1 INFO-VAX	Sun, 09 Jul 2000	Volume 2000 : Issue 380       Contents:" Re: Dead DECsystem 3000-500 power?7 Re: Disk I/O Performance (was Re: OpenVMS loses big...) 7 Re: Disk I/O Performance (was Re: OpenVMS loses big...) 7 Re: Disk I/O Performance (was Re: OpenVMS loses big...)   MIN BOOT problem - openVMS 5.5-2$ Re: MIN BOOT problem - openVMS 5.5-2$ Re: MIN BOOT problem - openVMS 5.5-2$ Re: MIN BOOT problem - openVMS 5.5-2$ Re: MIN BOOT problem - openVMS 5.5-2$ Re: MIN BOOT problem - openVMS 5.5-2$ Re: MIN BOOT problem - openVMS 5.5-2$ Re: MIN BOOT problem - openVMS 5.5-2$ Re: MIN BOOT problem - openVMS 5.5-2$ Re: MIN BOOT problem - openVMS 5.5-2: RE: Summer 2000 OpenVMS TIMES Now Available on CPQ Website: Re: Summer 2000 OpenVMS TIMES Now Available on CPQ WebsiteD UNICENTER PERFORMANCE DATA COLLECTOR - DOES NOT INSTALL ON VMS 7.2!!  F ----------------------------------------------------------------------  # Date: Sat, 08 Jul 2000 22:24:39 GMT # From: Mark Sterk <strong@chello.nl> + Subject: Re: Dead DECsystem 3000-500 power? ) Message-ID: <3967A91C.913888E4@chello.nl>   
 Thomas wrote:   = > I have an "old" DECsystem 3000-500 that powers up for about = > five to ten seconds, starts spinning up its internal drives > > and such, but after the LED display hits "46", the red light@ > next to the 0/1 power switch comes on and the machine seems to1 > lose power to everything except that red light. ? > Even if I turn off the power switch, that red light stays on.  > = > Is this the way the system tells me that there is something @ > horribly wrong with the power supply and that I should replace > it?   > (yes, I have no documentation) >  > Thomas  H Check if all the fans are running and all the cables are connected, thisF machine has a protection that shuts the power down after a few seconds if one of them fails.     
 Greetings,   Mark   ------------------------------  % Date: Sat, 08 Jul 2000 21:46:28 -0500 , From: "Glenn C. Everhart" <Everhart@GCE.com>@ Subject: Re: Disk I/O Performance (was Re: OpenVMS loses big...)' Message-ID: <3967A133.560E7812@GCE.com>   A Where the disk in question supports tagged queuing (and presuming @ a recent VMS; support was introduced in 6.2), the drive can have@ many operations inside the drive, not queued anywhere in the OS.C Drives capable of this kind of work typically rearrange I/O so that @ rotational and seek delays are minimized. So long as you can getD the commands to the disk fast enough (which depends on many factors;C VMS can saturate SCSI busses and is pretty good at getting commands G to drives, but interface chips differ and of course if you issue single E block $qio, you're adding overhead that this effect may not help) the  I/O will stream.  E This is a reason to look for qualified disks. Lots of disks out there ? have no clue how to do tagged queueing. Some claim to do it but > do bizarre things so whatever support is there cannot be used.. Qualification tests this behavior, and others.  A Command delays can be at SCSI level, in the conversation with the C control processor, in port driver control logic, or at class driver D and above ($qio or the like).  If you don't measure, it is difficultC to know where this is. Once VMS passes an IRP to a class driver, it 3 knows nothing about what goes on below that level.    D While on this topic, remember that ACP_DATACHECK can cause slowdownsG in SCSI processing by forcing data compare on metadata writes. I always G turn this off, because it not only prevents tagged queueing across such < operations, but forces a rotational delay in rereading data.     > = > Richard B. Gilbert <DRAGON@compuserve.com> wrote in message 3 > news:200007062103_MC2-AB62-6E5A@compuserve.com... N > >         But simply flooding the disk with I/O requests is not the solutionH > > either.  You need to flood the disk with BIG I/O requests.  Four I/OL > > requests for 125 blocks each will run a lot faster than 500 requests for > > one block each.  > B > Only because SCSI command processing happens to take longer thanN > transferring a single block does:  otherwise, the request size is irrelevantI > to streaming bandwidth, as long as you keep at least one request always  > queued at the drive. > N > And in the context of this discussion, 125 blocks is not particularly large,L > and in fact is within the VMS QIO size limit that has been brought up as aG > possible problem (which it isn't, at least in this context, since for N > sequential access 125-block requests will stream at the maximum data rate ofK > the drive, because the transfer time *does* exceed the command-processing 1 > time and thus the queue can be kept non-empty).  > I > >         I've seen an Oracle import running like a dog with a peak I/O L > > request queue on a disk of 498 requests and an average of about 90.  