1 INFO-VAX	Wed, 23 Jul 2003	Volume 2003 : Issue 403       Contents: Re: A Caching Question Re: A Caching Question Ask VMS 5 Re: BACKUP/IGNORE=INTERLOCK (was: Re: OpenVms Backup) 5 Re: BACKUP/IGNORE=INTERLOCK (was: Re: OpenVms Backup) 5 Re: BACKUP/IGNORE=INTERLOCK (was: Re: OpenVms Backup) : Re: BasicPlus Uncompiler (Was Re: PDP-11 OS Release Dates): Re: BasicPlus Uncompiler (Was Re: PDP-11 OS Release Dates): Re: BasicPlus Uncompiler (Was Re: PDP-11 OS Release Dates): Re: BasicPlus Uncompiler (Was Re: PDP-11 OS Release Dates): Re: BasicPlus Uncompiler (Was Re: PDP-11 OS Release Dates)P DEC-AXPVMS-VMS731_CDRECORD-V0100--4.PCSI v. DEC-AXPVMS-VMS731_PCSI-V0100--4.PCSI Re: DECnet over IP Re: DHCP startup problems  Re: DHCP startup problems G Re: Does RT-11 run on the PDP-11/70?  (was Re: PDP-11 OS Release Dates) G Re: Does RT-11 run on the PDP-11/70?  (was Re: PDP-11 OS Release Dates) G Re: Does RT-11 run on the PDP-11/70?  (was Re: PDP-11 OS Release Dates) F Re: Does RT-11 run on the PDP-11/70? (was Re: PDP-11 OS Release Dates)F Re: Does RT-11 run on the PDP-11/70? (was Re: PDP-11 OS Release Dates) Gartner on HP Services Re: HP FUDBusting  Re: HP FUDBusting  Re: HP FUDBusting P Re: IDC reports PC Market Share: Dell over HP by a nose; both pulling away from G Re: Intel's had Dirty Bombs Anthrax and everything! (Was: Re: HP World: P Re: Intel's had Dirty Bombs Anthrax and everything! (Was: Re: HP World: Why Alph# Re: Mozilla... the new NULL process  Re: New OpenVMS-to-Itanium FAQ Re: New OpenVMS-to-Itanium FAQ RE: New OpenVMS-to-Itanium FAQ6 Re: Non-OpenVMS ethernet access to Printserver 17/600?/ Re: Obsolete cross compilers: where did you go?  Re: OpenVms Backup Re: OpenVms Backup Re: OpenVms Backup Re: OpenVms Backup Re: OpenVms Backup Re: OpenVMS Cluster overhead Packed decimal arithmetic in C" RE: Packed decimal arithmetic in C" RE: Packed decimal arithmetic in C" RE: Packed decimal arithmetic in C" RE: Packed decimal arithmetic in C" Re: Packed decimal arithmetic in C Re: PDP-11 OS Release Dates J Re: PL/I to C and C++ translation services, process overview now availableJ RE: PL/I to C and C++ translation services, process overview now available Poor disk IO.  How to fix? Re: Poor disk IO.  How to fix? Re: Poor disk IO.  How to fix? RE: Poor disk IO.  How to fix? Re: Poor disk IO.  How to fix? Re: Poor disk IO.  How to fix? Re: Sun and SCO   Terminate with extreme prejudice$ Re: Terminate with extreme prejudice$ Re: Terminate with extreme prejudice$ Re: Terminate with extreme prejudice$ Re: Terminate with extreme prejudice  F ----------------------------------------------------------------------    Date: 22 Jul 2003 12:06:55 -07001 From: keithparris_NOSPAM@yahoo.com (Keith Parris)  Subject: Re: A Caching Question = Message-ID: <cf15391e.0307221106.4ae53cd9@posting.google.com>   l "Michael D. Ober" <obermd-@-alum-mit-edu-nospam> wrote in message news:<Z1bTa.6$kf.16533@news.uswest.net>...H > bascially, the built in cache in VMS is too small and not real bright.  $ This situation has changed with XFC.   ------------------------------  % Date: Tue, 22 Jul 2003 14:55:39 -0600 6 From: "Michael D. Ober" <obermd-@-alum-mit-edu-nospam> Subject: Re: A Caching Question 1 Message-ID: <gxhTa.177$bG2.58776@news.uswest.net>    Thats good to hear. > "Keith Parris" <keithparris_NOSPAM@yahoo.com> wrote in message7 news:cf15391e.0307221106.4ae53cd9@posting.google.com... C > "Michael D. Ober" <obermd-@-alum-mit-edu-nospam> wrote in message * news:<Z1bTa.6$kf.16533@news.uswest.net>...J > > bascially, the built in cache in VMS is too small and not real bright. > & > This situation has changed with XFC.   ------------------------------  # Date: Tue, 22 Jul 2003 21:02:17 GMT , From: "James Gessling" <jgessling@yahoo.com> Subject: Ask VMS; Message-ID: <tDhTa.3$K82.556177@newssvr21.news.prodigy.com>   % Looking for answers to VMS questions? 2 Try:    http://www3.compaq.com/support/ask/askvms/  H Someone posted this as apart of a reply, but I thought it deserved a topI level mention.  I'm embarassed to say I had never heard of it.  It's like C the old DSNlink that's been webified.  Very useful. (Except for the 
 irritationD of getting into HP's generic search site when responses are listed.)  H Just for fun, I tryed to see how far back the articles go.  A search for? "VMS 3.4" found this from 1984 regarding copying VMS files to a  TOPS-20.  L http://h18000.www1.hp.com/support/asktima/operating_systems/COS-TOPS20000-00 274-01.html   L Can anyone beat that?  (Sorry to so frivilous, but as I sit here in a mostly empty K office building with the cubes being disassembled  around me, it's hard not  to. look for a diversion from the current madness)  
 Regards,  Jim    ------------------------------    Date: 22 Jul 2003 15:40:45 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) > Subject: Re: BACKUP/IGNORE=INTERLOCK (was: Re: OpenVms Backup)3 Message-ID: <j97zlgFkrXmF@eisner.encompasserve.org>   ` In article <3F1D4181.EEB2EF1C@ost.cdrh.fda.gov>, Jonathan Boswell <jsb@ost.cdrh.fda.gov> writes: > Rob.Buxton@wcc.govt.nz wrote: I >> Now compared to some I'm a newcomer, only been involved with VMS since 4 >> 4.4, but I've never had to recover a System Disk. > Q > True Story time.  (I go back to 1979, but I cannot remember what version of VMS J > that was.  I do remember having to type some of those darn backwards RSXP > commands like PIP, and I once sysgen'd myself into oblivion before discovering- > autogen, if in fact it existed that early.)   C    IIRC autogen was circa VMS 4.0 (1984).  1979 was more likely 1.x   N > The only time I have ever had to recover a system disk was by performing theO > Digital-approved standalone backup!  So, those who say there are risks to the R > /ignore=interlock must also realize that there are risks to standalone backup as  > well, given human fallibility. > O > I was attempting to de-fragment my system disk when I committed the following O > blunder.  I backed up the WRONG DISK due to a single-character typo, and then H > went blind to that mistake and restored the image over my system disk.  F    I used to try very hard to convince the operations staff to use theE    write lock switch on the source disk when doing standalone DSC or  G    standalone backup of the system disk.  Somehow we only lost a couple C    system disks when changing over from DSC to backup, and actually -    had other backups to fall further back on.   C    Anyone see where the write-lock switch is on a Seagate Baracuda?    ------------------------------  % Date: Tue, 22 Jul 2003 16:13:58 -0500 / From: Chris Scheers <chris@applied-synergy.com> > Subject: Re: BACKUP/IGNORE=INTERLOCK (was: Re: OpenVms Backup)3 Message-ID: <3F1DA916.47FF4977@applied-synergy.com>    Bob Koehler wrote: > b > In article <3F1D4181.EEB2EF1C@ost.cdrh.fda.gov>, Jonathan Boswell <jsb@ost.cdrh.fda.gov> writes:! > > Rob.Buxton@wcc.govt.nz wrote: K > >> Now compared to some I'm a newcomer, only been involved with VMS since 6 > >> 4.4, but I've never had to recover a System Disk. > > S > > True Story time.  (I go back to 1979, but I cannot remember what version of VMS L > > that was.  I do remember having to type some of those darn backwards RSXR > > commands like PIP, and I once sysgen'd myself into oblivion before discovering/ > > autogen, if in fact it existed that early.)  > E >    IIRC autogen was circa VMS 4.0 (1984).  1979 was more likely 1.x  > P > > The only time I have ever had to recover a system disk was by performing theQ > > Digital-approved standalone backup!  So, those who say there are risks to the T > > /ignore=interlock must also realize that there are risks to standalone backup as" > > well, given human fallibility. > > Q > > I was attempting to de-fragment my system disk when I committed the following Q > > blunder.  I backed up the WRONG DISK due to a single-character typo, and then J > > went blind to that mistake and restored the image over my system disk. > H >    I used to try very hard to convince the operations staff to use theF >    write lock switch on the source disk when doing standalone DSC orI >    standalone backup of the system disk.  Somehow we only lost a couple E >    system disks when changing over from DSC to backup, and actually / >    had other backups to fall further back on.  > E >    Anyone see where the write-lock switch is on a Seagate Baracuda?   D Well for a ST34371W (Baracuda 4LP) (RZ29B?), it is 5th jumper on J2.  E It can be wired to a switch.  I don't know if you have to power cycle # the drive for it to notice changes.   G ----------------------------------------------------------------------- $ Chris Scheers, Applied Synergy, Inc.  C Voice: 817-237-3360            Internet: chris@applied-synergy.com     Fax: 817-237-3074    ------------------------------  + Date: Wed, 23 Jul 2003 01:42:39 +0000 (UTC) 7 From: moroney@world.std.spaamtrap.com (Michael Moroney) > Subject: Re: BACKUP/IGNORE=INTERLOCK (was: Re: OpenVms Backup)( Message-ID: <bfkp6f$o1j$1@pcls4.std.com>  / Jonathan Boswell <jsb@ost.cdrh.fda.gov> writes:   M >The only time I have ever had to recover a system disk was by performing the N >Digital-approved standalone backup!  So, those who say there are risks to theQ >/ignore=interlock must also realize that there are risks to standalone backup as  >well, given human fallibility.  ... P >hardware failure.  I was so sure of my commands that it took me half an hour ofP >befuddlement before a backup/list of the tape incontrovertibly showed my deviceO >name typo.  The dia0/dib0 console device names didn't help, since they are all Q >DIAn devices to VMS.  (My system disk was dia0 from the console on one node, but  >dib0 from the other.)  H I think it was V5 when VMS went from calling DSSI disks DIAx/DIBx to allJ DSSI disks being DIAx:  (this was necessary for cluster consistancy, can'tJ have one node calling a disk DIA0: and another calling the same disk DIB0:# like would happen in this cluster!)  --   -Mike    ------------------------------  % Date: Tue, 22 Jul 2003 11:46:19 -0600 . From: Sam Seiber <Sam.Seiber1234@sbseiber.com>C Subject: Re: BasicPlus Uncompiler (Was Re: PDP-11 OS Release Dates) ( Message-ID: <3F1D786B.3E68@sbseiber.com>   Kevin Handy wrote: >  > Sam Seiber wrote:  > > Kevin Handy wrote:G > > If the Bill Lortz you are talking about is who I think he is, he is 2 > > easy enough to find at http://people.yahoo.comF > > Use CA for the state.  He was a good friend of mine in high school > > ('75), andI > > spent the early part of his career programming PDP-11's.  He knew his D > > RSTS/E.  Pretty sure it has to be the same guy.  I havn't talkedB > > to him for a few years, so I am not sure if he is still in CA. > > Worth a try at least.  > >  > D > This one would have been from Idaho, and spent a lot of time here.D > Last I heard was that he was in Tennessee, but that was many years > ago.  A Yep, Skyline High school, Idaho Falls, ID.  As of about 3-5 years E he was in the LA area of CA.  I would try the number for Bill C Lortz = as shown at people.yahoo.com.  That is what I did a few years ; ago when I got ahold of him.  It may take a few tries as he  is pretty busy.   
