0 INFO-VAX	Fri, 12 Jan 2001	Volume 2001 : Issue 24      Contents:' "OpenVMS + Active Directory" be careful  ACP_DATACHECK, was: Faster gzip  Apache web server configuration # Re: Apache web server configuration D are there any non-commercial VMS users in or near Frankfurt, Germany4 RE: Compaq: A simple, affordable clustering solution4 RE: Compaq: A simple, affordable clustering solution4 Re: Compaq: A simple, affordable clustering solution4 RE: Compaq: A simple, affordable clustering solution4 Re: Compaq: A simple, affordable clustering solution4 Re: Compaq: A simple, affordable clustering solution4 Re: Compaq: A simple, affordable clustering solution( Re: CSWS (Apache) update 1.0-1 available3 Re: DCPS printer support (was RE: LN06 replacement) 3 Re: DCPS printer support (was RE: LN06 replacement) 3 Re: DCPS printer support (was RE: LN06 replacement)  dec 3000 300 crashing  Re: dec 3000 300 crashing  Re: Dec 3000 300 help (humour) RE: Dec 3000 300 help (humour)/ Re: DECNet dyslexia (was: Re: Version dyslexia) # Re: Ebay dead as a post - yet again # Re: Ebay dead as a post - yet again / eBay extended downtime (was Re: eBay SPAMARAMA) 3 Re: eBay extended downtime (was Re: eBay SPAMARAMA) 3 Re: eBay extended downtime (was Re: eBay SPAMARAMA) 3 Re: eBay extended downtime (was Re: eBay SPAMARAMA) 3 Re: eBay extended downtime (was Re: eBay SPAMARAMA) * Re: Faster gzip, thank you Paul Repacholi!* Re: Faster gzip, thank you Paul Repacholi!* Re: Faster gzip, thank you Paul Repacholi!* Re: Faster gzip, thank you Paul Repacholi!* Re: Faster gzip, thank you Paul Repacholi!$ Re: Gigabit Ethernet and VMS 6.2-1H3 Re: GS160 hardware question  Re: GS160 hardware question  Re: GZIP enhancements  RE: Identify que entry # Re: Identify que entry # Just what's wrong with eBay? Kitchen sink. ( not emacs! ) Re: LN29 escape sequences  Loads of Alphas (700+)...  Re: Loads of Alphas (700+)...  Re: Loads of Alphas (700+)... * Minimum cluster size in a 36GB striped set. RE: Minimum cluster size in a 36GB striped set. Re: Minimum cluster size in a 36GB striped set. RE: Minimum cluster size in a 36GB striped set. Re: Minimum cluster size in a 36GB striped set. RE: Minimum cluster size in a 36GB striped set. Re: Minimum cluster size in a 36GB striped set. RE: Minimum cluster size in a 36GB striped set. Re: Minimum cluster size in a 36GB striped set. RE: Minimum cluster size in a 36GB striped set mntverifytimeout Re: mntverifytimeout$ Re: Modifications to directory files$ Re: Modifications to directory files; Non-Disclosure Agreements (was: OpenVMS + Active Directory) ? Re: Non-Disclosure Agreements (was: OpenVMS + Active Directory) 5 OpenGL on a Compaq XP1000 w/ Elsa Gloria Synergy card 9 Re: OpenGL on a Compaq XP1000 w/ Elsa Gloria Synergy card 9 Re: OpenGL on a Compaq XP1000 w/ Elsa Gloria Synergy card < Re: OpenVMS and DII COE (was Re: IBM + Linux + US$ 1 bilion) Re: OpenVMS on CNET  RE: OpenVMS on CNET / Oracle and OpenVMS - Oops, I did it again ! ! ! 3 Re: Oracle and OpenVMS - Oops, I did it again ! ! ! 3 Re: Oracle and OpenVMS - Oops, I did it again ! ! !  Positive Gartner VMS article! powerchute for VMS AXP don't work % Re: powerchute for VMS AXP don't work % Re: powerchute for VMS AXP don't work % Re: powerchute for VMS AXP don't work % Re: powerchute for VMS AXP don't work % Re: powerchute for VMS AXP don't work  Re: QuixoticVMS? Re: QuixoticVMS?@ RAH as default behavior?  (Was: Faster gzip, thank you Paul R.!) Re: Repair costs for TK-70?  Re: Retired Oldtimer Re: Retired Oldtimer Re: Retired Oldtimer= re: Simply Marvelous (was Re: DS20 vs DS20E. Was: Sun Cluster = re: Simply Marvelous (was Re: DS20 vs DS20E. Was: Sun Cluster = Re: Simply Marvelous (was Re: DS20 vs DS20E. Was: Sun Cluster  Re: SPXgt video board 
 tcpiptrace Re: tcpiptrace+ Re: Today's Nugget of Worthless Information  Transaction Server for OpenVMS Re: UCX Routing Problem  Re: UCX Routing Problem  Re: UCX Routing Problem  Re: UCX Routing Problem ' Re: VMS performance data in HTML format / Re: VMS wishlist [was Re: Identify que entry #] / Re: VMS wishlist [was Re: Identify que entry #] & Re: voor de belgische Alpha gebruikers; Windows driver for LPS20 and LPS17 (Windows 2000 compliant) ? Re: Windows driver for LPS20 and LPS17 (Windows 2000 compliant)   F ----------------------------------------------------------------------  % Date: Fri, 12 Jan 2001 11:28:13 +0100 ' From: Theo Jakobus <jakobus@iaf.fhg.de> 0 Subject: "OpenVMS + Active Directory" be careful* Message-ID: <3A5EEA4D.2DC345C7@iaf.fhg.de>   Chuck Chopp wrote: >  > JF Mezei wrote:  > R > > Is there a quick explanation in laymen's terms of what that "active directory" > > really is ? O > > (eg: all marketing fluff removed, just the technical explanation of what it  > > really is).  > O > In a word, it is an implementation of LDAP.  There is more to it, but that is  > basically it.  >   H The "active directory" under Windows 2000 has problems with LDAP serversH provided by SUN, IBM and Novell. Windows 2000 has also problems with DNSA servers. More information is given in: "Windows 2000: A Threat to F Internet Diversity and Open Standards?" David Chadwick, IEEE COMPUTER,: August 2000, http://computer.org/computer/co2000/r8toc.htm For online reading: H http://dlib.computer.org/co/books/co2000/pdf/r8107.pdf membership of the# IEEE Computer Society is necessary.      Regards, --    ; *********************************************************** ; *                                                         * ; *  Theo Jakobus                                           * ; *  Fraunhofer-Institut fuer Angewandte Festkoerperphysik  * ; *  Tullastr. 72                                           * ; *  D-79108 Freiburg                                       * ; *  Germany                                                * ; *  Phone:   +49-(0)761-5159-325                           * ; *  FAX :    +49-(0)761-5159-200                           * ; *  e-mail:  jakobus@iaf.fhg.de                            * ; *  http://www.iaf.fhg.de                                  * ; *                                                         * ; ***********************************************************    ------------------------------   Date: 12 Jan 2001 18:24:35 GMT2 From: mathog@seqaxp.bio.caltech.edu (David Mathog)( Subject: ACP_DATACHECK, was: Faster gzip+ Message-ID: <93ni53$3v@gap.cco.caltech.edu>   a In article <93le66$sam@gap.cco.caltech.edu>, mathog@seqaxp.bio.caltech.edu (David Mathog) writes:   > Glenn Everhart kindly suggested that I try changing the SYSGENP parameter ACP_DATACHECK on my DS10 from the default value of 2 to 0.  ApparentlyC the default value for this causes the system to reread certain data G from a disk even when the SCSI device has already reported a successful  write.  For details see:  *   $ mc sysgen help sys_param acp_datacheck  D Changing this parameter had no effect on run times on RAMdisks (dataD not shown) but it did result in a 5-15% improvement in run times forC programs which created 80 files.  Since it's none too likely that a F SCSI device will report a valid read when the read has in fact failed,@ this check is probably unnecessary in most instances and so the D parameter change relativel safe.  As Glenn points out, all it can doB is tell you that your directory is _already_ corrupted so that you> don't make it even worse later, and it's only ever going to beB triggered by a SCSI disk that is so badly broken that it sends the@ wrong status signal for a failed write, as opposed to failing to6 write, which is the usual failure mode for SCSI disks.  B Ok, enough with the preliminaries, here are the results using the  "mybenchmark" test suite from:  <   http://seqaxp.bio.caltech.edu/pub/SOFTWARE/mybenchmark.zip   $ @make fastrms  $ @testsplit  /                DS10  DS10          DS10         /                Linux VMS           VMS          /                -     FASTRMS       FASTRMS      /                U2W   U2W           U2W          L ACP_DATACHECK  -     2             0              [        files           ]N                (s)   (s)  dev ops  (s)  dev ops   created  opened read written  H MAKETEST:      .3     1.77   147    1.77 147       1        1      0   1H MYSTART:       .0      .02     5     .02   5       0        0      0   0H MYOPENIN:      .0      .04     8     .04   7       0        1      0   0H MYREAD:        .1      .47   138     .56 136       0        1      1   0H MYOPEN:        .2     9.05   575    8.76 572      80        1      1   0H MYSPLIT:       .5    16.62  1087   13.97 890      80        1      1  80H MYSPLITBINARY: .4    14.38   915   13.57 890      80        1      1  80  L Shutting off ACP_DATACHECK had no effect on the first 4 tests, it only makesG a difference when files are created and/or written.  There's about a 5% H improvement in file creation speed and an additional 10% improvement forG writes on those files (fopen/fprint) or 2% (open/write).  Note that we  H can deduce from the difference between MYREAD and MYOPEN that each file F creation takes (8.76-.56)/80 = .103 seconds.  The same operation on a C RAMdisk takes (.64-.3)/80 = .0043 seconds.  So there's roughly 100  H msecs of disk action associated with each file creation.  Plenty of roomD for improvement there - file caching would conceivably reduce those F 80 rewrites of the directory file to just 1!  We can also see that theD time to write all this data (1.77secs) and to read it (.56) plus theC file creation time for 80 files (8.76 - .56 = 8.2) is less than the M time to do all 3 actions (13.97) seconds. That is, 13.97 - 1.77 - 8.76 = 3.44 F seconds.  That will correspond to disk head motion (since all the read# /write time is already included).     F The biggest benefit of this SYSGEN parameter change appears to be thatD it makes fopen/fprint IO run at exactly the same speed as open/writeC IO. Notice that the two MYSPLIT variants even do the same number of = operations in this mode.  I checked and it really was running E MYSPLITBINARY in one case and MYSPLIT in the other.  Perhaps somebody A in the C RTL group can explain to us how it is that changing this A SYSGEN parameter equalizes C's formatted and binary write speeds.   B FASTRMS is a big help, and changing ACP_DATACHECK is a minor help F (if your software is creating a lot of files), but the fundamental RMS@ factor of 2 or 3 performance lag of RMS vs. Linux remains.  (AndF that's being generous and assuming that XFC will deliver RAMdisk levelF performance right out of the gate.)  DII/COE is good idea but one mustF seriously question why anybody would port any IO intensive apps to VMSF using that API if, after doing so, the result was always software thatA ran at least 2 to 3 times slower than it did on other platforms.  D Without XFC (or a RAMdisk) the factor is to 20 or 30.  Those are theD sorts of performance ratios that are more likely to drive migrations in the other direction.    Regards,   David Mathog mathog@seqaxp.bio.caltech.edu ? Manager, sequence analysis facility, biology division, Caltech     ------------------------------  % Date: Fri, 12 Jan 2001 17:28:10 +0000 " From: Nic P Clews <nclews@csc.com>( Subject: Apache web server configuration@ Message-ID: <OF1E536286.F1F871FB-ON802569D2.005C8B31@eu.csc.com>   Hi,   5 I've installed the Apache server (CSWS 1.0-1) onto an 7 Alpha running 7.2-1, and got it configured and working, 6 that is it correctly serves the "success" page and its documentation.  , I've edited the APACHE$ROOT:[CONF]HTTPD.CONF9 file to change the root page. I've actually changed it to    /www_root/documents   0 which as you know is the OSU webserver documents4 default location. This also works without a problem.  1 What I have done with the OSU server is to map to 4 other disks and directories, create a rooted logical- and put it in the HTTP_PATHS.CONF for the OSU  server.   ) e.g. a http://node.domain/fred can map to + /fredroot/files, where fredroot is a rooted * logical, say $1$dia1:[fred.] and the files- are therefore seen in fredroot:[files]. Works  fine under OSU.   4 The question is, how do I get Apache to do the same?   i.e.    The Apache root document maps to /www_root/documents / where www_root is defined to $1$dia2:[osu.] and + documents is obviously www_root:[documents]   , A http browser request to http://node.domain1 gives me the index.html from the above directory.  This is OK.   , How do I ADD a mapping for another location?   i.e. a http browser request to http://node.domain/arbitaryname  to map to something else.   4 I do not want to have arbitaryname as a subdirectory1 of the root document, I want to be able to map it , to somewhere else, e.g. $1$dia3:[test.files]  , Therefore in the Apache HTTPD.CONF I want to- map "arbitaryname" to (say) /test/files where ( test is a rooted logical $1$dia3:[test.]  * Hopefully I've explained what I want to do. clearly enough, I've checked the documentation1 and not found a way of doing it. Virtual hosts is * not an option (we want only one DNS entry), and the rewrite rules do not look as if they