TheJ > > disk had no lack of work to do but it was not accomplishing very much. > N > If it was not transferring at maximum bandwidth, it was because the requestsL > weren't physically sequential on the drive.  Or because they were so smallK > that the command processing overhead exceeded the transfer time (as noted M > above), though in that case you would see the queue form in the driver, not  > at the disk. >  > - bill   ------------------------------  % Date: Sat, 08 Jul 2000 21:59:45 -0400 2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>@ Subject: Re: Disk I/O Performance (was Re: OpenVMS loses big...)7 Message-ID: <200007082200_MC2-AB97-2000@compuserve.com>   :         The I/O Request Queue that I described was show by $ MONITOR DISK /ITEM=3DQUEUE> So it's the queue that VMS saw rather than what the drive saw.  G Oracle 8 was importing a table, with a couple of indices.  There was no H other activity on the system except my process running monitor.  The DBAC says that the table file was being written sequentially; if we were J bouncing the heads around for every block written, I could understand its=   taking a long time. # Message text written by "Bill Todd" < >Richard B. Gilbert <DRAGON@compuserve.com> wrote in message1 news:200007081106_MC2-AB92-E1C8@compuserve.com... 6         Does the HSZ50 controller queue them that way?  F It's somewhat hard to believe that any competently-designed controller6 wouldn't allow its disks to stream data if they could.  G Your example of Oracle 8 KB transfers falls somewhere in the grey area, J though:  on a modern disk, getting 8 KB off the platter takes only 0.5 ms= . J or less, which may well be less than the time it takes to process the nex= t J SCSI command (the newest, fastest SCSI standards may finally have found a=  J way to speed up command processing, which at least until recently has bee= n B stuck at SCSI-1 levels despite the bandwidth improvements in otherJ respects):  if so, then the disk will miss a rev once in a while, but eac= h D time it does, it will give a few more commands the opportunity to beJ processed, so it won't miss a rev on *every* 8 KB request (and the reques= t - queue at the disk shouldn't grow very large).   F With a long queue, it sounds a lot more like the requests are randomlyJ distributed around the disk, rather than physically contiguous.  Since ea= chH request will then take a few ms. to complete (due to seek and rotationalJ latency), a lengthy queue can indeed develop.  However (duh...), IIRC SCS= I J drives don't normally support more than 256 requests concurrently queued = atJ the drive - so the example you gave sounds as if the queuing was in fact = in the driver.   J The difference between Oracle and BACKUP is likely that Oracle just submi= tsA disk requests from a large number of threads or processes without 
 attemptingH to coordinate them (trusting the underlying OS and disk facilities to doH so), whereas BACKUP is submitting a single, multi-buffered read or writeJ stream to the disk and just submits requests fast enough to keep the queu= e G from becoming empty (since it usually takes only a few buffers to reach  optimal performance).    - bill     BACKUP puts a J bigger load on the disks than anything else I recall seeing and I've neve= r , seen BACKUP develop a queue of 499 requests.  # Message text written by "Bill Todd" < >Richard B. Gilbert <DRAGON@compuserve.com> wrote in message0 news:200007071930_MC2-AB88-465@compuserve.com...H >         Don't forget that there is rotational latency for each request evenE > if the file is being written sequentially and the heads are already $ > positioned at the proper cylinder.  J No, there is not (not on normal SCSI drives, anyway).  As long as the wri= teG requests are queued to the drive asynchronously, such that the drive is H always in possession of the next request when the current one completes, the 9 data will be streamed onto the disk without interruption.  <  <    ------------------------------  $ Date: Sun, 9 Jul 2000 01:45:27 -0400' From: "Bill Todd" <billtodd@foo.mv.com> @ Subject: Re: Disk I/O Performance (was Re: OpenVMS loses big...)( Message-ID: <8k93e4$41u$1@pyrite.mv.net>  ; Richard B. Gilbert <DRAGON@compuserve.com> wrote in message 1 news:200007082200_MC2-AB97-2000@compuserve.com... < >         The I/O Request Queue that I described was show by > $ MONITOR DISK /ITEM=QUEUE@ > So it's the queue that VMS saw rather than what the drive saw. > I > Oracle 8 was importing a table, with a couple of indices.  There was no J > other activity on the system except my process running monitor.  The DBAE > says that the table file was being written sequentially; if we were K > bouncing the heads around for every block written, I could understand its  > taking a long time.   