1 INFO-VAX	Fri, 01 Sep 2006	Volume 2006 : Issue 479       Contents: Re: Alpha remembrance day  Re: Blue Ray Discs Re: Blue Ray DiscsC Re: Boot failure after AUTOGEN in OpenVMS 8.3 Alpha (Fresh Install)  Creating a bit bucket for mail" Re: Creating a bit bucket for mail" Re: Creating a bit bucket for mail( Re: Distributed Netbeans 5.0 field test?( Re: Distributed Netbeans 5.0 field test?, Re: Mystery multiple TCP connections dropped, Re: Mystery multiple TCP connections droppedA Re: openvms - java - timezone - daylight savings - what the heck! P OT - running IE in Windows (was:Re: Thoughts on the book: DEC is dead, long liveP RE: OT - running IE in Windows (was:Re: Thoughts on the book: DEC is dead, long P Re: OT - running IE in Windows (was:Re: Thoughts on the book: DEC is dead, long P Re: OT - running IE in Windows (was:Re: Thoughts on the book: DEC is dead, long < PERL with OpenVMS & OSU Web Server - Latest Version of PERL?@ Re: PERL with OpenVMS & OSU Web Server - Latest Version of PERL? Problem during boot  Re: Problem during boot % problem with file creation time stamp ) Re: problem with file creation time stamp ) Re: problem with file creation time stamp  Re: Shadow set problem4 Re: Thoughts on the book: DEC is dead, long live DEC4 Re: Thoughts on the book: DEC is dead, long live DEC4 Re: Thoughts on the book: DEC is dead, long live DEC RE: VMS in The DA . Re: We're not dead yet, we're feeling better !. Re: We're not dead yet, we're feeling better !. RE: We're not dead yet, we're feeling better !  F ----------------------------------------------------------------------   Date: 1 Sep 2006 03:23:29 -0700 - From: "Andrew" <andrew_harrison@symantec.com> " Subject: Re: Alpha remembrance dayB Message-ID: <1157106209.460402.229920@74g2000cwt.googlegroups.com>   hoffman@hp.nospam wrote: > L > In article <1157022773.778593.4430@m73g2000cwd.googlegroups.com>, "Andrew=( " <andrew_harrison@symantec.com> writes: > |> > |> Hoff Hoffman wrote: > |> > Andrew wrote: > |> >L > |> > > Does it really. Windows 2003 does we know that but can you  provid= e a 7 > |> > > Microsoft reference for 64GB support in Win2K.  > |> >L > |> >    Here are two such references, to Microsoft Windows 2000 with 64 G=
 B support,L > |> > currently available at the Microsoft web site, with the requisite qu= otations* > |> > from these pages directly provided: > |> >L > |> > http://www.microsoft.com/presspass/press/2001/oct01/10-02mecreevespr= .mspx  > |> >L > |> >    "In particular, Windows 2000 Datacenter Server allows enterprises=  to scale upL > |> > by supporting as many as 32 processors and up to 64 gigabytes of mem= ory."  > |> >L > |> > http://www.microsoft.com/technet/prodtechnol/windows2000serv/evaluat= e/servroi.mspx > |> >L > |> >    "Microsoft has addressed these specific issues in Windows 2000 Da= tacenterL > |> > Server and Windows 2000 Advanced Server with improved Symmetric Mult= iprocessing L > |> > and Enterprise Memory Architecture (EMA). By utilizing Intel's Physi= cal Address L > |> > Extension (PAE) on Intel 32 bit (IA-32) platforms, Windows 2000 Data= center> > |> > Server supports up to 64 gigabytes of physical memory." > |> > > |> >    -- > |> >L > |> >    Your favorite search engine can locate rather more of these.  Man= y more.  > |>J > |> Sure they can but they are all based on Microsoft Press and MarketingH > |> material. The Windows Server Hardware Developers guide on the otherI > |> hand lists Windows 2000 Datacenter Server as supporting 32GB and the J > |> documentation for PAE which is the mechanism used by Intel to support > |> more than 4GB.  > |>I > |> http://www.microsoft.com/whdc/system/platform/server/PAE/PAEdrv.mspx  > |> > F >   I'm discussing VIRTUAL addressing and PHYSICAL addressing, and youD > claimed there was no use for extended VIRTUAL addressing, and then- > discussions of limited PHYSICAL addressing.  >    B No you introduced VIRTUAL addressing. I was talking about PHYSICAL> addressing. Although some of the limits that apply to PHYSICAL9 addressing obviously apply to VIRTUAL addressing as well.   I >   But it matters not, as OpenVMS has found use for more than 32 bits of H > physical address space and more than 32 bits of virtual address space,G > and applications have directly benefited from this -- and starting as H > soon as OpenVMS offered it -- whether the application code was updatedG > to use more than 32 bits of virtual address space -- and however many I > bits of physical addressing are supported by the box is of significance H > only in its need to be "enough" -- there can  be less or more physicalB > address space than virtual space, and (with the exception of the= > operating system, drivers, diagnostics, and a few very rare K > applications) never even consider the physical addressing implementation.  > G >   You said there was no proof, and I pointed to OpenVMS Alpha itself. B > Tru64 UNIX is another consumer of 64-bit virtual, too.  That theG > AlphaServer GS1280 is only the high-end of the series, 64-bit virtual F > is used throughout the line, and applications on the low-end 100 MHzH > EV4 Alpha can and do directly or indirectly use it, so long as they're' > Tru64 UNIX, or OpenVMS V7.0 or later.  >   F Of course 64bit is used throughput the line, Alpha is 64bit only.  TheE question is not if its is available but if it is usefull to consumers  of Alpha systems.   G >   And it's not the 64th bit virtual addressing that's truely all that D > interesting, either, it's the 33rd or the 34th bit that matters to  G Of course most current 64bit systems do not actually allow you to use a F full flat 64bit address space. Why would they when the limittations onG the availability of physical memory fall well short of this. The IBM P5  595 supports 2TB of RAM,  G Occasionally the odd poster to this group chimes in with a x86-64 isn't C a real 64bit CPU because it only supports 40bit of PHYSICAL memory, A EMT64 when it was first released only supported 36bit addressing.   D > most current and existing 64-bit applications.  As for OpenVMS, S2F > space is increasingly heavily used, and specifically to offload dataB > (and more recently code) and free up room for application usage. > I >   And applications indirectly and directly benefit from 64-bit virtual.  >  > 	--  > I >   Here's a reference to 64 GB PHYSICAL addressing on Windows 2000, BTW:  > L > http://msdn.microsoft.com/library/default.asp?url=3D/library/en-us/dnge= nlib/html/awewindata.asp > D >   Apparently Microsoft has various documentation and specificationH > discrepencies in this area, at least for their cited PHYSICAL address.C > VIRTUAL addresing is 32 bits up until specific variant updates to > > Windows XP, and then only on specific IA-32e or AMD64 boxes. >   G I prefer the technical documentation which is clear, you are welcome to   use pres s releases as a source.   Regards  Andrew Harrison   F >   Windows applications haven't particularly made the transition pastH > 32-bit virtual, but other systems have.  Linux and Linux applications,F > for instance, has had 64-bit virtual for quite some time.  On Alpha,: > and on IA-32e, AMD64 and IA-64 boxes, among other boxes. >=20 >     ------------------------------  $ Date: Fri, 1 Sep 2006 09:46:51 +02007 From: "Eberhard Heuser" <vaxinf@chemie.uni-konstanz.de>  Subject: Re: Blue Ray Discs 0 Message-ID: <012001c6cd9a$c9c6bb60$05072286@vg2>   Tom,  8 I've just released DVDwrite for OpenVMS V6.0 that writes9 single Layer Blue-Ray disks (BD-R and BD-RE). You can use > the scsi or the ide connector to read those disks. I've tested9 my program with a Pioneer Blue-Ray burner but others will 