 Sam Seiber   ------------------------------  % Date: Tue, 22 Jul 2003 15:13:02 -0400 0 From: Christopher McNabb <cmcnabb@n0sp4m.vt.edu>C Subject: Re: BasicPlus Uncompiler (Was Re: PDP-11 OS Release Dates) : Message-ID: <pan.2003.07.22.19.12.57.398101@n0sp4m.vt.edu>  5 On Tue, 22 Jul 2003 11:46:19 -0600, Sam Seiber wrote:     C > Yep, Skyline High school, Idaho Falls, ID.  As of about 3-5 years G > he was in the LA area of CA.  I would try the number for Bill C Lortz ? > as shown at people.yahoo.com.  That is what I did a few years = > ago when I got ahold of him.  It may take a few tries as he  > is pretty busy.  >  > Sam Seiber  I Well, if someone can contact Bill Lortz and get permission, I will gladly ; host the unbac sources on my machine here at Virginia Tech.    ------------------------------  % Date: Tue, 22 Jul 2003 14:32:14 -0600  From: Kevin Handy <kth@srv.net> C Subject: Re: BasicPlus Uncompiler (Was Re: PDP-11 OS Release Dates) & Message-ID: <3F1D9F4E.2010801@srv.net>   Sam Seiber wrote:    > C > Yep, Skyline High school, Idaho Falls, ID.  As of about 3-5 years G > he was in the LA area of CA.  I would try the number for Bill C Lortz ? > as shown at people.yahoo.com.  That is what I did a few years = > ago when I got ahold of him.  It may take a few tries as he  > is pretty busy.  >  > Sam Seiber  A Yup. That's the one. Didn't know where he ended up, and no search @ turned up anything useful when I tried it a couple of years ago.   ------------------------------  # Date: Wed, 23 Jul 2003 00:23:33 GMT 8 From: "Jerome H. Fine" <jhfineb9rv@b9rvnospamcompsys.to>C Subject: Re: BasicPlus Uncompiler (Was Re: PDP-11 OS Release Dates) 4 Message-ID: <3F1DD552.80EBD232@b9rvnospamcompsys.to>   >Christopher McNabb wrote:  8 > >On Tue, 22 Jul 2003 02:55:42 +0000, Don Stokes wrote:* > >> Galen <gspamtackett@yahoo.com> wrote:D > >>Somebody used to distribute a Basic-Plus uncompiler, which was aJ > >>pretty cool and magical tool to impress other computer lab staff with.$ > >We had one of these under RSTS/E.L > Just looking on my pdp-11/83 I see that I have the source for a Basic-PlusI > un-compiler.  It consistes of two files located in PPN [111,111] titled H > unbac.bas and unbac1.bas.  According to the copyright statement in the1 > files, they were written in 1979 by Bill Lortz.    Jerome Fine replies:  B I used RSTS/E many, many years ago - probably around 1975 to 1979.< There was a COMPILE command which saved the source file as aC *.BAC file as opposed to the normal *.BAS file.  When saved in that D form, there was a definite advantage, but I can't remember what that) advantage was.  Can anyone else remember?   ? However, whenever I did what was usually done with a *.BAS file F (to look at the source) to a *.BAC file, the source code also appeared( automatically.  Can anyone confirm this?  ? Also, what am I remembering when I relate this?  The *.BAC file 4 doesn't sound like a truly compiled and linked form!   Sincerely yours,   Jerome Fine  --4 To obtain the original e-mail address, please remove5 the ten characters which immediately follow the 'at'. 8 If you attempted to send a reply and the original e-mail7 address has been discontinued due a high volume of junk 5 e-mail, then the semi-permanent e-mail address can be 7 obtained by replacing the four characters preceding the . 'at' with the four digits of the current year.   ------------------------------  % Date: Tue, 22 Jul 2003 22:00:59 -0400   From: John Santos <JOHN@egh.com>C Subject: Re: BasicPlus Uncompiler (Was Re: PDP-11 OS Release Dates) 5 Message-ID: <1030722215251.1344B-100000@Ives.egh.com>   * On Wed, 23 Jul 2003, Jerome H. Fine wrote:   > >Christopher McNabb wrote: > : > > >On Tue, 22 Jul 2003 02:55:42 +0000, Don Stokes wrote:, > > >> Galen <gspamtackett@yahoo.com> wrote:F > > >>Somebody used to distribute a Basic-Plus uncompiler, which was aL > > >>pretty cool and magical tool to impress other computer lab staff with.& > > >We had one of these under RSTS/E.N > > Just looking on my pdp-11/83 I see that I have the source for a Basic-PlusK > > un-compiler.  It consistes of two files located in PPN [111,111] titled J > > unbac.bas and unbac1.bas.  According to the copyright statement in the3 > > files, they were written in 1979 by Bill Lortz.  >  > Jerome Fine replies: > D > I used RSTS/E many, many years ago - probably around 1975 to 1979.> > There was a COMPILE command which saved the source file as aE > *.BAC file as opposed to the normal *.BAS file.  When saved in that F > form, there was a definite advantage, but I can't remember what that+ > advantage was.  Can anyone else remember?  > A > However, whenever I did what was usually done with a *.BAS file H > (to look at the source) to a *.BAC file, the source code also appeared* > automatically.  Can anyone confirm this? > A > Also, what am I remembering when I relate this?  The *.BAC file 6 > doesn't sound like a truly compiled and linked form! >  > Sincerely yours, > 
 > Jerome Fine   > Basic-Plus is a push-pop compiler (like UCSD Pascal and Java.)  : When source code is read in (from a .BAS file) or typed in: interactively, it is parsed and compiled into instructions< for a virtual stack-oriented (push/pop) system.  The COMPILE> command saves the push/pop code in a disk file (.BAC file), so: it can be reloaded and re-executed without being parsed or8 compiled.  (This makes it *much* faster, a fraction of a: second to load and start executing a .BAC file vs. 10's of seconds for a .BAS file.)    --   John Santos  Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------  + Date: Tue, 22 Jul 2003 21:32:35 -0500 (CDT)  From: sms@antinode.orgY Subject: DEC-AXPVMS-VMS731_CDRECORD-V0100--4.PCSI v. DEC-AXPVMS-VMS731_PCSI-V0100--4.PCSI ) Message-ID: <03072221323521@antinode.org>   A    Not a big deal, but it appears that someone's been cutting and @ pasting without reading carefully.  (Too much automation, or notG enough?)  When I tried to install the new CDRECORD patch without having 7 VMS731_UPDATE-V0100 installed first, the complaint was:    --------  7 The following product will be installed to destination: G     DEC AXPVMS VMS731_CDRECORD V1.0        DISK$VMS073ALP:[VMS$COMMON.]   ! *** KIT WILL NOT BE INSTALLED ***   ,                         vvvvvvvvvvvvvvvvvvvv? In order to install the VMS731_CPU2608-V0100 kit, you must have ,                         ^^^^^^^^^^^^^^^^^^^^> the VMS731_UPDATE-V0100 and  VMS731_PCSI-V0100 (or later) kits1 installed.  Installed kits can be verified with a  PROD SHO HISTORY command.    --------  F    I know better than to waste my time installing VMS731_CPU2608-V0100G for my AlphaStation 200 4/233, but it was a bit startling to be accused  of doing so.  C    How much more would it cost to spell out "PRODUCT SHOW HISTORY"?   H ------------------------------------------------------------------------  4    Steven M. Schweda               (+1) 651-699-98183    382 South Warwick Street        sms@antinode,org     Saint Paul  MN  55105-2547    ------------------------------  % Date: Wed, 23 Jul 2003 13:42:51 +1200 / From: Martin Hunt <martin.hunt@fairfaxnz.co.nz>  Subject: Re: DECnet over IP 8 Message-ID: <6sprhvs1ng285830npbn4enp15i8n70ugo@4ax.com>  E On Fri, 18 Jul 2003 07:28:02 GMT, peter@langstoeger.at (Peter 'EPLAN'  LANGSTOEGER) wrote:   k >In article <pl8ehvkb85qug7bcnhid4hgghna86tq3dr@4ax.com>, Martin Hunt <martin.hunt@fairfaxnz.co.nz> writes: E >>My employer would like to stop running DECnet over its WAN, so I am  >>trying out DECnet over IP. > D >Nice idea. This keeps you the advantages of DECnet (remote terminalL >configuration, lexical functions like F$FILE and F$SEARCH, wildcard supportH >for COPY/DELETE and so on) while bringing in the advantages of TCPIP... >  > K >3) Clear you CDI cache to get rid of all your real DECnet node names there M >to prevent SET HOST trying the wrong connection (which has to timeout then).  > 2 >	NCL>FLUSH SESSION CONTROL NAMING CACHE ENTRY "*"  F That's what I was missing. After flushing, I was able to SET HOST, andD it went over IP. Now, I just need to check out this back translation stuff.   > @ >Check with SYS$SYSTEM:CDI_CACHE_DUMP.EXE before and afterwards.     ---  Martin Hunt  Systems Administrator  Fairfax New Zealand Limited 
 Wellington New Zealand    ------------------------------  # Date: Tue, 22 Jul 2003 18:52:47 GMT 6 From: peter@langstoeger.at (Peter 'EPLAN' LANGSTOEGER)" Subject: Re: DHCP startup problems5 Message-ID: <3KfTa.313771$1F6.3168275@news.chello.at>   V In article <3F1D58BF.B0B7C84C@istop.com>, JF Mezei <jfmezei.spamnot@istop.com> writes:! >Peter 'EPLAN' LANGSTOEGER wrote: ) >> Why QIO ? A simple NCL solves it, too.  > N >I doubt that the DHCP client really wants to spawn a suprocess that has fancyJ >code to start either NCL or NCP to find out the ethernet address and thenN >parse the various output formats to get it into a usable form. QIO is the way >to go.   H You only need a QIO if you create your own DHCP client (which we weren'tG talking about). For us managers a NCL is sufficient to get the infos we # need to interpret the sniffer logs.   Q >> No, it's the opposite. The DHCP client obtains/uses the DECnet address instead K >> of the Hardware address while the rest of the IP stack uses the Hardware 
 >> Address > M >TCPIP 5.3 on VAX uses the DECNET address. (the real decnet, not that phase 5 N >stuff) (decnet started early in boot, TCPIP started late in boot). TCPIP SHOW6 >ARP from another node reveals an aa-00 ether address.  G Yes, this is the case with DECnet-Phase4 as I wrote more than once now. J But with DECnet-Phase5, TCPIP uses the hardware address, while DECnet-PlusH uses the DECnet Mac Address in parallel. And this was what I wrote here.  L >You'd have to use ethermon to monitor the outbound DHCP request to see whatN >ethernet address is used in the *contents* of the packet. ( the actual "from"I >in the ethernet packet is not a reliable piece of information since DHCP G >packets can be relayed in which case, the ethernet "from" becomes that 
 >of a router.    Yup.  L >> I did, or to wait for a fixed versions of the DHCP client (one which uses1 >> the same address as the rest of the IP stack).  > K >I haven't tried the DHCP client. However, in the RFC, it is mentioned that L >~often~, the DHCP client needs to execute before the IP stack is up. (since1 >the IP stack would have no IP address etc etc).   > M >Does the DHCP client on VMS have the ability to run by itself, or must it be ' >started only after TCPIP has started ?   J It is the duty of the TCPIP product to start the DHCP Client. I don't careK if it will run by itself, if it is not documented. I do care however to get J all parameters set the DHCP server provides or my system needs for running
