1 INFO-VAX	Mon, 28 Jul 2003	Volume 2003 : Issue 413       Contents:( Re: Debugging inbount telnet connections Re: DHCP startup problems  Re: DHCP startup problems  Re: duplicating system disks Re: duplicating system disks Re: duplicating system disks Re: DVD compatibility with VMSP Re: IDC reports PC Market Share: Dell over HP by a nose; both pulling away from P Re: IDC reports PC Market Share: Dell over HP by a nose; both pulling away from P Re: IDC reports PC Market Share: Dell over HP by a nose; both pulling away from 2 Re: Migrate obsolete VAX/VMS SW to OpenVMS Itanium2 Re: Migrate obsolete VAX/VMS SW to OpenVMS Itanium2 Re: Migrate obsolete VAX/VMS SW to OpenVMS Itanium Re: NFS Authentication Re: PDP-11 OS Release Dates  Re: PDP-11 OS Release Dates  Re: PDP-11 OS Release Dates  Re: PDP-11 OS Release Dates  Re: PDP-11 OS Release Dates  Re: PDP-11 OS Release Dates  Re: PPPOE? When? How? $ Re: Terminate with extreme prejudice  F ----------------------------------------------------------------------  # Date: Sun, 27 Jul 2003 19:00:51 GMT ; From: "John Gemignani, Jr." <jon-nope@thiswontworkossc.net> 1 Subject: Re: Debugging inbount telnet connections ; Message-ID: <DjVUa.1411$gn6.355468@news1.news.adelphia.net>   7 "JF Mezei" <jfmezei.spamnot@istop.com> wrote in message # news:3F234667.2CE30F80@istop.com...  > "John Gemignani, Jr." wrote:H > > I'm not sure exactly where you want to place this ... before sending data4 > > into TELNET, or when sending data out of TELNET. >  > I > No, I just want to very easily see on a terminal what the remote telnet  clientI > is sending, and use that terminal's keyboard to send characters to that  remote device. > K > The TCPIP stack on the other device is poorly documented, so I need to be  ableJ > to send stuff to it in variou schunks to see how it behaves. (The writes arteJ > easier since you know exactly how many bytes to write, but the reads are not I > so obvious since you don't know how many bytes you need to read, unless  you do5 > single character IO which adds tremendous overhead.   J You need this just for debugging, I assume.  You're looking to see how the other K end (the client) sends and receives data?  You would be watching the entire 
 connection from start to finish.   L If this is true, then you can accept the incoming TELNET connection yourself on an G alternate port, then create a new connection to the actual TELNET port. 
 Your readsL from one side would write to the other side.  You could display what you see	 on either L side (although since VMS does the echoing, if you are auditing you only want
 to display- what's coming out of the TELNET server side).   K Note that you will see incoming and outgoing TELNET protocol, including the 
 conversionK of the outgoing newlines to NVT sequences, the doubling of IAC (0xFF), etc.   K Since you are dealing with sockets issuing $QIOs with multiple byte buffers  (not single J character I/O) will give you optimized I/O especially for the TN out side.	 The TN in = side usually doesn't receive more than a character at a time.   I The Kerberos TELNET server works very much this way.  Since encryption is  somewhatH expensive it was not done in the driver (not to mention the fact that we only have the KRB I SHR image to work with, not objects that can be linked in with TNDRIVER).   F The TCPIP$TELNET_SERVER image accepts connections, puts it through the kerberosD engine, and the characters that come out are sent to an FTA.  Output characters make the 
 reverse trip.   I You could even make this more fun by using SMG to transliterate sequences  and perhaps H switch between the input and output streams.  You could also use special
 keys with SMG . to allow your input to be directed either way.  8 I know, it's turning into a project by itself, isn't it?   -John    ------------------------------  # Date: Mon, 28 Jul 2003 04:27:38 GMT ; From: "John Gemignani, Jr." <jon-nope@thiswontworkossc.net> " Subject: Re: DHCP startup problems; Message-ID: <_C1Va.4632$gn6.523154@news1.news.adelphia.net>   7 "JF Mezei" <jfmezei.spamnot@istop.com> wrote in message # news:3F1D54E0.B5279B83@istop.com...  > "John E. Malmberg" wrote: L > > There appears to be a way to tell the DHCP server what hostname you haveL > > so that it can update a DNS server.  I have not looked at how to do that > > either.  >  > K > Yep, with TCPIP 5.3, the DHCP server has the ability. You can give it the I > equivalent of a template to use for a host name and it will dynamically L > generate a host name which gets transmitted to the DNS server and will get  > zapped once the lease expires. > J > I haven't tested static allocations to see if you can have a static host nameJ > associated with a static IP address (associated to a MAC address), but IJ > supposed it would also be possible. (although in that case, while it may helpC > centralise management, I am not sure it would really be worth it.   L It also requires the proper version of DNS to support the dynamic update.  IL know that this was done for Alpha, but it may not be for VAX; I will need to check.   -John    ------------------------------  % Date: Mon, 28 Jul 2003 00:58:56 -0400 * From: JF Mezei <jfmezei.spamnot@istop.com>" Subject: Re: DHCP startup problems) Message-ID: <3F24AD81.61431F4D@istop.com>    "John Gemignani, Jr." wrote:N > It also requires the proper version of DNS to support the dynamic update.  IN > know that this was done for Alpha, but it may not be for VAX; I will need to > check.  L Dynamic updates work for the DNS server that shipped with TCPIP Services 5.33 VAX. And the DHCP server for same also supports it.    ------------------------------  # Date: Sun, 27 Jul 2003 18:37:03 GMT % From: "Mike Naime" <mnaime@kc.rr.com> % Subject: Re: duplicating system disks : Message-ID: <jZUUa.41702$o27.846590@twister.rdc-kc.rr.com>  L At work we currently have about 70 VMS OS disk mirrors in our 150 Terra ByteJ SAN.  We use HSG mirroring for redundancy, not host based shadowing.  WhenI we want to make a new cluster, we copy from our CLONE system the OS disk. F (BACKUP/IMAGE/IGNORE=INTER SYS$SYSDEVICE: {target disk})   This way, IF already have the correct OS patches and Licenses that the cluster willI require.  I then have about 1 hours work (including re-boots) per root to B set the  new cluster specific information on that particular root.L Compaq/HP support is fine with this method since we have all of the requiredK licenses on file.  I figure that it saves me about 2 days worth of setup on  each new OS disk.    I have to modify: 
 MODPARAMS.DAT  Cluster config TCPIP 
 Net$configure  DECNET_REGISTER  SMTP SNMP	 Timezone.     3 I currently keep a copy of 7.2-1h1, 7.3, and 7.3-1.   7 Paul Repacholi <prep@prep.synonet.com> wrote in message ' news:87y8yrj0q7.fsf@prep.synonet.com... L > helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) writes:  > ? > > There must be many folks who have multiple system disks for B > > redundancy, which should be identical except for NODE-SPECIFICE > > stuff.  Rather than upgrading, installing layered products on etc D > > ALL disks, it would make more sense to do it just on one "master@ > > disk" then make copies of this for other system disks (quite2 > > comfortably if all system disks are shadowed). > ! > > Does anyone actually do this?  > F > Yep. I used to do it with a truckload of layered products, then nukeA > off the ones that where not licenced on the system in question.  > E > There is a nasty if you boot onto an empty [SYSn] where it will use H > the PARAMS.DAT from syscommon, even if it is a satelite boot :( [SYS0]F > is a double trap, as that will be booted if you bootflags are empty. >  > --> > Paul Repacholi                               1 Crescent Rd.,9 > +61 (08) 9257-1001                           Kalamunda. B >                                              West Australia 6076, > comp.os.vms,- The Older, Grumpier Slashdot0 > Raw, Cooked or Well-done, it's all half baked.H > EPIC, The Architecture of the future, always has been, always will be.   ------------------------------  % Date: Mon, 28 Jul 2003 02:27:44 -0000 ! From: Z  <zarlenga@conan.ids.net> % Subject: Re: duplicating system disks / Message-ID: <vi92h09vvvkdfe@corp.supernews.com>   $ Mike Naime <mnaime@kc.rr.com> wrote:H : (BACKUP/IMAGE/IGNORE=INTER SYS$SYSDEVICE: {target disk})   This way, I ... N : Compaq/HP support is fine with this method since we have all of the required  @ They're fine with backing up the system disk while the system is running with /ignore=interlock?    I don't think so.    ------------------------------    Date: 27 Jul 2003 23:16:51 -0500+ From: young_r@encompasserve.org (Rob Young) % Subject: Re: duplicating system disks 3 Message-ID: <AlkxL6rnuIvK@eisner.encompasserve.org>   S In article <vi92h09vvvkdfe@corp.supernews.com>, Z  <zarlenga@conan.ids.net> writes: & > Mike Naime <mnaime@kc.rr.com> wrote:J > : (BACKUP/IMAGE/IGNORE=INTER SYS$SYSDEVICE: {target disk})   This way, I > ... P > : Compaq/HP support is fine with this method since we have all of the required > B > They're fine with backing up the system disk while the system is! > running with /ignore=interlock?  >  > I don't think so.   7 	Sure.  You do want to back up your system disk between  	semi-annual downtimes.    				Rob    ------------------------------  % Date: Sun, 27 Jul 2003 21:26:24 -0700 7 From: David Spencer <spencer@spaamfree.pageweavers.com> ' Subject: Re: DVD compatibility with VMS B Message-ID: <270720032126240838%spencer@spaamfree.pageweavers.com>  t In article <013101c34de7$d34e1600$0501a8c0@jo>, eberhard heuser-hofmann <vaxinf@chclu.chemie.uni-konstanz.de> wrote:   > David, > 4 > if you speek of "discs", do you mean CDs or DVDs?? > 