 work, too.    If you wand to know more look at   http://www.dvdwrite.de   Hope this helps  Eberhard ----- Original Message -----  * From: "Tom Linden" <tom@kednos-remove.com> To: <Info-VAX@Mvb.Saic.Com> ' Sent: Thursday, August 31, 2006 9:01 PM  Subject: Blue Ray Discs     4 > Are these going to be available for OpenVMS Alpha? > Tom    ------------------------------  % Date: Fri, 01 Sep 2006 08:26:18 -0400 + From: Steve Matzura <number6@speakeasy.net>  Subject: Re: Blue Ray Discs 8 Message-ID: <sc9gf2195n4imaipspm1ldc4fdbv8obaj5@4ax.com>  0 On Thu, 31 Aug 2006 14:31:22 -0700, "Tom Linden" <tom@kednos-remove.com> wrote:J >I brought it up because i happened to see a News story about TDK new 200GK >disks.  I don't have a specific req't but it would be nice to get rid of    >tapes >for backup.  B 200gb per BD platter?  Gee, they were talking about 50 or 75.  1009 would require four layers.  200 would probably have to be C double-sided.  If these things aren't any cheaper than seventy-five D cents a gig, then they're no bargain except that they're immune fromD magnetics.  But what about minor flexing of the media?  All it wouldC take is a small crack in some under-layer to ruin it, and who knows B what knock-on effect there might be with reflections through otherF layers on the same surface.  "Handle with care" will in all likelihood; be the order of the day.  Permanent cartridges would not be > unappealinng.  They'd keep dirt out, keep the media from being@ accidentally bent, might fix a whole host of potential problems.   ------------------------------   Date: 1 Sep 2006 07:04:16 -0700 / From: "Volker Halle" <volker_halle@hotmail.com> L Subject: Re: Boot failure after AUTOGEN in OpenVMS 8.3 Alpha (Fresh Install)C Message-ID: <1157119456.778372.104460@i42g2000cwa.googlegroups.com>   @ This issue has been discussed in the OpenVMS ITRC forum over at:  L http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=1056315  < The high value of MIN_MAXPROCESSCNT=32000 seems to have been instrumental in this problem.   D Note that a symbol error in OpenVMS Alpha V8.3 AUTOGEN has also been discovered:   8 Workaround: change references from WSMAX_MAX to WSMAX_MX   --- , Volker Halle, Invenate GmbH, OpenVMS Support  # An OpenVMS crashdump analysis a day # makes the Windows headaches go away    ------------------------------  % Date: Fri, 01 Sep 2006 03:19:26 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> ' Subject: Creating a bit bucket for mail , Message-ID: <44F7DEE5.43EFF7B4@teksavvy.com>  D Seems I have caught the same disease as Mr SMS. My system is gettingG lots of non delivery messages for messages apparently sent by my system E from non-existant username on my machine. Some of those usernames are > "complex" such as BELLGLOBAL.COM.POSTED@chocolate.com (replace chocolate.com with my domain).  E For simple usernames, I can just create an empty distribution list in D the TCPIP$SMTP_COMMON directory. But more complex username require a different tactic.   H I was thinking about creating a "Bitbucket" account whose default deviceC is NLA0: and default directory is "[EMPTY]" and then using MAIL SET G FORWARD for those accounts to point to BITBUCKET. However, MAIL refuses < to accept delivery to a user whose SYS$LOGIN is NLA0:[EMPTY]    F Are there any suggestion on how to create some username where any mailB received to it will simply be ignored ?  Perhaps writing a foreignG transport that does nothing except acknowledge receipt of message ? Any 
 other ideas ?   E In recent weeks, I have been getting lots of such messages which deal D with a rather small set of usernames, so creating default forwardingD entries of that small set of usernames would solve a large number of problem messages.    ------------------------------   Date: 1 Sep 2006 00:24:57 -0700 / From: "David B Sneddon" <dbsneddon@bigpond.com> + Subject: Re: Creating a bit bucket for mail A Message-ID: <1157095496.959859.8850@i42g2000cwa.googlegroups.com>    JF Mezei wrote: F > Seems I have caught the same disease as Mr SMS. My system is gettingI > lots of non delivery messages for messages apparently sent by my system G > from non-existant username on my machine. Some of those usernames are @ > "complex" such as BELLGLOBAL.COM.POSTED@chocolate.com (replace  > chocolate.com with my domain). > G > For simple usernames, I can just create an empty distribution list in F > the TCPIP$SMTP_COMMON directory. But more complex username require a > different tactic.  > J > I was thinking about creating a "Bitbucket" account whose default deviceE > is NLA0: and default directory is "[EMPTY]" and then using MAIL SET I > FORWARD for those accounts to point to BITBUCKET. However, MAIL refuses > > to accept delivery to a user whose SYS$LOGIN is NLA0:[EMPTY] >  > H > Are there any suggestion on how to create some username where any mailD > received to it will simply be ignored ?  Perhaps writing a foreignI > transport that does nothing except acknowledge receipt of message ? Any  > other ideas ?  > G > In recent weeks, I have been getting lots of such messages which deal F > with a rather small set of usernames, so creating default forwardingF > entries of that small set of usernames would solve a large number of > problem messages.   8 Try creating user BITBUCKET with a real device/directory/ and use the deliver package to do what you want    Dave   ------------------------------  * Date: Fri, 1 Sep 2006 15:09:47 +0000 (UTC)< From: gartmann@nonsense.immunbio.mpg.de (Christoph Gartmann)+ Subject: Re: Creating a bit bucket for mail ) Message-ID: <ed9ifr$bkh$1@news.BelWue.DE>   \ In article <44F7DEE5.43EFF7B4@teksavvy.com>, JF Mezei <jfmezei.spamnot@teksavvy.com> writes:E >Seems I have caught the same disease as Mr SMS. My system is getting H >lots of non delivery messages for messages apparently sent by my systemF >from non-existant username on my machine. Some of those usernames are? >"complex" such as BELLGLOBAL.COM.POSTED@chocolate.com (replace  >chocolate.com with my domain).  > F >For simple usernames, I can just create an empty distribution list inE >the TCPIP$SMTP_COMMON directory. But more complex username require a  >different tactic. > I >I was thinking about creating a "Bitbucket" account whose default device D >is NLA0: and default directory is "[EMPTY]" and then using MAIL SETH >FORWARD for those accounts to point to BITBUCKET. However, MAIL refuses= >to accept delivery to a user whose SYS$LOGIN is NLA0:[EMPTY]  >  > G >Are there any suggestion on how to create some username where any mail C >received to it will simply be ignored ?  Perhaps writing a foreign H >transport that does nothing except acknowledge receipt of message ? Any >other ideas ? > F >In recent weeks, I have been getting lots of such messages which dealE >with a rather small set of usernames, so creating default forwarding E >entries of that small set of usernames would solve a large number of  >problem messages.  J What about creating a real username with an existing device and directory?H Then create a self-submitting batch job that deletes the mail at regular