 will help.  $ Has anyone with experience of Apache( configuration managed this, or can help?# Can this be achieved under UNIX, if % so can someone perhaps supply me with * a copy of the HTTPD.CONF file (oops should be lowercase!)?   . We're trialing/comparing the Apache web server3 to OSU. Any assistance or help gratefully received.   ( Regards, Nic Clews CSC Computer Sciences nclews at csc dot com    ------------------------------  % Date: Fri, 12 Jan 2001 12:43:58 -0500 " From: Dan Sugalski <dan@sidhe.org>, Subject: Re: Apache web server configuration: Message-ID: <5.0.2.1.0.20010112124315.01f4a850@24.8.96.48>  - At 05:28 PM 1/12/01 +0000, Nic P Clews wrote: - >How do I ADD a mapping for another location?  >  >i.e. a http browser request to   >http://node.domain/arbitaryname >to map to something else. > 5 >I do not want to have arbitaryname as a subdirectory 2 >of the root document, I want to be able to map it- >to somewhere else, e.g. $1$dia3:[test.files]  > - >Therefore in the Apache HTTPD.CONF I want to . >map "arbitaryname" to (say) /test/files where) >test is a rooted logical $1$dia3:[test.]   7 The Alias directive will do this, IIRC. Something like:   ! Alias /test/files/ /$1$dia3/test/    should do what you want.   					Dan  I --------------------------------------"it's like this"------------------- 2 Dan Sugalski                          even samurai? dan@sidhe.org                         have teddy bears and even ;                                       teddy bears get drunk    ------------------------------   Date: 12 Jan 2001 11:13:29 GMT* From: helbig@astro.rug.nl (Phillip Helbig)M Subject: are there any non-commercial VMS users in or near Frankfurt, Germany . Message-ID: <93mosp$j1g$3@info.service.rug.nl>  G Having just moved to Frankfurt to take up a new (VMS) job, I'd like to  F hear from other non-commercial (e.g. hobbyist, academic,...) users of H VMS and/or |d|i|g|i|t|a|l| hardware (other than the few I know already)  from Frankfurt or nearby.   D Da ich gerade nach Frankfurt gezogen bin um eine neue (VMS-) Arbeit E aufzunehmen, wre interessant zu wissen, ob es dort oder in der Nhe  F andere nicht-kommerzielle (z.B. Hobby-, akademische) Benutzer von VMS H und/oder |d|i|g|i|t|a|l| Hardware gibt (auer denjenigen, die ich schon  kenne).    ------------------------------  % Date: Fri, 12 Jan 2001 06:54:56 -0300 ) From: fabio_compaq@ep-bc.petrobras.com.br = Subject: RE: Compaq: A simple, affordable clustering solution L Message-ID: <OFDE677BB6.C9835C35-ON032569D2.003537C8@ep-bc.petrobras.com.br>   Hi    H I am not saying that OpenVMS is a legacy OS..... but what people are sa= ying inH the market....and the development people here mainly....they dont want = to makeH investments anymore in OVMS ... probably the company will buy a GS-80 t= o H mantain the "legacy applications"  running  with better performance dur= ing  the H next 2-3 years - the time of development of the SAP... if the things ch= ange inE these 2-3 years ,may be we can have a SAP under OVMS... it=B4s really 	 important H to port and I think Compaq should - we have an experession in Brazil - = open the legs9  to SAP, or Peoplesoft, or BAAN, to port and ERP to OVMS.XH Would be expensive for Compaq make this investment ? Why not use a part=  ofeE the US$ 4 bilion of sales of OVMS products to improve the market ....   H If AS-400 still being used and is considered a "legacy" operating syste= m !l ! !u  C There are a lot of installations of SAP under AS-400 in Brazil.....u  H I dont like IBM and their arrogance .... but if Compaq is making a lot = of
 agreementsH with them.... if they are selling StorageWorks and Compaq will sale Sha= rkso andnF Tivoli TSM . . .  why not ask a help for them to OpenVMS too ! ! ! :-)   Regardsd   FC    3 In my opinion legacy are DOS, OS/2, Mumps, etc ....t                  < "Main, Kerry" <Kerry.Main@Compaq.com> em 12/01/2001 03:05:29      H       "'fabio_compaq@ep-bc.petrobras.com.br'" <fabio_compaq@ep-bc.petro= bras.com.br>       Info-VAX@Mvb.Saic.Com6      = Assunto: RE: Compaq: A simple, affordable clustering solutione     Fabio,  H >>> 1 - Digital grew up under the umbrella of OpenVMS - Digital was 80%=  H OVMS, 10 % Unix and 10% NT. The Alpha processor came under the OVMS too=  !E And now OVMS  become a legacy operating system which Compaq must giveeH support to the next 15 years - sounds forced - and sounds it is becomin= g a( military OS. <<A  E Given that the UNIX architecture ia approximately 10 years older thant. OpenVMS, are you saying UNIX is also "legacy"?  E Given Windows NT/W2K is still a 32bit HW/OS platform (in this day anda age?),H that only supports primitive fail-over clustering, are you saying Windo= ws NT and W2K is also "legacy"?k  2 Lets determine what you mean by the term "legacy".  D Lets see now, OpenVMS V7.3 (FT2 now, FCS in March timeframe) has the
 following: - hot swap/add CPU'sE - full 64bit HW/SW with directly addressable VLM capabilities of muche. greater than 4Gb limitations of 32bit servers.H - fully load balanced clustered support of up to 96 individual system n= odesF (a few in the 150 node range have been run successfully, but "only" 96 nodes  are officially supported.   7 How many nodes are supported in your favourite cluster?   H - supports multiple instances of the OS on the same physical HW. This i= ss theEH Galaxy feature. Each OS instance can be setup to dynamically share CPU'= s6H between OS instances based on rules such as time of day, performance le= velsE or even manually drag-n-drop. Want to share the CPU's on the same boxpF between your Apache Web server and your back end database depending on loadsf" ect ? You can do this with Galaxy.H - clustered Apache web server such that multiple Apache instances all h= aveiH the same file names, logical names and all can access the same files at=  theF same time. Need more CPU's? Simply add them into the cluster with ZERO  application availability impact.H - supports multiple cluster interconnects (including GBE) that enable t= heH cluster interconnects to load balance the cluster traffic over multiple=   interconnects.H - single system image (common system disk) such that every  system in t= he< cluster can be updated at the same time with one OS upgrade.H - no matter what system the user or app connection connects to, the vie= w ofH all the data, system logicals, and queues (print and batch) in the clus= tere is 100% the same.wA - can shut down individual cluster nodes in the cluster with zeroSH application availability impact (even with apps using stateful connecti= ons  like the TP world)H - cluster batch and print queues that will direct batch jobs to the lea= st busy server or print queue.l  > Ok, now what systems or architectures do you think are legacy?   Regards,    
 Kerry Main Senior Consultant0 Compaq Canada Inc. Professional Services( Voice: 613-592-4660. Fax  :  819-772-7036 Email: Kerry.Main@Compaq.com     -----Original Message-----) From: fabio_compaq@ep-bc.petrobras.com.bro, [mailto:fabio_compaq@ep-bc.petrobras.com.br] Sent: January 11, 2001 2:05 PM To: Info-VAX@Mvb.Saic.ComC= Subject: Re: Compaq: A simple, affordable clustering solutione     It is sad to know . . .i  H 1 - Digital grew up under the umbrella of OpenVMS - Digital was 80% OVM= S, 10 % Unix and 10% NT.eH The Alpha processor came under the OVMS too ! And now OVMS  become a le= gacy operating systemH which Compaq must give support to the next 15 years - sounds forced - a= nd sounds it is becoming a- military OS.  )    2 - There is no SAP to run under OVMS.k  .    3 - There is no Netscape Web Sever anymore.  H    4 - There is no Lotus Notes or similar messaging / groupware system.=   Please dont remember All-in-1.  F    5 - There is no improvment on Decwindows or a new kind of graphical	 terminal.i  H    6 - There is no FULL integration with Windows NT / 2000. NT of cours= e is the best choice of OSi         to integrate with OVMS.w  H    7 - There is not a strong / popular usergroup to disseminate the OS.=    F    8 -  The net companies dont develop decnet routers anymore. What is
 decnet ???  F    9 - The soft companies dont port the last version of their softs to OVMS. Example: BEA Tuxedo.  H    10 - We: OpenVMS addicted / ambassadors / affiliates are looking lik= en Don Quijotes de La Mancha ....    
    See you      Fabio Cardoso        A andrew harrison <andrew.nospam@uk.sun.com> em 11/01/2001 15:18:22o             Info-VAX@Mvb.Saic.Comm      = Assunto: Re: Compaq: A simple, affordable clustering solution-     "Main, Kerry" wrote: >*
 > Ahh Andrew,* >* > Back in style .. >   = Its nice to see that you forgot to answer the question again.a   > re: Oracle and UNIX  > H > The reality is that almost all Customers have requirements to integra= teB > UNIX/NT (and yes, in a great deal of Customer situations today -	 OpenVMS).FH > Large Customers also do not want to deal with multiple vendors unless=   theyH > absolutely have to .. They are tired of the "pointing fingers" scenar= ios  .. >   D Interesting, Sun has a program called the vendor integration programE it gives you one point of call for say Oracle problems on Sun and theaB calls are resolved regardless of who is actaully at fault. We have the same arrangement with MS.   4 Perhaps you are refering to your own experience !!!!  D > >>> Why would Oracle be interested in working with the Compaq UNIX
 > Company,<<<  >hE > With Compaq, these Customers can choose Oracle and then decide whatF platformH > they feel is best to deploy it on (UNIX/OpenVMS/NT) and still deal wi= th onel	 > vendor.m  3 Kerry I am sure that you know what the flaw in this 5 argument is, you cannot be that deluded. Compaq isn't 5 one company or at least it appears to be incapable ofa9 behaving as one company. It is from a platform standpoints< four companies. Tru64, OpenVMS, NT and Tandem. There appears9 to be virtually no cooperation between the entities (witht- the exception of the UNIX and OpenVMS folks).p  5 The single most common complaint on this newsgroup isc8 that Compaq seems unable to behave as one Company unless& that one company sells NT/Intel boxes.   > / > With Sun, the answer is UNIX or UNIX or UNIX.  > H > What is Sun's position with respect to Windows NT or W2K integration = ?t Think,H > this is not an important consideration (especially at dept level) for=   many > Customers today? >G  = We support the NT file and print services natively in SolarissF and can be a PDC or a BDC. We can also host active directory, its beenH ported to Solaris (we were the UNIX reference port) and we can also run=  E NDS if people are using that as a name service and unsurprisingly NDS > runs rather better on a Solaris box (roughly 10x the number of
 transactions)h4 than an equivalently configured proliant running NT.  C SCO also support the Tarantella Server on Solaris this allows us touC support Win32 apps on Sun Workstations/SunRays using Tarantella and . we also have the native Citrix client as well.  D We also of course provide a cross platform Office suite (StarOffice)C which lets people have a compatible office suite running on Solaris  and on Win32/Linux etc.2  @ This is before you start lookig at our ability to support things< like ASP's, FrontPage server extensions, Notes/Domino Server
 IE4/5 etc.  C Whats the OpenVMS possition with respect to NT/Win2K integration ??Y  @ It seems to me that in the light of the non answer to the Active: Directory Support for OpenVMS question currently posted toB this thread that you are as usual in need of being better informed before jumping in head first.     A > Well, for one reason, Compaq does not directly make a competingI
 middlewareH > (IPlanet) or support its biggest competitor in the database world (DB= 2) likeF > Sun does. Do you think press releases like the following impress the Oracle > folks:E > <http://www.sun.com/smi/Press/sunflash/9905/sunflash.990504.2.html>  >/  D You forgot to answer the question again. We know that you don't haveD any competing middleware or DBMS's you sold them long ago. So why isA it that despite this lack of conflict with the Oracle sales forcenA OpenVMS is such a poorly supported and (sold by Oracle) platform..  A Could it be that the Oracle sales forces care more about actually.B having something to sell than a little bit of competition for part+ of the product portfolio that do they sell.y     regards  Andrew Harrison  Enterprise IT Architecte                   =    ------------------------------   Date: 12 Jan 2001 12:40:17 GMT* From: helbig@astro.rug.nl (Phillip Helbig)= Subject: RE: Compaq: A simple, affordable clustering solution8. Message-ID: <93mtvh$kv2$1@info.service.rug.nl>  
 In articleC <910612C07BCAD1119AF40000F86AF0D805284BE5@kaoexc3.kao.cpqcorp.net>,w. "Main, Kerry" <Kerry.Main@compaq.com> writes:    > Fabio, > I > >>> 1 - Digital grew up under the umbrella of OpenVMS - Digital was 80%rK > OVMS, 10 % Unix and 10% NT. The Alpha processor came under the OVMS too !sG > And now OVMS  become a legacy operating system which Compaq must givesL > support to the next 15 years - sounds forced - and sounds it is becoming a > military OS. <<o > G > Given that the UNIX architecture ia approximately 10 years older thane1 > OpenVMS, are you saying UNIX is also "legacy"? p > N > Given Windows NT/W2K is still a 32bit HW/OS platform (in this day and age?),N > that only supports primitive fail-over clustering, are you saying Windows NT > and W2K is also "legacy"?t > 4 > Lets determine what you mean by the term "legacy". > F > Lets see now, OpenVMS V7.3 (FT2 now, FCS in March timeframe) has the > following: > - hot swap/add CPU'sG > - full 64bit HW/SW with directly addressable VLM capabilities of mucho0 > greater than 4Gb limitations of 32bit servers.M > - fully load balanced clustered support of up to 96 individual system nodes2N > (a few in the 150 node range have been run successfully, but "only" 96 nodes > are officially supported.  > 9 > How many nodes are supported in your favourite cluster?t > N > - supports multiple instances of the OS on the same physical HW. This is theJ > Galaxy feature. Each OS instance can be setup to dynamically share CPU'sM > between OS instances based on rules such as time of day, performance levels G > or even manually drag-n-drop. Want to share the CPU's on the same boxnN > between your Apache Web server and your back end database depending on loads$ > ect ? You can do this with Galaxy.L > - clustered Apache web server such that multiple Apache instances all haveM > the same file names, logical names and all can access the same files at the H > same time. Need more CPU's? Simply add them into the cluster with ZERO" > application availability impact.K > - supports multiple cluster interconnects (including GBE) that enable thelI > cluster interconnects to load balance the cluster traffic over multiple  > interconnects.K > - single system image (common system disk) such that every  system in theh> > cluster can be updated at the same time with one OS upgrade.M > - no matter what system the user or app connection connects to, the view ofsL > all the data, system logicals, and queues (print and batch) in the cluster > is 100% the same. C > - can shut down individual cluster nodes in the cluster with zero L > application availability impact (even with apps using stateful connections > like the TP world)K > - cluster batch and print queues that will direct batch jobs to the leastt > busy server or print queue.i > @ > Ok, now what systems or architectures do you think are legacy?  G It all depends on the connotations.  "Legacy" can be positive, such as .H "the legacy of JFK".  Actually, it just means "been around a long time"." A Mercedes is a legacy automobile.   ------------------------------  % Date: Fri, 12 Jan 2001 12:35:41 +0000d0 From: andrew harrison <andrew.nospam@uk.sun.com>= Subject: Re: Compaq: A simple, affordable clustering solution * Message-ID: <3A5EFA1D.B863264A@uk.sun.com>   "David J. Dachtera" wrote: >  > andrew harrison wrote: > >c > > "David J. Dachtera" wrote: > > >  > > > Let's see here now...s > > >  > > > Mark Berryman wrote: > > > >n5 > > > > I have a good friend who works for Oracle ...l > > >  > > > ...and...  > > >n > > > andrew harrison wrote: > > > >s6 > > > > Sorry, check with your local Oracle sales rep. > > >m$ > > > Andrew: Comprende Angles(sp?)? > > >T > >O > > Oh I understand completely,e > 