 the stack.  I The TCPIP SET NAME/PATH=(dom1,dom2,dom3,...) is still not possible by the I DHCP Client of TCPIP V5.3 (and V5.4 beta - but MIGHT become possible with I V5.5 at earliest as a private extension, since the DHCP RFC still doesn't I offer a standard parameter for this) so one has to enter such commands to G TCPIP$SYSTARTUP.COM bypassing the advantages of doing this centrally on  the DHCP server.  D The TCPIP DHCP client is working otherwise (like SET NAME/SERVER= or+ SET NAME/DOMAIN= and so on) perfectly well.   6 >In your case, is the DHCP server under your control ?  & Not really. My department, but not me.   --   Peter "EPLAN" LANGSTOEGER % Network and OpenVMS system specialist  E-mail  peter@langstoeger.atF A-1030 VIENNA  AUSTRIA              I'm not a pessimist, I'm a realist   ------------------------------  % Date: Tue, 22 Jul 2003 19:51:44 -0400 * From: JF Mezei <jfmezei.spamnot@istop.com>" Subject: Re: DHCP startup problems) Message-ID: <3F1DCDEC.5048A32D@istop.com>     Peter 'EPLAN' LANGSTOEGER wrote:8 > >In your case, is the DHCP server under your control ? > ( > Not really. My department, but not me.  J If you could get them to increase logging level and then have your VMS box> issue a DHCP request, it may reveal eaactly what is happening.   ------------------------------    Date: 22 Jul 2003 14:26:32 -05004 From: kaplow_r@encompasserve.org.TRABoD (Bob Kaplow)P Subject: Re: Does RT-11 run on the PDP-11/70?  (was Re: PDP-11 OS Release Dates)3 Message-ID: <HWr$5ll7cn5f@eisner.encompasserve.org>   r In article <3f19da65_1@news.iprimus.com.au>, "Christine Ricketts/Andrew Stewart" <u1276a@uxnxixtxe.com.au> writes:6 > "Douglas A. Gwyn" <DAGwyn@null.net> wrote in message, > news:4b2dndNoSs4WUYWiXTWJjQ@comcast.com... >> Rob Warnock wrote: L >> > And OS-8 could even run with *only* DECtape as the "system disk", too!! >>7 >> I once ran RT-11 off DECtape on a PDP-11/70 in DEC's & >> Marlboro facility.  It was amusing. >> > A > Interesting.  The RT-11 V5.5 SPD does not mention the PDP-11/70 @ > yet p257 of the RT-11 V5.5 Mini Reference Manual has bit 14 ofA > the Configuration Word 2 is set if it is a PDP-11/70 processor.   J As many have already noted, RT-11 was never supported on the 11/70. But inL my experience it worked fine. There was even an 11/70 console light blinking' pattern built into the RT-11 idle loop.   H [Each PDP-11 OS had a unique blinken lights pattern on the 11/70. I long! since forgot what they all were.]   1 	Bob Kaplow	NAR # 18L	TRA # "Impeach the TRA BoD" & 		>>> To reply, remove the TRABoD! <<<K Kaplow Klips & Baffle:	http://nira-rocketry.org/LeadingEdge/Phantom4000.pdf L     www.encompasserve.org/~kaplow_r/    www.nira-rocketry.org    www.nar.org  O  Save Model Rocketry from the HSA!   http://www.space-rockets.com/congress.html    ------------------------------  # Date: Wed, 23 Jul 2003 00:23:53 GMT 8 From: "Jerome H. Fine" <jhfineb9rv@b9rvnospamcompsys.to>P Subject: Re: Does RT-11 run on the PDP-11/70?  (was Re: PDP-11 OS Release Dates)4 Message-ID: <3F1DD567.75666ADF@b9rvnospamcompsys.to>   >Bob Koehler wrote:   : > >In article don@news.daedalus.co.nz (Don Stokes) writes:K > > Quite a lot of things were.  RT-11 volumes were (a) simple, (b) kind of F > > a lowest common denominator; ever pdp11/VAX platform support RT-119 > > volumes to some extent, e.g. through FLX or EXCHANGE. 9 >    VMS' EXCHANGE utility runs just fine on Alphas, too.    Jerome Fine replies:  3 The last time I used EXCHANCE on a large hard drive 3 that was being used under RT-11 (actually a magneto 0 optical media for the Sony SMO S-501 which holds, 295 MBytes per side), EXCHANGE allowed me to. look at ONLY the first partition on the drive.  7 In order to look at the other 8 RT-11 partitions, I was 8 forced to abandon the ability to look at RT-11 partition3 zero and place the VMS file structure there - which 8 wasted the first 32 MBytes of the drive.  In those first2 32 MBytes, I managed (over a period of a few days)5 to figure out how to set up VMS files of 65536 blocks 6 in the exact locations of the other 7 RT-11 partitions1 PLUS one additional file of about 47000 blocks at 3 the end of the media where the last RT-11 partition  was located.  1 Since only about 6 of the 9 RT-11 partitions were 4 actually being used, it was not a problem to give up2 RT-11 partition zero.  Under RT-11, partition zero/ was always left untouched and each of the RT-11 - partitions that were needed by the VMS system 2 were copied to the next RT-11 partition up - where& the successive VMS files were located.  4 I thought the solution rather elegant, but maybe NOW- EXCHANGE allows the user to specify the RT-11 0 partition number as opposed to using ONLY  RT-11" partition zero.  Does anyone know?  2 Also, now that RT-11 is able to handle dates up to' 2099, does EXCHANGE also know about the , new layout of the DATE word for RT-11 files?   Sincerely yours,   Jerome Fine  --4 To obtain the original e-mail address, please remove5 the ten characters which immediately follow the 'at'. 8 If you attempted to send a reply and the original e-mail7 address has been discontinued due a high volume of junk 5 e-mail, then the semi-permanent e-mail address can be 7 obtained by replacing the four characters preceding the . 'at' with the four digits of the current year.   ------------------------------  # Date: Wed, 23 Jul 2003 03:12:25 GMT * From: Michael Ross <abaddon@attglobal.net>P Subject: Re: Does RT-11 run on the PDP-11/70?  (was Re: PDP-11 OS Release Dates)7 Message-ID: <62789131a4960a272be481f908c0246d@TeraNews>   E On 22 Jul 2003 14:26:32 -0500, kaplow_r@encompasserve.org.TRABoD (Bob  Kaplow) wrote:  I >[Each PDP-11 OS had a unique blinken lights pattern on the 11/70. I long " >since forgot what they all were.]  E IIRC, one had a 'snake chasing its tail' pattern - several LEDs would D light in sequence, starting on the left, running across to the right4 and disappearing, then starting on the left again...   ?RSTS???   Mike http://www.corestore.org! The avalanche has already started ) It is too late for the pebbles to vote...    ------------------------------    Date: 22 Jul 2003 15:41:59 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) O Subject: Re: Does RT-11 run on the PDP-11/70? (was Re: PDP-11 OS Release Dates) 3 Message-ID: <LWKIMF4Jke71@eisner.encompasserve.org>   S In article <3f1e5a0c.1174277@news.supernews.com>, ian@hammo.com (paramucho) writes:  > B > Well, that's wierd because the RT-11 RL01/RL02 device driver hadH > extensive (and extremely horrible) modifications to support the 11/780: > console disk. Perhaps it just didn't *say* it was RT-11. >   F    Not suprizing tofin that end up in the RT-11 kernel, but the 11/7803    console did not have the full RT-11 command set.    ------------------------------   Date: 22 Jul 2003 20:45:14 GMT2 From: "Zane H. Healy" <healyzh@shell1.aracnet.com>O Subject: Re: Does RT-11 run on the PDP-11/70? (was Re: PDP-11 OS Release Dates) , Message-ID: <bfk7oq02blh@enews3.newsguy.com>  1 In comp.os.rsts John Santos <JOHN@egh.com> wrote: F > I think you're thinking of the 85x0/8700/88x0's, which had Pros withI > P/OS as consoles.  I can't remember if they were 350's, 380's, or both.   D > I think the 8600/8650 had a PDP-11 console with an RL02 instead of3 > a floppy.  No idea what O/S (if any) was on them.   K That's what I was thinking of.  The consoles started as 350's and then when G they ran out of them switched to 380's.  I've got the console (380) and I d|i|g|i|t|a|l badge off of one, that a coworkers Dad scrapped a few years  ago.   			Zane    ------------------------------    Date: 22 Jul 2003 11:23:09 -07001 From: keithparris_NOSPAM@yahoo.com (Keith Parris)  Subject: Gartner on HP Services = Message-ID: <cf15391e.0307221023.4863cf3a@posting.google.com>   C Gartner recently released a report on HP Services entitled "Company C Profile: HP Services' Strategy Moves Forward" in which they comment B favorably on the results to date and on the outlook going forward.  D I was pleased to see the number of times the word "OpenVMS" occurredC in this Gartner report.  The report notes HP has 5,300 VMS experts.   . http://www.gartner.com/reprints/hp/113746.html   ------------------------------  % Date: Tue, 22 Jul 2003 15:19:57 -0400 0 From: "Brian Tillman" <Tillman@sparkingwire.com> Subject: Re: HP FUDBusting$ Message-ID: <3f1d8e60$1@news.si.com>  @ >I guess people who're primarly programmers prefer Unix, whereas administrators prefer VMS   K Perhaps among the programmers you know, but that's a self-selected list and  very atypical. --  I Brian Tillman         Internet: Brian.Tillman at smiths-aerospace dot com 5 Smiths Aerospace  Addresses modified to prevent SPAM. D 3290 Patterson Ave. SE, MS 1B3 Replace "at" with "@", "dot" with "." Grand Rapids, MI 49512-1991 8        This opinion doesn't represent that of my company   ------------------------------    Date: 22 Jul 2003 20:56:51 -00004 From: Doc.Cypher <Use-Author-Address-Header@[127.1]> Subject: Re: HP FUDBusting5 Message-ID: <20030722205651.4071.qmail@gacracker.org>   F On Tue, 22 Jul 2003, "Brian Tillman" <Tillman@sparkingwire.com> wrote:A >>I guess people who're primarly programmers prefer Unix, whereas  >administrators prefer VMS > L >Perhaps among the programmers you know, but that's a self-selected list and >very atypical.    Quite.  N When you've horribly mangled something and have a versioning filesystem you'll be thankful.     Doc. --  K OpenVMS.         Eight out of ten hackers prefer *other* operating systems. K [New PGP Key - Get via finger]        http://deathrow.vistech.net/BOFH/doc/    ------------------------------  % Date: Tue, 22 Jul 2003 19:08:08 -0500 ( From: Michael Rice <marice@whiteice.com> Subject: Re: HP FUDBusting/ Message-ID: <vhrkfiqjgnrq17@corp.supernews.com>   . On 7/21/2003 8:58 AM, Stefaan A Eeckels wrote: [snip] > < > I did some work on VMS way back in 1986-1987, but I'm more9 > of a Unix man by heart (I guess people who're primarly  > > programmers prefer Unix, whereas administrators prefer VMS). >  [snip]  ? After 10+ years of writing software for VMS, Unix and Windows,  F everything from device drivers to GUIs, I'd definitely put VMS at the E top of the "joy to program on" list.  The overall consistency of the  E system services, RTL, LP libraries, etc. makes VMS infinitely better  E than Unix or Windows.  Additionally, you can study the internals and  I they actually make sense!  I love writing code for almost all platforms,  H but from a software engineering perspective, VMS is far superior to the  others.t   ------------------------------    Date: 22 Jul 2003 21:08:13 -0500+ From: young_r@encompasserve.org (Rob Young) Y Subject: Re: IDC reports PC Market Share: Dell over HP by a nose; both pulling away from  3 Message-ID: <1DNVB083waab@eisner.encompasserve.org>e  h In article <Gc6Ta.10472$8g6.62275@news1.news.adelphia.net>, "Mark Buda" <buda_NO@SPAM.yahoo.com> writes:M > "Andrew Harrison SUNUK Consultancy" <Andrew_No.Harrison_No@nospamn.sun.com>1= > wrote in message news:bfh172$m94$1@new-usenet.uk.sun.com... - >> UNIX number one                        NOT0- >> PC business returning to profitability NOTI- >> Ditto Enterprise Division              NOT - >> Merger complete                        NOTn- >> Its good to be second                  NOTD > , > It's good to own Sun stock             NOT  C 	NOT - is right.  And how telling a comment prior to Sun financialsw 	at close of day.l  9 	Again, Sun is quickly approaching Apple status regardingD 	relevancy - not very:  A http://biz.yahoo.com/rb/030722/markets_stocks_afterthebell_2.html)  M Network computer maker Sun Microsystems Inc. (NasdaqNM:SUNW - News), however,nN fell after it reported a slight quarterly profit but said revenue fell for the> ninth consecutive quarter on weak demand for computer servers.  K Its shares slipped to $4.31 in the extended session from $4.77 at Tuesday'sP close.   				Robe   ------------------------------    Date: 22 Jul 2003 15:55:10 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) P Subject: Re: Intel's had Dirty Bombs Anthrax and everything! (Was: Re: HP World:3 Message-ID: <3GUKgxDHKl0L@eisner.encompasserve.org>m  V In article <3F1D5A99.5CA60890@istop.com>, JF Mezei <jfmezei.spamnot@istop.com> writes: > P > Unfortunatly, shortly after that, Compaq fired its own ad agency and hired theA > one that had served Digital and it was all downhill from there.   