L Given that this was the VMS queue, there's no congestion point comparable toK that of SCSI command acceptance (at something 1 per ms.) to limit the speed J with which the queue depth can increase.  So if, for example, Oracle builtJ up the entire table structure in memory and then just submitted 8 KB writeL requests as fast as it could form them from the pre-built in-memory data, itK could potentially build up quite a request back-log even if the disk was in 6 fact streaming them to the platters at peak bandwidth.  L Or the VMS file containing the Oracle table may have been heavily-fragmentedK below the Oracle level, though fragmentation sufficient to slow things down K as much as you suggest by saying "The disk had no lack of work to do but it H was not accomplishing very much" seems somewhat unlikely.  Or Oracle mayK have been rebuilding the table logically rather than physically, meaning it I had to update the table indexing information for each page inserted, plus J the per-tuple indexing information for each *tuple* inserted (*that* wouldK slow you down more than a bit):  if the table was not being imported from a G copy with absolutely identical characteristics, including page size and H indexed columns, at least *some* such index-building would be necessary,J though there are ways to cut it down drastically as long as the insertionsC aren't being made to a table while other accesses to it may also be 
 occurring.  K There are too many possibilities to make speculation here very useful.  And L it's even possible that the HSZ *is* incompetent at handling data-streaming,L though I certainly wouldn't place that near the top of the probability list.   - bill   ------------------------------  # Date: Sun, 09 Jul 2000 02:26:34 GMT 0 From: gilley@nospam.bravewc.com (Charles Gilley)) Subject: MIN BOOT problem - openVMS 5.5-2 = Message-ID: <SmR95.3177$451.91241@news-west.usenetserver.com>   O On a rather aged VAX today, I ran into a bit of a problem booting the machine.  J Now I'll admit right now that my VMS is rusty. At boot, this machine wouldL stop at the # prompt.  Hmmm, I thought, it is set to minimum boot.  Okay, I M can fix that.  So, setting the bflg to 1, I popped into sysboot to change the N value of STARTUP_P1.  But it isn't set to MIN, it's blanks.  Huh?  So, I went > on and ran startup.com, but why is my machine acting this way?  1 Thanks for any pointers, this one has me stumped.   - chg (watch the antispam in the email address)    ------------------------------  % Date: Sat, 08 Jul 2000 22:41:07 -0400 , From: Howard S Shubs <hshubs@mindspring.com>- Subject: Re: MIN BOOT problem - openVMS 5.5-2 > Message-ID: <hshubs-F6915C.22410708072000@news.mindspring.com>  > In article <SmR95.3177$451.91241@news-west.usenetserver.com>, 1 gilley@nospam.bravewc.com (Charles Gilley) wrote:   G >On a rather aged VAX today, I ran into a bit of a problem booting the  
 >machine. K >Now I'll admit right now that my VMS is rusty. At boot, this machine would K >stop at the # prompt.  Hmmm, I thought, it is set to minimum boot.  Okay,   >I  K >can fix that.  So, setting the bflg to 1, I popped into sysboot to change   >theJ >value of STARTUP_P1.  But it isn't set to MIN, it's blanks.  Huh?  So, I  >went ? >on and ran startup.com, but why is my machine acting this way?   I Setting the boot flag to 1 just does a conversational boot, giving you a  H SYSBOOT> prompt.  You then have to explicitly set STARTUP_P1 as desired.  . Maybe I'm misunderstanding what you're saying.   --   Howard S Shubs, the Denim Adept    ------------------------------  % Date: Sat, 08 Jul 2000 23:15:36 -0400 , From: Howard S Shubs <hshubs@mindspring.com>- Subject: Re: MIN BOOT problem - openVMS 5.5-2 > Message-ID: <hshubs-9655B3.23153608072000@news.mindspring.com>  > In article <a%R95.3246$451.94619@news-west.usenetserver.com>, 1 gilley@nospam.bravewc.com (Charles Gilley) wrote:   @ >Whoops, I'll be more explicit.  Setting bflg does get you into  >conversational L >boot mode.  And that turns on/off as expected.  But I cannot get VMS to do  >a  F >normal boot.  STARTUP_P1 is alwas "    ", but it still boots minimum. > + >Is that clearer?  Sorry for the confusion.   G It says it's booting minimum?  I know it may actually say that when it  F says it's not starting the network because the machine is doing a MIN * boot.  How do you know it's coming up MIN?   --   Howard S Shubs, the Denim Adept    ------------------------------  # Date: Sun, 09 Jul 2000 03:09:33 GMT 0 From: gilley@nospam.bravewc.com (Charles Gilley)- Subject: Re: MIN BOOT problem - openVMS 5.5-2 = Message-ID: <a%R95.3246$451.94619@news-west.usenetserver.com>   N Whoops, I'll be more explicit.  Setting bflg does get you into conversational M boot mode.  