 > You do, eh?i >  > So let me get this straight: > C > Kerry quotes one Oracle employee, who he probably knows and whosesE > opinion and information he respects, so you refer him to someone he-# > doesn't know and may not respect.: >  > Makes sense to me... (???!!!)@ >   9 Really, how many of these "I know an Oracle employee and -= they say that oracle are really interested in openVMS" posts P; have we seen over the last two years. Probably one a month.   8 These unatributable sources have never turned out to be ; accurate in the past what leads you to the conclusion that .& this source will be any more accurate.  7 In the mean time people posting to this group are stillb3 complaining that Oracle don't mention OpenVMS as a f5 supported platform, OpenVMS isn't a tier one port andt3 the majority of Oracle's apps don't run on OpenVMS.t  5 Don't you think that deeds speek rather more stronglye than words.    regards  Andrew Harrison  Enterprise IT Architectd   ------------------------------  % Date: Fri, 12 Jan 2001 12:53:49 +0000-  From: steven.reece@quintiles.com= Subject: RE: Compaq: A simple, affordable clustering solutionsH Message-ID: <OFC747FFD5.92806684-ON802569D2.00465DA8@qedi.quintiles.com>   Kerry Main replied to Fabio :  >>>eD >8 -  The net companies dont develop decnet routers anymore. What is >decnet ???9  G While I agree DECnet is an excellent protocol, the industry has adoptednH TCPIP as the standard internetworking protocol. Compaq OpenVMS is movingE towards TCPIP as well. However, that does not mean DECnet will not be ' supported for a long time ..it will be.  <<<   K But Fabio - there's DECnet over TCP/IP if you make the move to DECnet-Plus.|I It solves your problems of not being able to run DECnet over your network K because the PHNM (pointy haired network managers) want only TCP/IP on theirz network and it works very well.e  J I have some machines in the distant corners of my company's network that IG can communicate with using DECnet over TCP/IP and in the case of one of G them I use DQS to do local printing at that site as well.  Until I told I him, the network manager on that site didn't even know that I was runningc= DECnet into his network - he could only see TCP/IP packets!!!l   Steve.   ------------------------------  % Date: Fri, 12 Jan 2001 12:49:18 -0600 1 From: "Dave Gudewicz" <david.gudewicz@abbott.com>O= Subject: Re: Compaq: A simple, affordable clustering solution 8 Message-ID: <93njc3$ikr$1@fizban.fizban.pprd.abbott.com>  ) Saw a bumper sticker recently in Florida.t    Charlton Heston is my President.   Dave...   : "JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in message& news:3A5B7D74.80151E18@videotron.ca... > Robert Deininger wrote:iI > > I think it was Sony where Charlton Heston (small stockholder, but big J name) used this method.  He simply stood up at the annual meeting and read, out the lyrics to a horribly offensive song, > H > I was expecting him to take out his guns. His association with the gun lobby 8 > has erase any other image of mayt have had of the guy.   ------------------------------    Date: 13 Jan 2001 00:57:13 +0800, From: Paul Repacholi <prep@prep.synonet.com>= Subject: Re: Compaq: A simple, affordable clustering solutionb0 Message-ID: <87g0ior9na.fsf@k9.prep.synonet.com>  9 "David J. Dachtera" <djesys.nospam@earthlink.net> writes:   J > Hhmmm... Sounds (to me) like yet another reason to finally break throughG > the "Intel barrier". Convince Compaq (or someone else :-)) to finallyiJ > build an Intel mobo that provides the support VMS requires, then finallyH > complete Emerald (better ten(10) years late than never, I suppose). IfI > the parnter is not Compaq, consider licensing the mobo design to the Q,B > IBM, or whoever...  B Ah, like getting VMS to work on the Samsung UP1100 and UP2000 MBs,8 then getting Linux say working on them with Thuderbergs?  + God presserve us from VMS on x86, please...e   --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda.u@                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.   ------------------------------    Date: 13 Jan 2001 01:02:22 +0800, From: Paul Repacholi <prep@prep.synonet.com>= Subject: Re: Compaq: A simple, affordable clustering solutions0 Message-ID: <87bstcr9ep.fsf@k9.prep.synonet.com>  - "Main, Kerry" <Kerry.Main@compaq.com> writes:@  M >    6 - There is no FULL integration with Windows NT / 2000. NT of course isa/ > the best choice of OS to integrate with OVMS.r >  > Sure there is. Check out:r6 > <http://www.openvms.compaq.com/pathworks/as72a.html>8 > <http://www.openvms.compaq.com/pathworks/as73new.html>@ > <http://www.openvms.compaq.com/pathworks/pw_win2000_stat.html>  C No, the 'integration' is lated out in the VMS SPD. Just need to get  the price right.  H >    8 -  The net companies dont develop decnet routers anymore. What is > decnet ??? > I > While I agree DECnet is an excellent protocol, the industry has adoptedmJ > TCPIP as the standard internetworking protocol. Compaq OpenVMS is movingG > towards TCPIP as well. However, that does not mean DECnet will not bet* > supported for a long time ..it will be.   I DECnet has the odd chariteristic that it works. And keeps working, unlike F another suite I could mention, if BGP lets this post get to the world!     > M >   9 - The soft companies dont port the last version of their softs to OVMS.  > Example: BEA Tuxedo. > H > Not sure what the latest is on this one but I know there is some stuff8 > cooking here. Also check out ISV's like the following: > < N > <http://www.attunity.com/content/newsevents/detail.asp?catid=6&scatid=20&o=1 > 07&y=01/01/2000&h=1>! > <http://www.cognos.com/adtpci/>e > N >    10 - We: OpenVMS addicted / ambassadors / affiliates are looking like Don > Quijotes de La Mancha .... > G > Hey, I started on Digitals first RISC systems (PDP8's only had 6 mainrM > instructions.. don't get much more reduced than that) and I still feel like > > a spring chicken .. ok, so not all of my feathers are there. > M > Bottom line is that while the marketing needs to be improved (and I suspectaH > some stuff might be cooking here), based on only some of the technicalH > features I outlined in my previous thread to you, folks that are doing: > OpenVMS today do not need to take a back seat to anyone. > 
 > Regards, >  > Kerry Main > Senior Consultantc > Compaq Canada Inc. > Professional Servicess > Voice: 613-592-4660a > Fax  :  819-772-7036 > Email: Kerry.Main@Compaq.com >  >  > -----Original Message-----+ > From: fabio_compaq@ep-bc.petrobras.com.brh. > [mailto:fabio_compaq@ep-bc.petrobras.com.br]  > Sent: January 11, 2001 2:05 PM > To: Info-VAX@Mvb.Saic.Comn? > Subject: Re: Compaq: A simple, affordable clustering solution, >  >  > It is sad to know . . .e > K > 1 - Digital grew up under the umbrella of OpenVMS - Digital was 80% OVMS,  > 10 % Unix and 10% NT.lM > The Alpha processor came under the OVMS too ! And now OVMS  become a legacyt > operating systemK > which Compaq must give support to the next 15 years - sounds forced - andy > sounds it is becoming a  > military OS. > + >    2 - There is no SAP to run under OVMS.? > 0 >    3 - There is no Netscape Web Sever anymore. > I >    4 - There is no Lotus Notes or similar messaging / groupware system.u  > Please dont remember All-in-1. > H >    5 - There is no improvment on Decwindows or a new kind of graphical > terminal.0 > M >    6 - There is no FULL integration with Windows NT / 2000. NT of course is, > the best choice of OSl! >         to integrate with OVMS.6 > I >    7 - There is not a strong / popular usergroup to disseminate the OS.c > H >    8 -  The net companies dont develop decnet routers anymore. What is > decnet ??? > H >    9 - The soft companies dont port the last version of their softs to > OVMS. Example: BEA Tuxedo. > J >    10 - We: OpenVMS addicted / ambassadors / affiliates are looking like  > Don Quijotes de La Mancha .... >  >  >    See you >  >    Fabio Cardoso >  >  >  > C > andrew harrison <andrew.nospam@uk.sun.com> em 11/01/2001 15:18:22r >  >  >  >       Info-VAX@Mvb.Saic.Comy >  >  > ? > Assunto: Re: Compaq: A simple, affordable clustering solutions >  >  > "Main, Kerry" wrote: > >s > > Ahh Andrew,n > >s > > Back in style .. > >o > ? > Its nice to see that you forgot to answer the question again.m >  > > re: Oracle and UNIXr > >aK > > The reality is that almost all Customers have requirements to integratesD > > UNIX/NT (and yes, in a great deal of Customer situations today - > OpenVMS). I > > Large Customers also do not want to deal with multiple vendors unless  > theyL > > absolutely have to .. They are tired of the "pointing fingers" scenarios > .. > >n > F > Interesting, Sun has a program called the vendor integration programG > it gives you one point of call for say Oracle problems on Sun and theSD > calls are resolved regardless of who is actaully at fault. We have > the same arrangement with MS.  > 6 > Perhaps you are refering to your own experience !!!! > F > > >>> Why would Oracle be interested in working with the Compaq UNIX > > Company,<<<m > >AG > > With Compaq, these Customers can choose Oracle and then decide what4
 > platformK > > they feel is best to deploy it on (UNIX/OpenVMS/NT) and still deal withh > oneo > > vendor.d > 5 > Kerry I am sure that you know what the flaw in thisl7 > argument is, you cannot be that deluded. Compaq isn'ta7 > one company or at least it appears to be incapable ofl; > behaving as one company. It is from a platform standpointh> > four companies. Tru64, OpenVMS, NT and Tandem. There appears; > to be virtually no cooperation between the entities (withP/ > the exception of the UNIX and OpenVMS folks).  > 7 > The single most common complaint on this newsgroup isr: > that Compaq seems unable to behave as one Company unless( > that one company sells NT/Intel boxes. >  > >l1 > > With Sun, the answer is UNIX or UNIX or UNIX.f > >oJ > > What is Sun's position with respect to Windows NT or W2K integration ? > Think I > > this is not an important consideration (especially at dept level) forl > many > > Customers today? > >o > ? > We support the NT file and print services natively in SolariswH > and can be a PDC or a BDC. We can also host active directory, its beenI > ported to Solaris (we were the UNIX reference port) and we can also runzG > NDS if people are using that as a name service and unsurprisingly NDS @ > runs rather better on a Solaris box (roughly 10x the number of > transactions)a6 > than an equivalently configured proliant running NT. > E > SCO also support the Tarantella Server on Solaris this allows us toaE > support Win32 apps on Sun Workstations/SunRays using Tarantella andi0 > we also have the native Citrix client as well. > F > We also of course provide a cross platform Office suite (StarOffice)E > which lets people have a compatible office suite running on Solariso > and on Win32/Linux etc.  > B > This is before you start lookig at our ability to support things> > like ASP's, FrontPage server extensions, Notes/Domino Server > IE4/5 etc. > E > Whats the OpenVMS possition with respect to NT/Win2K integration ??  > B > It seems to me that in the light of the non answer to the Active< > Directory Support for OpenVMS question currently posted toD > this thread that you are as usual in need of being better informed > before jumping in head first.. >  > C > > Well, for one reason, Compaq does not directly make a competingi > middlewareK > > (IPlanet) or support its biggest competitor in the database world (DB2)  > likeH > > Sun does. Do you think press releases like the following impress the > Oracle