G    As in the 4 page add I just found from HP, stating that for reliable-<    computing they recommend Windows 2003 on Intel platforms.  E    If that's what HP thinks is reliable computing, I should at least v"    buy it from a different vendor.   ------------------------------  % Date: Tue, 22 Jul 2003 12:48:48 -0400 ( From: David Froble <davef@tsoft-inc.com>Y Subject: Re: Intel's had Dirty Bombs Anthrax and everything! (Was: Re: HP World: Why Alphe, Message-ID: <3F1D6AF0.1000509@tsoft-inc.com>  O I've really got better things to do, but, some things irritate me to the point t of distraction.e   Bill Todd wrote:  F > "Fred Kleinsorge" <my-last-name@stardotzko.dec.com> wrote in message+ > news:PhVSa.509$hU.317@news.cpqcorp.net...- > H >>Now, now.  Bill is the Ralph Nader of the computer industry.  He is anJ >>altruistic and brilliant analyst who fights only for Truth, Justice, and >> > thel > % >>Ame... ah... the... well - his way.o    Q I sincerely hope you are not a Ralph Nader fan.  Ralph certainly has no interest hQ in the truth, just is getting his way.  A know-nothing that proclaims himself an  ' expert on everything.  A rabble rouser.     K > Since I really do believe that the values which I espouse are pretty muchnN > traditional American values (competence, honesty, integrity, accountability,L > competition), you don't have to worry too much about which formulation you > choose above.s     <snip>    ) >>Bill is not just PO'ed about the world.. >> > B > Nope - mostly about cHumPaq's behavior since June, 2001, and theI > reprehensible behavior of the Republican party since 1998 (with special N > emphasis on that of the current Administration since Sepember, 2001).  Don't/ > have much problem with the rest of the world.u    C Let's leave politics out of this, and stick to the topic of Compaq.>   >  > % >>Bill is not just PO'ed about Alpha.7= >>Bill is not just PO'ed about every CEO since Ken was fired.e >> > I > I liked what Pfeiffer was doing.  If you can't understand the degree of N > incompetence displayed by Palmer and Capellas, then you're as incompetent as > they were. >  >  >>Bill is not just bored.  >> > J > I'm seldom bored, Fred:  I wish keeping an eye on cHumPaq's toadies hereM > didn't take as much time as it does, but it's a commitment (you may need tofK > look that up if you understand it as poorly as your employer does) that In > took on voluntarily. >  > $ >>Bill does not just want attention. >> > H > Even if you don't understand much else, Fred, you really ought to haveH > caught on by now to the fact that I don't give a flying fuck what most& > people may choose to think about me. >  > = >>Bill is not a fictional character from Pinky and the Brain.d >> > 5 > Cartoons do seem to be about your speed these days.- >  > J >>He is not any of those things.  He only has the best interests of us all >> > at > K >>heart.  And when he runs the world, things will be better.  A lot better.s >>Really, really better. >> > J > Confused again, Fred?  I have no desire to run the world - just a strongL > opposition to seeing it run by the kind of sleazes you work for.  And I'veM > had sufficient success in my efforts in this particular area to consider iti > worthwhile to continue.r    O It's come down to personalities and calling names.  Now, I refuse to spend the nO time going back and finding some posts, but my memory does seem to recall that:o  M Fred was sorry to see Alpha go (not likely if he thought it a dud) but there o3 wasn't much he could do about it, and life goes on.e  K DEC/Compaq made public a white paper on how EPIC just wasn't too good of a eM concept, and how Alpha (and others) had much more potential.  (A white paper  H that might not be so easy to find on the HP/Compaq web site these days.)  L VMS **PROFITS** were more than enough, by themselves, to fund ongoing Alpha 6 development, should the owners of Alpha want to do so.  Q Wildfire systems, while good, didn't seem to live up to the capabilities of EV6. wB   Seems system implementation people may have miscalculated a bit.  O Bill was upset about the end of Alpha, but it wasn't his product and his place tL to be more than upset.  However, the spin that Compaq attempted to place on K their decision was full of more holes than Swiss cheese.  Bill is entirely  K within his right to object to lying to customers and the public, and doing sL nothing more than exposing how ridiculous the story was, and the lies being K told.  Hey, if you don't want to see public reaction, don't make something cQ public.  Compaq made public statements, and public comment on such statements is  J allowable, regardless of how much Compaq doesn't appriciate such comment. , Bill's story hasn't changes much in 2 years.   Some of the holes:  K Some HP employees now publically embrace the questionable story put out by tO Compaq.  There may be more self interest for them doing so than the actions of >L others.  While I can understand such self interest, I can also recognize it.  N Compaq used statements from those who may have had differences with the Alpha L CPU developers to support it's story, but not one word from those Alpha CPU P developers, the very ones who you'd expect to have the most relavent statements L to make.  Now there were some rumors of off-the-record statements from some Q Alpha CPU developers, it's understandable if they fear to dispute the 'official'  P statements of their employer, who can exert enormous financial pressure on such N employees.  I'd think that if the Alpha CPU developers agreed with the story, * there would have not been any such rumors.  O Now these people who declared IA-64 surperior to Alpha, and on who's judgement hP Compaq made a major business decision, they'll have a significant place in HP's Q future plans I'd assume?  Can't lose such prestigeous, knowledgable, and capable n folks, can they?  P The latest Alpha, EV7, seems unable to keep up with IBM's Power series of CPUs, N in contrast to the past where it would match and many times surpassed the IBM Q product.  Since there is similar concepts in both products, it's hard to imagine oO that such lag is because of the concepts in the Alpha design.  The most likely tN reason then is lack of effort to make EV7 the best it can be.  Since Alpha is N still being sold to customers, and since buyers of EV7 are some times looking M for maximum performance, it brings into question HP's honesty and commitment   with such customers.  J Recently there have been statements about whether it would have been cost P effective to continue with Alpha.  To me this is a statement that Compaq wasn't Q interested in competing with it's competitors, and instead was more than willing  N to drop out of the competition.  I personally find this to be contrary to the N concept of running a business.  In the high tech field, you run a business to N provide the 'better mousetrap' in order to out-sale the competition.  If this P was the case, why wouldn't Compaq just announce that they weren't interested in O developing CPUs?  Answer, because then they wouldn't appear to be competitive, h5 and sales would go to those still willing to compete.   O This then explains the 'big lie'.  Compaq didn't want to compete, but couldn't  O appear to be non-competitive.  So, you take the products of others, and try to aE convince the public (potential customers) that what you are doing is iN competitive.  The problem is, regardless of the merits of IA-64, Compaq, (now K HP), isn't competing.  They're just taking whatever Intel will supply, and  O packaging it.  Therefore, if Intel isn't winning the competition, then HP will nP be seen as less desirable than those who are surpassing the Intel product.  The O only thing HP can do is to attempt to project the image that IA-64 is superior mO to all the competition.  Unfortunately for them, their story has holes.  Their  O response, the above nonsense of trying to deflect those who have been sticking lO their arms through the holes and saying "see, nothing there" with name calling -	 and such.-  I Let's see, I've heard that you can fool some of the people all the time. <P There's my friend Rob, who extolled the virtues of Alpha, and specifically with O respect to IA-64, right up to the beginning of the 'big lie', and now not only  P extols the virtues of IA-64 against Alpha, but also any other competitors.  Did 0 anything change, other than Rob's self interest?  I I'm having a bit of difficulty with Rob's perspective.  Compaq went to a  N competitor, and was convinced by this competitor that Compaq couldn't compete P with them, and would Compaq just give up and accept whatever the competitor was P willing to sell them.  Oh, by the way, the competitor would be happy to take on F those Compaq employees who just couldn't match the performance of the L competitor's developers.  I have yet to hear why the competitor would waste Q money on these 'inferior' developers who could never match what the competitor's oN developers were claiming they would produce in the future.  There is also the O issue of relative past performance of the competitor's developers and Compaq's  P 'inferior' developers, those same developers who had been producing the world's M fastest CPUs, while the competitor's developers had been rather unsuccessful.2  P So Ok, I've offered Rob the chance, under 'heavy NDA', to see my claims of soon N producing starships.  I've claimed that they'll scale well, and if he buys 2, M he'll reach his destination twice as fast.  But I have yet to see any orders 0$ from him.  Where are the orders Rob?    K Instead of trying to make fun of Bill, could his detractors just show some  N evidence of him being wrong?  Just address the issues?  Maybe name a customer P that was promised that Alpha NT was the way to go, that was promised that Alpha 7 had a future, who will trust the commitments of Compaq?    Dave   -- e4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com6 T-Soft, Inc.  170 Grimplin Road  Vanderbilt, PA  15486   ------------------------------  % Date: Tue, 22 Jul 2003 14:10:42 -0500e/ From: Chris Scheers <chris@applied-synergy.com>r, Subject: Re: Mozilla... the new NULL process3 Message-ID: <3F1D8C32.AB83000D@applied-synergy.com>h   Colin Blake wrote: > F > To be chewing up 1 second every 15-20, you must be using a real slowE > system. If this is unacceptable you can adjust how enthusiastic the G > polling is. There's a previous thread which describes how to do this.hI > Search Google Groups for vms_poll_timer_min and a date of October 2001.e   Colin,  ? I have some AST based code that I use to "poll" modems that has G extremely low overhead and latency.  I suspect that it could be adaptedv
 for use here.s  5 Are the sources somewhere where I could look at them?   - PS: eMail to your address is getting bounced.2  G -----------------------------------------------------------------------0$ Chris Scheers, Applied Synergy, Inc.  C Voice: 817-237-3360            Internet: chris@applied-synergy.com b   Fax: 817-237-3074m   ------------------------------  % Date: Tue, 22 Jul 2003 10:53:15 -0700t* From: Alan Frisbie <Abuse@Flying-Disk.com>' Subject: Re: New OpenVMS-to-Itanium FAQ/. Message-ID: <3F1D7A0B.6080407@Flying-Disk.com>   Tom Linden wrote:c  8 >>BTW, is it PL/1 or PL/I ? I thought it was the latter.  