 > eberhard > ----- Original Message -----; > From: "David Spencer" <spencer@spaamfree.pageweavers.com>  > To: <Info-VAX@Mvb.Saic.Com> ' > Sent: Thursday, July 17, 2003 5:54 AM ) > Subject: Re: DVD compatibility with VMS  >  > : > > In article <vhac5k4cge0m2c@news.supernews.com>, Island  > <dbturner@islandco.com> wrote: > > N > > > Can anyone give me a list of DVD's and CDRW that are compatible with VMS > > > (and T64 if possible)  > > > N > > > I am getting asked quite often for these and have not been able to quote > > > N > > > I know it comes down to the 512K block issue, but that is something BELL/ > > > Micro and other distys have no idea about  > > N > > David, I've had a lot of success using a Pioneer DVD-303S SCSI drive in myH > > PWS 500au. I've created ISO9660 discs on my Mac and read them on VMSJ > > without any difficulty. The drive itself has a jumper for 512K blocks. > That's > > my 2 cents worth...  > >  > >   > > -- Dave Spencer, PageWeavers  / I meant DVD-R discs. It also reads CDs as well.      -- Dave    ------------------------------    Date: 27 Jul 2003 12:52:02 -07001 From: keithparris_NOSPAM@yahoo.com (Keith Parris) Y Subject: Re: IDC reports PC Market Share: Dell over HP by a nose; both pulling away from  = Message-ID: <cf15391e.0307271152.37fa684f@posting.google.com>    Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> wrote in message news:<bfivq7$cjv$1@new-usenet.uk.sun.com>... 9 > The merger isn't complete it will only be complete when 7 > all the product lines have merged and that is nowhere  > near being the case.  @ I disagree.  There are some cases where product lines may, quite5 intentionally and quite logically, never be combined.   F When I go into an electronics retailer, like CompUSA, Circuit City, orA Best Buy, I typically see a full range of PCs under the HP brand, C another full line under the Compaq brand, a set of high-end systems E under the Sony brand, and a set of low-end systems under a brand such E as eMachines.  (It's encouraging to me to see that both HP and Compaq C systems are competitive or superior in price and features with both > the low-end and high-end competition.)  It would be foolish toA eliminate either the HP or Compaq PC brands -- we'd lose half our  retailer shelf space.    ------------------------------  # Date: Mon, 28 Jul 2003 04:24:59 GMT L From: winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston - SSRL Admin Cmptg Mgr")Y Subject: Re: IDC reports PC Market Share: Dell over HP by a nose; both pulling away from  6 Message-ID: <00A2380D.FF3AAC80@SSRL.SLAC.STANFORD.EDU>  V In article <3F243237.329A08DC@istop.com>, JF Mezei <jfmezei.spamnot@istop.com> writes: >Keith Parris wrote:A >> the low-end and high-end competition.)  It would be foolish to D >> eliminate either the HP or Compaq PC brands -- we'd lose half our >> retailer shelf space. > N >But retailers will get wise and consider HP and Compaq to be the same and not  >give it twice the shelf space.   ? Why?  If they're selling enough of them, what's to get wise to?    > L >HP will be under trememndous pressure to truly combine its wintel junk into< >one product line to reduce costs and compete against Dell.  > F >Customers will eventually stop buying products with a dead brand nameM >(Compaq). Until then, they may be confused by the difference in brand names.  >   O Customers happily bought "Packard-Bell" computers for years.   That was a brand E name that had been dead for years until some guys bought the name and  started making cheap computers.   F Rolls-Royce continued making Bentleys for years, even though they wereB essentially the same cars.  And how about Ford-Lincoln-Mercury, or Chrysler-Plymouth?    K >Yes, HP is currently playing the "double the shelf space"  exactly because M >integration isn't 100% complete and it can still play that game. (especially G >if Compaq products are still distributed different from HP products).    K Coca-Cola is playing the multiply-the-shelf-space-as-much-as-possible game  M with Coke, Cherry Coke, Vanilla Coke, Lemon Coke (yuck!) and diet versions of H each one, not to mention Sprite Remix.  That strategy seems to have beenK working for -- uh -- 19 years or so; I think it's been that long since the   New Coke debacle.  > N >Another reason I think HP chose to keep the Compaq brand around, at least forK >now, is that it realised that abruptly widthdrawing the Compaq brand might M >have some impact on wintel server cutsomers who were loyal to Compaq and who I >may then feel left out on their own, making itceasier for them to choose F >another vendor. Keeping the Compaq brand and slowly integrating thoseN >customers into the HP world (sales, support etc) will make it much easier for& >HP to then pitch HP branded products.  O Sure.  They bought Compaq and hoped to keep Compaq's desktop customers (and the M enterprise customers too, but the desktop customers were likelier, since they H hadn't been screwed over as much as the enterprise customers).  Is there anything wrong with that?    > L >I give the Compaq brand perhaps one more year before one sees a significant