And that turns on/off as expected.  But I cannot get VMS to do a  E normal boot.  STARTUP_P1 is alwas "    ", but it still boots minimum.   * Is that clearer?  Sorry for the confusion.  N In article <hshubs-F6915C.22410708072000@news.mindspring.com>, Howard S Shubs  <hshubs@mindspring.com> wrote:? >In article <SmR95.3177$451.91241@news-west.usenetserver.com>,  2 >gilley@nospam.bravewc.com (Charles Gilley) wrote: > H >>On a rather aged VAX today, I ran into a bit of a problem booting the  >>machine.  L >>Now I'll admit right now that my VMS is rusty. At boot, this machine wouldL >>stop at the # prompt.  Hmmm, I thought, it is set to minimum boot.  Okay,  >>I L >>can fix that.  So, setting the bflg to 1, I popped into sysboot to change  >>the K >>value of STARTUP_P1.  But it isn't set to MIN, it's blanks.  Huh?  So, I   >>went  @ >>on and ran startup.com, but why is my machine acting this way? > J >Setting the boot flag to 1 just does a conversational boot, giving you a I >SYSBOOT> prompt.  You then have to explicitly set STARTUP_P1 as desired.  > / >Maybe I'm misunderstanding what you're saying.  >    ------------------------------  # Date: Sun, 09 Jul 2000 03:37:27 GMT 0 From: gilley@nospam.bravewc.com (Charles Gilley)- Subject: Re: MIN BOOT problem - openVMS 5.5-2 = Message-ID: <mpS95.3293$451.96863@news-west.usenetserver.com>   I There isn't anythign there in terms of process support.  STARTUP has not  O executed.  No drives, etc.  It just boots to the # prompt.  I confess again my  K VMS is rusty, so my use of the term MIN may be incorrect.  But if I recall  M correctly, booting conversationally, then minimum yielded a # prompt w/o any  J password.  This was the approved way of accessing a system w/o the system  password, etc.  N In article <hshubs-9655B3.23153608072000@news.mindspring.com>, Howard S Shubs  <hshubs@mindspring.com> wrote:? >In article <a%R95.3246$451.94619@news-west.usenetserver.com>, l2 >gilley@nospam.bravewc.com (Charles Gilley) wrote: >dA >>Whoops, I'll be more explicit.  Setting bflg does get you into M >>conversational eM >>boot mode.  And that turns on/off as expected.  But I cannot get VMS to do e >>a G >>normal boot.  STARTUP_P1 is alwas "    ", but it still boots minimum.T >>, >>Is that clearer?  Sorry for the confusion. >VH >It says it's booting minimum?  I know it may actually say that when it G >says it's not starting the network because the machine is doing a MIN o+ >boot.  How do you know it's coming up MIN?5 >2   ------------------------------  % Date: Sat, 08 Jul 2000 23:43:28 -04002, From: Howard S Shubs <hshubs@mindspring.com>- Subject: Re: MIN BOOT problem - openVMS 5.5-2o> Message-ID: <hshubs-C92B7F.23432808072000@news.mindspring.com>  > In article <mpS95.3293$451.96863@news-west.usenetserver.com>, 1 gilley@nospam.bravewc.com (Charles Gilley) wrote:a  J >There isn't anythign there in terms of process support.  STARTUP has not G >executed.  No drives, etc.  It just boots to the # prompt.  I confess .
 >again my L >VMS is rusty, so my use of the term MIN may be incorrect.  But if I recall J >correctly, booting conversationally, then minimum yielded a # prompt w/o  >any eK >password.  This was the approved way of accessing a system w/o the system m >password, etc.h  C You mean the $ prompt, I assume.  You've modified the startup file n6 setting to be OPA0: and you need to switch it back to ! SYS$SYSTEM:STARTUP.COM  I betcha.m   -- m Howard S Shubs, the Denim Adeptr   ------------------------------  # Date: Sun, 09 Jul 2000 03:51:48 GMTs0 From: gilley@nospam.bravewc.com (Charles Gilley)- Subject: Re: MIN BOOT problem - openVMS 5.5-2a= Message-ID: <UCS95.3322$451.98978@news-west.usenetserver.com>e  ! Okay, thanks.  I'll check it out.   N In article <hshubs-C92B7F.23432808072000@news.mindspring.com>, Howard S Shubs  <hshubs@mindspring.com> wrote:? >In article <mpS95.3293$451.96863@news-west.usenetserver.com>, /2 >gilley@nospam.bravewc.com (Charles Gilley) wrote: >DK >>There isn't anythign there in terms of process support.  STARTUP has not qH >>executed.  No drives, etc.  It just boots to the # prompt.  I confess  >>again my eM >>VMS is rusty, so my use of the term MIN may be incorrect.  But if I recall bK >>correctly, booting conversationally, then minimum yielded a # prompt w/o   >>any L >>password.  This was the approved way of accessing a system w/o the system  >>password, etc. >aD >You mean the $ prompt, I assume.  You've modified the startup file 7 >setting to be OPA0: and you need to switch it back to s" >SYS$SYSTEM:STARTUP.COM  I betcha. >'   ------------------------------  $ Date: Sat, 8 Jul 2000 22:58:07 -0500) From: "John E. Malmberg" <wb8tyw@qsl.net>s- Subject: Re: MIN BOOT problem - openVMS 5.5-2t- Message-ID: <smftmjc3nu62@corp.supernews.com>   - Charles Gilley <gilley@bravewc.company> wrotenB in message news:SmR95.3177$451.91241@news-west.usenetserver.com...C > On a rather aged VAX today, I ran into a bit of a problem bootingoA > the machine.  