 > > folks:G > > <http://www.sun.com/smi/Press/sunflash/9905/sunflash.990504.2.html>c > >' > F > You forgot to answer the question again. We know that you don't haveF > any competing middleware or DBMS's you sold them long ago. So why isC > it that despite this lack of conflict with the Oracle sales forceiC > OpenVMS is such a poorly supported and (sold by Oracle) platform.u > C > Could it be that the Oracle sales forces care more about actuallyiD > having something to sell than a little bit of competition for part- > of the product portfolio that do they sell.A >  > 	 > regardsU > Andrew Harrison  > Enterprise IT Architectr >  >  >  >  >    -- y< Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.   ------------------------------    Date: 13 Jan 2001 00:49:17 +0800, From: Paul Repacholi <prep@prep.synonet.com>1 Subject: Re: CSWS (Apache) update 1.0-1 availablet0 Message-ID: <87ofxcra0i.fsf@k9.prep.synonet.com>  + Rick Barry <barry@star.zko.dec.com> writes:   M > If you have an application that needs to access pages protected by a SYSUAFaS > password over an external network connection, consider putting those pages behindiP > an SSL sonnection to encrypt the username/password information protect againstJ > password sniffing. (Come to think of it, we should highlight this in our! > documentation -- thanks, Dan!).t  G Or do what WASD does, and require special set-up to allow VMS passwordsdC to be used at all. Opps, make that VMS username/passwords! You needy; to use an identifer to have sysuaf work over non-SSL links.e   -- A< Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda.e@                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.   ------------------------------   Date: 12 Jan 2001 16:03:04 GMT2 From: mathog@seqaxp.bio.caltech.edu (David Mathog)< Subject: Re: DCPS printer support (was RE: LN06 replacement), Message-ID: <93n9ro$opi@gap.cco.caltech.edu>  g In article <3A5E476A.142F31EC@earthlink.net>, "David J. Dachtera" <djesys.nospam@earthlink.net> writes:y >Paul Anderson wrote:e >> yD >> In article <3A5D1F27.DBF1A6BB@earthlink.net>, "David J. Dachtera"' >> <djesys.nospam@earthlink.net> wrote:o >>  F >> > Well, if extensible printer support is not in the next release, I% >> > sure hope it's in the one after!= >> eH >> It's not in the next release.  It's been thought about but due to the8 >> amount of work involved, there are no definite plans. >p  # MAKE SOME!!!!!!!!!!!!!!!!!!!!!!!!!!t  F I can't name another platform with such limited printer support.  And G there's just no reason for it other than inertia.   Compaq doesn't make K printers anymore so there's no reason for your product to be so picky aboutnJ printers.  The older Pathworks/Mac printer methodology used PPD files, andI I've yet to find a postscript printer for which there wasn't a PPD file. gH When we the MSAP method to print to everything I never had the problems I that I have with DCPS.  (Admittedly, it wouldn't do as much, but it wouldiI always print a .txt or .ps file, which is all we needed 99% of the time.) H Compaq still owns that code (much as I've tried to break it out into theJ public domain!) so why not bolt it onto the side of DCPS?  I know it won'tG do as much as DCPS can now with the supported printers, but things liketK NUMBER_UP could certainly run on top of it.  Printers supported under that  D method might have a type of FOREIGN=DEV:[DIR]foo.ppd, and DCPS wouldL set up the printer by interrogating the ppd file instead of conversing with  the printer.  I And for that matter, since Compaq won't release the Appletalk stack can'tAI you at least fold it into DCPS?  It's _much_ easier to configure printersaG on Appletalk than it is on TCP/IP.   In fact, most of the printers here D ONLY have appletalk addresses and even the PCs are running ethertalk software to talk to them.e  D >Given the number of new printer models / features that appear every9 >month, I'd say it's time to start making definite plans.W   I emphatically agree.9   >eI >If not, may as well drop the product altogther (mature product support).: >A  L Besides the fact that then we couldn't print from VMS at all, except in the 1 crudest possible manner, and with even more work?e   Regards,   David Mathog mathog@seqaxp.bio.caltech.eduo? Manager, sequence analysis facility, biology division, Caltech b   ------------------------------  % Date: Fri, 12 Jan 2001 11:46:24 -0500t0 From: Paul Anderson <paul.r.anderson@compaq.com>< Subject: Re: DCPS printer support (was RE: LN06 replacement)C Message-ID: <paul.r.anderson-546233.11462412012001@news.compaq.com>   - In article <93n9ro$opi@gap.cco.caltech.edu>, m$ mathog@seqaxp.bio.caltech.edu wrote:  % > MAKE SOME!!!!!!!!!!!!!!!!!!!!!!!!!!s > G > I can't name another platform with such limited printer support. And eC > there's just no reason for it other than inertia. Compaq doesn't hG > make printers anymore so there's no reason for your product to be so r > picky about printers.g  F We're adding about two dozen printer models from Genicom, HP, Lexmark G and Xerox (including Tektronix) in the next release.  This is a direct W8 result of the fact that Compaq no longer sells printers.  H There is a technical reason DCPS is picky about printers:  its design.  G DCPS now depends on bidirectional communication with a printer and its eG data does not come from PPD files.  It is a SMOP to change DCPS to get  ? its data from PPD files.  It is a SMOP to change DCPS to use a  G unidirectional approach, although some features would be lost.  All of  E this would take time and money, and a belief that the time and money a  spent would be worth the effort.  G > The older Pathworks/Mac printer methodology used PPD files, and I've aF > yet to find a postscript printer for which there wasn't a PPD file. A > When we the MSAP method to print to everything I never had the nG > problems that I have with DCPS. (Admittedly, it wouldn't do as much, iG > but it would always print a .txt or .ps file, which is all we needed mG > 99% of the time.) Compaq still owns that code (much as I've tried to oD > break it out into the public domain!) so why not bolt it onto the D > side of DCPS? I know it won't do as much as DCPS can now with the G > supported printers, but things like NUMBER_UP could certainly run on pG > top of it. Printers supported under that method might have a type of C? > FOREIGN?V:[DIR]foo.ppd, and DCPS would set up the printer by  E > interrogating the ppd file instead of conversing with the printer. f  G Printer data is hardcoded into DCPS.  The interrogation validates user eF requests for information that can't be determined by static data; for F example, "there's no letter paper loaded in that tray" types of stuff.  F > And for that matter, since Compaq won't release the Appletalk stack > > can't you at least fold it into DCPS? It's _much_ easier to F > configure printers on Appletalk than it is on TCP/IP. In fact, most F > of the printers here ONLY have appletalk addresses and even the PCs 1 > are running ethertalk software to talk to them.c  F I agree that AppleTalk is much easier to configure than TCP/IP.  DCPS B never has a problem over AppleTalk, while TCP/IP sometimes causes C problems, usually because of its lack of required bidirectionality..  I I also would love a supported AppleTalk stack.  I'm not sure DCPS is the   place to put it though..  D We in DCPS-land fully understand the need by some to extend DCPS to 4 support additional printers in the way you describe.   Paul   -- a,    Paul Anderson, OpenVMS Engineering (DCPS),    Compaq Computer Corporation, Littleton MA   ------------------------------   Date: 12 Jan 2001 17:04:33 GMT' From: david20@alpha2.mdx.ac.uk (D.Webb) < Subject: Re: DCPS printer support (was RE: LN06 replacement)0 Message-ID: <93ndf1$aat$1@aquila.news.mdx.ac.uk>  a In article <93n9ro$opi@gap.cco.caltech.edu>, mathog@seqaxp.bio.caltech.edu (David Mathog) writes:oh >In article <3A5E476A.142F31EC@earthlink.net>, "David J. Dachtera" <djesys.nospam@earthlink.net> writes: >>Paul Anderson wrote: >>> E >>> In article <3A5D1F27.DBF1A6BB@earthlink.net>, "David J. Dachtera" ( >>> <djesys.nospam@earthlink.net> wrote: >>> G >>> > Well, if extensible printer support is not in the next release, In& >>> > sure hope it's in the one after! >>> I >>> It's not in the next release.  It's been thought about but due to the 9 >>> amount of work involved, there are no definite plans.2 >> > $ >MAKE SOME!!!!!!!!!!!!!!!!!!!!!!!!!! >oK >public domain!) so why not bolt it onto the side of DCPS?  I know it won'tDH >do as much as DCPS can now with the supported printers, but things likeL >NUMBER_UP could certainly run on top of it.  Printers supported under that E >method might have a type of FOREIGN=DEV:[DIR]foo.ppd, and DCPS would M >set up the printer by interrogating the ppd file instead of conversing with a
 >the printer.: >w    L And if you don't need to converse with the printer for setup information youB might be able to provide a base level of support for LPD printers.    
 David Webb VMS and Unix team leader CCSS Middlesex University   ------------------------------  % Date: Sat, 13 Jan 2001 05:52:31 +1300 9 From: "Antony Wardle" <antony.wardle@noospammm.met.co.nz>r Subject: dec 3000 300 crashing0 Message-ID: <tyG76.674$Jy6.30866@ozemail.com.au>   Hi.g  @ Well, I have managed to keep my alpha, but i can't get it to go.  E It keeps bug checking and restarting before I can see whats going on.c    ; Once when it started up it complained about the memory, butu% doing a power off/on got rid of that.o  J I have tried 7.2, then 6.2. 6.2 it gets as far as the configuring devices.  ' It has a couple of rz26's and an rrd43.B  9 I have tried linux as well, and it crashes with an error:    ?06 DBL MCHK- PC= 00000000.00060401 PSL= 000000000.00000007s     Any ideas welcome.   cheers   antony   ------------------------------    Date: 12 Jan 2001 18:27:29 +0100G From: Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de>p" Subject: Re: dec 3000 300 crashingH Message-ID: <y4u274llz2.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>  K A double machine check usually indicates severe hardware trouble - it meansdL that the processor got a machine check (something's wrong with the hardware-, indication) while it was handling the first.   	Jan   ------------------------------   Date: 12 Jan 2001 11:08:45 GMT* From: helbig@astro.rug.nl (Phillip Helbig)' Subject: Re: Dec 3000 300 help (humour) . Message-ID: <93mojt$j1g$2@info.service.rug.nl>  ; > My wife is unreasonable like that, too.  She wouldn't leta> > me get a three-phase power feed to the house so that I could > bring home a VAX 6640. >  > Women, hrrumph.  ;^)  G Just tell her that all the neighbours will be jealous of her wonderful o  new super-duper washing machine.   ------------------------------  % Date: Fri, 12 Jan 2001 17:28:16 +0100_6 From: "Dijk, Jeroen van" <Jeroen.vandijk@getronics.nl>' Subject: RE: Dec 3000 300 help (humour) M Message-ID: <2795B75EF003D311801A00A0C906B511BF1000@cucexec.gbc.getronics.nl>   2 Check this page for a real VMS wife and VAX kidds.  & http://24.191.50.156/pix/FAMILY.HTMLX;     -----Original Message-----6 From: helbig@astro.rug.nl [mailto:helbig@astro.rug.nl]% Sent: donderdag 11 januari 2001 21:08t To: Info-VAX@Mvb.Saic.Come' Subject: Re: Dec 3000 300 help (humour)r    ? In article <878zojgdgn.fsf@k9.prep.synonet.com>, Paul RepacholiT  <prep@prep.synonet.com> writes:   ( > Wolfgang Rupp <rupp@chello.at> writes: > 2 > > [1] Unixish OSes, I admit. One Alpha runs VMS.G > > [2] PDP-11, MicroVAX-II, DECsystem5500, MV2000, VS3100, two Alphas, C > > a Pro/380 (completely useless, and perfectly working) and a PC.c > D > You have it easy. KL-10E, 11/70, 11/84, ah, some 11/73s, 3 uV-IIs,G > x uV2000s, n 3100s, 6520, 2 M600s, Prioris, noname, 4400, 4200, 3200.aK > I'm not even going to try to remember how many 11/34s. And a couple of 8ss? > Should be a Rainbow, DECmates, and a Vaxmate here somnewhere.c  = VAXman used to have some wonderful pictures of his VAXcave at       http://24.191.50.156/pix/  G but they seem to have disappeared.  (I think he mentioned here a while n@ back that they needed to be updated, as he had much more stuff.)   ------------------------------  % Date: Sat, 13 Jan 2001 06:34:32 +0000 + From: "antonio.carlini" <arcarlini@iee.org>u8 Subject: Re: DECNet dyslexia (was: Re: Version dyslexia)' Message-ID: <3A5FF6F8.FF6974E5@iee.org>t   John Santos wrote: > ? > If I'm understanding you correctly, I should install this kit = > on DECnet V7.2 and that will upgrade it to V7.2-1, and thent? > install the V7.2-1 ECO02, which will give me the latest ECO's 	 > for it?c  0 You understand me correctly but you also need to, understand that I don't have either of these/ kits handy any more nor do I have a V7.2 system ) to try this out on so I'm just guessing !    > E > I will try this out tonight.  I thought I had also tried to install B > the ECO01 kit, but I may have just assumed they were cumulative,> > that ECO02 replaced ECO01...  (I actually was doing the V7.2  5 ECOs have always been cumulative in my experience but : sometimes one ECO has another as a prerequisite. (Although: the only example I can recall is one of the Alpha X25 kits' that needs one of the X25 WANDD kits). p  9 It will certainly be interesting to know if the ECO1 doesh: install V7.2-1: if it doesn't then I have no idea what the! name of the correct kit would be.C   Antonioe   -- e   ---------------e- Antonio Carlini             arcarlini@iee.orge   ------------------------------  # Date: Fri, 12 Jan 2001 07:52:13 GMT ) From: Dave Rodgers <rodgersd@my-deja.com>u, Subject: Re: Ebay dead as a post - yet again) Message-ID: <93md3c$pss$1@nnrp1.deja.com>   F All this chum in the water and still no siting of Andrew the Sun-fish!   The horror!f     Sent via Deja.comn http://www.deja.com/   ------------------------------  % Date: Fri, 12 Jan 2001 08:06:06 -0500i0 From: David Beatty <David.Beatty.NOSPAM@sas.com>, Subject: Re: Ebay dead as a post - yet again2 Message-ID: <7QBfOovAQgkj8mqGhWfrxuo3LJJK@4ax.com>  F He's still out there, just not responding to this thread.  He's havingE way too much fun with Kerry.  It's been confirmed they had a hardwareNF problem with their Sun system, just not exactly what kind of hardware  problem.   David R. Beattyn  E On Fri, 12 Jan 2001 07:52:13 GMT, Dave Rodgers <rodgersd@my-deja.com>  wrote:  G >All this chum in the water and still no siting of Andrew the Sun-fish!r >e >The horror! >a >w >Sent via Deja.com >http://www.deja.com/h   ------------------------------  % Date: Fri, 12 Jan 2001 11:17:54 +0000e$ From: Steve.Spires@yellowpages.co.uk8 Subject: eBay extended downtime (was Re: eBay SPAMARAMA)/ Message-ID: <002569D2.003DF2B6.00@quegw01.btyp>r  L Contact:   Tel: 3063  -  IS - Infrastructure, 1st Floor, Bridge Street Plaza    K I shouldn't worry about getting any today - eBay have extended their Fridaya' downtime by FIVE HOURS for maintenance.   G They now have the systems down for SEVEN hours once a week, every week.x  ! Sorry, but that sucks big time...u   Steve Spires        @ Paul Repacholi <prep@prep.synonet.com> on 11/01/2001 08:02:19 PM    To:        Info-VAX@Mvb.Saic.Com+ cc:         (bcc: Steve Spires/YellowPages)eM From:      Paul Repacholi <prep@prep.synonet.com>, 11 January 2001, 8:02 p.m.    Re: eBay SPAMARAMA        6 "Terry C. Shannon" <terryshannon@mediaone.net> writes:  N > Bloody right, mate! Of course, there is a solution of sorts: don't patronize > eBay!o  H Ah, but it's all right now. Sun has sold e-Bay a special 'spam reduction system'.   --< Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.   ------------------------------    Date: 12 Jan 2001 09:16:36 -0500* From: young_r@eisner.decus.org (Rob Young)< Subject: Re: eBay extended downtime (was Re: eBay SPAMARAMA)+ Message-ID: <GnSEQzHosQrM@eisner.decus.org>   V In article <002569D2.003DF2B6.00@quegw01.btyp>, Steve.Spires@yellowpages.co.uk writes:N > Contact:   Tel: 3063  -  IS - Infrastructure, 1st Floor, Bridge Street Plaza >  > M > I shouldn't worry about getting any today - eBay have extended their Fridayo) > downtime by FIVE HOURS for maintenance.  > I > They now have the systems down for SEVEN hours once a week, every week.e > # > Sorry, but that sucks big time...a >   = 	This is a one shot deal (extra downtime) to fix the "broken"U> 	storage stuff that they "put off for the holidays."  I'm sure? 	they will return to their much shorter downtimes (not countingc- 	crashes... but who's counting?) next Friday.   ; 	I believe they are due back up in 45 minutes (10 Eastern).    				Roba   ------------------------------  % Date: Fri, 12 Jan 2001 14:24:28 +0000-  From: steven.reece@quintiles.com< Subject: Re: eBay extended downtime (was Re: eBay SPAMARAMA)H Message-ID: <OF611201C7.99AB450F-ON802569D2.004F0629@qedi.quintiles.com>  G But Rob, isn't that like "The check's due to arrive in the post any day 