 > It is PL/I.1  < I could never figure out if it was pronounced "Pee Ell Eye",5 or if the company was pronounced "One Bee Emm".   :-)e   Alan   ------------------------------  % Date: Tue, 22 Jul 2003 14:32:24 -0400 # From: sol gongola <sol@adldata.com> ' Subject: Re: New OpenVMS-to-Itanium FAQo+ Message-ID: <3F1D8338.33E8C9C2@adldata.com>    It's a roman numeral   Alan Frisbie wrote:' >  > Tom Linden wrote:a > : > >>BTW, is it PL/1 or PL/I ? I thought it was the latter. >  > > It is PL/I.n > > > I could never figure out if it was pronounced "Pee Ell Eye",7 > or if the company was pronounced "One Bee Emm".   :-)i >  > Alan   ------------------------------    Date: 22 Jul 2003 15:46:10 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) ' Subject: RE: New OpenVMS-to-Itanium FAQ 3 Message-ID: <Tto3+r0Jp6Z1@eisner.encompasserve.org>c  _ In article <CIEJLCMNHNNDLLOOGNJIMENCHJAA.tom@kednos.com>, "Tom Linden" <tom@kednos.com> writes:  >>From: Simon Clubleyt7 >>[mailto:clubley@remove_me.eisner.decus.org-Earth.UFP]e >>8 >>BTW, is it PL/1 or PL/I ? I thought it was the latter. > 
 > It is PL/I.n  A    IIRC it is spelled PL/I and pronounced pea ell one.  Didn't we--    go throught that just a couple months ago?:  B    We have the same problem with the NSSC-I.  One contractor calls3    it the nissick one, another calls it the niscky.g   ------------------------------  % Date: Tue, 22 Jul 2003 18:53:55 -0400r* From: Bill Valentine-Cooper <wvc@nwcs.com>? Subject: Re: Non-OpenVMS ethernet access to Printserver 17/600?r8 Message-ID: <vfdrhvcc6uld8n97r2h0n22d7aak7eih7i@4ax.com>  F On Tue, 22 Jul 2003 15:06:12 GMT, Paul Anderson <paul.anderson@hp.com> wrote:  > >In article <1vaqhvcmkpm5uvhajjde53radbdvshfu0s@4ax.com>, Bill' >Valentine-Cooper <wvc@nwcs.com> wrote:u >lE >> Does anyone know how to add the Digital Network Port monitor to an 5 >> existing Windows 2000 machine thru registry magic?N >1H >If the PrintServer is already booted from somewhere, can't you just useF >LPD from Windows to print?  (Assuming your flavor of Windows has been  >blessed with LPD capabilities.)  F To the best of my knowledge, no. I can't find a port that's up for LPDD -- nobody's home on 515. There is something listening on 170, but it, doesn't respond in any way I can understand.   >n >Paul    ------------------------------  % Date: Tue, 22 Jul 2003 15:46:21 -0400d0 From: "Brian Tillman" <Tillman@sparkingwire.com>8 Subject: Re: Obsolete cross compilers: where did you go?$ Message-ID: <3f1d9490$1@news.si.com>  : >I know of a VMS customer in this area still using SD-Ada.G >I believe it is supported by EDS, who have no intention of porting it.h > C >So that customer is stuck on VAX, and perhaps cannot even considerr> >VAX emulators due to certification issues for their industry.  D Exactly.  Even if one finds acceptable compilers to use on the newerL architecture, it can cost hundreds of thousands, if not millions, of dollarsL to recertify the software produced by the new compiler even if the source is
 identical. -- eI Brian Tillman         Internet: Brian.Tillman at smiths-aerospace dot comr5 Smiths Aerospace  Addresses modified to prevent SPAM.kD 3290 Patterson Ave. SE, MS 1B3 Replace "at" with "@", "dot" with "." Grand Rapids, MI 49512-1991e8        This opinion doesn't represent that of my company   ------------------------------  # Date: Tue, 22 Jul 2003 17:43:48 GMT-3 From: hammond@not@peek.ssr.hp.com (Charlie Hammond)m Subject: Re: OpenVms BackupD0 Message-ID: <oJeTa.626$l12.378@news.cpqcorp.net>  * In article <jymTI9hn5Coz@elias.decus.ch>, + p_sture@elias.decus.ch (Paul Sture) writes:d  f >In article <KyWSa.521$5X.420@news.cpqcorp.net>, hammond@not@peek.ssr.hp.com (Charlie Hammond) writes:  M >> In a production environment, a backup strategy that uses /INGORE=INTERLOCKvM >> is bogus.  Period.  The cost of lost information will usually exceed by a h/ >> LARGE amount the cost of doing backup right.i > C >As others have said later in this thread, 24 x 7 availability is aO" >direct conflict with this stance.  E Yes it is.  Such an environment needs to use a transaction managementlD or database management system that supports either "on line backup".E Another method is to have an off line copy of the data that is rolled( forward from a journal..  = >No insult taken here. I see this as a chance to give OpenVMSdA >Engineering some feedback on one of the issues we face out here.   D OpenVMS Engineering is well aware that the OpenVMS Backup utility is1 not a solution to backup in a 24 x 7 environment!   K Solutions along the lines outlined above are available from dbms vendors.  -G They cost money, but they work.  If OpenVMS Engineering is to solve theeE same problem again, we would expect it to cost money too, not be parts of the O/S.r   -- SJ       Charlie Hammond -- Hewlett-Packard Company -- Ft Lauderdale  FL  USAF           (hammond@not@peek.ssr.hp.com -- remove "@not" when replying)J       All opinions expressed are my own and not necessarily my employer's.   ------------------------------  % Date: Tue, 22 Jul 2003 15:34:03 -0400(0 From: "Brian Tillman" <Tillman@sparkingwire.com> Subject: Re: OpenVms Backupa$ Message-ID: <3f1d91ae$1@news.si.com>  H >>...how many have used /ign=inter and ended up with an UNBOOTABLE disk? > % >An unbootable disk is probably rare.CK >More common is an inconsistent or completely invalid (unusable) data file.   I But the files susceptible to being inconsistent are quite rare (unless andI Oracle or other database system is involved) and those that do happen areCE easily remedied simply by recreating the file (like the queue managern files).a -- hI Brian Tillman         Internet: Brian.Tillman at smiths-aerospace dot comt5 Smiths Aerospace  Addresses modified to prevent SPAM.ED 3290 Patterson Ave. SE, MS 1B3 Replace "at" with "@", "dot" with "." Grand Rapids, MI 49512-1991.8        This opinion doesn't represent that of my company   ------------------------------  # Date: Tue, 22 Jul 2003 20:05:19 GMT.3 From: hammond@not@peek.ssr.hp.com (Charlie Hammond)r Subject: Re: OpenVms Backupl/ Message-ID: <3OgTa.643$E42.76@news.cpqcorp.net>-  % In article <3f1d91ae$1@news.si.com>, M2 "Brian Tillman" <Tillman@sparkingwire.com> writes:   [wrt /IGNORE=INTERLOCK]t  J >But the files susceptible to being inconsistent are quite rare (unless an1 >Oracle or other database system is involved) ...e  ' Data in RMS files is equally vunerable.A  D >                                   ... and those that do happen areF >easily remedied simply by recreating the file (like the queue manager >files).  3 Not necessarily true for data files -- dbms or RMS.    --  J       Charlie Hammond -- Hewlett-Packard Company -- Ft Lauderdale  FL  USAF           (hammond@not@peek.ssr.hp.com -- remove "@not" when replying)J       All opinions expressed are my own and not necessarily my employer's.   ------------------------------   Date: 22 Jul 2003 17:55 CDT ' From: carl@gerg.tamu.edu (Carl Perkins)  Subject: Re: OpenVms Backups- Message-ID: <22JUL200317553736@gerg.tamu.edu>e  V In article <oJeTa.626$l12.378@news.cpqcorp.net>, hammond@not@peek.ssr.hp.com writes...+ }In article <jymTI9hn5Coz@elias.decus.ch>,  , }p_sture@elias.decus.ch (Paul Sture) writes: } g }>In article <KyWSa.521$5X.420@news.cpqcorp.net>, hammond@not@peek.ssr.hp.com (Charlie Hammond) writes:e } N }>> In a production environment, a backup strategy that uses /INGORE=INTERLOCKN }>> is bogus.  Period.  The cost of lost information will usually exceed by a 0 }>> LARGE amount the cost of doing backup right. }>D }>As others have said later in this thread, 24 x 7 availability is a# }>direct conflict with this stance.t } F }Yes it is.  Such an environment needs to use a transaction managementE }or database management system that supports either "on line backup".iF }Another method is to have an off line copy of the data that is rolled }forward from a journal. } > }>No insult taken here. I see this as a chance to give OpenVMSB }>Engineering some feedback on one of the issues we face out here. } E }OpenVMS Engineering is well aware that the OpenVMS Backup utility ise2 }not a solution to backup in a 24 x 7 environment! } L }Solutions along the lines outlined above are available from dbms vendors.  H }They cost money, but they work.  If OpenVMS Engineering is to solve theF }same problem again, we would expect it to cost money too, not be part }of the O/S. }  }-- K }      Charlie Hammond -- Hewlett-Packard Company -- Ft Lauderdale  FL  USA   H That the DBMS vendors have solved the problem for their DBMS/DB is fine.  4 But it does nothing for backing up your system disk.   --- Carl   ------------------------------    Date: 22 Jul 2003 18:15:56 -07001 From: keithparris_NOSPAM@yahoo.com (Keith Parris)  Subject: Re: OpenVms BackupD= Message-ID: <cf15391e.0307221444.36e1d5f3@posting.google.com>   [ JF Mezei <jfmezei.spamnot@istop.com> wrote in message news:<3F1D5CB3.B72A9987@istop.com>...cP > Question: if you have a shadow set, and you remove one member at some "random"O > time. Are you assured that this now-independant member has a 100% safe image,sH > or is it possible that some indexed file would have inconsistent indexG > information because the drive was dismounted in the middle of a writem$ > requiring an index split etc etc ?  > The situation with breaking a member out of a shadow set seemsC significantly different to me than using BACKUP /IGNORE=INTERLOCK. 3F Backup sweeps across the disk, picking things up from different placesD at different points over a period of time, increasing the chance forB inconsistencies; in dismounting a member of a shadowset, you get aB snapshot of the contents of the disk volume as of a point in time.  D Both the VMS Files-11 code and RMS use "careful write ordering" withE the intent of ensuring that the file system metadata and internal RMS.= file metadata, respectively, are always left in a recoverable2D (readable) state during updates.  RMS also does this during a bucketE split.  So the result of dismounting a single shadowset member should) always be readable, in theory.  F Note that unless you're doing some sort of journaling, and either backC out or roll forward any in-progress transactions, there may be some A inconsistencies among your data items for transactions which wereNF being processed at the time the snapshot was taken.  (So quiescing the= application is probably a very good idea before dismounting a>D shadowset member for backups.)  But the metadata should be left in a6 state such that you can get to your files and records.   ------------------------------    Date: 22 Jul 2003 12:21:27 -07001 From: keithparris_NOSPAM@yahoo.com (Keith Parris) % Subject: Re: OpenVMS Cluster overheadP< Message-ID: <cf15391e.0307221121.ae14c01@posting.google.com>  { andrew.rycroft@intrinsitech.com (Andrew Rycroft) wrote in message news:<58ba0101.0307220546.4c818de6@posting.google.com>...lH > We have a Oracle/Rdb application which we are thinking of putting intoG > a two node cluster with gigabit interconnect to enhance availability.bC > But the question we need to address is whether by clustering thiseC > application the overall performance will be negatively impacted ?d  E If you simply added a 2nd node and formed a cluster, but continued toaA run your application on the 1st node, you'd likely not notice any # difference from being in a cluster.s  F If you put half of the workload onto the new 2nd node, two things will happen:rE 1) One of the nodes will be the lock-master for a given resource/lockeF tree, and users of that same resource/lock tree on the other node willF have their lock requests satisfied in perhaps 200 microseconds insteadF of perhaps 5 microseconds for a lock request on the same node which is? mastering the lock tree.  