 >phasing out.   K Why?  It makes sense to get rid of it from the enterprise space as fast as  K possible, since HP has a better rep there, but why should it phase out when  the products sell? > M >If HP wanted shelf space, perhaps it should resurrect the Digital brand name 0 >and then it could have tripple the shelf space.  M Ha-ha.  Doesn't work because consumer PC purchasers never heard of Digital -  M no TV advertising, remember - while a lot of money and effort was poured into ! building a Compaq brand identity.    -- Alan    --  O =============================================================================== 0  Alan Winston --- WINSTON@SSRL.SLAC.STANFORD.EDUM  Disclaimer: I speak only for myself, not SLAC or SSRL   Phone:  650/926-3056 M  Paper mail to: SSRL -- SLAC BIN 99, 2575 Sand Hill Rd, Menlo Park CA   94025 O ===============================================================================    ------------------------------  % Date: Sun, 27 Jul 2003 16:12:56 -0400 * From: JF Mezei <jfmezei.spamnot@istop.com>Y Subject: Re: IDC reports PC Market Share: Dell over HP by a nose; both pulling away from  ) Message-ID: <3F243237.329A08DC@istop.com>    Keith Parris wrote: @ > the low-end and high-end competition.)  It would be foolish toC > eliminate either the HP or Compaq PC brands -- we'd lose half our  > retailer shelf space.   M But retailers will get wise and consider HP and Compaq to be the same and not  give it twice the shelf space.    K HP will be under trememndous pressure to truly combine its wintel junk into ; one product line to reduce costs and compete against Dell.    E Customers will eventually stop buying products with a dead brand name L (Compaq). Until then, they may be confused by the difference in brand names.  J Yes, HP is currently playing the "double the shelf space"  exactly becauseL integration isn't 100% complete and it can still play that game. (especiallyF if Compaq products are still distributed different from HP products).   M Another reason I think HP chose to keep the Compaq brand around, at least for J now, is that it realised that abruptly widthdrawing the Compaq brand mightL have some impact on wintel server cutsomers who were loyal to Compaq and whoH may then feel left out on their own, making itceasier for them to chooseE another vendor. Keeping the Compaq brand and slowly integrating those M customers into the HP world (sales, support etc) will make it much easier for % HP to then pitch HP branded products.   K I give the Compaq brand perhaps one more year before one sees a significant  phasing out.  L If HP wanted shelf space, perhaps it should resurrect the Digital brand name/ and then it could have tripple the shelf space.    ------------------------------  % Date: Sun, 27 Jul 2003 21:24:32 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net> ; Subject: Re: Migrate obsolete VAX/VMS SW to OpenVMS Itanium ' Message-ID: <3F24895F.26409557@fsi.net>    sms@antinode.org wrote:  > 3 > From: "David J. Dachtera" <djesys.nospam@fsi.net>  > L > > > > BUT - take that same 5% and raise it to the second power, it becomes' > > > > 25%, the third power: 125%, ...  > > > I > > > Bad math.  5% raised to the second power is .25%, raised to the 3rd  > > > power is .0125%. > > K > > Try again, using integers instead of fractions. Then you getthe correct  > > results. > H >    It must be difficult for anyone so stupid and/or ignorant of simpleF > arithmetic to cope in the field of computers.  5% = 5/100.  Per centE > _means_ per centum, that is, divided by 100.  5% _is_ a fraction (a 2 > proper fraction, less than one), not an integer. > = >    It ill behooves one to be so adamant when he's so wrong.  > D >    Please pardon this Popeye moment.  I tried, I _really_ tried toH > ignore the original blunder, and I succeeded, but the "correction" was > just too much for me.   C Well, I don't know how it works in your world, but when I have five H pennies, then acquire four more groups of five (and then have five timesD what I once had, I get 25 pennies, not little bits of what were onceH whole coins adding up to less than what I started with. When I have fiveF groups of 25 pennies (five times more than I had on the last round), IF have $1.25, not less than when I only had five groups of five, or even one group of five.  H You're applying the rules of math to the example improperly. The error IB originally made was saying 5%**3, which does not match the penniesD example. That would be (5**3)/100. To me, 5 is an integer as is 1 (aB penny - single coin, undivided amd undiminished), and a penny is 1F one-hundredth of a dollar just as 1% is 1 one-hundredth of x. Thus, weE use integer math to arrive at the correct number of units, then apply  the denominator.  C Two wrongs don't make a right (except in Redmond), so not sure what  point you're chasing here.  : We now return you to your regularly scheduled newsgroup...   --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------  + Date: Sun, 27 Jul 2003 23:56:56 -0500 (CDT)  From: sms@antinode.org; Subject: Re: Migrate obsolete VAX/VMS SW to OpenVMS Itanium ) Message-ID: <03072723565680@antinode.org>   C    Please pardon this waste of group space.  Private e-mail failed.       Mr. Dachtera:  ?    Were you born that way, or is it an acquired characteristic?   * > [...]  To me, 5 is an integer as is 1 (aD > penny - single coin, undivided amd undiminished), and a penny is 1H > one-hundredth of a dollar just as 1% is 1 one-hundredth of x. Thus, weG > use integer math to arrive at the correct number of units, then apply  > the denominator.  B    5 (an integer) is different from 5% (a proper fraction).  1% isF 1/100, not x/100.  (x% is x/100.)  Yes, if you change the problem, you% can expect to get a different answer.   G    You wrote "[...] take that same 5% and raise it to the second power, F it becomes 25%".  This is simply wrong, whether you believe it or not.> When you were (properly) corrected, you reasserted this error.  