 now"?  :-)  @ Rob Young (young underscore r at decus dot decus dot org) wrote:A >>>  This is a one shot deal (extra downtime) to fix the "broken"gB      storage stuff that they "put off for the holidays."  I'm sureC      they will return to their much shorter downtimes (not countingn1      crashes... but who's counting?) next Friday.t  ?      I believe they are due back up in 45 minutes (10 Eastern).a                       Robo <<<    ------------------------------  % Date: Fri, 12 Jan 2001 12:31:41 -0500e2 From: norm lastovica <norman.lastovica@oracle.com>< Subject: Re: eBay extended downtime (was Re: eBay SPAMARAMA)) Message-ID: <3A5F3F7D.764EBBF@oracle.com>   % Steve.Spires@yellowpages.co.uk wrote:  > N > Contact:   Tel: 3063  -  IS - Infrastructure, 1st Floor, Bridge Street Plaza > M > I shouldn't worry about getting any today - eBay have extended their Fridayr) > downtime by FIVE HOURS for maintenance.t > I > They now have the systems down for SEVEN hours once a week, every week.  > # > Sorry, but that sucks big time...t  9 	hummmm - if I read you correctly, they build in over 4% d? (assuming my math is close) downtime.  Less than 96% available.y   >  > Steve Spires > B > Paul Repacholi <prep@prep.synonet.com> on 11/01/2001 08:02:19 PM > " > To:        Info-VAX@Mvb.Saic.Com- > cc:         (bcc: Steve Spires/YellowPages)eO > From:      Paul Repacholi <prep@prep.synonet.com>, 11 January 2001, 8:02 p.m.e >  > Re: eBay SPAMARAMA > 8 > "Terry C. Shannon" <terryshannon@mediaone.net> writes: > P > > Bloody right, mate! Of course, there is a solution of sorts: don't patronize	 > > eBay!h > J > Ah, but it's all right now. Sun has sold e-Bay a special 'spam reduction
 > system'. >  > --> > Paul Repacholi                               1 Crescent Rd.,9 > +61 (08) 9257-1001                           Kalamunda.pB >                                              West Australia 60760 > Raw, Cooked or Well-done, it's all half baked.   -- h> norman lastovica / oracle rdb engineering / usa / 610.696.4685   ------------------------------  # Date: Fri, 12 Jan 2001 17:43:53 GMTi4 From: "Terry C. Shannon" <terryshannon@mediaone.net>< Subject: Re: eBay extended downtime (was Re: eBay SPAMARAMA)< Message-ID: <tnH76.57697$1t.2837837@typhoon.ne.mediaone.net>  ? "norm lastovica" <norman.lastovica@oracle.com> wrote in messageS# news:3A5F3F7D.764EBBF@oracle.com...  >  >n' > Steve.Spires@yellowpages.co.uk wrote:n > >oJ > > Contact:   Tel: 3063  -  IS - Infrastructure, 1st Floor, Bridge Street PlazaR > >pH > > I shouldn't worry about getting any today - eBay have extended their Friday+ > > downtime by FIVE HOURS for maintenance.m > >,K > > They now have the systems down for SEVEN hours once a week, every week.i > >C% > > Sorry, but that sucks big time...i >r9 > hummmm - if I read you correctly, they build in over 4%hA > (assuming my math is close) downtime.  Less than 96% available.e >n  J Sounds like a Windoze98 infrastructure. Who is to blame I do not know, but? for sure the site is characterized by more finger-pointing thano fault-tolerance!   ------------------------------  # Date: Fri, 12 Jan 2001 11:21:30 GMTh= From: system@SendSpamHere.ORG (Brian Schenkenberger, VAXman-)h3 Subject: Re: Faster gzip, thank you Paul Repacholi!x0 Message-ID: <009F5FE8.5DABCE5E@SendSpamHere.ORG>  V In article <3A5E32A5.43FA904C@GCE.com>, "Glenn C. Everhart" <Everhart@GCE.com> writes:: >This is an area that's been a bit of a pet peeve of mine.9 >It should be possible to make a process default to using > >readahead/writebehind and have the result basically invisible: >to programs; the RMS semantics otherwise are not altered. >-7 >Brian Schenkenberger did a few experiments with an RMSm; >call intercept to set the RAB bits needed forcibly though,s8 >and found that the overhead of the intercept he had was, >enough to make speed advantage often small.  H Just let me clarify the way the interection was accomplished.  The firstI trial had each RMS service intercepted such that the ROPs were set on thenH way in and reset on the way out.  This, it was completely transparent toH the applicationt that it was being done.  The overhead of saving the ROP6 bits going in and restoring going out was excessive.    I Later, intercerpting just the $CONNECT and having it set the ROP bits foraG all time did show remarkable performance increases for tests performed.d Glenn and I left it at that.   --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COMh            gO city, n., 1. a place where trees are cut down and streets are named after them.i   ------------------------------    Date: 12 Jan 2001 08:37:20 -0500* From: young_r@eisner.decus.org (Rob Young)3 Subject: Re: Faster gzip, thank you Paul Repacholi!.+ Message-ID: <7E8qFjb9vOuJ@eisner.decus.org>l  a In article <93le66$sam@gap.cco.caltech.edu>, mathog@seqaxp.bio.caltech.edu (David Mathog) writes:   J > I used DEQ=4096 throughout, and that's really too big for MYSPLIT, whereJ > the resultant files in this test are only 204 blocks.  Changing DEQ fromM > 4096 to 256 in MYSPLIT reduced the run time to 15.53 seconds and the devicem8 > operations to 944, which was a very small improvement.  ? 	I have found through trial and error that the best I could get G 	performance wise was to set DEQ = 2048 (largish files).  Above that ,  C 	my timings actually trailed off.  This has been a few years... butp( 	it seems you are seeing the same thing.   	Anyone explain such an effect?    				RobD   ------------------------------    Date: 12 Jan 2001 16:44:47 +0100G From: Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de>a3 Subject: Re: Faster gzip, thank you Paul Repacholi!eH Message-ID: <y47l40epw0.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>  ? system@SendSpamHere.ORG (Brian Schenkenberger, VAXman-) writes:f  K > Later, intercerpting just the $CONNECT and having it set the ROP bits for I > all time did show remarkable performance increases for tests performed.r  L So all one really needs is to patch the $CONNECT entry point in RMS.EXE withA that one (or few, on Alpha) instruction that sets those two bits?e   	Jan   ------------------------------  # Date: Fri, 12 Jan 2001 17:01:33 GMTi= From: system@SendSpamHere.ORG (Brian Schenkenberger, VAXman-)h3 Subject: Re: Faster gzip, thank you Paul Repacholi!s0 Message-ID: <009F6017.DEC82CBF@SendSpamHere.ORG>   In article <y47l40epw0.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>, Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de> writes:@ >system@SendSpamHere.ORG (Brian Schenkenberger, VAXman-) writes: >rL >> Later, intercerpting just the $CONNECT and having it set the ROP bits forJ >> all time did show remarkable performance increases for tests performed. >2M >So all one really needs is to patch the $CONNECT entry point in RMS.EXE withDB >that one (or few, on Alpha) instruction that sets those two bits?  8 In theory, that's all I did.  It's not a rocket science!   --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COM5            CO city, n., 1. a place where trees are cut down and streets are named after them.l   ------------------------------    Date: 12 Jan 2001 18:29:11 +0100G From: Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de> 3 Subject: Re: Faster gzip, thank you Paul Repacholi!hH Message-ID: <y4r928llw8.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>  ? system@SendSpamHere.ORG (Brian Schenkenberger, VAXman-) writes:   O > >So all one really needs is to patch the $CONNECT entry point in RMS.EXE with!D > >that one (or few, on Alpha) instruction that sets those two bits?: > In theory, that's all I did.  It's not a rocket science!   OK.t  K In email, Glen mentioned that David might try setting ACP_DATACHECK to 0 tofL avoid data checks on metadata writes by the XQP. THis should improveopen and
 close timing.a   	Jan   ------------------------------   Date: 12 JAN 2001 16:28:58 GMT4 From: karcher@thuria.waisman.wisc.edu (Carl Karcher)- Subject: Re: Gigabit Ethernet and VMS 6.2-1H3t6 Message-ID: <12JAN01.16285828@thuria.waisman.wisc.edu>  ? In a previous article, dashw459@aol.comeatspam (Doug W.) wrote:e ->  P ->Is this true?  I heard VMS does not yet support jumbo frames, but I assumed it/ ->could at least get up to FDDI sized frames.  n  L True. There's another issue as well: the Gigabit Ethernet hardware path mustJ also support jumbo frames from end to end. Many switches, hubs and routers8 don't and it's unlikely to be a simple firmware upgrade.   ------------------------------  % Date: Fri, 12 Jan 2001 09:16:34 -0500u2 From: norm lastovica <norman.lastovica@oracle.com>$ Subject: Re: GS160 hardware question* Message-ID: <3A5F11C2.7366E8E8@oracle.com>   andrew harrison wrote:  	.s 	. 	.: > Compaq have used the wrong system latency numbers for an9 > E10K and by doing so (they are much too high) show that0; > the GS's average memory latency will be better than Sun'sh9 > providing ~50% of memory acceses are local. In fact the27 > Sun's memory latency is ~440ns not 600 and because ofh4 > that the GS would actaully need to have ~90% local> > memory accesses to have the same average latency as an E10K.  4 	is this 440ns for corrupted data or for uncorrupted8 data.  I can not keep track of where the data corruption occurs in the E10K.t   ------------------------------  % Date: Fri, 12 Jan 2001 09:20:05 -0500t2 From: norm lastovica <norman.lastovica@oracle.com>$ Subject: Re: GS160 hardware question* Message-ID: <3A5F1295.2EC85F39@oracle.com>   Wolfgang Rupp wrote: > < > Larry Kilgallen <Kilgallen@eisner.decus.org.nospam> wrote:G > > Oracle Rdb is a VMS-specific database, but I think one nugget theiraF > > developer presented at the most recent Los Angeles DECUS SymposiumJ > > might be of interest. He said that in their situation data on the diskL > > is compressed, but when it comes time to fetch it from a buffer in "far"H > > memory it must be decompressed, and it gets decompressed into "near"H > > memory.  Thus he said the many subsequent accesses far outnumber theL > > one "decompress" access and one potential "compress" access (just before > > writing to disk).c > G > > So for his database engine the NUMA considerations were not a major.K > > issue.  As somebody said earlier, it depends on the exact software thatrI > > is being used.  In this case it is not the operating system software.  > H > Ahh... I thought that *something* might surface. I have no idea if theG > Oracle 8 in question (for Tru64) supports that behaviour. From what IsD > heard in the meetings I attended, this has not been mentioned likeH > above. Maybe because it is a feature of Oracle on VMS? I will point myE > colleagues to this, they can investigate that further. I am networkl6 > admin, not a db person. They are ignorant of Usenet. > H > What comes to mind here, though, is what happens when the decompressedF > data in "near" memory becomes larger than the available near memory?I > Maybe this is truely too OS dependent to be asked in comp.os.vms. OTOH,u> > out of near mem is out of near mem, hardware wise, isn't it?  : 	I was likely the Rdb developer in question...  Basically,? we end up doing NUMA friendly things in a couple cases.  when ae? compressed record is read from the database, it is uncompressedoA into a process local VM buffer.  this is assumed to be a home-rad- buffer. < 	further, when using the row cache feature, since the cachesB themselves will likely be spread among the rads, we also copy dataB from the cache upon access into local per-process VM.  Originally,C this wasn't done because of NUMA, but we realized later that it wast' ultimately a good thing to do for NUMA>u5 	I have no idea what Oracle Oracle is up to, however.O -- h> norman lastovica / oracle rdb engineering / usa / 610.696.4685   ------------------------------    Date: 12 Jan 2001 08:53:45 -05002 From: malmberg@eisner.decus.org (John E. Malmberg) Subject: Re: GZIP enhancements+ Message-ID: <lgeUS5YAjob0@eisner.decus.org>   g In article <3A5E829F.6D1DA45A@earthlink.net>, "David J. Dachtera" <djesys.nospam@earthlink.net> writes:e > "John E. Malmberg" wrote:p >>L >> The GZIP 1.2.4 for OpenVMS VAX 5.5-2 (with AACRTL060) and higher has beenB >> available for some time from ftp://ftp.qsl.net/pub/wb8tyw/gzip/ >>$ >> And should support the following: >>C >> GZIP [/stdout] [/decompress] [/force] [/help] [/list] [/license]p: >>      [/no-name] [/name] [/quiet] [/recursive (ignored)]? >>      [/suffix] [/test] [/verbose] [/version] [/fast] [/best]a >o > DJAS01::DDACHTERA$ gzip/help1 > /help: non-translatable vms error code: 0x186D4n- > %rms-f-syn, file specification syntax error- <snip>I > There is no .HLP file, AFAIK, and the above display makes no mention ofg, > anything other than the UN*X-like options. >oL >> I think I modified the help message to display that information, and also >> the .HLP file.l > E > Not per the above, and I don't remember ever seeing a .HLP file for  > GZIP.w  ; The functionality is available only in GZIP downloaded fromaL the ftp://ftp.qsl.net/pub/wb8tyw/gzip/ as pointed out in the previous posts.  I I just checked the kit, and the .HLP file is there with my edits, but theo< embedded help is still the same as per all versions of GZIP.  K I do not remember where I got the .HLP file from that I updated.  It may behK from the LINUX distribution, or I may have edited the .INFO files supplied.    -Johnr wb8tyw@qsl.network Personal opinion only    ------------------------------  % Date: Fri, 12 Jan 2001 09:58:54 -0500A4 From: "Bochnik, William J" <BochnikWJ@bernstein.com>! Subject: RE: Identify que entry # J Message-ID: <2B37459189B0D211BE710000F8EF9D8508908B18@nts0147.beehive.com>  J Unless you're lucky enough to have accounting running (which will give youL the required info) you could look at the modification times of ALL log filesH on the system, and compare them to the time you killed the job - the mod, time should reflect when you killed the job.   William J. Bochnik   Systems Programmer   --------------------------     original message:I From: bawilhelm@my-deja.com  Sent: January 11, 2001 3:48 PM To: Info-VAX@mvb.saic.com  Subject: Identify que entry #     C I just issued the command DEL/ENTRY=20 instead of DEL/ENTRY=28.  IsoF there any way I can determine what Entry 20 was.  I did not receive an# error message saying "no such job".P   We are running VMS 7.1   Thanks  