Now I'll admit right now that my VMS is rusty. At B > boot, this machine would stop at the # prompt.  Hmmm, I thought,E > it is set to minimum boot.  Okay, I  can fix that.  So, setting thew9 > bflg to 1, I popped into sysboot to change the value ofn? > STARTUP_P1.  But it isn't set to MIN, it's blanks.  Huh?  So,S> > I went  on and ran startup.com, but why is my machine acting > this way?i  L The "#" prompt is not from any VMS default prompts I am aware of.  It soundsB like something is taking over the console during the boot process.  G If you set the value of STARTUP_P1 to be "MIN" then you can examine the-I SYSTART*.COM files in the SYS$MANAGER: directory to see what it really is  doing.  ! What specific model of VAX is it?   C > But if I recall correctly, booting conversationally, then minimums= > yielded a # prompt w/o any password.  This was the approvedf9 > way of accessing a system w/o the system password, etc.r  E Not any version of VMS that I remember.  Even a minimum boot requires  a login.  L What you may be refering to is SET/STARTUP OPA0: and that still prompts with a "$".   -Johne wb8tyw@qsl.network   ------------------------------  # Date: Sun, 09 Jul 2000 03:58:39 GMTs0 From: gilley@nospam.bravewc.com (Charles Gilley)- Subject: Re: MIN BOOT problem - openVMS 5.5-2 = Message-ID: <kJS95.3342$451.98978@news-west.usenetserver.com>e  K Gad, not even in my wildest imagination would I expect such rapid feedback.uN Thanks.  Well, the system is an old 4000-300, and older VT something, and the H # I refer to may have been in fact the $ sign.  The screen is dirty and M slightly fuzzy (it's my defense and I'm sticking to it).  At the moment, the u OPA0: theory is in the lead.  A In article <smftmjc3nu62@corp.supernews.com>, "John E. Malmberg" t <wb8tyw@qsl.net> wrote:p. >Charles Gilley <gilley@bravewc.company> wroteC >in message news:SmR95.3177$451.91241@news-west.usenetserver.com...sD >> On a rather aged VAX today, I ran into a bit of a problem bootingB >> the machine.  Now I'll admit right now that my VMS is rusty. AtC >> boot, this machine would stop at the # prompt.  Hmmm, I thought, F >> it is set to minimum boot.  Okay, I  can fix that.  So, setting the: >> bflg to 1, I popped into sysboot to change the value of@ >> STARTUP_P1.  But it isn't set to MIN, it's blanks.  Huh?  So,? >> I went  on and ran startup.com, but why is my machine acting_ >> this way? >pM >The "#" prompt is not from any VMS default prompts I am aware of.  It soundsOC >like something is taking over the console during the boot process.w > H >If you set the value of STARTUP_P1 to be "MIN" then you can examine theJ >SYSTART*.COM files in the SYS$MANAGER: directory to see what it really is >doing.e >A" >What specific model of VAX is it? >iD >> But if I recall correctly, booting conversationally, then minimum> >> yielded a # prompt w/o any password.  This was the approved: >> way of accessing a system w/o the system password, etc. >@F >Not any version of VMS that I remember.  Even a minimum boot requires	 >a login.  > M >What you may be refering to is SET/STARTUP OPA0: and that still prompts withe >a "$".t >y >-John >wb8tyw@qsl.networkl >  >l >t   ------------------------------  % Date: Sun, 09 Jul 2000 00:25:48 -0400K, From: Howard S Shubs <hshubs@mindspring.com>- Subject: Re: MIN BOOT problem - openVMS 5.5-2h> Message-ID: <hshubs-184B5F.00254809072000@news.mindspring.com>  > In article <kJS95.3342$451.98978@news-west.usenetserver.com>, 1 gilley@nospam.bravewc.com (Charles Gilley) wrote:o  C >Gad, not even in my wildest imagination would I expect such rapid n
 >feedback.   Sometimes we're quick.   -- i Howard S Shubs, the Denim Adepto   ------------------------------  % Date: Sat, 08 Jul 2000 15:18:33 -0400 + From: "Main, Kerry" <Kerry.Main@compaq.com>rC Subject: RE: Summer 2000 OpenVMS TIMES Now Available on CPQ WebsitenJ Message-ID: <910612C07BCAD1119AF40000F86AF0D805284503@kaoexc4.kao.dec.com>   Bill,g   Always the positive outlook :-)s  I >>> "I think [VMS is] almost uniquely qualified to service the demands ofeI e-business," said Capellas, conveniently ignoring the fact that while VMSmL indeed provides in many cases uniquely superior under-pinnings, a lot of theH super-structure required by e-business is available only on more popularK platforms, a situation which even with major efforts by Compaq won't changeeD immediately - assuming that such efforts are in fact made at all.