Your users may or may not notice this  difference.t= 2) The number of users on the first node will go down by 1/2, C potentially improving their response times.  But whether or not youHD notice this will depend on how heavily-loaded the system was before.  D A lot depends on where your performance bottleneck currently is.  IfD you presently have a shortage of CPU or memory, adding a second nodeF will likely result in significantly-better performance for your users.D  If your current bottleneck is in the storage subsystem, performanceB is likely to be exactly the same, and would get worse if you added additional workload.   ------------------------------  # Date: Tue, 22 Jul 2003 18:13:37 GMTB, From: "James Gessling" <jgessling@yahoo.com>' Subject: Packed decimal arithmetic in C : Message-ID: <l9fTa.199$BX4.147@newssvr24.news.prodigy.com>  G I'm trying to help out a programmer who would like to do this.  We havee? BASIC decimal data stored in a C array that he would like to doTA arithmetic on.  (It's in a C array to get around BASIC's 65k bytea limit for arrays).  > Anyone know of library functions (or anything) callable from C that can handle this?t   Thanks,  Jim   ------------------------------  % Date: Tue, 22 Jul 2003 14:54:51 -0700y# From: "Tom Linden" <tom@kednos.com>n+ Subject: RE: Packed decimal arithmetic in Ce9 Message-ID: <CIEJLCMNHNNDLLOOGNJIEEOCHJAA.tom@kednos.com>l  J PL/I doesn't have the 65K limit and does packed decimal with precision, p,7 and scale factor q such that 0<= p <= 31 and 0<= q <= p   A You could have the BASIC program  call a PL/I procedure to do the  calculations   >-----Original Message----- 2 >From: James Gessling [mailto:jgessling@yahoo.com]& >Sent: Tuesday, July 22, 2003 11:14 AM >To: Info-VAX@Mvb.Saic.Com( >Subject: Packed decimal arithmetic in C >i >dH >I'm trying to help out a programmer who would like to do this.  We have@ >BASIC decimal data stored in a C array that he would like to doB >arithmetic on.  (It's in a C array to get around BASIC's 65k byte >limit for arrays).c >a? >Anyone know of library functions (or anything) callable from Ca >that can handle this? >-
 >Thanks,  Jime >e >2 >---' >Incoming mail is certified Virus Free.b; >Checked by AVG anti-virus system (http://www.grisoft.com).1A >Version: 6.0.502 / Virus Database: 300 - Release Date: 7/18/20030 >2 ---f& Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).@ Version: 6.0.502 / Virus Database: 300 - Release Date: 7/18/2003   ------------------------------  # Date: Tue, 22 Jul 2003 22:27:12 GMTa# From: hoff@hp.nospam (Hoff Hoffman)-+ Subject: RE: Packed decimal arithmetic in Ct/ Message-ID: <4TiTa.662$%r2.13@news.cpqcorp.net>   1 From: James Gessling [mailto:jgessling@yahoo.com]R? :Anyone know of library functions (or anything) callable from Cd3 :that can handle this [packed decimal format math]?   F   Packed decimal math is slow in general, and conversions are requiredI   for most math operations; OpenVMS performs math using native datatypes.eJ   Convert to a native integer format, and do the math.  Then convert back.D   Use lib$cvt_dx_dx, or such, with the packed-decimal string type.    D   You could also call a COBOL or PL/I routine, languages which have (   integrated support for packed decimal.  A   If it were my code, I'd move it out of packed decimal entirely.s    N  ---------------------------- #include <rtfaq.h> -----------------------------K     For additional, please see the OpenVMS FAQ -- www.hp.com/go/openvms/faq N  --------------------------- pure personal opinion ---------------------------E         Hoff (Stephen) Hoffman   OpenVMS Engineering   hoff[at]hp.comh   ------------------------------    Date: 22 Jul 2003 15:34:31 -07001 From: nothome@spammers.are.scum (Malcolm Dunnett)o+ Subject: RE: Packed decimal arithmetic in C - Message-ID: <S3r46X$zyYbw@malvm7.mala.bc.ca.>o  0 In article <4TiTa.662$%r2.13@news.cpqcorp.net>, )     hoff@hp.nospam (Hoff Hoffman) writes:e   > H >   Packed decimal math is slow in general, and conversions are requiredK >   for most math operations; OpenVMS performs math using native datatypes.   7     Packed decimal *IS* a native data type on a VAX :-)t  =     (well, he never actually said what platform he was using)s   ------------------------------  % Date: Tue, 22 Jul 2003 16:49:25 -0700c# From: "Tom Linden" <tom@kednos.com>g+ Subject: RE: Packed decimal arithmetic in C-9 Message-ID: <CIEJLCMNHNNDLLOOGNJIOEOFHJAA.tom@kednos.com>O   >-----Original Message----- + >From: Hoff Hoffman [mailto:hoff@hp.nospam] % >Sent: Tuesday, July 22, 2003 3:27 PMo >To: Info-VAX@Mvb.Saic.Com, >Subject: RE: Packed decimal arithmetic in C >p > 2 >From: James Gessling [mailto:jgessling@yahoo.com]@ >:Anyone know of library functions (or anything) callable from C4 >:that can handle this [packed decimal format math]? >fG >  Packed decimal math is slow in general, and conversions are requiredoJ >  for most math operations; OpenVMS performs math using native datatypes.K >  Convert to a native integer format, and do the math.  Then convert back.0C >  Use lib$cvt_dx_dx, or such, with the packed-decimal string type.  >BD >  You could also call a COBOL or PL/I routine, languages which have) >  integrated support for packed decimal.  >iB >  If it were my code, I'd move it out of packed decimal entirely.  L Sometimes you can't do that.  Try telling a bank or insurance company to useH float binary instead of scaled fixed decimal or float decimal. (and BTW, that8 is why financial institutions stick with PL/I and Cobol)  J Even though they aren't in HW on Alpha they are implemented for PL/I using very) efficient algorithms, nines and excess 3.y   >- >-1 > ---------------------------- #include <rtfaq.h>: >-----------------------------L >    For additional, please see the OpenVMS FAQ -- www.hp.com/go/openvms/faq3 > --------------------------- pure personal opinionl >---------------------------F >        Hoff (Stephen) Hoffman   OpenVMS Engineering   hoff[at]hp.com >e >---' >Incoming mail is certified Virus Free. ; >Checked by AVG anti-virus system (http://www.grisoft.com).fA >Version: 6.0.502 / Virus Database: 300 - Release Date: 7/18/2003b >m ---d& Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).@ Version: 6.0.502 / Virus Database: 300 - Release Date: 7/18/2003   ------------------------------  % Date: Tue, 22 Jul 2003 17:08:32 -0700-8 From: Randy Park <rjpark.nospaam@mindspring.nospaam.com>+ Subject: Re: Packed decimal arithmetic in Ce8 Message-ID: <s0krhvck2kavrifme869s962rc3685c1in@4ax.com>  2 On Tue, 22 Jul 2003 18:13:37 GMT, "James Gessling" <jgessling@yahoo.com> wrote:  H >I'm trying to help out a programmer who would like to do this.  We have@ >BASIC decimal data stored in a C array that he would like to doB >arithmetic on.  (It's in a C array to get around BASIC's 65k byte >limit for arrays).i >s? >Anyone know of library functions (or anything) callable from Ce >that can handle this? >p
 >Thanks,  Jim  >   < Arrays in BASIC can be more than 65K bytes.  Arrays in BASIC> can contain more than 65K elements.  A single character string cannot be more than 65K bytes.  7 None of this would prevent BASIC from processing packedt: decimal data.  BASIC supports the packed decimal data type
 just fine.  < It sounds like there is some confusion about how the data is9 actually stored (on disk?) and how BASIC can process that29 data.  I've got 25 years experience in DEC BASIC and I'veR< yet to encounter data I couldn't handle in BASIC.  Sometimes@ I've had to call a library routine to handle the tough stuff and* sometimes I've bit twiddled things myself.  < Maybe if you are more specific on the nature of the data and? how it is stored, someone can  tell you how it can be processedyC in BASIC.  Is the data in a disk file, or is it being passed to thef5 BASIC routine from a routine in a different language.e   ------------------------------    Date: 22 Jul 2003 13:49:44 -05004 From: kaplow_r@encompasserve.org.TRABoD (Bob Kaplow)$ Subject: Re: PDP-11 OS Release Dates3 Message-ID: <8x8rIGP$sa5a@eisner.encompasserve.org>   ` In article <OI1Ta.6421$9f7.757998@news02.tsnz.net>, don@news.daedalus.co.nz (Don Stokes) writes:? > In article <bdc65a53.0307180649.4a7d5d60@posting.google.com>,q' > Galen <gspamtackett@yahoo.com> wrote:DB >>Somebody used to distribute a Basic-Plus uncompiler, which was aH >>pretty cool and magical tool to impress other computer lab staff with. > D > We had one of these under RSTS/E.  The company had installed a fewH > tur(n)key systems with smallish disks (MicroPDP-11s, mainly) where the  K I thought I was the originator of that. For many years, back around the mideG 80s, my .sig or mail header or whatever was "There is no N in Turnkey!"-  G > source to the accounting package on them wasn't provided.  DECOMP (Or.7 > DCOMP, I forget now) worked wonders on these cases.  d  1 	Bob Kaplow	NAR # 18L	TRA # "Impeach the TRA BoD"l& 		>>> To reply, remove the TRABoD! <<<K Kaplow Klips & Baffle:	http://nira-rocketry.org/LeadingEdge/Phantom4000.pdfeL     www.encompasserve.org/~kaplow_r/    www.nira-rocketry.org    www.nar.org  O  Save Model Rocketry from the HSA!   http://www.space-rockets.com/congress.htmld   ------------------------------  % Date: Tue, 22 Jul 2003 22:31:18 +0200n2 From: "Robert Boers" <robert.boers@softresint.com>S Subject: Re: PL/I to C and C++ translation services, process overview now availableI7 Message-ID: <3f1d9f14$0$22104$5402220f@news.sunrise.ch>   F Just to note that we deliver the translated PL/I code according to theF former Digital Engineering C/C++ coding standards, or whatever inhouseK standards the customer requests. The translator does a (very) big part, but1 the remainder is done manually.'  D In other words, we deliver a complete translation service, includingD applying available or developed application/compatibility tests. The. translator itself is not available externally.   Regards, Robert Boers.  : "Keith A. Lewis" <lewis@spyder.mitre.org> wrote in message& news:bfjmkk$5$1@newslocal.mitre.org...1 > "Tom Linden" <tom@kednos.com> writes in articlerD <CIEJLCMNHNNDLLOOGNJIAENCHJAA.tom@kednos.com> dated Tue, 22 Jul 2003 07:49:11 -0700:cJ > >The resultant code will not be of a nature suitable to be maintained byE > >programmers, so how do you modify the code?  Do you go back to the  originalJ > >and retranslate it?  If you want to move from PL/I to C or C++ then the best* > >approach is to reengineer from scratch. > H > That's a problem with any translator, especially one which goes from aI > higher-level language to a lower-level one.  We played with an in-housenI > PL/I->C++ translator, and from what I heard, it produced unmaintainablew > code.> > G > On the other hand, we did develop a PL/I->C data structure translatore whichiI > has been very useful in our mixed-code environment.  Even set up MMS tooL > convert certain .DCL files into .H on a regular basis.  The result is thatJ > our application can share large data structures between PL/I and C code.J > Of course, the advanced stuff like REFER() doesn't translate, so we keep) > that out of modules we intend to share.e >yJ > If anybody wants a copy of the data structure translator, drop me a line atG > my slightly munged e-mail address below.  The one in the header won'tp work.t > - > --Keith Lewis              klewis$mitre.org @ > The above may not (yet) represent the opinions of my employer.   ------------------------------  % Date: Tue, 22 Jul 2003 15:17:46 -0700 # From: "Tom Linden" <tom@kednos.com>aS Subject: RE: PL/I to C and C++ translation services, process overview now available 9 Message-ID: <CIEJLCMNHNNDLLOOGNJICEODHJAA.tom@kednos.com>n  B Of course the downside for the customer with this approach is thatF he ends up with C code rather than PL/I and it will take several yearsD for the code to stabilize and it will have reduced functionality and never be as robust.   A I'm not sure how many companies would turn over their proprietary F code to a third party to translate.  There were several companies thatH tried this with PL/I and Cobol in the early 80s and they all went out of( business, we're still selling compilers.     >-----Original Message-----e8 >From: Robert Boers [mailto:robert.boers@softresint.com]% >Sent: Tuesday, July 22, 2003 1:31 PM  >To: Info-VAX@Mvb.Saic.ComF >Subject: Re: PL/I to C and C++ translation services, process overview >now available >f >sG >Just to note that we deliver the translated PL/I code according to the G >former Digital Engineering C/C++ coding standards, or whatever inhousedL >standards the customer requests. The translator does a (very) big part, but  >the remainder is done manually. >lE >In other words, we deliver a complete translation service, includingcE >applying available or developed application/compatibility tests. Thep/ >translator itself is not available externally.c >, >Regards, Robert Boers.s >a; >"Keith A. Lewis" <lewis@spyder.mitre.org> wrote in messageo' >news:bfjmkk$5$1@newslocal.mitre.org...h2 >> "Tom Linden" <tom@kednos.com> writes in articleE ><CIEJLCMNHNNDLLOOGNJIAENCHJAA.tom@kednos.com> dated Tue, 22 Jul 2003o >07:49:11 -0700:K >> >The resultant code will not be of a nature suitable to be maintained byiF >> >programmers, so how do you modify the code?  Do you go back to the	 >original K >> >and retranslate it?  If you want to move from PL/I to C or C++ then thee >bestb+ >> >approach is to reengineer from scratch.m >>I >> That's a problem with any translator, especially one which goes from aeJ >> higher-level language to a lower-level one.  We played with an in-houseJ >> PL/I->C++ translator, and from what I heard, it produced unmaintainable >> code. >>H >> On the other hand, we did develop a PL/I->C data structure translator >whichJ >> has been very useful in our mixed-code environment.  Even set up MMS to> >> convert certain .DCL files into .H on a regular basis.  The >result is thatuK >> our application can share large data structures between PL/I and C code.eK >> Of course, the advanced stuff like REFER() doesn't translate, so we keep * >> that out of modules we intend to share. >>K >> If anybody wants a copy of the data structure translator, drop me a lines >atnH >> my slightly munged e-mail address below.  The one in the header won't >work. >>. >> --Keith Lewis              klewis$mitre.orgA >> The above may not (yet) represent the opinions of my employer.e >  >  >---' >Incoming mail is certified Virus Free.a; >Checked by AVG anti-virus system (http://www.grisoft.com).-A >Version: 6.0.502 / Virus Database: 300 - Release Date: 7/18/2003t >a ---f& Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).@ Version: 6.0.502 / Virus Database: 300 - Release Date: 7/18/2003   ------------------------------  % Date: Tue, 22 Jul 2003 22:08:27 +0200r< From: Michael Lemke <ai26.nospam@sternwarte.uni-erlangen.de># Subject: Poor disk IO.  How to fix? : Message-ID: <3F1D99BB.FD6D081D@sternwarte.uni-erlangen.de>  F On our 3 node Alpha cluster I notice terrible disk IO perfromance whenE compared to say, a Linux PC.  A simple tar xvf ...  takes 40 sec realaG time but only 1 sec CPU time.  The same operation on an 800MHz Linux PC * finishes in 1.7 sec with 0.1 sec CPU time.  G What I am really concerned with is the elapsed time.  I notice this badsH performance with all sorts of applications, it is not restricted to tar.  9 Exiting MAIL and purging the spam also takes a long time.m   Exact values for the tar test:  > CPUTime: 1020 ms  Elapsed: 00:00:39.17  DIO: 2741  BufIO: 2342    H The Alpha is a "COMPAQ Professional Workstation XP1000" with 1GB memory.  E The disk I tried this on was a locally attached, cluster wide mountede
 SCSI disk:   A400> sh dev disk15/fu  @ Disk A400$DKA400:, device type DEC RZ2CC-KA, is online, mounted,
 file-orientedoG     device, shareable, served to cluster via MSCP Server, error loggingw is     enabled.  <     Error count                    0    Operations completed 5624571     Owner process                 ""    Owner UICe [SYSTEM]0     Owner process ID        00000000    Dev Prot S:RWPL,O:RWPL,G:R,W:RWPL6     Reference count               53    Default buffer size                 5123     Total blocks             8380080    Sectors per  track                   168l2     Total cylinders             2495    Tracks per cylinder                  20  7     Volume label            "DISK15"    Relative volumew number                0n3     Cluster size                   9    Transaction  count                    36o=     Free blocks              2105460    Maximum files allowed  423620-     Extend quantity                5    Mount ! count                           3 2     Mount status              System    Cache name "_A400$DKA0:XQPCACHE" F     Extent cache size             64    Maximum blocks in extent cache 210546B     File ID cache size            64    Blocks currently in extent cache   7137>     Quota cache size               0    Maximum buffers in FCP cache       12020     Volume owner UIC           [1,1]    Vol Prot S:RWCD,O:RWCD,G:RWCD,W:RWCDe  G   Volume Status:  ODS-2, subject to mount verification, file high-water  marking,!       write-back caching enabled.s'   Volume is also mounted on X1, MELMAC.i      G Cluster interconnect is 100MBit switched Ethernet.  The other two nodesaG are a "Digital AlphaPC 164 500 MHz" and a "AlphaStation 400 4/233".  At = the time of the test nothing else was running in the cluster.n    H This looks to me like some misconfiguration of our cluster.  Do you have  H any suggestions what might cause this?  Where to look?  The cluster grew  B out of a pure VAX cluster over time so one thing I am wondering isA whether it inherited some way too small tuning parameters or usero( quotas.  AUTOGEN is happy the way it is.     Thanks for any suggestions,v Michaela   ------------------------------    Date: 22 Jul 2003 15:27:09 -0500+ From: young_r@encompasserve.org (Rob Young) ' Subject: Re: Poor disk IO.  How to fix?t3 Message-ID: <EwZhyvZOID2c@eisner.encompasserve.org>r  y In article <3F1D99BB.FD6D081D@sternwarte.uni-erlangen.de>, Michael Lemke <ai26.nospam@sternwarte.uni-erlangen.de> writes:oH > On our 3 node Alpha cluster I notice terrible disk IO perfromance whenG > compared to say, a Linux PC.  A simple tar xvf ...  takes 40 sec realsI > time but only 1 sec CPU time.  The same operation on an 800MHz Linux PCo, > finishes in 1.7 sec with 0.1 sec CPU time. > I > What I am really concerned with is the elapsed time.  I notice this badvJ > performance with all sorts of applications, it is not restricted to tar. >   ^ http://groups.google.com/groups?selm=8esfrh%24rc9%40gap.cco.caltech.edu&oe=UTF-8&output=gplain   			Rob   ------------------------------  % Date: Tue, 22 Jul 2003 13:57:00 -0700 + From: "Barry Treahy, Jr." <Treahy@MMaz.com>l' Subject: Re: Poor disk IO.  How to fix? ' Message-ID: <3F1DA51C.6000204@MMaz.com>s   Michael Lemke wrote:  G >On our 3 node Alpha cluster I notice terrible disk IO perfromance whenrF >compared to say, a Linux PC.  A simple tar xvf ...  takes 40 sec realH >time but only 1 sec CPU time.  The same operation on an 800MHz Linux PC+ >finishes in 1.7 sec with 0.1 sec CPU time.- >-H >What I am really concerned with is the elapsed time.  I notice this badI >performance with all sorts of applications, it is not restricted to tar.- >  - > F First, you can't compare software designed to run in Unix with how it G performs on VMS, the environments are not the same.  The fact that tar -C works on VMS is most likely an effort of ported functionality with  H little to no concern over performance.    The fact that your VMS system B is taking 40 seconds with only a second of CPU indicates your are L bottle-necked elsewhere; Be that I/O or memory (working set configurations).  : >Exiting MAIL and purging the spam also takes a long time. >e >Exact values for the tar test:  > ? >CPUTime: 1020 ms  Elapsed: 00:00:39.17  DIO: 2741  BufIO: 2342v >e >  d >t [snip]  G Oddly enough here, you did exclude what the page faults were.  If your h? working sets are hosed, you would be causing VMS to do tons of gH unnecessary paging turning image activation and termination.  Presuming I it is the same system and same user account, that could explain the poor 3D elapsed time performance with both the VMS Mail command and the tar 
 program...  H >Cluster interconnect is 100MBit switched Ethernet.  The other two nodesH >are a "Digital AlphaPC 164 500 MHz" and a "AlphaStation 400 4/233".  At> >the time of the test nothing else was running in the cluster. >wI You will NEVER get native disk performance over 100Mb Ethernet and since iC the last time I checked, Linux cannot perform clusters in the same eD capacity of VMS, to compare VMS cluster I/O over a NIC with Linux's 9 local I/O, well you are not comparing apples to apples...:  I >This looks to me like some misconfiguration of our cluster.  Do you havetI >any suggestions what might cause this?  Where to look?  The cluster grewnC >out of a pure VAX cluster over time so one thing I am wondering isTB >whether it inherited some way too small tuning parameters or user) >quotas.  AUTOGEN is happy the way it is.t >sE Bah!  Autogen is a sanity check but can be both your best friend and  H your worst enemy!  If you don't know anything about performance tuning, G hire a consultant or attend a VMS Performance and Tuning course (if HP K2 still offers those), it would be money well spent!  E My supposition is that first, you have unrealistic expectations when CI comparing dissimilar hardware architectures and configurations, and that sB your system's memory is most likely still configured for VAX (ie. 6 smaller working sets, max virtual address space, etc).   Anyway, that's my 2 cents...   Barryn   -- I  @ Barry Treahy, Jr  *  Midwest Microwave  *  Vice President & CIO   A E-mail: Treahy@mmaz.com * Phone: 480/314-1320 * FAX: 480/661-7028b   ------------------------------  % Date: Tue, 22 Jul 2003 16:53:42 -0400c From: koskaj@bender.com-' Subject: RE: Poor disk IO.  How to fix?,< Message-ID: <03072216534224.f956.58533216@alaxp3.bender.com>  ] >From:	SMTP%"ai26.nospam@sternwarte.uni-erlangen.de"  "Michael Lemke" 22-JUL-2003 16:25:19.68c >To:	Info-VAX@Mvb.Saic.Com! >Subj:	Poor disk IO.  How to fix?0 >2G >On our 3 node Alpha cluster I notice terrible disk IO perfromance whencF >compared to say, a Linux PC.  A simple tar xvf ...  takes 40 sec realH >time but only 1 sec CPU time.  The same operation on an 800MHz Linux PC+ >finishes in 1.7 sec with 0.1 sec CPU time.i >tH >What I am really concerned with is the elapsed time.  I notice this badI >performance with all sorts of applications, it is not restricted to tar.r > : >Exiting MAIL and purging the spam also takes a long time.   cut/trim...w  H >  Volume Status:  ODS-2, subject to mount verification, file high-water	 >marking,e  K Is file high-water marking needed for your setup?  It has some I/O penalty   involved with it.-   :) jck   ------------------------------    Date: 22 Jul 2003 15:45:37 -0500+ From: young_r@encompasserve.org (Rob Young):' Subject: Re: Poor disk IO.  How to fix?y3 Message-ID: <DXJY9YWz2eMJ@eisner.encompasserve.org>o  y In article <3F1D99BB.FD6D081D@sternwarte.uni-erlangen.de>, Michael Lemke <ai26.nospam@sternwarte.uni-erlangen.de> writes:hH > On our 3 node Alpha cluster I notice terrible disk IO perfromance whenG > compared to say, a Linux PC.  