 intervals.   Regards,    Christoph Gartmann     --  E  Max-Planck-Institut fuer      Phone   : +49-761-5108-464   Fax: -452   ImmunbiologieI  Postfach 1169                 Internet: gartmann@immunbio dot mpg dot de   D-79011  Freiburg, Germany 9                http://www.immunbio.mpg.de/home/menue.html    ------------------------------   Date: 1 Sep 2006 07:47:20 -0700 ) From: "Sue" <susan_skonetski@hotmail.com> 1 Subject: Re: Distributed Netbeans 5.0 field test? B Message-ID: <1157122040.215312.36800@i42g2000cwa.googlegroups.com>    Just came out like 2 minutes ago  > OpenVMS Engineering is pleased to announce the availability ofC Distributed NetBeans Version T5.0 FT1 for OpenVMS Alpha and OpenVMS  I64.  > Distributed NetBeans allows you to run the NetBeans integrated? development environment (IDE) on your desktop system to develop + applications on a remote HP OpenVMS system.   B This latest version supports NetBeans 5.0 and NetBeans 5.5 Beta 2.  1 Features of Distributed NetBeans T5.0 FT1 include & *	Support for Integrity Server Systems   *	FTP FilesystemF        Distributed NetBeans uses FTP to access your remote filesystem.  4 *	C/C++, Cobol, Fortran, Basic, and Pascal LanguagesB        Distributed NetBeans contains support for the C/C++, Cobol,$ Fortran, Basic, and Pascal languages;        in addition to  DCL support, and EDT keypad support.    *	MMS and BASHF        MMS and BASH shell script files are recognized as text files in Distributed NetBeans. You can G        create, edit, and remotely execute MMS make files and BASH shell  scripts from within the IDE.   *	Server Setup CommandD        The server setup command property allows you to specify a DCL  command string to be executed byA        the remote IDE User server during its startup. The command  specified is passed to DCL forD        execution. You can specify a different setup command for each remote server you create.   E The NetBeans IDE, Distributed NetBeans client, and the IDE Server are ? provided free of charge.  Support is provided with your OpenVMS A operating system support contract and by the Distributed NetBeans  development team.   E For more information or to download the Distributed NetBeans T5.0 FT1  kit, please visit   C http://h71000.www7.hp.com/openvms/products/ips/netbeans/distnb.html     
 Best Regards, ) The Distributed NetBeans Engineering Team      sean@obanion.us wrote:F > Back in July, Meg Watson of Distributed NetBeans Engineering stated:8 > "We are about to release Distributed NetBeans T5.0..."> > I had asked to be in the field test, but have heard nothing.F > I also have a client that is looking forward to the Cobol support in > 5.0. > G > Has anybody gotten the field test?  How's it going (that you can talk 	 > about)?  >  >  >  > Sean   ------------------------------   Date: 1 Sep 2006 09:31:12 -0700  From: sean@obanion.us 1 Subject: Re: Distributed Netbeans 5.0 field test? B Message-ID: <1157128272.886287.97140@p79g2000cwp.googlegroups.com>  ; I'd like to claim that means I'm JUST ahead of the curve...   = I also got the email notice last night, so I think I'm on the  distribution list.# I'll probably work on it next week.    Thanks!      Sean  
 Sue wrote:" > Just came out like 2 minutes ago > @ > OpenVMS Engineering is pleased to announce the availability ofE > Distributed NetBeans Version T5.0 FT1 for OpenVMS Alpha and OpenVMS  > I64. > @ > Distributed NetBeans allows you to run the NetBeans integratedA > development environment (IDE) on your desktop system to develop - > applications on a remote HP OpenVMS system.  > D > This latest version supports NetBeans 5.0 and NetBeans 5.5 Beta 2. > 3 > Features of Distributed NetBeans T5.0 FT1 include ( > *	Support for Integrity Server Systems >  > *	FTP FilesystemH >        Distributed NetBeans uses FTP to access your remote filesystem. > 6 > *	C/C++, Cobol, Fortran, Basic, and Pascal LanguagesD >        Distributed NetBeans contains support for the C/C++, Cobol,& > Fortran, Basic, and Pascal languages= >        in addition to  DCL support, and EDT keypad support.  >  > *	MMS and BASHH >        MMS and BASH shell script files are recognized as text files in > Distributed NetBeans. You can I >        create, edit, and remotely execute MMS make files and BASH shell  > scripts from within the IDE. >  > *	Server Setup CommandF >        The server setup command property allows you to specify a DCL" > command string to be executed byC >        the remote IDE User server during its startup. The command   > specified is passed to DCL forF >        execution. You can specify a different setup command for each > remote server you create.  > G > The NetBeans IDE, Distributed NetBeans client, and the IDE Server are A > provided free of charge.  Support is provided with your OpenVMS C > operating system support contract and by the Distributed NetBeans  > development team.  > G > For more information or to download the Distributed NetBeans T5.0 FT1  > kit, please visit  > E > http://h71000.www7.hp.com/openvms/products/ips/netbeans/distnb.html  >  >  > Best Regards, + > The Distributed NetBeans Engineering Team  >  >  > sean@obanion.us wrote:H > > Back in July, Meg Watson of Distributed NetBeans Engineering stated:: > > "We are about to release Distributed NetBeans T5.0..."@ > > I had asked to be in the field test, but have heard nothing.H > > I also have a client that is looking forward to the Cobol support in > > 5.0. > > I > > Has anybody gotten the field test?  How's it going (that you can talk  > > about)?  > >  > >  > >  > > Sean   ------------------------------   Date: 1 Sep 2006 07:09:35 -0700 $ From: "AEF" <spamsink2001@yahoo.com>5 Subject: Re: Mystery multiple TCP connections dropped C Message-ID: <1157119775.095336.249730@p79g2000cwp.googlegroups.com>    JF Mezei wrote:  > AEF wrote:L > > > This appears to be done in the Pascal code so I'll report back after I= > > > talk to our developer who is far more familiar with it.  > > J > > He says he just gives it the name of the driver and uses some standard	 > > call.  > G > That is very vague, especialy since "standard calls" can be in C (the 8 > socket routines), or the VMS specific $QIO interfaces.  3 Yes, it's vague. That's all I could get out of him.   % The code is written in Pascal, not C.    > D > Does it have any code to handle an accidental disconnection ? (ak,  E The code (VAX side) used to be able to handle only one connection. If ? that connection died, the program would exit. Now, it takes all " connections to die to do the same.  