<<<  L Michael Capellas is the CEO of a major multi-platform company, Do you reallyL expect him to come out and say "OpenVMS is clearly the solution to everyonesI problems.." ? Does Gerstner from IBM ever come out and say "MVS (pick any > IBM OS here) is clearly the solution for everyones problems.."   Nope.   L Specific to your concern about applications, Compaq just recently announced:    I 1.) the COE initiative which will allow easier porting and development ofaL applications for those that prefer a UNIX style of programming and a minimumL 15 year commitment to ongoing development for OpenVMS. A major investment by Compaq.q  K 2.) Compaq-Oracle partnership that will see OpenVMS as one of the platformspB supported for Oracles recently announced major change in their IAS? application strategy.  A major investment by Compaq and Oracle.c  L 3.) OpenVMS porting library to make it easier to move applications from UNIX to OpenVMS.d  H 4.) New beta kit of the JServ/JSSI Modules for Apache for OpenVMS Alpha.  H 5.) M16 beta kit of Mozilla for OpenVMS released two days after official Mozilla M16 release.  8 6.) JAVA 2 (Version 1.2.2-beta1) recently released beta.  F 7.) Cognos has recently released their web based products for OpenVMS.
 Reference: http://www.cognos.com/adtpci/   I 8.) The recent Customer / ISV references for OpenVMS and the new Alpha GSl Series:b3 http://www.openvms.digital.com/gsseries/quotes.htmlc  L There are also ongoing discussions with other (major) companies which should& also become public in the near future.  K So, while there is obviously lots of work to do, your statement "- assuming I that such efforts are in fact made at all.." clearly does not reflect theM work that is really going on.t  I This newsletters goal is to simply raise the awarenes of some of the goodeK things happening in the OpenVMS space with both current and new Customers. c  L Of course, if you are looking for something else, then you are free to trashI it and/or read between the lines as much as you want, but the Customers IcL have heard from enjoy reading what others are doing and what is happening in the OpenVMS space.   Regards,  
 Kerry Main Senior Consultant,
 Compaq Canadad Professional Servicesn Voice : 613-592-4660 FAX   : 819-772-7036 Email : kerry.main@compaq.com        -----Original Message-----, From: Bill Todd [mailto:billtodd@foo.mv.com]% Sent: Saturday, July 08, 2000 4:52 AMa To: Info-VAX@Mvb.Saic.CommC Subject: Re: Summer 2000 OpenVMS TIMES Now Available on CPQ Websitei    9 Terry C. Shannon <shannon@world.std.com> wrote in messaget  news:FxC1ut.4uq@world.std.com...C > Compaq today posted the third issue of its OpenVMS TIMES customervJ > newsletter. Volume 1 Number 3 is a 960KB PDF that can be downloaded from > . > http://www.openvms.digital.com/openvmstimes/  J Busy couple of days, but I finally got around to reading it - so here's my usual reality-check response:c  J Capellas pigeon-holed VMS as 'by definition' operating in mission-criticalL situations in his so-called interview (I mean, *I* understand the differenceC between a *real* interview and a staged conversation with soft-balltH questions and set-up answers, and I assume most of the rest of you do asK well, so why insult our intelligence?).  While it's nice to emphasize VMS'seI strengths in this area, a Compaq strategy to sell VMS *only* when nothingeL else will do the job is not going to give it the 'renaissance' that a lot ofJ us might hope for - and if Capellas actually sees VMS the way he described= it, then we can't hope for much help from him in this regard.a  E "I think [VMS is] almost uniquely qualified to service the demands ofeI e-business," said Capellas, conveniently ignoring the fact that while VMSaL indeed provides in many cases uniquely superior under-pinnings, a lot of theH super-structure required by e-business is available only on more popularK platforms, a situation which even with major efforts by Compaq won't changeRA immediately - assuming that such efforts are in fact made at all.V  F "Now it's a question of taking the fight to the streets and taking ourK message to the market" would be a more stirring phrase if I hadn't heard it I (plus a couple of other statements in the article) *verbatim* a month agohF from a different source in the same context:  at least one of them wasK offering someone else's words as his own (OK, call me naive:  I still don't  approve of it).p  G I'd have a lot more faith in these guys if their public demeanor didn'tiB remind me so strongly of politicians or used-car salesmen far moreD interested in making a sale than in satisfying a need.  Until CompaqI understands what customers need and tries to provide it, I fear they willBG continue to get clobbered by other companies more in tune with reality:eI enterprise computing is serious business, and purchase decisions are madeeK accordingly.  