 Brent Wilhelmc System Admin/DBA Salisbury State University   ------------------------------  % Date: Fri, 12 Jan 2001 14:08:04 -0500 & From: "Syltrem" <syltrem@videotron.ca>! Subject: Re: Identify que entry #i7 Message-ID: <3uI76.2544$aR2.84610@wagner.videotron.net>e  K If the job was still pending or holding, you will have no trace of it being A deleted. Accounting will not record it as no process was running.    -- Syltrema3 http://pages.infinit.net/syltrem (OpenVMS Web Site)         ? "Bochnik, William J" <BochnikWJ@bernstein.com> wrote in message-D news:2B37459189B0D211BE710000F8EF9D8508908B18@nts0147.beehive.com...L > Unless you're lucky enough to have accounting running (which will give youH > the required info) you could look at the modification times of ALL log files5J > on the system, and compare them to the time you killed the job - the mod. > time should reflect when you killed the job. >o > William J. Bochnik >  > Systems Programmer >  > -------------------------- >v >x > original message:  > From: bawilhelm@my-deja.comi  > Sent: January 11, 2001 3:48 PM > To: Info-VAX@mvb.saic.coms > Subject: Identify que entry #  >r >bE > I just issued the command DEL/ENTRY=20 instead of DEL/ENTRY=28.  IsiH > there any way I can determine what Entry 20 was.  I did not receive an% > error message saying "no such job".e >m > We are running VMS 7.1 >u > Thanks >h > Brent Wilhelmo > System Admin/DBA > Salisbury State University >k >    ------------------------------    Date: 12 Jan 2001 11:26:28 -0500* From: young_r@eisner.decus.org (Rob Young)% Subject: Just what's wrong with eBay?u+ Message-ID: <P6nfCNUxAWIc@eisner.decus.org>a  > 	What's wrong?  Well they probably fixed some of it today, BUT, 	here is what it looked like even yesterday:  2 http://www.techweb.com/wire/story/TWB20010111S0021  I                   Indeed, five years after the site's launch, the company J                   hasn't installed redundant storage hardware to take overH                   if the main storage system fails, said one former eBay?                   staff member who asked not to be identified. h  E                   The IT department alerted management to the risk ofa?                   storage failure a year ago and recommended an<F                   upgrade to redundant storage arrays, but the expense4                   wasn't approved, the source said.    ---u  < 	Talk about a recipe for disaster.  Meg is too cheap.  Looks: 	like she signed off on redundant storage recently though.  ? 	Sheesh, couple crashing UE10000s with crashing storage... whatc< 	a nightmare, eh?  And Meg tries to make it look like it wasC 	a problem "for the holidays" when I pointed out they had a storageeB 	event in August 1999.  In a way, you gotta feel kind of sorry forE 	Sun as it becomes more apparent a "flagship" customer really doesn'tk; 	understand what uptime is all about AND is too cheap to dolH 	anything about it.  But don't feel too sorry for Sun.  Their UE10000 is( 	still a nightmare with ecahce blowouts.  ; 	The article is a very good read as there are a ton of gems ( 	regarding just what is wrong with eBay.   				Robc   ------------------------------    Date: 13 Jan 2001 02:20:43 +0800, From: Paul Repacholi <prep@prep.synonet.com>% Subject: Kitchen sink. ( not emacs! )h3 Message-ID: <87y9wgpr7o.fsf_-_@k9.prep.synonet.com>l  9 "David J. Dachtera" <djesys.nospam@earthlink.net> writes:n  , > fabio_compaq@ep-bc.petrobras.com.br wrote: > > F > > Do you know if OpenVMS will support Microsoft Active Directory ???  E > With appropriate layered software, OpenVMS will support the kitchen0 > sink!0  
 "With ..."  % Not on LDAP and the like I know, but;   A Go to http://www.ntp.org.  NTP is Network Time Protocol, a methodr? for disseminating accurate time across a net work. Pretty basicl stuff knowing what day it is.9  ? Go to the software link. Compare what is listed for VMS and for1	 gateshit.i  = Yes, VMS does have NTP. Running on the machine I'm typing on.R= But you can't run it as a primary, or from any external clock ? except DTSS. The hoard of windows stuff, it pretty much 3 truckc= loads of shit. There is almost nothing there that isn't in NT76 NTP or a win client. But notice where the numbers are.   Same old story isn't it...   -- :< Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda.o@                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.   ------------------------------  % Date: Fri, 12 Jan 2001 09:48:18 -0500t0 From: Paul Anderson <paul.r.anderson@compaq.com>" Subject: Re: LN29 escape sequencesC Message-ID: <paul.r.anderson-4FFEE3.09481812012001@news.compaq.com>e  E In article <OF180105A1.4EE6B9C8-ON882569D2.0000F587@foundation.com>, k" Shane.F.Smith@Healthnet.com wrote:  G > I've been asked to try and locate the escape sequences understood by - > an LN29 printer.  C Are you sure you mean an LN29 printer?  I've never heard of such a n0 printer.  Is it a line, matrix or laser printer?   Paul   -- a,    Paul Anderson, OpenVMS Engineering (DCPS),    Compaq Computer Corporation, Littleton MA   ------------------------------  % Date: Fri, 12 Jan 2001 12:47:28 -0000 > From: Nicholas  Mark de Smith <Nicholas.MarkdeSmith@liffe.com>" Subject: Loads of Alphas (700+)...M Message-ID: <FCB2387D053DD31180B200805FEA4C3F24F001@ntprdex5.admin.liffe.com>g  G Not that I use them much anymore, but in this week trade rags in the UKeF there is an article about the Sanger Centre (public genome sequencing)  in Cambridge (the original one).  ' Aparently, they have over 700 Alphas...h   Nick nick at desmith dot neta   ------------------------------  # Date: Fri, 12 Jan 2001 15:25:46 GMTo4 From: "Terry C. Shannon" <terryshannon@mediaone.net>& Subject: Re: Loads of Alphas (700+)...< Message-ID: <_lF76.44815$1M.9924454@typhoon.ne.mediaone.net>  J "Nicholas Mark de Smith" <Nicholas.MarkdeSmith@liffe.com> wrote in messageG news:FCB2387D053DD31180B200805FEA4C3F24F001@ntprdex5.admin.liffe.com...eI > Not that I use them much anymore, but in this week trade rags in the UK H > there is an article about the Sanger Centre (public genome sequencing)" > in Cambridge (the original one). > ) > Aparently, they have over 700 Alphas...  >V  H Indeed they do. The Sanger Centre is the UK's answer to Celera Genomics.   ------------------------------  % Date: Sat, 13 Jan 2001 06:26:13 +0000-+ From: "antonio.carlini" <arcarlini@iee.org> & Subject: Re: Loads of Alphas (700+)...' Message-ID: <3A5FF505.3C81739E@iee.org>p   "Terry C. Shannon" wrote:b > L > "Nicholas Mark de Smith" <Nicholas.MarkdeSmith@liffe.com> wrote in messageI > news:FCB2387D053DD31180B200805FEA4C3F24F001@ntprdex5.admin.liffe.com...tK > > Not that I use them much anymore, but in this week trade rags in the UK J > > there is an article about the Sanger Centre (public genome sequencing)$ > > in Cambridge (the original one). > >v+ > > Aparently, they have over 700 Alphas...m > >m > J > Indeed they do. The Sanger Centre is the UK's answer to Celera Genomics.  ' And, IIRC, Celera are similarly stuffedr to the gills with Alphas.    Antoniog     -- y   --------------- - Antonio Carlini             arcarlini@iee.org    ------------------------------  % Date: Fri, 12 Jan 2001 11:37:28 -0500-+ From: Clayton Kroh <ckroh@MAIL.CLARION.EDU>23 Subject: Minimum cluster size in a 36GB striped sets+ Message-ID: <3A5F32C8.E6E4CD67@clarion.edu>c   Environment:   Alpha 4100, OpenVMS 7.1, HSJ50  Z On the HSJ50 I created a 36 GB striped set using two 18 GB drives. By default, the clusterX size is set to 69 when the drive is initialized so when I copied 7 GB of data to the newP striped set it ended up taking 17 GB (ouch!) leaving only 19 GB free. I tried toS initialize the striped set with initialize/cluster_size=(anything < 69)  $2$dua903:y5 userss  and it will not accept anything less than 69.t  W Is this truly that smallest cluster size I can use due to the geometry of the two 18 GB1W drives? This striped set contains mostly very small mail files for our students so this ( seems like it is a major waste of space.  T Could I stripe four 9 GB drives and be able to use a smaller cluster size to be more& efficient and not waste so much space?  Z When I backup this new 36 GB striped set, will I actually be copying 17 GB of data to tapeZ (1/2 of which would just be empty space because of the large cluster size) or will it onlyG copy the true 7 GB of real data? (Meaning will I need one tape or two.)a     Thanks in advanced.    -- Clayton Kroh Manager, Enterprise Services Clarion University of PA ckroh@clarion.eduE (814) 393-1673   ------------------------------  % Date: Fri, 12 Jan 2001 16:47:26 +0000t5 From: "Steeples, Oliver" <Oliver.Steeples@compaq.com>m7 Subject: RE: Minimum cluster size in a 36GB striped set(N Message-ID: <F498D199EDB12D468CD2C66680D3080116C36F@reoexc04.emea.cpqcorp.net>  K Have you tried an init on the HSJ50 setting the chunksize to a lower value?2   Regards8   Oliver Steeples  Storage Technical Specialist! Technical Customer Support Center  Compaq Computers Limited       -----Original Message-----2 From: Clayton Kroh [mailto:ckroh@MAIL.CLARION.EDU]& Sent: Friday, January 12, 2001 4:37 PM To: Info-VAX@Mvb.Saic.ComV3 Subject: Minimum cluster size in a 36GB striped setp     Environment:   Alpha 4100, OpenVMS 7.1, HSJ50  E On the HSJ50 I created a 36 GB striped set using two 18 GB drives. Bye default, the clusterH size is set to 69 when the drive is initialized so when I copied 7 GB of data to the newnG striped set it ended up taking 17 GB (ouch!) leaving only 19 GB free. I. tried toG initialize the striped set with initialize/cluster_size=(anything < 69)-
 $2$dua903:5 userss  and it will not accept anything less than 69.w  I Is this truly that smallest cluster size I can use due to the geometry ofg
 the two 18 GB1F drives? This striped set contains mostly very small mail files for our students so this( seems like it is a major waste of space.  L Could I stripe four 9 GB drives and be able to use a smaller cluster size to be more & efficient and not waste so much space?  J When I backup this new 36 GB striped set, will I actually be copying 17 GB of data to tape J (1/2 of which would just be empty space because of the large cluster size) or will it only.G copy the true 7 GB of real data? (Meaning will I need one tape or two.)0     Thanks in advanced.<   -- Clayton Kroh Manager, Enterprise Services Clarion University of PA ckroh@clarion.edu8 (814) 393-1673   ------------------------------  % Date: Fri, 12 Jan 2001 12:18:07 -0500 + From: Clayton Kroh <ckroh@MAIL.CLARION.EDU>o7 Subject: Re: Minimum cluster size in a 36GB striped setm* Message-ID: <3A5F3C4F.53CAAC2@clarion.edu>  W No I didn't. The chunksize is 256 on both the old 8 GB stripe set (two 4 GB drives) andnZ the new 36 GB stripe set (two 18 GB drives) so I didn't know if changing it would make any difference.h       "Steeples, Oliver" wrote:.  M > Have you tried an init on the HSJ50 setting the chunksize to a lower value?r > 	 > Regardse >t > Oliver Steepless > Storage Technical Specialist# > Technical Customer Support Center  > Compaq Computers Limited >m > -----Original Message-----4 > From: Clayton Kroh [mailto:ckroh@MAIL.CLARION.EDU]( > Sent: Friday, January 12, 2001 4:37 PM > To: Info-VAX@Mvb.Saic.Coma5 > Subject: Minimum cluster size in a 36GB striped set: >  > Environment: >M  > Alpha 4100, OpenVMS 7.1, HSJ50 >nG > On the HSJ50 I created a 36 GB striped set using two 18 GB drives. ByS > default, the clusterJ > size is set to 69 when the drive is initialized so when I copied 7 GB of > data to the newcI > striped set it ended up taking 17 GB (ouch!) leaving only 19 GB free. Ie