F > something akin to a mailbox or some AST that gets triggered when theI > link is abnormally disconnected ? Also, is this at the server or client I > side (aka: is that the node that receives calls from the other node, or 5 > the node which initiates a call to a remote node) ?   7 The MicroVAX initiates the connection to the multifeed.    > J > And does the other node at the other end detect the disconnection at the > same time as this node ?  E Same times as best as I can tell. The Multifeeds (Windows boxes) were E not on any time service. I just checked the times now and the primary B multifeed is 1m46s slow! The connections dropped on the multifeeds? between 2 and 8 seconds after they did on the MicroVAX systems. G Allowing for time drift of the Windows servers' clocks between then and @ now could account for that I think! (My MicroVAX systems are allB running NTP and synched to an Internet time server.) So within allD possible errors I'd say everything happened within the same 1-second period or less.   I > Also, you need to know if any router in between acts as NAT/PAT router. D > Such routers can get messed up in certain circumstances and drop a > virtual link between 2 nodes.    Our network guy says no.   Thanks again for your efforts.   AEF    ------------------------------   Date: 1 Sep 2006 07:13:33 -0700 $ From: "AEF" <spamsink2001@yahoo.com>5 Subject: Re: Mystery multiple TCP connections dropped A Message-ID: <1157120013.258615.69400@i3g2000cwc.googlegroups.com>    Doug Phillips wrote: > AEF wrote: > > JF Mezei wrote:  > > > AEF wrote:G > > > > Being that this happened on two MicroVAX systems, three Windows I > > > > multifeed systems, with all six connections between them dropping H > > > > within a tenth of a second of each other, I can't see any of theN > > > > systems being at fault. It must be some external entity. But why these( > > > > two pages and not the other two? > > >  > > > M > > > If pages 1 and 2 are sent at 08:00 and pages 3 and 4 are sent at 08:01, M > > > and a problem happens between 07:59 and 08:00:58, then failure would be N > > > noticed for pages 1 and 2, but by the time the systems try to send pagesE > > > 3 and 4, the link si re-established and those attempts succeed.  > > C > > But this all happened withing the same 0.1-second period in two G > > separate incidents! (+/- any NTP-error time differences between the 7 > > system times of the two involved MicroVAX systems).  > > I > > > From the application point of view, you need to look at how the app L > > > reacts to essentially the ethernet plug being pulled from the machine. > > C > > The same as when the problem happened, though there is a 5 min. H > > time-out between when the plug is pulled and when the error happens.F > > The 0.1-second time windows occuring on two separate days makes me' > > doubt that a time-out was involved.  > > N > > > (equivalent of a switch going down etc etc). Does the app rely on TNxxx:C > > > devices or are they simple GB devices ? How is the TCPIP link > > > > established by the application ? What options are used ? > > J > > This appears to be done in the Pascal code so I'll report back after I; > > talk to our developer who is far more familiar with it.  > > N > > > You may wish to look at your TCPIP stack's documemtation to see if thereN > > > are logicals you xcan set to lengthen any timeout values for the type ofL > > > connexction you are using, which would allow your apps to survive some& > > > sporadic switch/router problems. > > G > > Unless both pages were sending time-stamps withing 0.1 sec. of each I > > other, it looks like a bona fide close connection command occurred at I > > the multifeed end, except how can three separate boxes conspire to do J > > that to the same two ports withing 0.1 sec. of each other? The timeout. > > is already 5 min. for a broken connection. > >  > E > Maybe a restatement of the problem would help. Is this about right?  > 4 >  - There are four MicroVax systems (MV) in London.G >  - There are four MultiFeed systems (MF); one in London, three in NY. : >  - Each MV has a TELNET connection open to all four MFs.B >  - On Tuesday morning two of the MVs lost communication with NY.+ >  - All other connections remained intact. 3 >  - Restarting the application restored the links. H >  - On Thursday morning this happened again, within an hour of the same > time as on Tuesday.   E This is correct. Also, the page jobs run on the MicroVAX systems. The E multifeed program (which is manually configured to accept connections 2 on various TCP ports) runs on the Windows systems.   >  > If so: > < > First: Why did only those two MVs lose connection with NY?  
 I don't know.    > Second: What caused the loss?   
 I don't know.    > G > Maybe you'll never know the second, but if you identify the first you 4 > can take steps to prevent it from happening again. > I > As I suggested before; these two systems probably share something along @ > the link between their London location and NY that you haven'tH > identified. Time to pull out (or create) your infrastructure diagrams,I > validate all the boxes, trace the lines and lightning bolts and explore 
 > any clouds.   ) I'll have to check with London personnel.    Thanks for your help.    AEF  > C > Increasing some timeouts might just mask the problem for a while.    ------------------------------   Date: 1 Sep 2006 08:45:11 -0700  From: n.rieck@sympatico.caJ Subject: Re: openvms - java - timezone - daylight savings - what the heck!B Message-ID: <1157125511.768972.274740@i3g2000cwc.googlegroups.com>  $ Troy, here is something else to try.  . $ @SYS$COMMON:[JAVA$142.COM]JAVA$142_SETUP.COM $ java  -fullversion java full version "1.4.2-4.p2" $   @ I thought I was running the latest version of JAVA142 but wasn'tG becuase a dash 5 version is now available. (BTW - it mentions something C about fixing an Austrailian Standard/Daylight problem. This doesn't C affect you on the Pacific coast but could if your minor version was  much lower than mine)   
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.8 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html   ------------------------------  % Date: Fri, 01 Sep 2006 06:42:53 -0400 - From: bradhamilton <bradhamilton@comcast.net> Y Subject: OT - running IE in Windows (was:Re: Thoughts on the book: DEC is dead, long live * Message-ID: <44F80EAD.3000309@comcast.net>   Main, Kerry wrote: >> -----Original Message----- 6 >> From: Doug Phillips [mailto:dphill46@netscape.net]    [...]   7 >> You apparently haven't actually looked at the "5-20   >> *security* patches"> >> that come out every month for Windows. Most are for desktop> >> applications, and most of the ones that do apply to server  >> are specific ( >> to certain applications or functions. >  > Like IE?   > G > Ever try removing IE from any Windows system - workstation or server?  > * > It's a great exercise - try it sometime. >  > :-)   ) "Strawman" - it *is* easy to "remove" IE:   . - Run any number of other browsers besides IE.  F - It's easy to remove all IE shortcuts and icons from the desktop and H menu bars, and to change browser preferences, such that the only way to G run IE is to hunt it down and deliberately let it kill your system.	