C    Perhaps the group would benifit if you confined your superfluous G and/or mistaken contributions toa maximum of, say, one per day, instead H of more than one per topic.  I know _I_ would appreciate such restraint.  H ------------------------------------------------------------------------  4    Steven M. Schweda               (+1) 651-699-98183    382 South Warwick Street        sms@antinode,org     Saint Paul  MN  55105-2547    ------------------------------  % Date: Mon, 28 Jul 2003 01:18:42 -0400 * From: JF Mezei <jfmezei.spamnot@istop.com>; Subject: Re: Migrate obsolete VAX/VMS SW to OpenVMS Itanium ) Message-ID: <3F24B222.FD15C41B@istop.com>    sms@antinode.org wrote: I >    You wrote "[...] take that same 5% and raise it to the second power, H > it becomes 25%".  This is simply wrong, whether you believe it or not.@ > When you were (properly) corrected, you reasserted this error.  + 5% of 5% is indeed 0.0025   ( 0.05 * 0.05 )   L However, in the Dacthera context, if , in a group of 100, you have 5 personsK who each tell a secret to another 5 persons, you will end up with 25 people  knowing the secret.    ------------------------------  # Date: Sun, 27 Jul 2003 23:45:56 GMT ; From: "John Gemignani, Jr." <jon-nope@thiswontworkossc.net>  Subject: Re: NFS Authentication ; Message-ID: <UuZUa.1528$gn6.442732@news1.news.adelphia.net>   A "George Cornelius" <cornelius@encompasserve.org> wrote in message - news:Q1wek$cvPmFL@eisner.encompasserve.org... C > In article <yvLSa.6013$KZ.2787814@news1.news.adelphia.net>, "John 7 Gemignani, Jr." <jon-nope@thiswontworkossc.net> writes:  > > E > > "George Cornelius" <cornelius@encompasserve.org> wrote in message 1 > > news:z9JHvIbBjcWp@eisner.encompasserve.org... L > >> You may want to find a way to use real UID's and GID's.  TCPIP ServicesI > >> (UCX) uses the -1 and -2 values as special cases and that may be the L > >> source of your difficulties.  I normally use a client such as Chameleon > K > No endorsement of Chameleon implied.  I have had some problems with it as J > well, and have heard that there are others that may work better with VMS% > these days (Hummingbird, perhaps?).  > F > >> NFS which at initial connect actually supplies a VMS username and passwordJ > >> to a separate 'PCNFS' authentication daemon which then uses the proxyL > >> database in reverse to supply the PC with a UID/GID pair to be used forH > >> subsequent file accesses.  [There doesn't seem to be any particular reason@ > >> for doing this, though, other than adding an extra layer of authenticationF > >> to give the illusion that the flimsy NFS security is really doing something.]  >  > [...]  > H > >> I have found TCPIP 5 proxies to be even more difficult to deal with thanL > >> those of version 4, for they seem to require a unique GID for any given UID,K > >> meaning you probably have to have your UID/GID definitions coordinated  among K > >> all prospective  client machines or you will have proxies that fail to  loadB > >> (getting 'duplicate key' errors). My problem may be that I am
 completelyJ > >> misunderstanding the purpose of providing a GID as part of the proxy,H > >> but it certainly is an ordeal for me to get the proxies right these days.  > K > > Use George's scheme; either PCNFS (to get a UID/GID pair) or find a way  toL > > assign them.  Using PCNFS means that you need a username and password to get 4 > > the pair, which actually is a good way to do it. > H > Hey, I'm now a Unix system administrator as well [yesterday I couldn't spelK > sysadmin, now I _are_ one! ]  and PCNFS daemons are getting hard to find. L > My complaint is not about UCX or the PCNFS layer, just that the underlyingG > NFS authentication - the UID/GID/hostname triple - is extremely weak, J > especially given that in the Unix world UID/GID pairs are not privileged > information. > J > With regard to uniqueness of UID/GID in proxies, I had problems going toH > UCX/TCPIP Services V5 in that I might have been able to have these twoC > proxies under V4, but they would seem to be in conflict under V5:  >  > 9 >     APP1MGR       ND      1001   10  host1.mydomain.com  > 
 >      and > 9 >     APP2USR       ND      1001   37  host2.mydomain.com  > L > Actually, I believe there may be a dual conflict in that UID 1001 probablyL > needs to be associated with the same VMS username as well as with the same > GID in the two proxies.   G     I wrote the new proxy for the V5.1+ NFS as I needed access to proxy  information *     from directly out of the kernel layer.  K     The old behavior was actually a ~bug~ that was capitalized on, that is,  two H     different UIDs on one client being able to map to the same VMS user. The L     other issue was two VMS users being able to translate to the same UID on the L     same host.  Clearly you were at the mercy of the proxy record's position in theE     proxy cache, and this caused problems with customers complaining.   I     In the new scheme, there are uniqueness rules -- one VMS user maps to  one -     UID/GID combination on a particular host.    > I > For PCNFS this is OK but for casual access from Linux desktops, etc., I  had L > been used to not having to coordinate UID's and GID's with the rest of the > world. > I > Also, the change does not seem to be documented anywhere - while V4 was D > possibly more relaxed than the documented behavior, V5 errs in the	 direction 6 > of enforcing an overly rigid a set of mapping rules. >   *     I believe that this was release noted.  E > I suppose the fact that the proxy database must map uniquely in the  reverse G > direction[1], VMS_username/remote_hostname to UID/GID, is part of the  reasonK > for the rigid mapping, but it doesn't seem to explain it all.  Perhaps it D > is the structure of the in-memory lookup tables, a structure quite possiblyJ > ported over fresh under V5 from Tru64 Unix.  Whatever it is, the product no5 > longer supports the degrees of freedom it once did.   I     There is no port here.  The code is specific to VMS.  Yes, it is more  rigid,L     but while your application may be more lax about file access, others are more>     stringent and those customers prefer the current behavior.  L     If you can suggest some particular solution for what you're looking for, please<     do post it.  The change was really not made haphazardly.  	     -John    ------------------------------  # Date: Sun, 27 Jul 2003 20:55:28 GMT 4 From: Brian Inglis <Brian.Inglis@SystematicSw.ab.ca>$ Subject: Re: PDP-11 OS Release Dates8 Message-ID: <eoe8iv8rtmd3lcio38s4gq24q860usa242@4ax.com>  A On Thu, 17 Jul 2003 14:35:35 GMT in alt.sys.pdp11, "Kelvin Smith"  <fcs_smith1111@snet.net> wrote:    > = >"Thomas Dickey" <dickey@saltmine.radix.net> wrote in message % >news:bf13le$8pa$1@news1.radix.net... > >> In comp.os.vms Kelvin Smith <fcs_smith1111@snet.net> wrote:H >> > Well, at least one reason for the short variable names was so there	 >would be L >> > no confusion between variable names and keywords, an ambiguity that hasI >> > tripped up more than one programmer since variable names got longer.  >ThereL >> > would also have been space considerations; it's easy to forget now just >howI >> > frightfully expensive RAM was back then ($1 or more per byte, IIRC).  >Most of >>- >> iirc, it was "core" back then (late 60's). 8 >> "RAM" wasn't a commonly-used term til the early 70's. >> >> (still expensive, of course)  > G >It was still typically called core memory in the late 1970s when I got L >started with PDPs; I don't remember if it was actually physical core memoryE >(i.e., the ferrite donuts with wires running through) or it had been M >replaced with semiconductors by then. I forget that when I'm conversing in a L >group like this, everyone actually understands what terminology like "core"' >means (and perhaps even prefers such).   ? Semiconductor memory was referred to as MOS memory when it came 2 out: core was ferrite torii on boards in drawers.   9 Thanks. Take care, Brian Inglis 	Calgary, Alberta, Canada  --  F Brian.Inglis@CSi.com 	(Brian dot Inglis at SystematicSw dot ab dot ca),     fake address		use address above to reply   ------------------------------  % Date: Sun, 27 Jul 2003 20:58:35 -0400 . From: Glenn Everhart <Everhart-nospam@gce.com>$ Subject: Re: PDP-11 OS Release Dates+ Message-ID: <bg1qka$dr2$1@bob.news.rcn.net>   C There was still a fight for memory by some of the makers of core as B late as 1975 or 76. The original DEC MOS memory for pdp11 had someD batches that had flaky errors and at that time the process was stillF (in retrospect) settling down. Core meanwhile was quite mature by thenF and had the advantage of being non volatile. If you used core for yourG low memory you could get your OS to come back up after a power restore   sometimes without disk access.  H The term "RAM" got popular only as micros got more prominent, which was G a couple years later. The front panel of an old Altair or Imsai looked  F impressive but 256 bytes of memory with no peripherals or some of the @ crud that was being tried (remember "kansas city standard" data H recording and its second-harmonic problems?) was not very attractive to # those who wanted to write programs.   H People who wanted to use BASIC in those days often found ways to get to D a pdp10 terminal. That's one reason Bill Gates transliterated pdp10 H BASIC (listings were readily available from DEC back then) as the story  goes.   I Back then of course you'd spend a few thou to buy 16K words (16 bits) of  I memory. It was common to refer to memory as "core" back then, whether it  F was made of MOS or ferrite, so pervasive had core become for a while. D After all you wanted to make sure nobody confused it with disk. The E older machines that used disk or drums for their main memory were of  * course not much remembered by the mid 70s.     Brian Inglis wrote: C > On Thu, 17 Jul 2003 14:35:35 GMT in alt.sys.pdp11, "Kelvin Smith" ! > <fcs_smith1111@snet.net> wrote:  >  > > >>"Thomas Dickey" <dickey@saltmine.radix.net> wrote in message& >>news:bf13le$8pa$1@news1.radix.net... >>> >>>In comp.os.vms Kelvin Smith <fcs_smith1111@snet.net> wrote: >>> G >>>>Well, at least one reason for the short variable names was so there  >>