A simple tar xvf ...  takes 40 sec reallI > time but only 1 sec CPU time.  The same operation on an 800MHz Linux PCe, > finishes in 1.7 sec with 0.1 sec CPU time. >   / 	To be a little more thorough.  You can do the b  + 	$ SET RMS/buffer=132/block=64 /extend=2000i  < 	and that will help a bit.  This thread shows how gunzip was 	sped up by Paul Repacholi:s  ^ http://groups.google.com/groups?selm=93le66%24sam%40gap.cco.caltech.edu&oe=UTF-8&output=gplain  @ 	Now you know where the problem lies and if you tweek up the tarA 	sources, how to speed it up.  Check - maybe someone has already	-> 	done it , or someone else might chime in knowing where a sped) 	up tar lives (if there is such a thing).c   				RobV   ------------------------------    Date: 22 Jul 2003 16:01:18 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)(' Subject: Re: Poor disk IO.  How to fix?a3 Message-ID: <fELSxhvn27vz@eisner.encompasserve.org>p  y In article <3F1D99BB.FD6D081D@sternwarte.uni-erlangen.de>, Michael Lemke <ai26.nospam@sternwarte.uni-erlangen.de> writes:.H > On our 3 node Alpha cluster I notice terrible disk IO perfromance whenG > compared to say, a Linux PC.  A simple tar xvf ...  takes 40 sec realrI > time but only 1 sec CPU time.  The same operation on an 800MHz Linux PCc, > finishes in 1.7 sec with 0.1 sec CPU time. > I > What I am really concerned with is the elapsed time.  I notice this badhJ > performance with all sorts of applications, it is not restricted to tar.  H    When VMS says it's on the disk, the IT IS ON THE DISK.  Most other OS'    are happy just to leave it in cache.e  H    You can get around this with language specific buffering controls, orC    with generic RMS buffering if you're happy to emulate the risks.n  D    Most folks who port utilities to VMS don't know or care about the    tradeoff.   ------------------------------    Date: 22 Jul 2003 15:29:46 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)u Subject: Re: Sun and SCO3 Message-ID: <4jPXVjhbY4yr@eisner.encompasserve.org>i  q In article <uqpPmMr5zJvw@eisner.encompasserve.org>, koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:t > In article <bfgvpg$lik$1@new-usenet.uk.sun.com>, Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> writes: >> ,A >> You didn't consider that the deal could simply be a technologyn@ >> licensing deal that has nothing to do with SCO's ongoing case >> against IBM ??? > 
 >    Nope.  E    And the July 21 interview with Scott McNealy in eWeek supports ourd1    skepticism on Sun's true feelings about Linux.t   ------------------------------  + Date: Tue, 22 Jul 2003 18:49:55 +0000 (UTC)b$ From: THIS$THAT <nospam@spamcop.net>) Subject: Terminate with extreme prejudicei/ Message-ID: <bfk10j$j7h$1@grandcanyon.binc.net>m  @ The VMS box is the NFS server.  A Solaris box is the NFS client.  @ Locking failed when two IMAP clients tried to open same mailbox.  A Closed the windows in which they were running (Solaris).  The NFSM@ server tcpip$nfs_server.exe maxed 100% CPU.  Can't kill it - get? %SYSTEM-F-NODELETE, object cannot be deleted.  Can only suspend 
 it at kernel.o  C On Solaris, can't forcibly unmount the NFS share (umount -f hangs).3? A client process in kernel waiting for that NFS i/o to completed2 exists but won't die when killed, even with pstop.  @ So the VMS NFS server's looping and can't be killed, and the NFS< client, on Solaris won't die until it returns out of kernel.  A How do you terminate a VMS process with extreme prejudice, or fors< that matter a Solaris one (other than rebooting both boxes)?  & VMS V7.3, TCPIP V5.1 ECO 4, Solaris 8.   ------------------------------  # Date: Tue, 22 Jul 2003 19:06:19 GMTs# From: hoff@hp.nospam (Hoff Hoffman)d- Subject: Re: Terminate with extreme prejudicem0 Message-ID: <LWfTa.635$CT1.109@news.cpqcorp.net>  V In article <bfk10j$j7h$1@grandcanyon.binc.net>, THIS$THAT <nospam@spamcop.net> writes:  B :Closed the windows in which they were running (Solaris).  The NFSA :server tcpip$nfs_server.exe maxed 100% CPU.  Can't kill it - get @ :%SYSTEM-F-NODELETE, object cannot be deleted.  Can only suspend :it at kernel.  C   The usual reason the PCB$V_NODELET bit is set is because deletingVC   the process will lead to system instability or will likely resultcA   in a system crash, or because the process is critical to systemm
   operations.   C   The usual causes of looping involve a kernel-mode data structure  @   corruption, a lost I/O or lost quota value, or a coding error.C   If this is a lost I/O, AMDS or Availability Manager (if availablenE   to you in this configuration) might be able to reset this and allowK?   the process to exit -- but I would not count on that working.l  B :How do you terminate a VMS process with extreme prejudice, or for= :that matter a Solaris one (other than rebooting both boxes)?   C   Assuming restarting TCP/IP Services does not clear this -- and itvE   probably won't -- reboot the box.  When things go strange like this E   within kernel code, partial or half-solutions tend to lead to, um, a   unplanned downtime.t  ' :VMS V7.3, TCPIP V5.1 ECO 4, Solaris 8.r  A   I will assume you have the XFC ECO and other mandatory ECOs forVC   the OpenVMS release.  If you don't, well, start there.  Off-hand,.B   I don't recall the current TCP/IP Services ECO version for V5.1;2   I'm running V5.3 with ECO on most local servers.     Best and fastest to reboot.e  N  ---------------------------- #include <rtfaq.h> -----------------------------K     For additional, please see the OpenVMS FAQ -- www.hp.com/go/openvms/faqlN  --------------------------- pure personal opinion ---------------------------E         Hoff (Stephen) Hoffman   OpenVMS Engineering   hoff[at]hp.comt   ------------------------------  % Date: Tue, 22 Jul 2003 16:23:35 -0400 ! From: "JustMe" <no.body@no.where> - Subject: Re: Terminate with extreme prejudice-9 Message-ID: <Q1hTa.8947$FV6.491788@news20.bellglobal.com>p  ; There IS an ECO 5, with the following in the release notes:cL ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ ^^^^^^     ECO 5 updateso     ------------- #     ECO H 17-JUL-2002 Alpha and VAXc  
      Problem:y>      Performance problems when NFS serving between OpenVMS and      Sun Solaris.s        Deliverables:!      TCPIP$CFS_SHR.EXE   V5.1-15Ha$      TCPIP$NFS_SERVER.EXE   V5.1-15X&      TCPIP$NFS_SERVICES.EXE   V5.1-15X        Reference:a2      PTR 70-5-1924 / CFS.88652 / Req Id: SE_G028512      PTR 70-5-2039 / CFS.91749 / Req Id: SE_G03896L ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ ^^^^^^  0 "Hoff Hoffman" <hoff@hp.nospam> wrote in message* news:LWfTa.635$CT1.109@news.cpqcorp.net...9 In article <bfk10j$j7h$1@grandcanyon.binc.net>, THIS$THATi <nospam@spamcop.net> writes:  B :Closed the windows in which they were running (Solaris).  The NFSA :server tcpip$nfs_server.exe maxed 100% CPU.  Can't kill it - geth@ :%SYSTEM-F-NODELETE, object cannot be deleted.  Can only suspend :it at kernel.  C   The usual reason the PCB$V_NODELET bit is set is because deletingrC   the process will lead to system instability or will likely resultnA   in a system crash, or because the process is critical to systeml
   operations.g  B   The usual causes of looping involve a kernel-mode data structure@   corruption, a lost I/O or lost quota value, or a coding error.C   If this is a lost I/O, AMDS or Availability Manager (if availablesE   to you in this configuration) might be able to reset this and allowN?   the process to exit -- but I would not count on that working.i  B :How do you terminate a VMS process with extreme prejudice, or for= :that matter a Solaris one (other than rebooting both boxes)?f  C   Assuming restarting TCP/IP Services does not clear this -- and it-E   probably won't -- reboot the box.  When things go strange like thistD   within kernel code, partial or half-solutions tend to lead to, um,   unplanned downtime.   ' :VMS V7.3, TCPIP V5.1 ECO 4, Solaris 8..  A   I will assume you have the XFC ECO and other mandatory ECOs forsC   the OpenVMS release.  If you don't, well, start there.  Off-hand,hB   I don't recall the current TCP/IP Services ECO version for V5.1;2   I'm running V5.3 with ECO on most local servers.     Best and fastest to reboot.u  &  ---------------------------- #include' <rtfaq.h> ----------------------------- K     For additional, please see the OpenVMS FAQ -- www.hp.com/go/openvms/faqa*  --------------------------- pure personal# opinion ---------------------------'E         Hoff (Stephen) Hoffman   OpenVMS Engineering   hoff[at]hp.comu   ------------------------------  + Date: Tue, 22 Jul 2003 21:54:49 +0000 (UTC)e$ From: THIS$THAT <nospam@spamcop.net>- Subject: Re: Terminate with extreme prejudicet/ Message-ID: <bfkbr9$m18$1@grandcanyon.binc.net>l  
 Hoff replied: B >In article <bfk10j$j7h$1@grandcanyon.binc.net>, THIS$THAT writes: >cC >:How do you terminate a VMS process with extreme prejudice, or for > >:that matter a Solaris one (other than rebooting both boxes)? >)D >  Assuming restarting TCP/IP Services does not clear this -- and it$ >  probably won't -- reboot the box.  < 	No, shutting it down using the tcpipconfig menu is what was> 	tried originally but tcpip$nfs_shutdown.com hung in TCPIP$UCP+ 	(at the tcpip unmap * /noconfirm command).s  C >  I don't recall the current TCP/IP Services ECO version for V5.1;l3 >  I'm running V5.3 with ECO on most local servers.   = 	Anyone else in Europe waiting a long time for their hobbyistp; 	CDs?  Sent Montagar CWO for the media 8 weeks ago tomorrowo< 	in order to get TCP/IP Services V5.3 but nothing's arrived. >- >  Best and fastest to reboot.  ' 	Appreciate the fast response, respect.s   ------------------------------  + Date: Tue, 22 Jul 2003 22:47:09 +0000 (UTC)9$ From: THIS$THAT <nospam@spamcop.net>- Subject: Re: Terminate with extreme prejudice / Message-ID: <bfketd$mq1$1@grandcanyon.binc.net>o  # "JustMe" <no.body@no.where> writes:d > < >There IS an ECO 5, with the following in the release notes:M >^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^a >    ECO 5 updates >    -------------$ >    ECO H 17-JUL-2002 Alpha and VAX >- >     Problem:? >     Performance problems when NFS serving between OpenVMS andn >     Sun Solaris. >e >     Deliverables:e" >     TCPIP$CFS_SHR.EXE   V5.1-15H% >     TCPIP$NFS_SERVER.EXE   V5.1-15Xt' >     TCPIP$NFS_SERVICES.EXE   V5.1-15Xa >i >     Reference:3 >     PTR 70-5-1924 / CFS.88652 / Req Id: SE_G02851 3 >     PTR 70-5-2039 / CFS.91749 / Req Id: SE_G03896rM >^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^G  < 	Thank you for that timely reminder.  To be honest, I forget= 	why I didn't go through with the update - possibly because am9 	reboot would have been inconvenient.  But again, thanks.a  % Directory SYS$SYSDEVICE:[SYS$SCRATCH]   ' DEC-AXPVMS-TCPIP_ECO-V0503-181-4.PCSI;1tF                       102064  13-SEP-2002 12:06:17.99  (RWED,RWED,RE,)2 DEC-AXPVMS-TCPIP_ECO-V0503-181-4.PCSI-DCX_AXPEXE;1B                        58016  18-SEP-2002 18:04:31.72  (RWD,RWD,,)  O ----------------------------------- ----------- -------------------------------rH PRODUCT                             KIT TYPE    OPERATION  DATE AND TIMEO ----------------------------------- ----------- -------------------------------mO DEC AXPVMS TCPIP_ECO V5.1-154       Patch       Install    01-JUN-2002 18:35:18tO DEC AXPVMS TCPIP_ECO V5.1-153       Patch       Install    23-APR-2002 22:02:16-O DEC AXPVMS TCPIP V5.1-15            Full LP     Install    12-DEC-2001 10:18:41oO ----------------------------------- ----------- -------------------------------4   ------------------------------   End of INFO-VAX 2003.403 ************************