:-)    [...]    ------------------------------  $ Date: Fri, 1 Sep 2006 09:18:28 -0400' From: "Main, Kerry" <Kerry.Main@hp.com> Y Subject: RE: OT - running IE in Windows (was:Re: Thoughts on the book: DEC is dead, long  T Message-ID: <FA60F2C4B72A584DBFC6091F6A2B8684019BC809@tayexc19.americas.cpqcorp.net>   > -----Original Message-----9 > From: bradhamilton [mailto:bradhamilton@comcast.net]=20 ! > Sent: September 1, 2006 6:43 AM  > To: Info-VAX@Mvb.Saic.Com A > Subject: OT - running IE in Windows (was:Re: Thoughts on the=20 # > book: DEC is dead, long live DEC)  >=20 > Main, Kerry wrote: > >> -----Original Message----- : > >> From: Doug Phillips [mailto:dphill46@netscape.net]=20 >=20 > [...]  >=20; > >> You apparently haven't actually looked at the "5-20=20  > >> *security* patches"@ > >> that come out every month for Windows. Most are for desktopB > >> applications, and most of the ones that do apply to server=20 > >> are specific * > >> to certain applications or functions. > >=20 > > Like IE?=20  > >=20A > > Ever try removing IE from any Windows system - workstation=20  > or server? > >=20, > > It's a great exercise - try it sometime. > >=20 > > :-)  >=20+ > "Strawman" - it *is* easy to "remove" IE:  >=200 > - Run any number of other browsers besides IE. >=20J > - It's easy to remove all IE shortcuts and icons from the desktop and=20@ > menu bars, and to change browser preferences, such that the=20 > only way to=20@ > run IE is to hunt it down and deliberately let it kill your=20
 > system.	:-)  >=20 > [...]  >=20  D You are talking about ways to "hide" I.E. Actually removing from the< system is akin to a brain transplant in terms of complexity.  2 Btw, error message of the day .. Could not resist-  E "The system has detected the mouse has been moved. The system must be - rebooted for this operation to be completed."    :-)   
 Kerry Main Senior Consultant  HP Services Canada Voice: 613-592-4660  Fax: 613-591-4477  kerryDOTmainAThpDOTcom (remove the DOT's and AT)=20  4 OpenVMS - the secure, multi-site OS that just works.   ------------------------------  % Date: Fri, 01 Sep 2006 19:18:06 +0200 / From: Paul Sture <paul.sture.nospam@hispeed.ch> Y Subject: Re: OT - running IE in Windows (was:Re: Thoughts on the book: DEC is dead, long  J Message-ID: <paul.sture.nospam-60BDC4.19180601092006@mac.sture.homeip.net>  G In article <tLGdne4TGp1Sw2XZnZ2dnUVZ_vqdnZ2d@metrocastcablevision.com>, *  Bill Todd <billtodd@metrocast.net> wrote:  # > http://www.litepc.com/xplite.html    Thanks for that link Bill.   --  
 Paul Sture   ------------------------------  % Date: Fri, 01 Sep 2006 12:24:46 -0400 ( From: Bill Todd <billtodd@metrocast.net>Y Subject: Re: OT - running IE in Windows (was:Re: Thoughts on the book: DEC is dead, long  G Message-ID: <tLGdne4TGp1Sw2XZnZ2dnUVZ_vqdnZ2d@metrocastcablevision.com>    Main, Kerry wrote: >> -----Original Message----- 8 >> From: bradhamilton [mailto:bradhamilton@comcast.net] " >> Sent: September 1, 2006 6:43 AM >> To: Info-VAX@Mvb.Saic.Com@ >> Subject: OT - running IE in Windows (was:Re: Thoughts on the $ >> book: DEC is dead, long live DEC) >> >> Main, Kerry wrote:  >>>> -----Original Message----- 8 >>>> From: Doug Phillips [mailto:dphill46@netscape.net]  >> [...] >>9 >>>> You apparently haven't actually looked at the "5-20   >>>> *security* patches"@ >>>> that come out every month for Windows. Most are for desktop@ >>>> applications, and most of the ones that do apply to server  >>>> are specific * >>>> to certain applications or functions.
 >>> Like IE?   >>> ? >>> Ever try removing IE from any Windows system - workstation  
 >> or server? , >>> It's a great exercise - try it sometime. >>>  >>> :-) , >> "Strawman" - it *is* easy to "remove" IE: >>1 >> - Run any number of other browsers besides IE.  >>I >> - It's easy to remove all IE shortcuts and icons from the desktop and  ? >> menu bars, and to change browser preferences, such that the   >> only way to  ? >> run IE is to hunt it down and deliberately let it kill your   >> system.	:-) >> >> [...] >> > F > You are talking about ways to "hide" I.E. Actually removing from the> > system is akin to a brain transplant in terms of complexity.  ! http://www.litepc.com/xplite.html    - bill   ------------------------------  % Date: Fri, 01 Sep 2006 11:20:43 -0400 + From: Chuck Aaron <caaron@ceris.purdue.edu> E Subject: PERL with OpenVMS & OSU Web Server - Latest Version of PERL? 4 Message-ID: <ed9j4b$pq$1@mailhub227.itcs.purdue.edu>  @ What is the latest of version of PERL that runs with openvms 8.2 and OSU WebServer 3.10a?   Thanks,  Chuck    ------------------------------  % Date: Fri, 01 Sep 2006 18:01:38 +0200 / From: Paul Sture <paul.sture.nospam@hispeed.ch> I Subject: Re: PERL with OpenVMS & OSU Web Server - Latest Version of PERL? J Message-ID: <paul.sture.nospam-DE8600.18013801092006@mac.sture.homeip.net>  4 In article <ed9j4b$pq$1@mailhub227.itcs.purdue.edu>,-  Chuck Aaron <caaron@ceris.purdue.edu> wrote:   B > What is the latest of version of PERL that runs with openvms 8.2 > and OSU WebServer 3.10a? > 	 > Thanks,  > Chuck   I Sorry, dunno about OSU, but for OVMS V8.2, the latest version of Perl is   5.8.6   G http://h71000.www7.hp.com/openvms/products/ips/apache/csws_modperl.html   H Please note - the CSWS_PERL package there is as add-on to standard Perl H to give MOD_PERL functionality to CSWS (Apache). You should ask the OSU * folks how this fits in with OSU WebServer.   --  
 Paul Sture   ------------------------------   Date: 1 Sep 2006 01:41:38 -0700 ! From: "alok" <alok.net@gmail.com>  Subject: Problem during bootA Message-ID: <1157100096.661947.48890@74g2000cwt.googlegroups.com>    Hi  B I have a VAX system that has been running fine for long years. But today it has got a problem during boot. it shows   @ % delete - W- file not deleted. It was indicating that the file:C SYS$SYSROOT:[SYS$STARTUP]ESS$LAST_STARTUP.TMP could not be deleted.    Also  G DCL-E_OPENIN, error opening SYS$SYSROOT : [SYSMGR]ESS$LAST_STARTUP.LOG;  as input   2 -RMS-E-FLK, file currently locked by another user.  C I noticed that the protection of the file: ESS$LAST_STARTUP.TMP was C (RWED,RWED,RE,RE).   I changed it to (RWED,RWED,RWED,RWED).  On the  next reboot, the Vax normally.  , How can the protaction of file be changed  ?   ------------------------------   Date: 1 Sep 2006 09:11:20 -0700  From: bill@wcschmidt.com  Subject: Re: Problem during bootC Message-ID: <1157127080.095763.299220@p79g2000cwp.googlegroups.com>   B The foolowing are not protection issues, as a open file can not beG deleted. These files are created by the system boot process, either you C can ignore these messages or you can review the boot procedures and > re-write them to delete the files later in the startup process       alok wrote:  > Hi > D > I have a VAX system that has been running fine for long years. But > today it has got a problem > during boot. it shows  > B > % delete - W- file not deleted. It was indicating that the file:E > SYS$SYSROOT:[SYS$STARTUP]ESS$LAST_STARTUP.TMP could not be deleted.  >  > Also > I > DCL-E_OPENIN, error opening SYS$SYSROOT : [SYSMGR]ESS$LAST_STARTUP.LOG;  > as > input  > 4 > -RMS-E-FLK, file currently locked by another user. > E > I noticed that the protection of the file: ESS$LAST_STARTUP.TMP was E > (RWED,RWED,RE,RE).   I changed it to (RWED,RWED,RWED,RWED).  On the   > next reboot, the Vax normally. > . > How can the protaction of file be changed  ?   ------------------------------   Date: 1 Sep 2006 06:44:08 -0700 ! From: "alok" <alok.net@gmail.com> . Subject: problem with file creation time stampC Message-ID: <1157118248.603656.257250@h48g2000cwc.googlegroups.com>    Hi  C is it possible in VAX VMS that only the file creation time stamp is  changed but not the  content of file.  C Suppose a Log file is created on year 1995 by a command file. Since D then command file loggs some messages in that log file . but today iG noticed that the File creation time stamp has changed to today's date & 5 still the content of file is there . Is it possible ?    ------------------------------   Date: 1 Sep 2006 10:38:13 -0500 - From: Kilgallen@SpamCop.net (Larry Kilgallen) 2 Subject: Re: problem with file creation time stamp3 Message-ID: <v3wCe4EwnEZW@eisner.encompasserve.org>   g In article <1157118248.603656.257250@h48g2000cwc.googlegroups.com>, "alok" <alok.net@gmail.com> writes:  > Hi > E > is it possible in VAX VMS that only the file creation time stamp is  > changed but not the  > content of file.  9 It is possible to recreate a file with the same contents.   @ It is possible to change the creation time stamp with a program.   ------------------------------   Date: 1 Sep 2006 09:18:46 -0700  From: bill@wcschmidt.com2 Subject: Re: problem with file creation time stampC Message-ID: <1157127526.059101.153600@h48g2000cwc.googlegroups.com>   D Is it possible for you to be more specific on what log files you areF questioning ? After a reboot the system may generate new log files forE system processes and the content can be the same,  did you also check & for previous versions of these files ?     alok wrote:  > Hi > E > is it possible in VAX VMS that only the file creation time stamp is  > changed but not the  > content of file. > E > Suppose a Log file is created on year 1995 by a command file. Since F > then command file loggs some messages in that log file . but today iI > noticed that the File creation time stamp has changed to today's date & 7 > still the content of file is there . Is it possible ?    ------------------------------  # Date: Fri, 01 Sep 2006 16:23:58 GMT 1 From: Keith Parris <keithparris_NOSPAM@yahoo.com>  Subject: Re: Shadow set problem - Message-ID: <y8ZJg.83$rZ.18@news.cpqcorp.net>   / Phillip Helbig---remove CLOTHES to reply wrote: F > In article <ecudi7$jue$3@online.de>, helbig@astro.multiCLOTHESvax.de5 > (Phillip Helbig---remove CLOTHES to reply) writes:  I >>Personally, I specify all members at startup with neither of the above  J >>commands.  I could be bitten by the following situation (each member of ; >>the shadow set has a direct connection to only one node):  >># >>A goes down, B is up-to-date copy 
 >>B goes down . >>A comes up, doesn't see B, A becomes currentI >>B comes up, its disk gets overwritten by a full copy from A, losing the 6 >>  data written between A going down and B going down >>I >>Having said that, this has never happened to me.  Obviously, the above  H >>scenario would occur only in the case of severe problems, and in that I >>case I am going to have an ugly transition anyway, so the data loss is  K >>probably not that big of an issue.  With systems set to HALT at startup,  D >>obviously the above cannot happen unless one makes a mistake when G >>booting by hand.  With systems set to BOOT or RESTART on startup, it  G >>could, but it means that B would have to go down during A's reboot.    >>Very unlikely. > F > I should have mentioned that my hobbyist cluster is immune to such aH > scenario under normal circumstances, such as multiple power outages atI > short intervals.  In the startup sequence, each machine waits until all J > the nodes it expects are there (by periodically checking logicals in theE > cluster table) or until after a long-enough timeout before mounting < > disks with members directly connected to different nodes.  > A > I have a non-system disk with SYSUAF etc on it so I mount it in E > SYLOGICALS.COM.  What used to happen is that after a cluster reboot G > (power outage, say), one node would come up quick enough to mount it  K > before all the members were MSCP served to all nodes.  Later on, another  J > node would mount it and a full shadow copy would result.  I implemented F > the checks to avoid this, but they avoid the above scenario as well.  I Some new qualifers have been added to the MOUNT command for Shadowing in  I recent years to address many of the issues you're trying to deal with in   your custom startup procedures:   I The /POLICY=REQUIRE_MEMBERS qualifier prevents accidentally mounting the  D shadowset before all the members are accessible/visible and thereby C losing some of your redundancy and requiring a full-copy operation  @ afterward to bring the tardy member back in. If all the members D specified in the list in your MOUNT command aren't there, the MOUNT 1 fails and you can detect that and retry it later.   ? The /POLICTY=VERIFY_LABEL qualifier prevents ever accidentally  H overwriting a disk, particularly the dangerous case where you overwrite C a disk with newer information with the data form a disk with older  H information (a so-called "wrong-way" shadow copy). If this qualifier is F present, Shadowing will check and ensure that the volume label on the E target disk is SCRATCH_DISK or it will not overwrite the disk with a  E full-copy operation. This implies the system manager must explicitly  F (manually) INITIALIZE the disk with a label of SCRATCH_DISK before it % can ever be overwritten by Shadowing.    ------------------------------   Date: 1 Sep 2006 03:59:49 -0700 - From: "Andrew" <andrew_harrison@symantec.com> = Subject: Re: Thoughts on the book: DEC is dead, long live DEC B Message-ID: <1157108389.707216.16840@m73g2000cwd.googlegroups.com>   Doug Phillips wrote: > Andrew wrote:  > > Bill Todd wrote: > > > JF Mezei wrote: N > > > > Consider as that recently as the introduction of DS15/25 machines, theM > > > > argument was made by DECies that they cannot compete because they are 3 > > > > costlier to build that equivalent machines.  > > > B > > > Cannot compete with *what*?  They certainly couldn't competeM > > > head-to-head with machines using x86 processors that *sold* for a small M > > > fraction of what Alpha *cost* to build, and they were never designed to J > > > (Alpha having long since given up that fight).  But they were pricedF > > > reasonably for the people who needed something that x86 couldn'tI > > > provide, and engineering likely sweated bullets to accomplish that.  > >  > > Not quite. > > J > > The DS15 with 1x 1GHz Alpha  with 1GB of RAM and 1 internal drive with > > Tru64 costs $10900 > >  > E > For VMS base, make that closer to $12,000. Now, to bring it in line G > with what comes with Tru64 we need to add unlimited users for $12,000 G > more and NAS-300 for $16,643. Or maybe we only need 32 users for only D > $6,739 and NAS-250 for only $12,231 additional. No doubt we forgotH > something, like the fact that we need to exist in a Microsoft network,I > so better allocate a few thousand more. Anyway, consider a pretty basic % > DS15 will likely cost over $30,000.  > G > Can you get it cheaper? Sure. But not enough to make anyone choose to 1 > move away from another smaller system platform.   G Quite, anyone suggesting that the DS15 and DS25 are or were competitive F compared with other non commodity low end servers has clearly not beenC awake for the last 3 years either with or without OpenVMS licenses,  ditto for the DS10/DS20.  F The DS25 with 2 CPU's (the max) and 8GB of RAM, 1 disk and OpenVMS 8.2G costs $77K.  A Sun V440 with 4 x CPU's and 8GB of RAM and Solaris costs 9 $23K,  an IBM P5 550 in a similar config also costs $23K.      > I > IA64 licensing is finally in line with reality; hope it's not too late.  >   F The HP rx series running OpenVMS are considerably cheaper for the sameD number of CPU's/RAM when compared with the DS range running OpenVMS,B they are also faster since the EV7/EV7+ never made it into the DS.   regards  Andrew Harrison  > I > > The IBM P5 505 1 x 1.65 GHz POWER 5, 1GB of RAM and 2 internal drives  > > and AIX costs $3500. > > H > > The Sun V210 1 x 1.34 GHz SPARC 1 GB or RAM and 1 internal drive and > > Solaris costs $3400. > > J > > The HP rx1620 with 1.3 GHz Itanium II and 1 GB of RAM, 1 internal disk > > and HP-UX costs $6300  > > K > > The Sun, HP rx1620 and IBM p5 505 are all faster than the DS on SPECint  > > and SPECfp.  > > H > > The DS is the only single CPU system all the others support 2 CPU's.I > > The DS supports the smallest amount or RAM 4GB and while it has 4 PCI E > > slots more than any of the other units it has a poorer basic spec 5 > > (ethernet etc). It is 3RU and the others are 1RU.  > > K > > The 3RU vs 1RU is a pretty serious issue which would rule the DS out of K > > many HPC bids simply on the space it would require compared with POWER. " > > Itanium  and x86 alternatives. > > F > > For the price of 1 DS you could buy 3x IBM or SPARC boxes and they= > > would deliver >3x the throughput in the same form factor.  > >  > I > Or, if you were running OpenVMS on, say an AS1000, and were looking for G > an upgrade to the DS15 level, the 6 to 8 times price difference could A > justify the cost of moving to one of the non-OpenVMS platforms.  > I > Not everyone can move to IA64 now because some software isn't yet there I > (or at least, not ready for prime-time) or might never be. Much of that D > software is now and has been available for the aforementioned *nix= > platforms. Pretty obvious stuff, except to DECHPQ, I guess.    ------------------------------  % Date: Fri, 01 Sep 2006 12:13:06 -0400 ( From: Bill Todd <billtodd@metrocast.net>= Subject: Re: Thoughts on the book: DEC is dead, long live DEC G Message-ID: <-sydnQ962s2OwWXZnZ2dnUVZ_qGdnZ2d@metrocastcablevision.com>    Doug Phillips wrote:   ...   I >> The DS15 with 1x 1GHz Alpha  with 1GB of RAM and 1 internal drive with  >> Tru64 costs $10900  >> > , > For VMS base, make that closer to $12,000.  H Say, what?  Last I knew, the least expensive DS10s cost only about half D that - and the DS15 doesn't seem to be much more than a DS10 with a H minor face lift (still EV6x-based, at any rate, so there's no "But EV7s > cost *so much* more to produce!" blather to have to discount).  , Can you say 'cash cow'?  I knew you could...   - bill   ------------------------------   Date: 1 Sep 2006 10:20:41 -0700 - From: "Andrew" <andrew_harrison@symantec.com> = Subject: Re: Thoughts on the book: DEC is dead, long live DEC B Message-ID: <1157131241.050990.79840@m73g2000cwd.googlegroups.com>   Bill Todd wrote: > Doug Phillips wrote: >  > ...  > K > >> The DS15 with 1x 1GHz Alpha  with 1GB of RAM and 1 internal drive with  > >> Tru64 costs $10900  > >> > > . > > For VMS base, make that closer to $12,000. > I > Say, what?  Last I knew, the least expensive DS10s cost only about half E > that - and the DS15 doesn't seem to be much more than a DS10 with a I > minor face lift (still EV6x-based, at any rate, so there's no "But EV7s @ > cost *so much* more to produce!" blather to have to discount).  F The DS10 started at about $3.5K but that was for a system with 64MB of0 RAM and Linux.  The OpenVMS version was 2x that.  E You can lalso save money by buying a Linux DS15. But its still nearly E 2x the price of a IBM or Sun 1U server. 2x the price particularly for ( high density servers is not competitive.   Andrew > . > Can you say 'cash cow'?  I knew you could... >  > - bill   ------------------------------  $ Date: Fri, 1 Sep 2006 09:14:29 -0400' From: "Main, Kerry" <Kerry.Main@hp.com>  Subject: RE: VMS in The DAT Message-ID: <FA60F2C4B72A584DBFC6091F6A2B8684019BC80A@tayexc19.americas.cpqcorp.net>   > -----Original Message-----$ > From: bill@triangle.cs.uofs.edu=20A > [mailto:bill@triangle.cs.uofs.edu] On Behalf Of Bill Gunshannon   > Sent: August 31, 2006 10:46 PM > To: Info-VAX@Mvb.Saic.Com  > Subject: Re: VMS in The DA >=20 > In article=20 @ > <FA60F2C4B72A584DBFC6091F6A2B8684019BC4D3@tayexc19.americas.cp
 > qcorp.net>, , > 	"Main, Kerry" <Kerry.Main@hp.com> writes: > >=20 > >=20A > > Including many other big name chip manufacturing companies=20 
 > (and I mean  > > big)...  > >=20 > >:-) >=20H > Of course, you can't mention their names or you would have to kill us.? > You know, eventually people are going to stop believing in=20  > these phantom 8 > VMS users.  Oh wait, most of the industry already has. >=20  , Whether you believe it or not is irrelevant.  G The fact is that if you were to think of the majority of the really top B chip manufacturers, while all will have various OS's for differentG things, you would find OpenVMS being used for the more serious back end H stuff. These are environments where down time is measured in hundreds ofC thousands of $'s per hour in very HA 24x7 operational environments.   < You can throw mud at this all you want. It is what it is.=20   Regards   