 > tried toI > initialize the striped set with initialize/cluster_size=(anything < 69)n > $2$dua903:7 > userss  and it will not accept anything less than 69.t > K > Is this truly that smallest cluster size I can use due to the geometry oft > the two 18 GBNH > drives? This striped set contains mostly very small mail files for our > students so this* > seems like it is a major waste of space. >iN > Could I stripe four 9 GB drives and be able to use a smaller cluster size to	 > be morei( > efficient and not waste so much space? >8L > When I backup this new 36 GB striped set, will I actually be copying 17 GB > of data to tapedL > (1/2 of which would just be empty space because of the large cluster size) > or will it onlylI > copy the true 7 GB of real data? (Meaning will I need one tape or two.)e >r > Thanks in advanced.e >n > -- > Clayton Kroh > Manager, Enterprise Services > Clarion University of PA > ckroh@clarion.edue > (814) 393-1673   -- Clayton Kroh Manager, Enterprise Services Clarion University of PA ckroh@clarion.edud (814) 393-1673   ------------------------------  % Date: Fri, 12 Jan 2001 17:24:05 +0000,5 From: "Steeples, Oliver" <Oliver.Steeples@compaq.com>o7 Subject: RE: Minimum cluster size in a 36GB striped set N Message-ID: <F498D199EDB12D468CD2C66680D3080116C371@reoexc04.emea.cpqcorp.net>  J It might be worth trying it if you can.  The downside is you will lose the5 data on the disk when you do the init dx/chunksize=xx        -----Original Message-----2 From: Clayton Kroh [mailto:ckroh@MAIL.CLARION.EDU]& Sent: Friday, January 12, 2001 5:18 PM To: Steeples, Oliver Cc: Info-VAX@Mvb.Saic.Comu7 Subject: Re: Minimum cluster size in a 36GB striped sety    K No I didn't. The chunksize is 256 on both the old 8 GB stripe set (two 4 GBv drives) andoK the new 36 GB stripe set (two 18 GB drives) so I didn't know if changing it- would make any difference.J       "Steeples, Oliver" wrote:m  F > Have you tried an init on the HSJ50 setting the chunksize to a lower value? > 	 > Regardsb >i > Oliver Steeples  > Storage Technical Specialist# > Technical Customer Support Centere > Compaq Computers Limited >s > -----Original Message-----4 > From: Clayton Kroh [mailto:ckroh@MAIL.CLARION.EDU]( > Sent: Friday, January 12, 2001 4:37 PM > To: Info-VAX@Mvb.Saic.Comt5 > Subject: Minimum cluster size in a 36GB striped seth >o > Environment: >I  > Alpha 4100, OpenVMS 7.1, HSJ50 >hG > On the HSJ50 I created a 36 GB striped set using two 18 GB drives. By  > default, the clusterJ > size is set to 69 when the drive is initialized so when I copied 7 GB of > data to the newlI > striped set it ended up taking 17 GB (ouch!) leaving only 19 GB free. Ii
 > tried toI > initialize the striped set with initialize/cluster_size=(anything < 69)o > $2$dua903:7 > userss  and it will not accept anything less than 69.i >IK > Is this truly that smallest cluster size I can use due to the geometry of  > the two 18 GB H > drives? This striped set contains mostly very small mail files for our > students so this* > seems like it is a major waste of space. >hK > Could I stripe four 9 GB drives and be able to use a smaller cluster size  to	 > be moret( > efficient and not waste so much space? >iL > When I backup this new 36 GB striped set, will I actually be copying 17 GB > of data to tapehL > (1/2 of which would just be empty space because of the large cluster size) > or will it onlyaI > copy the true 7 GB of real data? (Meaning will I need one tape or two.)l >i > Thanks in advanced.- >- > -- > Clayton Kroh > Manager, Enterprise Services > Clarion University of PA > ckroh@clarion.edut > (814) 393-1673   -- Clayton Kroh Manager, Enterprise Services Clarion University of PA ckroh@clarion.eduv (814) 393-1673   ------------------------------    Date: 12 Jan 2001 18:25:30 +0100G From: Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de>l7 Subject: Re: Minimum cluster size in a 36GB striped setdH Message-ID: <y4wvc0lm2d.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>  N You need to upgrade to V7.2, which will allow you to use smaller cluster sizesM on large disks because the size of BITMAP.SYS can be larger than 255 blocks. VB IIRC, that requires it to be an ODS-5 volume (there were also someJ improvements that apply to ODS-2 volumes). The alternative is to break theI whole thing into smaller logical volumes, for which there are a number ofo possible approaches.   	Jan   ------------------------------    Date: 12 Jan 2001 13:18:07 -0500* From: young_r@eisner.decus.org (Rob Young)7 Subject: RE: Minimum cluster size in a 36GB striped setr+ Message-ID: <HidzOjjEhZ9n@eisner.decus.org>t   In article <F498D199EDB12D468CD2C66680D3080116C36F@reoexc04.emea.cpqcorp.net>, "Steeples, Oliver" <Oliver.Steeples@compaq.com> writes:   HSJ50 manual naming...   	Make that:l  " 	HSJ50 and HDS50 Array Controllers 	CLI Reference Manualt   	HSJ50 Array Controller	 	Configuration Manual   A 	I've been working with these for many years and just now noticedg> 	the names buried on the front cover just above the jet (lower? 	right corner).  Now I know where to look for the names.  Can't ; 	believe I missed that but it doesn't jump right out at ya!t   				Robt   ------------------------------   Date: 12 Jan 2001 18:11:00 GMT' From: david20@alpha2.mdx.ac.uk (D.Webb)i7 Subject: Re: Minimum cluster size in a 36GB striped set 0 Message-ID: <93nhbk$bgb$1@aquila.news.mdx.ac.uk>  Y In article <3A5F32C8.E6E4CD67@clarion.edu>, Clayton Kroh <ckroh@MAIL.CLARION.EDU> writes:a
 >Environment:3 >  >Alpha 4100, OpenVMS 7.1, HSJ50  >u[ >On the HSJ50 I created a 36 GB striped set using two 18 GB drives. By default, the cluster/Y >size is set to 69 when the drive is initialized so when I copied 7 GB of data to the new Q >striped set it ended up taking 17 GB (ouch!) leaving only 19 GB free. I tried tohT >initialize the striped set with initialize/cluster_size=(anything < 69)  $2$dua903:6 >userss  and it will not accept anything less than 69. >0X >Is this truly that smallest cluster size I can use due to the geometry of the two 18 GBX >drives? This striped set contains mostly very small mail files for our students so this) >seems like it is a major waste of space.i >e  L Yes it is a major waste of space. Unfortunately the only way around it is toL upgrade to VMS 7.2 (or better still VMS 7.2-1) where these ridiculous limits are greatly relaxed. dM With VMS 7.2 and later  you can have a cluster size of 1 for a 36GB and much  
 larger disks.a    U >Could I stripe four 9 GB drives and be able to use a smaller cluster size to be more ' >efficient and not waste so much space?a >e  L No this would make no difference - what counts is the total size of the disk presented to VMS.   O On our Academic cluster I had a number of 18GB disks which (when I was running tK VMS 7.1) I partitioned so that they appeared as 4 separate approx 4GB diskse3 using the HSZ50 so as not to waste space like this.dL Now I am running VMS 7.2-1 this is no longer a problem so I have removed the- partitioning and they are back as 18GB disks.t    J By the way what would you do to protect this proposed stripe set of 4 9GB = drives. Would you be shadowing it with another set of disks ?oO 4 drives striped = 4 times the chance of losing all your data when a disk driverO goes bad. I would be looking at shadowing or one of the controller raid optionsa for resilience.   N For that matter do you really need to stripe these disks into one large disk ?D Would splitting the files across separate disks not provide adequateC performance with any disk problem only affecting part of the data ?-  O If you do create one large stripe-set how are you going to create the directory<N structure on this disk. If you create all the users accounts at the top-level G you may have performance problems because of the size of the top level m> directory. You may need to create a number of sub-directories I [.usr1], [.usr2], [.usr3]  etc and split your user accounts amongst them.r  N (since you mention student mail files I am assuming that the disk will containM a large number of student accounts each of which contains their mail.mai filet etc).w  p  
 David Webb VMS and Unix team leader CCSS Middlesex University    [ >When I backup this new 36 GB striped set, will I actually be copying 17 GB of data to tapee[ >(1/2 of which would just be empty space because of the large cluster size) or will it only H >copy the true 7 GB of real data? (Meaning will I need one tape or two.) >y >t >Thanks in advanced. >. >--h
 >Clayton Kroh  >Manager, Enterprise Services  >Clarion University of PAo >ckroh@clarion.edu >(814) 393-1673T >r >w   ------------------------------    Date: 12 Jan 2001 13:12:06 -0500* From: young_r@eisner.decus.org (Rob Young)7 Subject: RE: Minimum cluster size in a 36GB striped set + Message-ID: <2FbYWSwZkapP@eisner.decus.org>    In article <F498D199EDB12D468CD2C66680D3080116C36F@reoexc04.emea.cpqcorp.net>, "Steeples, Oliver" <Oliver.Steeples@compaq.com> writes:M > Have you tried an init on the HSJ50 setting the chunksize to a lower value?a   	????u > 	 > Regards  >  > Oliver Steeplesg > Storage Technical Specialist# > Technical Customer Support Centera > Compaq Computers Limited >    >  > Environment: >   > Alpha 4100, OpenVMS 7.1, HSJ50 > G > On the HSJ50 I created a 36 GB striped set using two 18 GB drives. Byt > default, the clusterJ > size is set to 69 when the drive is initialized so when I copied 7 GB of > data to the new I > striped set it ended up taking 17 GB (ouch!) leaving only 19 GB free. In
 > tried toI > initialize the striped set with initialize/cluster_size=(anything < 69)- > $2$dua903:7 > userss  and it will not accept anything less than 69.k > K > Is this truly that smallest cluster size I can use due to the geometry ofe > the two 18 GB7H > drives? This striped set contains mostly very small mail files for our > students so this* > seems like it is a major waste of space. > N > Could I stripe four 9 GB drives and be able to use a smaller cluster size to	 > be morel( > efficient and not waste so much space? > L > When I backup this new 36 GB striped set, will I actually be copying 17 GB > of data to tape1L > (1/2 of which would just be empty space because of the large cluster size) > or will it onlysI > copy the true 7 GB of real data? (Meaning will I need one tape or two.)  >  >  	a@ 	What are you trying to do?  Relocate all student directories to? 	one disk for manageability?  You realize with a stripe set youm@ 	have less MTBF?  One drive goes, you are pulling a backup tape.  @ 	The reason I ask: "what are you trying to do" is there are someE 	ways to go better than others.  If you are after better performance,lF 	you might want to take those 9 gig drives, create a RAID5 , partition? 	the RAID5 into 4 seperate units.  That should leave you with 4-B 	drives/disks at the OS level of approximately 5 gigs each and youF 	can then use a much smaller cluster size.  There are several benefits= 	to this.  You have data protection underneath (RAID5), greato? 	read response and with writeback caching turned on good write c
 	response.  @ 	Do you have the manuals?  HSJ50 and HSD50 Array Controllers and@ 	HSJ50 Array Controller?  (They must have been tipping the sauce7 	when they named those manuals ... *slightly* kidding).u   				Robr   ------------------------------   Date: 12 Jan 2001 18:28:17 GMT' From: david20@alpha2.mdx.ac.uk (D.Webb) 7 Subject: Re: Minimum cluster size in a 36GB striped setn0 Message-ID: <93nic1$bnl$1@aquila.news.mdx.ac.uk>   In article <y4wvc0lm2d.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>, Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de> writes:O >You need to upgrade to V7.2, which will allow you to use smaller cluster sizes N >on large disks because the size of BITMAP.SYS can be larger than 255 blocks. C >IIRC, that requires it to be an ODS-5 volume (there were also somenK >improvements that apply to ODS-2 volumes). The alternative is to break theBJ >whole thing into smaller logical volumes, for which there are a number of >possible approaches.n >a >	Jano  I No it does not require ODS-5. With VMS 7.2 and ODS-2 volumes you can set (I a smaller cluster-size. I can't remember the maximum disk size which will B allow a cluster-size of 1 but it is certainly in the 100GB+ range.  ? I have a number of 18GB ODS-2 volumes with a cluster size of 1.l     From VMS 7.2-1 HELP    V72_FEATURES     System_Management_Features       Bitmap_Limits_Increased_  E          In previous versions of OpenVMS, both storage and index filerH          bitmaps were limited to 255 blocks. This size, in turn, limitedC          a volume to approximately one million allocation units, or?F          clusters. Larger disks were required to have a larger clusterB          factor to accommodate the limit; for example, a 9 GB disk)          required a cluster factor of 18.n  F          Beginning with OpenVMS Version 7.2, the limits of storage and;          index file bitmaps have been increased as follows:t  !          Type of Bitmap     Limito  (          Storage bitmap     65535 blocks'          Index file bitmap  4095 blocksr  C          The increased bitmap limits have the following advantages:e  H          o  For the foreseeable future, they allow you to use space more)             efficiently with small files.1  H          o  They increase the number of files allowed on a volume to the>             architectural maximum of approximately 16 million.  H          The behaviors of the INITIALIZE and BACKUP commands reflect theJ          larger bitmap sizes. On OpenVMS Version 7.1 and earlier versions,G          you cannot mount a volume with bitmaps larger than 255 blocks.oH          For this reason, the default cluster factor that the INITIALIZEG          command chooses on Files-11 structure level 2 disks limits theID          storage bitmap to 255 blocks. However, now you can overrideH          the default with an explicit cluster factor to make the storage%          bitmap as large as you want.     
 David Webb VMS and Unix team leader CCSS Middlesex University   ------------------------------   Date: 12 Jan 2001 18:33:17 GMT' From: david20@alpha2.mdx.ac.uk (D.Webb)o7 Subject: RE: Minimum cluster size in a 36GB striped seta0 Message-ID: <93nild$bnl$2@aquila.news.mdx.ac.uk>   In article <F498D199EDB12D468CD2C66680D3080116C371@reoexc04.emea.cpqcorp.net>, "Steeples, Oliver" <Oliver.Steeples@compaq.com> writes:K >It might be worth trying it if you can.  The downside is you will lose thec6 >data on the disk when you do the init dx/chunksize=xx >p >?   NO.n  = The chunksize has nothing whatsoever to do with this problem.1  C A single 36GB disk would have exactly the same cluster-size limits.   