It would also certainly not have hurt to have had someone whoa@ understands that Galaxy has *nothing* to do with (heterogeneous)A inter-operability proof-read the article to avoid that particular  embarrassment.  A Moving right along, I suppose I shouldn't be any less exacting inO criticising Terry's article:  I "Compaq senior management clearly regards OpenVMS as a valuable strategic J asset, not a cash cow that's destined for a one-way trip to the abbatoir."G 'Clearly'?  Perhaps when we see evidence of a real advertising campaign F (instead of a couple of twitches indicating that what we thought was aI long-dead corpse might have a breath of life hiding away somewhere), realbF willingness to compete with other Compaq systems, and real developmentH commitments aimed at widening the long-narrowing gap between VMS and itsD inferior but faster-moving competitors, *then* there will be 'clear'I evidence to that effect.  Meanwhile, sounds more like clear evidence of a0 successful snow job.  L And "... it now appears that OpenVMS is poised for a renaissance and perhapsJ even a return to growth mode" - ?  Any definition of 'renaissance' in thisK context that did *not* minimally imply a 'return to growth mode' would be au rather strange one, IMO.  @ Terry's got a somewhat fine line to walk, especially in a CompaqH publication:  it's not wise to alienate your sources, and he's naturallyH enthusiastic after all his years with VMS to see a potential turn-aroundK occurring.  But there's a taint of fluff in this piece (accompanying a fairnK amount of non-fluff, I hasten to add) that won't enhance his reputation for A objectivity (though since, come to think of it, this may be a job F requirement for anyone claiming to be an 'industry analyst', it may be unfair to single him out).  L Ah, well.  The rest of the publication seemed mildly informative or at worstG innocuous.  Perhaps one should expect no better from something which isnH effectively a sales brochure - except in those areas, like the first two= articles, where it pretends to be something more substantial.u   - bill   ------------------------------  $ Date: Sat, 8 Jul 2000 18:22:05 -0400' From: "Bill Todd" <billtodd@foo.mv.com>dC Subject: Re: Summer 2000 OpenVMS TIMES Now Available on CPQ Websitei( Message-ID: <8k89er$5on$1@pyrite.mv.net>  4 Main, Kerry <Kerry.Main@compaq.com> wrote in messageD news:910612C07BCAD1119AF40000F86AF0D805284503@kaoexc4.kao.dec.com... > Bill,t > ! > Always the positive outlook :-)t  F Not positive, not negative:  just relatively unbiased.  No one pays orI otherwise influences me concerning my observations about Compaq, so whileaJ one can certainly question their accuracy at least they have the virtue ofL being dead-honest (though of course the only evidence you have of that is myK tendency to defend Compaq's behavior when it seems reasonable to me as wella3 as to criticize it when I feel that's appropriate).r   >nK > >>> "I think [VMS is] almost uniquely qualified to service the demands ofAK > e-business," said Capellas, conveniently ignoring the fact that while VMS J > indeed provides in many cases uniquely superior under-pinnings, a lot of the J > super-structure required by e-business is available only on more popularF > platforms, a situation which even with major efforts by Compaq won't changeF > immediately - assuming that such efforts are in fact made at all.<<< > G > Michael Capellas is the CEO of a major multi-platform company, Do youg reallyD > expect him to come out and say "OpenVMS is clearly the solution to	 everyonesoK > problems.." ? Does Gerstner from IBM ever come out and say "MVS (pick any @ > IBM OS here) is clearly the solution for everyones problems.." >s > Nope.s  H I fear you misread my meaning.  I was hardly surprised that Capellas wasE singing the praises of VMS in this particular forum (since my limited4I acquaintance with his pronouncements suggests that he usually tailors hisrG words to his audience), just commenting that his statement seemed a tadeK over-the-top given that in the area of supported applications VMS is almosteL uniquely *un*qualified to service the demands of e-business (compared to its more popular competitors).   >nC > Specific to your concern about applications, Compaq just recentlys