 >>would be >>K >>>>no confusion between variable names and keywords, an ambiguity that has H >>>>tripped up more than one programmer since variable names got longer. >> >>There  >>K >>>>would also have been space considerations; it's easy to forget now just  >> >>how  >>H >>>>frightfully expensive RAM was back then ($1 or more per byte, IIRC). >>	 >>Most of  >>- >>>iirc, it was "core" back then (late 60's). 8 >>>"RAM" wasn't a commonly-used term til the early 70's. >>>  >>>(still expensive, of course)  >>H >>It was still typically called core memory in the late 1970s when I gotM >>started with PDPs; I don't remember if it was actually physical core memory F >>(i.e., the ferrite donuts with wires running through) or it had beenN >>replaced with semiconductors by then. I forget that when I'm conversing in aM >>group like this, everyone actually understands what terminology like "core" ( >>means (and perhaps even prefers such). >  > A > Semiconductor memory was referred to as MOS memory when it came 4 > out: core was ferrite torii on boards in drawers.  > ; > Thanks. Take care, Brian Inglis 	Calgary, Alberta, Canada    ------------------------------   Date: 28 Jul 2003 00:56:43 GMT/ From: Thomas Dickey <dickey@saltmine.radix.net> $ Subject: Re: PDP-11 OS Release Dates* Message-ID: <bg1scb$4rc$3@news1.radix.net>  > In comp.os.vms Glenn Everhart <Everhart-nospam@gce.com> wrote:  K > Back then of course you'd spend a few thou to buy 16K words (16 bits) of lK > memory. It was common to refer to memory as "core" back then, whether it hH > was made of MOS or ferrite, so pervasive had core become for a while.   + perhaps you did (or recall people who did).D* I didn't, and did not know anyone who did.   --  = Thomas E. Dickey <dickey@radix.net> <dickey@herndon4.his.com>n http://dickey.his.comt ftp://dickey.his.com   ------------------------------  + Date: Mon, 28 Jul 2003 02:27:27 +0000 (UTC)t From: aek@spies.com (Al Kossow)I$ Subject: Re: PDP-11 OS Release Dates$ Message-ID: <bg21mf$1l6$1@spies.com>  F > By the way, RSX-15 was released late 1970, then was ported to the 11 > as RSX-11D (Hank Krejci).    >   ? I've just scanned a 1971 version of the RSX-15 reference manualh< www.spies.com/aek/pdf/dec/pdp15/DEC-15-GRQA-D_RSX15_1971.pdf  < Do you happen to know where what the roots of the RSX design: came from ? It doesn't seem to come from any DEC operating system from the 60's.D   ------------------------------  # Date: Mon, 28 Jul 2003 01:38:54 GMTn) From: "Dan Brevik" <d.brevik@comcast.net>e$ Subject: Re: PDP-11 OS Release Dates. Message-ID: <O8%Ua.143293$OZ2.27934@rwcrnsc54>  < "Thomas Dickey" <dickey@saltmine.radix.net> wrote in message$ news:bg1scb$4rc$3@news1.radix.net...@ > In comp.os.vms Glenn Everhart <Everhart-nospam@gce.com> wrote: >oL > > Back then of course you'd spend a few thou to buy 16K words (16 bits) ofL > > memory. It was common to refer to memory as "core" back then, whether itI > > was made of MOS or ferrite, so pervasive had core become for a while.t > - > perhaps you did (or recall people who did).n, > I didn't, and did not know anyone who did. >s > --   > Thomas E. Dickey  L Glenn is correct.  We called it core no matter what it was, then went thru aJ transition when we would say, "Core, er, I mean, ..."  Eventually the term- RAM settled in and won the hearts of newbies.   D By the way, RSX-15 was released late 1970, then was ported to the 11C as RSX-11D (Hank Krejci).  The first RSX-15 systems went to Dupont,a@ amongst others.  RSX-15 was very popular and along with MUMPS-15J extended the life of the PDP-15 far beyond the death time upper managementD decreed.  After RSX-11D came the re-implementation known as RSX-11M.F RSX-15 remained a realtime system, RSX-11D tended more and more towardC timesharing and RSX-11M brought RSX back to realtime as its primary 	 function.     
 Dan Brevik   ------------------------------  # Date: Mon, 28 Jul 2003 03:47:24 GMT ) From: "Dan Brevik" <d.brevik@comcast.net>g$ Subject: Re: PDP-11 OS Release Dates/ Message-ID: <g11Va.162679$ye4.109589@sccrnsc01>n  K "Al Kossow" <aek@spies.com> wrote in message news:bg21mf$1l6$1@spies.com...sH > > By the way, RSX-15 was released late 1970, then was ported to the 11 > > as RSX-11D (Hank Krejci).l > >  > A > I've just scanned a 1971 version of the RSX-15 reference manualc> > www.spies.com/aek/pdf/dec/pdp15/DEC-15-GRQA-D_RSX15_1971.pdf >t> > Do you happen to know where what the roots of the RSX design< > came from ? It doesn't seem to come from any DEC operating > system from the 60's.r >i >  > 1 Thanks for scanning it in!  Good to see it again.h  C In August, 2001, I had a get-together with Bruce Mitchell and Ralphe Stammerjohn atK Ralph's summer place in Wisconsin (near the Dells).  Bruce videotaped while  I dideL a mind-dump.  I traced the roots of RSX back to 1959 and showed when and whyJ it evolved as it did.  I have a copy of the three raw videotapes amd Bruce was going toH get them edited (they are a bit amateurish, to say the least.)  However, things keept
 coming up....   D RSX did not come from a DEC operating system, as you observed.  It's intellectualK precedants were a realtime executive writen by John Neblett (now retired ino
 Ashville, NC)uD for the RW-300 process control computer.  Thence to "The Synchronous