 Kerry Main Senior Consultant  HP Services Canada Voice: 613-592-4660  Fax: 613-591-4477  kerryDOTmainAThpDOTcom (remove the DOT's and AT)=20  4 OpenVMS - the secure, multi-site OS that just works.   ------------------------------  # Date: Fri, 01 Sep 2006 10:08:41 GMT 5 From: rdeininger@mindspringdot.com (Robert Deininger) 7 Subject: Re: We're not dead yet, we're feeling better ! [ Message-ID: <rdeininger-0109060608400001@dialup-4.233.149.231.dial1.manchester1.level3.net>   5 In article <44F7A82E.A04CADDC@teksavvy.com>, JF Mezei % <jfmezei.spamnot@teksavvy.com> wrote:   " >"johnhreinhardt@yahoo.com" wrote:H >> itself leaves many opportunities for downtime.  The key point was theJ >> "almost 7 years with a few power-cycles for cleaning".  That's a pretty >> good record.  > G >Nothing compared to my all mighty Microvax II which went 18 years with G >only the need to replace a power supply once. And it was still working A >fine when I decided to retire it prematurely to make space for a & >4000-600 that was slightly faster :-)  F Of course, folks demanding less expensive hardware should realize thatF less expensive will often mean less reliable.  Instead of 10-15 years,G 5-10 years may be more realistic.  Or maybe 3-5 years.  How low can you  go?   H Long-time VAX and Alpha customers are likely to be particularly startled by this industry trend.   D If you need a lower-cost, good-performance, non-mission-critical VMSC solution, HP is trying to accomodate you.  If you need decade-scale E uptimes, you need to exercise more care with modern hardware than you  might have in the past.   E On all but the very cheapest Integrity servers, you'll see (sometimes H optional) redundant power supplies and fans.  And of course with VMS andG most other OSs, you can configure your disks for redundancy.  Those are F all hardware components where you should expect failures well before aI typical VAX or even Alpha would have failed.  This is one of the costs of 	 progress.   B Typical system design goals these days include (not necessarily in priority order):' 1. Constant attention to lowering costs G 2. Reasonably good MTBF for components like power supplies -- and 10-15 J years MTBF for these components is now often considered unreasonable and aD waste of money.  And a (hypothetical) 5 year MTBF likely means a few failures in the first year.   P 4. Continued system operation with failure of single components, where possible.J 3. Easy accessibility of components, so things like fans can be changed byJ semi-skilled customers, with no tools, in a few minutes.  While the system" continues to operate, if possible.  F If your workload can't tolerate the occasional hour of downtime in theC first decade/year/month of your system's life, you SHOULD be paying H attention to these issues.  You can no longer take it for granted.  (YouE really never could, but some folks obviously do, and get lucky.  Luck # doesn't get you as far these days.)   H Configure systems with redundant power supplies.  Plug them all into theI grid, preferably not all on the same circuit.  Maintain a fast-turnaround F service contract, or stock your own spares.  Put a UPS on at least oneG power supply.  Watch for, and react to, the first failure -- the system I will likely keep running, but the second failure will take it down if the  first hasn't been fixed.   Similar words apply to fans.  B Also disks.  Use VMS volume shadowing, or hardware RAID, or both. I Configure on-line spare disks where possible.  Again, watch for and react  to failures.  J If you can't tolerate network failure, add an extra LAN card.  Put it on aH different PCI bus, if possible.  Obviously you should consider redundant hardware throughout your LAN.   / Repeat the previous paragraph for SAN hardware.     H Most of this is pretty obvious.  Most customers don't need to care.  ButG VMS customers are more likely to care than the average Joe IT consumer.    ------------------------------  * Date: Fri, 1 Sep 2006 12:32:45 +0000 (UTC)( From: m.kraemer@gsi.de (Michael Kraemer)7 Subject: Re: We're not dead yet, we're feeling better ! 5 Message-ID: <ed999d$qgd$1@lnx107.hrz.tu-darmstadt.de>   C In article <1157078982.614029.133450@m79g2000cwm.googlegroups.com>, = "johnhreinhardt@yahoo.com" <johnhreinhardt@yahoo.com> writes:  >  > D > Nothing ever has zero downtime without redundant parts.  A DS10 byG > itself leaves many opportunities for downtime.  The key point was the I > "almost 7 years with a few power-cycles for cleaning".  That's a pretty  > good record. >    well, I'm less than impressed., Any half-decent gear, be it DEC or non-DEC, % manufactured in the "good old times", ) i.e. 10 or even 15 years ago will show up  with similar numbers. + The majority of boxes in my vicinity bought 1 8 years ago worked since then without any repair, J and power cycling only applied when being moved from one room to the other or for scheduled power-outages. & (I'm not talking PCs here, of course).> The funny thing with the OP is that a major web-based supplier0 of alpha/VMS solutions is publicly stating that > a) they are incabable or unwilling to setup an alpha/VMS-based3    web-service with downtime less than about 3 days ! b) a Wintel box is better anyway.   9 I can hardly think of a more devastating "advertisement"  - for this group's favorite platform than this. 2 So what does that guy want to sell in the future ? Probably neither alpha nor VMS.    ------------------------------  $ Date: Fri, 1 Sep 2006 09:18:26 -0400' From: "Main, Kerry" <Kerry.Main@hp.com> 7 Subject: RE: We're not dead yet, we're feeling better ! T Message-ID: <FA60F2C4B72A584DBFC6091F6A2B8684019BC808@tayexc19.americas.cpqcorp.net>   > -----Original Message-----= > From: Michael Kraemer [mailto:kraemer@biors6a.gsi.de] On=20  > Behalf Of Michael Kraemer ! > Sent: September 1, 2006 8:33 AM  > To: Info-VAX@Mvb.Saic.Com 9 > Subject: Re: We're not dead yet, we're feeling better !  >=20E > In article <1157078982.614029.133450@m79g2000cwm.googlegroups.com>, ? > "johnhreinhardt@yahoo.com" <johnhreinhardt@yahoo.com> writes:  > >=20 > >=20F > > Nothing ever has zero downtime without redundant parts.  A DS10 by> > > itself leaves many opportunities for downtime.  The key=20 > point was the > > > "almost 7 years with a few power-cycles for cleaning". =20 > That's a pretty  > > good record. > >=20 >=20  > well, I'm less than impressed.0 > Any half-decent gear, be it DEC or non-DEC,=20' > manufactured in the "good old times", + > i.e. 10 or even 15 years ago will show up  > with similar numbers. - > The majority of boxes in my vicinity bought 3 > 8 years ago worked since then without any repair, B > and power cycling only applied when being moved from one room=20 > to the other! > or for scheduled power-outages. ( > (I'm not talking PCs here, of course).@ > The funny thing with the OP is that a major web-based supplier4 > of alpha/VMS solutions is publicly stating that=20@ > a) they are incabable or unwilling to setup an alpha/VMS-based5 >    web-service with downtime less than about 3 days # > b) a Wintel box is better anyway.  >=20  G Sounds like local office of vendor has Windows techies on staff and are F simply expressing their OS religion without taking into account all ofE the monthly security patching (with associated app testing required), C stability, availability issues. Course, the vendor is typically not J involved in the Operations end, so they do not view that as a big deal.=20  = > I can hardly think of a more devastating "advertisement"=20 / > for this group's favorite platform than this. 4 > So what does that guy want to sell in the future ?! > Probably neither alpha nor VMS.  >=20  H Since it sounds like the local office has Windows techies on staff, theyH will likely not want to sell Solaris, HP-UX, AIX, NSK, z/OS or any other higher end platform either.=20   >=20   ------------------------------   End of INFO-VAX 2006.479 ************************