 announced: >i >uK > 1.) the COE initiative which will allow easier porting and development ofwF > applications for those that prefer a UNIX style of programming and a minimumMK > 15 year commitment to ongoing development for OpenVMS. A major investment  by	 > Compaq.o  G We'll see.  If it's essentially a 'check the box' feature (as many haveMJ characterized the earlier POSIX support on VMS), then it simply representsL Compaq's willingness to do the minimum necessary to protect its existing (inI this case, government) revenue stream.  If it's a really well-integrated,sL full-featured Unix-style environment, then it will indeed constitute a major% investment in VMS's future viability.a   > C > 2.) Compaq-Oracle partnership that will see OpenVMS as one of theI	 platformscD > supported for Oracles recently announced major change in their IASA > application strategy.  A major investment by Compaq and Oracle.s  H I was not aware of the Compaq side of this investment (save for whateverF convincing it might have taken to get Oracle to include VMS):  is more detail available?    >YI > 3.) OpenVMS porting library to make it easier to move applications from- UNIX
 > to OpenVMS.-  K I seem to remember a comment that this library, while it might be generally ( useful, was part of the Mozilla(?) port.   > J > 4.) New beta kit of the JServ/JSSI Modules for Apache for OpenVMS Alpha. >eJ > 5.) M16 beta kit of Mozilla for OpenVMS released two days after official > Mozilla M16 release. >,: > 6.) JAVA 2 (Version 1.2.2-beta1) recently released beta. >mH > 7.) Cognos has recently released their web based products for OpenVMS. > Reference: > http://www.cognos.com/adtpci/  >tK > 8.) The recent Customer / ISV references for OpenVMS and the new Alpha GSr	 > Series: 5 > http://www.openvms.digital.com/gsseries/quotes.htmle  K Yup, some things are happening.  They run the gamut from "maybe too little,bG too late - but better late than never" to "hey, that's nice!", but stopw! short of anything truly dramatic.    >pG > There are also ongoing discussions with other (major) companies whichY should( > also become public in the near future. >sD > So, while there is obviously lots of work to do, your statement "- assumingK > that such efforts are in fact made at all.." clearly does not reflect the, > work that is really going on.   A I have to agree:  that off-hand comment was made without adequate 5 reflection, and does not do justice to the situation.s  J However, it remains true that there's no external way to discern whether aL 'VMS renaissance' is truly under way or whether we're just seeing a startledL Compaq realizing that its cash cow is still alive and had better be tossed aK half-bale of hay once in a while so they can keep milking it.  And the only.D real evidence will come if/when we begin to see real money and majorI additional resources devoted to VMS:  aggressive advertising that doesn'teH pull its punches to avoid anything that might suggest the limitations ofF other Compaq offerings ('cause if VMS can't take to the field at leastJ implicitly against Compaq's Windows and Unix offerings, it can't very wellG compete against the identical Windows and near-identical Unix platforms.L offered by the rest of the industry), plus aggressive VMS enhancements aimedC at leading the industry where it pretty clearly wants to go (tough,yJ competent VMS core, friendly Unix/Windows-style exterior wouldn't be a bad first approximation).t   >lK > This newsletters goal is to simply raise the awarenes of some of the goodeL > things happening in the OpenVMS space with both current and new Customers. > H > Of course, if you are looking for something else, then you are free to trashbK > it and/or read between the lines as much as you want, but the Customers IhK > have heard from enjoy reading what others are doing and what is happening  in > the OpenVMS space.  F As I said, the remainder of the newsletter was at worst innocuous, andK occasionally informative.  But the first two articles certainly appeared tomB be passing themselves off as 'industry insight' rather than simply9 interesting gossip - and I commented on them accordingly.0   - bill   >s
 > Regards, >h > Kerry Main > Senior Consultant, > Compaq Canadah > Professional Servicesl > Voice : 613-592-4660 > FAX   : 819-772-7036 > Email : kerry.main@compaq.com0   ------------------------------   Date: 8 Jul 2000 19:43:51 GMT. From: Dave smith (DaveSmith)M Subject: UNICENTER PERFORMANCE DATA COLLECTOR - DOES NOT INSTALL ON VMS 7.2!!s. Message-ID: <39677667_2@nnrp1.news.uk.psi.net>   Hello ,sH       The company I work for bought Unicenter Performance Analyzer/Data E Collectors 2.5 years ago for 20K plus from CA. (Just after CA boughtoH it from Digital - A Big mistake by Digital was to sell off PolyCenter!).B It is running on Alpha VMS 7.1. When I recently tried to install aK data collector on Alpha VMS V7.2 it failed with a 'version mismatch error' iI please relink. I contacted CA who told me that I either need to go to TNG,E or ADVISEIT. To upgrade to ADVISEIT it would cost 40K plus!!!!!. Thet8 TNG route wasn't worth looking at (very expensive!!!!). 3     Is there any way round my installation problem?    regardss Dave Smith     ------------------------------   End of INFO-VAX 2000.380 ************************