 Executive" byhJ me about 1963 for the TRW-330 process control computer.  Then I headed the project5A to write "Ops Control" for the BR-340 in 1964  (Dupont loved it).r Bunker-Ramo left theK process control business and I joined Honeywell where I managed the projectr	 for OLERToA on the DDP-516 (John Haynie and Walt Duncan, chief architechs and  developers).  I joinedI DEC in 1969 in marketing (all marketeers were engineers at that time) and  hired Sam ReesewJ part time to help me write a real time executive for the PDP-15. (Paid him $15,000).  I hadJ known Sam at Honeywell where he dashed off a small RT exec called Samtran. This gave me theK time to just sit back and think.  The basic specifications and architecture  came primarily out of me based on experience.  C My wife and I have a small website and on it you will find some olde
 documents. http://www.demillar.com/RSX/  J First, an outline I prepared before the aforementioned videotape was made. It's really 4 just notes by me but may contain things of interest.  J Next is a very early informal writeup very early in the effort, written at the urging of Bob McInnes. You will see RSX taking shape.  G Third is the oldest:  it is a list of macros for "DEX-15" which was the  first name for RSX-15.  ThisJ list is extremely early and changed quite a bit, but you can see RSX begin to form.  E Then there is a letter of reference writen for me by Bernie LaCroute.a  L The first released version of RSX-15 was provided with source.  On the title page is theiK statement that RSX-15 was programmed by  Sam Reese, Dan Brevik, Hank Krejci  and  Bernard LaCroute.-  >  You might also like to visit Bruce's site for some anecdotes. http://www.miim.com/G Click on "An RSX FAQ"   When you get there look at "General", "Software7 Generalities" and- "The Executive"w     Regards,  
 Dan Brevik   ------------------------------   Date: 27 Jul 2003 22:29:02 GMT2 From: Thierry Dussuet <thierry@squeeeez.no-ip.com> Subject: Re: PPPOE? When? How?3 Message-ID: <slrnbi8khf.21h.thierry@NEPTUNE.Family>/  9 In article <3F1CD160.C8101112@istop.com>, JF Mezei wrote:IJ >> You can get the various network boxes that provide a PPPoE client mode. > I > Yep, and for VMS it is the only way to go. And I'd much rather have PPPhM > support on all VMS platforms than have the VMS engineers develop some PPPoE 0 > software that would be totally useless on VMS. > K > If you operate a lan, you are far better off with a router doing the realpO > TCPIP to PPPOE thing translation. This way, all machines on your lan benefit,C; > and you isolate the stupid PPPoE layer from your own lan.  > K > As far as the router, just choose one that has telnet or even serial portpM > access. This enables you to configure and maintain the router from your VMS O > box. And usually the serial interfaces provide greater functionality than thet > web pages for a router.   M I agree partly -- the OS on these routers aren't always the best.  (Since thedN goal is mostly commercial, as with PPPoE).  Here, the most used ones are ZyXELM routers.  If you let them do NAT (or SUA) for you, and want to have more than J 100 parallel connections (which can get quite realistic sometimes), you'reM stuck.  I had such a router for a while, then switched it rather swiftly intocH bridge mode and let a "real" computer do the NAT,routing,PPPoE and such. Since then, no more problems.n   Thierryr   ------------------------------  # Date: Mon, 28 Jul 2003 04:23:59 GMT ; From: "John Gemignani, Jr." <jon-nope@thiswontworkossc.net>T- Subject: Re: Terminate with extreme prejudice:; Message-ID: <zz1Va.4630$gn6.521996@news1.news.adelphia.net>o  1 "THIS$THAT" <nospam@spamcop.net> wrote in message ) news:bfkbr9$m18$1@grandcanyon.binc.net...6 > Hoff replied: D > >In article <bfk10j$j7h$1@grandcanyon.binc.net>, THIS$THAT writes: > > E > >:How do you terminate a VMS process with extreme prejudice, or ford@ > >:that matter a Solaris one (other than rebooting both boxes)? > >tF > >  Assuming restarting TCP/IP Services does not clear this -- and it& > >  probably won't -- reboot the box. >0= > No, shutting it down using the tcpipconfig menu is what was:? > tried originally but tcpip$nfs_shutdown.com hung in TCPIP$UCPr, > (at the tcpip unmap * /noconfirm command). >:E > >  I don't recall the current TCP/IP Services ECO version for V5.1;,5 > >  I'm running V5.3 with ECO on most local servers.s >e> > Anyone else in Europe waiting a long time for their hobbyist< > CDs?  Sent Montagar CWO for the media 8 weeks ago tomorrow= > in order to get TCP/IP Services V5.3 but nothing's arrived.  > >   > >  Best and fastest to reboot. >n( > Appreciate the fast response, respect.  I What is the priority of the NFS server process that is looping?  NormallyrI NFS is at 8, but if it had an internal problem and attempted to exit withnJ NODELET set, the process is set to priority 0 and it goes into an infinite loop (VMS exit does this).  L I know that a few fixes have been put in TCPIP$NFS_SERVICES/TCPIP$NFS_SERVERK to address some errors.  Be sure that when you get your 5.3 that you updatepH to the latest ECO.  If there are any problems, contact me and I will see what I can do.   -John    ------------------------------   End of INFO-VAX 2003.413 ************************