIVER emulation, I'll give you that argument,+C     however that argument can only be based on new things to the BGxJ     emulation that is also used by applications.  But aside of that, thereF     isn't any Compaq code that is assimilated.  We follow our own roadH     maps, and do our own coding.  To say the definition of a third partyF     product is to be a step behind is completely insane and unfounded.  H >but what is going to happen when Compaq does bundle UCX with VMS.  WillI >Process dump Multinet like they did with Purveyor (another great Processe8 >Software product that we had to find a substitute for)?  I     I can't speak for the powers that be in Process Software, but I can'tpK     see any reason for Multinet or TCPware to be dropped.  As long as thereoH     are customers buying licenses and support, there will be a product. I     Purveyor is a much different story, as it was a port to VMS of our NT H     product that was quickly crushed when Microsoft came out with a freeH     webserver.  The NT market is vastly different than the VMS market inE     that with NT, there are many that will take a free product over aoH     strong and solid product.  So the NT product died, and unfortunately'     Purveyor for VMS went down with it.e  B     Believe me when I say Purveyor for VMS isn't something that weJ     [engineering] are happy to have gone, and if it was under our control,I     it would be available to the public as either a product or freeware. sI     However there are many legal issues with the source code that I can'ttK     even begin to comprehend, and until those are worked out, Purveyor willoL     stay a stone cast by former powers-that-be, that gets thrown back in our?     faces whenever someone is in need of slinging a little mud.o       						-Jeff-  I     Ooh.. this is one of those ones I really should put a disclaimer on. aI     The above is my statements, and shouldn't be taken as formal opinionsg4     of Process Software or Platinum Equity Holdings.   --/ Jeff Schreiber,            Process Software LLCd1 schreiber@mx.process.com   http://www.process.com61      TCPware, MultiNet & PMDF: Stronger than Evern      L   ------------------------------  % Date: Sun, 15 Oct 2000 16:25:32 -0400/- From: JF Mezei <jfmezei.spamnot@videotron.ca>s7 Subject: Re: time to consolidate the TCP/IP work on VMSw, Message-ID: <39EA12B5.77245880@videotron.ca>   Javier Henderson wrote:f? >         So, promote the use of NETLIB to use as a single API,w; > and let NETLIB worry about interfacing with the IP stack.     N Not good enough unless it comes standard with VMS with VMS-grade documentationJ and support. With the emphasis on DECnet going down and reliance on TCP-IPL going way up, one needs a rock solid , standard, fully documented stack thatL comes with every VMS, in the same way that DECnet came with every VMS system" (except the short lived MicroVMS).  N When competitors ship their systems with a standard TCPIP stack, it means that VMS should too.2   ------------------------------  % Date: Sun, 15 Oct 2000 16:38:57 -0400r+ From: Tim Shoppa <shoppa@trailing-edge.com> 7 Subject: Re: time to consolidate the TCP/IP work on VMSn1 Message-ID: <39E9DDA1.4AEC765D@trailing-edge.com>    JF Mezei wrote:. >  > Javier Henderson wrote:nA > >         So, promote the use of NETLIB to use as a single API,s= > > and let NETLIB worry about interfacing with the IP stack.w > P > Not good enough unless it comes standard with VMS with VMS-grade documentation > and support.  ? I've used NETLIB for many years in many different applications.n2 The documentation is as rock-solid as the product.  : The upside - for me - of NETLIB is that it presents a very8 VMS-centric way of accessing the TCP/IP stack.  i.e. theE calls comform to the standards of VMS, you get AST's for asynchronoush processing, etc.  ? Those who are used to doing things the Unix way - that is, with$F select()'s instead of proper asynchronous I/O interfaces - may be less? comfortable with NETLIB.  But after they "see the light" of hownG much easier async I/O is to do VMS-style, I'd hope they switch over :-)e  ? > With the emphasis on DECnet going down and reliance on TCP-IP N > going way up, one needs a rock solid , standard, fully documented stack that > comes with every VMS  ? NETLIB doesn't come pre-installed, but it *is* easy to install!  To quote from the overview:p      1.1    Overview  =           NETLIB provides a single programming interface that,<           can be used with almost any TCP/IP package for VMS@           systems. While every package provides a $QIO interface@           and most also provide a socket library, the advantages&           to using NETLIB instead are:  =           o  NETLIB allows VMS-style asynchronous programminga8              (using ASTs) without having to use the $QIO?              interface directly. The $QIO programming interface *              varies from vendor to vendor.  <           o  NETLIB selects its vendor-specific support code?              at run-time, allowing you to switch between TCP/IP.>              packages without re-linking your program. You canA              even have the same application, installed just once,'@              run over different TCP/IP packages at the same time7              on different nodes in a VMScluster system.   >           o  NETLIB provides a VMS common language environment=              style of programming interface, making it easier @              to use from other languages than a UNIX-type socket/              library written for C programmers.    Tim.   ------------------------------  # Date: Sun, 15 Oct 2000 21:50:57 GMT.$ From: plugh@NO.SPAM.PLEASE (Caveman)B Subject: Re: WAY OT: Ballistic (was Re: Halon dump - a data point)/ Message-ID: <5FpG5.616$7v5.70164@nnrp2.sbc.net>S  3 In article <39d110ed.6066996@news.cableinet.co.uk>,y&  <greenaum@BOLLOCKSyahoo.co.uk> wrote: >eE >Ballistic missiles are ballistic. They use rockets to launch them upoG >through the against-gravity part of the parabola, then point downwardsND >and free-fall towards the earth in a curve. This means that, by theG >time they're close enough to see, they're going far too fast to stop. .  F So, you're essentially saying that DoD doesn't have access to anything= which has more accelerative power than the earth's gravity?!?   D You might like to discuss that with John Force, the NHRA drag racer.  ? Some hunting rifles fire a projectile faster than the free-falli  velocity of a ballistic missile.  D Like most current problems, it's a matter of latency, not bandwidth.  D You can fire an "anti missile missile" which can easily intercept anH incoming ballistic missile well after it's close to target.  The problemE is the latency of the detection and decisionmaking process to launch, D and the difficulty of hitting such a small target.  It can, however,8 be done, and the proof-of-concept has been demonstrated.  C The real issue in this is whether it's diplomatically desirable forcC everyone to wear a gun on their hip or to simply agree not to shootE! at each other in the first place.      -- -D      "A great deal of intelligence can be invested in ignorance when%       the need for illusion is deep."e           -- Saul Bellow   ------------------------------  % Date: Sun, 15 Oct 2000 18:45:32 -0400 2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>" Subject: Re: Welcome to Encompass!7 Message-ID: <200010151845_MC2-B701-94BF@compuserve.com>w  J         Has it occurred to anyone that the DECUS board has been completel= ylE out of touch with the membership for years?  I'd say it dates back toi4 before the Philadelphia symposium that was canceled!  H         It seems as if all I've been hearing from DECUS for years now isB "NT", "Unix", and "crucify anyone who leaks any information to theJ membership"!!!   The NT and Unix part has been the official DEC/Compaq li= neF for quite a while, but the membership didn't say that the board had to@ follow it.   Crucifixion seems to have been their own invention.    ' Message text written by "Dave Gudewicz"s; >"JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in messagef& news:39E7573B.880A778B@videotron.ca... > Dave Gudewicz wrote: > >tJ > > I for one would like to have Encompass exist.  And if we don't vote f= or6 > > incorporation, it will cease to exist.  What then? >hJ > Sound to me like those stories we hear about the US government blocking=  aJ > whole series of "good" bills because of one bad bill in the lot, and so= me sortH > of requirement that all these bills/laws be voted on as a one package.  J DECUS/Encompass is run (now) by a bunch of volunteers.  Comparing it to t= he5 US government is interesting and I do see your point.k  J > Could DECUS not provide its members with multiple decisions and vote on=  : > individual issues instead of voting on a whole package ?  > Sure they could and with 20/20 hindsight, they probably would.  J > By having multiple questions, the answers would provide the board with = ar much/ > better guidance on what the membership wants.a   Agreed.n  D > For instance, if the membership had no problem with the concept ofJ > incorporation, but had serious problems with renaming DECUS to Encompas= s, thenJ > the board could still proceed with incorporation but keep the DECUS nam= e. Etce
 > etc etc.   Agreed.<   ------------------------------  % Date: Sun, 15 Oct 2000 19:04:13 -0500i/ From: Chris Scheers <chris@applied-synergy.com>,- Subject: Re: Why has Compaq retired DECnet ?? O Message-ID: <C57D24C2B7053B96.4C817A3FF10A6E42.30DFBA8FC3AEDADA@lp.airnews.net>0   Javier Henderson wrote:p > 1 > JF Mezei <jfmezei.spamnot@videotron.ca> writes:v >  > >e > > Phillip Helbig wrote:y@ > > > Note that clustering per se has NOTHING TO DO WITH DECNET. > >oJ > > But many tools commonly used to manage a cluster still rely on DECNET.R > > Although Monitor can now be coaxed into using TCPP instead of DECNET, but what > > about SYSMAN ? > 1 >         SYSMAN doesn't use DECnet, it uses SCS.e  B SYSMAN can be used to manage nodes outside of a cluster.  I alwaysD thought that SCS traffic was only used within a cluster, so I assume) that SYSMAN uses DECnet for this traffic.a    Please correct me if I am wrong.  G -----------------------------------------------------------------------p$ Chris Scheers, Applied Synergy, Inc.  C Voice: 817-237-3360            Internet: chris@applied-synergy.com t   Fax: 817-237-3074    ------------------------------   End of INFO-VAX 2000.578 ************************