 David Webb VMS and Unix team leader CCSS Middlesex University       >? >-----Original Message-----e3 >From: Clayton Kroh [mailto:ckroh@MAIL.CLARION.EDU]r' >Sent: Friday, January 12, 2001 5:18 PM  >To: Steeples, Oliver@ >Cc: Info-VAX@Mvb.Saic.Com8 >Subject: Re: Minimum cluster size in a 36GB striped set >1 >8L >No I didn't. The chunksize is 256 on both the old 8 GB stripe set (two 4 GB >drives) andL >the new 36 GB stripe set (two 18 GB drives) so I didn't know if changing it >would make anye >difference. >h >l >  >"Steeples, Oliver" wrote: > G >> Have you tried an init on the HSJ50 setting the chunksize to a lowern >value?o >>
 >> Regards >> >> Oliver Steeples >> Storage Technical Specialistu$ >> Technical Customer Support Center >> Compaq Computers Limitedr >> >> -----Original Message-----p5 >> From: Clayton Kroh [mailto:ckroh@MAIL.CLARION.EDU]m) >> Sent: Friday, January 12, 2001 4:37 PMt >> To: Info-VAX@Mvb.Saic.Com6 >> Subject: Minimum cluster size in a 36GB striped set >> >> Environment:  >>! >> Alpha 4100, OpenVMS 7.1, HSJ50X >>H >> On the HSJ50 I created a 36 GB striped set using two 18 GB drives. By >> default, the cluster K >> size is set to 69 when the drive is initialized so when I copied 7 GB ofi >> data to the newJ >> striped set it ended up taking 17 GB (ouch!) leaving only 19 GB free. I >> tried to J >> initialize the striped set with initialize/cluster_size=(anything < 69)
 >> $2$dua903:o8 >> userss  and it will not accept anything less than 69. >>L >> Is this truly that smallest cluster size I can use due to the geometry of >> the two 18 GBI >> drives? This striped set contains mostly very small mail files for ouru >> students so thisy+ >> seems like it is a major waste of space.d >>L >> Could I stripe four 9 GB drives and be able to use a smaller cluster size >to 
 >> be more) >> efficient and not waste so much space?  >>M >> When I backup this new 36 GB striped set, will I actually be copying 17 GB  >> of data to tapeM >> (1/2 of which would just be empty space because of the large cluster size)g >> or will it onlyJ >> copy the true 7 GB of real data? (Meaning will I need one tape or two.) >> >> Thanks in advanced. >> >> --e >> Clayton Krohl >> Manager, Enterprise Servicesa >> Clarion University of PA  >> ckroh@clarion.edu >> (814) 393-1673a >e >--t
 >Clayton Krohs >Manager, Enterprise Services  >Clarion University of PA  >ckroh@clarion.edu >(814) 393-1673, >t   ------------------------------  % Date: Fri, 12 Jan 2001 08:42:42 -08006& From: "Horst Hinz" <hhinz@bcmhs.bc.ca> Subject: mntverifytimeout45 Message-ID: <ZrG76.133705$47.2069454@news.bc.tac.net>o  G I am VERY new to VMS, we just lost our VMS admin so I'm struggling heregC quite a bit.  One of my disks has a mntverifytimeout as the status,eH obviously the disk is unavailable to me right now.  What does this errorL mean, and do I have to reset my VMS system to fix the problem (assuming it's not hardware related)?   Thanks,r   Horsto   ------------------------------  # Date: Fri, 12 Jan 2001 17:19:07 GMTe1 From: "Mark D. Jilson" <jilly@clarityconnect.com>R Subject: Re: mntverifytimeoutG2 Message-ID: <3A5F3DA3.2276DF84@clarityconnect.com>  G It means that VMS lost connection to the disk for longer than MVTIMEOUTtF amount of time.  Do a DISMOUNT/ABORT $n$ddcu: and then try to mount it? again making sure you use the /NOASSIST qualifier on the mount.u   Horst Hinz wrote:  > I > I am VERY new to VMS, we just lost our VMS admin so I'm struggling hereeE > quite a bit.  One of my disks has a mntverifytimeout as the status,nJ > obviously the disk is unavailable to me right now.  What does this errorN > mean, and do I have to reset my VMS system to fix the problem (assuming it's > not hardware related)? > 	 > Thanks,u >  > Horstm   -- rD Jilly	- Working from Home in the Chemung River Valley - Lockwood, NY0 	- jilly@clarityconnect.com			- Brett Bodine fan. 	- Mark.Jilson@Compaq.com			- since 1975 or so, 	- http://www.jilly.baka.com               -   ------------------------------   Date: 12 Jan 2001 12:42:57 GMT* From: helbig@astro.rug.nl (Phillip Helbig)- Subject: Re: Modifications to directory files . Message-ID: <93mu4h$kv2$2@info.service.rug.nl>  F In article <3Iv76.30207$Ip3.1156862@typhoon.mn.mediaone.net>, Ed Wilts <ewilts@mediaone.net> writes:   L > Sometime in recent history, I started noticing that the modification date M > of some of my directory files started changing.  This certainly never used fJ > to be the case, since it sure does screw up the backups.  I'm currently N > running V7.1-2 on AXP with most high-priority patches (sys-0200 excepted of 
 > course). > J > I haven't detected the pattern yet.  One of my developers detected that M > when you create enough files in a directory for it to be extended, the mod cM > date changes, but I'm sure other activities are doing it too since some of lH > my directories are fairly stagnant (like the dsnlink_new_dir that has # > version limits on its log files).  > J > Any guesses as what might be doing this?  I'm seeing random directories N > scattered around, including my own personal home directory that has had its K > directory modified at 1am when I've not been around so has had almost no -= > activity (perhaps a batch job triggered by Deliver though).0 > M > Before I scratch all my hair off, I'd like a few ideas as to what I should : > be looking for.r  E There was a change a while back so that ALL files in a directory are nI backed up when the modification date of the directory changes.  A common oD way to change this is to change the protection.  I was surprised at D first, but realise that it makes sense that this is the default.  I 7 believe there is a switch to restore the old behaviour.t  I Say a directory is unreadable and you backup [...]*.*.  It won't back up oH the files.  If the directory protection is changed so that you can read F it, then now you can back up the files---and should.  Not all backups # are done from accounts with BYPASS.,   ------------------------------   Date: 12 Jan 2001 18:42:36 GMT2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)- Subject: Re: Modifications to directory files 6 Message-ID: <93nj6s$775$1@mailint03.im.hou.compaq.com>  d In article <3Iv76.30207$Ip3.1156862@typhoon.mn.mediaone.net>, Ed Wilts <ewilts@mediaone.net> writes:J :...I started noticing that the modification date of some of my directory  :files started changing...> :...I'd like a few ideas as to what I should be looking for...  J   The directory revision date changes when you rename the directory, when H   a sufficient number of entries are added or removed or the file itselfG   is truncated -- enough shuffling of the contents that the header must K   be updated, and when you change an attribute of the file header or alter  ?   an ACL.  (OpenVMS does not update the date for every change.)   H   Expect to see the revision date change (in a future release) when you K   alter more than "n" entries and when the directory file grows and expect uK   to see the addition of a last-access date -- this part of various changesi   including the DII COE work.  t   N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------    Date: 12 Jan 2001 06:53:32 -05009 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen) D Subject: Non-Disclosure Agreements (was: OpenVMS + Active Directory)+ Message-ID: <15qlgOgM0CTW@eisner.decus.org>8  ` In article <NSeO79W4q35v@malvm1.mala.bc.ca>, nothome@spammers.are.scum (Malcolm Dunnett) writes:. > In article <3A5E7661.DD2AE52D@rtfmcsi.com>, 2 >     Chuck Chopp <ChuckChopp@rtfmcsi.com> writes: >> 8L >> I'd be interested in being able to integrate OpenVMS with NDS eDirectory.G >> It should be feasible to use the external authentication features to 5 >> validate usernames & passwords through eDirectory.n >> /@ >     Rumour has it that a more general facility to interface toC > external authenticators is coming, but finding out anything about B > how it will work appears next to impossible at this time. RuslanC > Laishev tried for several months last year to get an NDA executedeE > to get this information - as far as I know he was never successful.J  D Presumably if he were successful in executing an NDA (Non-DisclosureB Agreement) he would be prohibited from disclosing that fact by the terms of the Agreement.n  D At least that was the way the old DEC Non-Disclosure Agreements were worded.    ------------------------------    Date: 12 Jan 2001 07:51:42 -07001 From: nothome@spammers.are.scum (Malcolm Dunnett) H Subject: Re: Non-Disclosure Agreements (was: OpenVMS + Active Directory), Message-ID: <RSU+Bv2KZuxY@malvm1.mala.bc.ca>  , In article <15qlgOgM0CTW@eisner.decus.org>, ?     Kilgallen@eisner.decus.org.nospam (Larry Kilgallen) writes:r > F > Presumably if he were successful in executing an NDA (Non-DisclosureD > Agreement) he would be prohibited from disclosing that fact by the > terms of the Agreement.C > C    Good point. Before it was signed he was certainly free to say henF hadn't been able to get one though and that went on for quite a while.  B    All I can say for certain then is that he tried for quite a few? months and got nowhere, he may have ultimately been successful.h  C    The query I sent to the same folks at Compaq on the same subjectkF was never answered, but I got busy with other things and didn't pursue it.    ------------------------------  # Date: Fri, 12 Jan 2001 09:37:10 GMT  From: pdafniotis@hotmail.com> Subject: OpenGL on a Compaq XP1000 w/ Elsa Gloria Synergy card) Message-ID: <93mj86$vck$1@nnrp1.deja.com>f  G I have a Compaq XP1000 with 512MB RAM. I pursuaded management to buy itmG for engineering computations and OpenGL apps. Yes,it does have the ElsatD Gloria Synergy v7.03 card on it. Looking through deja I saw the MANY> posts that say that OpenGL and Elsa do not talk to each other.  7 My question: is this final or is there help on the way?   ) Thank you for reading this. Kind regards,  Petros ---t Dr. Petros DAFNIOTIS Senior Research Engineer pdafniotis@hotmail.com     Sent via Deja.comi http://www.deja.com/   ------------------------------   Date: 12 Jan 2001 16:33:13 GMT2 From: mathog@seqaxp.bio.caltech.edu (David Mathog)B Subject: Re: OpenGL on a Compaq XP1000 w/ Elsa Gloria Synergy card, Message-ID: <93nbk9$opi@gap.cco.caltech.edu>  H In article <93mj86$vck$1@nnrp1.deja.com>, pdafniotis@hotmail.com writes:H >I have a Compaq XP1000 with 512MB RAM. I pursuaded management to buy itH >for engineering computations and OpenGL apps. Yes,it does have the ElsaE >Gloria Synergy v7.03 card on it. Looking through deja I saw the MANY7? >posts that say that OpenGL and Elsa do not talk to each other.o  E The early blurbs for the DS10 and Elsa indicated that OpenGL would be G supported but it never materialized.  I cannot in all honesty recommend I that at this time anybody buy VMS systems for graphics work. For what youmI spent on that XP1000 you could have had a dual processor PIII NT box withoF not only a well supported OpenGL driver but also goodies like hardwareJ stereo, 2X or 4X AGP speed, AND had money left over to buy a DS10 with 512. Mb of memory to run number crunching jobs on.   K I'm not sure if other graphics cards work on that machine under VMS or not.tK Compaq's web site is typically cryptic on the matter.  Some indications for J Powerstorm 300 and 350 support exist on the Compaq web site and the CompaqI system configurator allows a Powestorm 300 for an XP1000, but if you lookn at:   2   http://www.support.compaq.com/open3d/swps300.htm  H they don't say a word about VMS.  So I can't tell you after 5 minutes ofK looking whether or not you can get OpenGL support for that card, on VMS, onyH that XP1000.  (Compaq's web site looks better these days, but it's stillI often ridiculously hard to get answers to these simple sorts of questionsr out of it.)k   Fred, what's the answer?  
 Good luck,   David Mathog mathog@seqaxp.bio.caltech.edu ? Manager, sequence analysis facility, biology division, Caltech     ------------------------------  % Date: Fri, 12 Jan 2001 12:39:07 -0500b5 From: "Fred Kleinsorge" <kleinsorge@star.zko.dec.com>tB Subject: Re: OpenGL on a Compaq XP1000 w/ Elsa Gloria Synergy card, Message-ID: <93nfjf$3ur9$1@lead.zk3.dec.com>  G You need to get a P300 or P350 card for OpenGL support.  Late this yeariK there will be another set of cards with OpenGL support.  The ELSA does not.mJ I could tell you how to get Software GL to work, but it is painfully slow.  G If you have to use the ELSA card and want to use GL, I'd get the MesaGLu freeware (and GLUT).   _Fredi  H pdafniotis@hotmail.com wrote in message <93mj86$vck$1@nnrp1.deja.com>...H >I have a Compaq XP1000 with 512MB RAM. I pursuaded management to buy itH >for engineering computations and OpenGL apps. Yes,it does have the ElsaE >Gloria Synergy v7.03 card on it. Looking through deja I saw the MANY ? >posts that say that OpenGL and Elsa do not talk to each other.a >a8 >My question: is this final or is there help on the way? >s* >Thank you for reading this. Kind regards, >Petrosr >--- >Dr. Petros DAFNIOTISl >Senior Research Engineerd >pdafniotis@hotmail.com  >  >e >Sent via Deja.com >http://www.deja.com/o   ------------------------------  % Date: Thu, 11 Jan 2001 15:22:50 -0500l5 From: "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> E Subject: Re: OpenVMS and DII COE (was Re: IBM + Linux + US$ 1 bilion) , Message-ID: <93n3nh$3qbb$1@lead.zk3.dec.com>  E Everyone take a deep breath.  If you don't know what COE is, then you K probably don't need to know.  If you do, then you probably already know thes important facts.  K - There are two COE platform types, one is POSIX-like systems, the other is  Windows.  K - Only a single vendor has every passed DII/COE certification: Compaq Tru64p UNIX..  A - All other implementations pass because they are the "reference"a implementations.  L - All vendors (including the reference vendors) will eventually have to pass for real.  There willnL   continue to be differences between POSIX-like and Windows implementations.  I - VMS is non-UNIX, and non-Windows, but is considered a POSIX-like system8 for certification6  L - If you have a requirement for COE compliance (which would imply you are in the defenseiI   industry) and need specific details, contact Hoff or I, and we will getn you in touch with thee   appropriate people.y  D - We are deep into the final stages prior to starting certification. Certification is not a simple,H   or quick process.  The certification will be on a "special" V7.2 based version of VMS, that0   eventually will fold back into the mainstream.  L - COE Segments are only "theoretically" portable, and even then, only at the
 source level. I   Segments tend to be built specifically for a target platform, and often  even use non-standarduK   vendor-specific interfaces.  But a correctly designed Segment, which usesc only the standard<>   interfaces for the portability environment will be portable.  L - The COE Kernel provides a common look & feel, and management interfaces so that theD   user will see pretty much the same thing regardless of the vendor.   _Freds  K Hoff Hoffman wrote in message <93l16v$gi8$1@mailint03.im.hou.compaq.com>...  > @ >In article <87g0iqau2x.fsf@k9.prep.synonet.com>, Paul Repacholi <prep@prep.synonet.com> writes:z/ >:"Main, Kerry" <Kerry.Main@compaq.com> writes:n >: >:> Check out:E >:> http://www.openvms.compaq.com/solutions/government/coe/index.htmlh >:F >:Well Kerry, if you do, and go look at the details, you find Solaris,* >:HPUX, NT!! but no VMS. Secure VMS, no... >l; >  The DII COE standards are directly based on Sun Solaris.  >iA >  AFAIK, OpenVMS is the first non-UNIX platform that is actively 0 >  working on DII COE evaluation and compliance. >o< >:Seems there is a way to go. The COE info would tell people >:to NOT use VMS at the moment.h >SG >  OpenVMS is currently implementing the DII COE requirements.  If/whena" >  compliance has been verified... >9H >  [AFAIK, Microsoft Windows NT was "grandfathered" into compliance, butG >  I am told that the next release of the DII COE requirements will not-0 >  include any "grandfathered" implementations.] > D >  Applications that are compliant with DII COE requirements will beD >  directly portable across any DII COE platform, including OpenVMS. > + > --------------------------- pure personalo# opinion ---------------------------o0 >   Hoff (Stephen) Hoffman   OpenVMS Engineering hoffman#xdelta.zko.dec.com >t   ------------------------------  % Date: Fri, 12 Jan 2001 06:40:31 -0300t) From: fabio_compaq@ep-bc.petrobras.com.br  Subject: Re: OpenVMS on CNETL Message-ID: <OF62091AEE.796BA34B-ON032569D2.0034E1A1@ep-bc.petrobras.com.br>  J I hope Compaq is not counting Pathworks licenses because we have more than one hundred ... / but the software is installed in ... none ! ! !    Regards    FC        E "Terry C. Shannon" <terryshannon@mediaone.net> em 11/01/2001 22:50:56i             Info-VAX@Mvb.Saic.Comd       Assunto: Re: OpenVMS on CNET      6 "Christof Brass" <brass@infopuls.com> wrote in message& news:3A620952.EF341C3F@infopuls.com...     >e@ > 10 years ago at a DEC event we were told that 16 Million users > were served by VMS.o  I The party line today is 10M to 12M users, but that apparently is based onvI license units or seats or what-have-you. Still, there are a lot of users.1  I If you haven't czeched out the Garter review on CNET, by all means do so.e
 You can go tou  K http://enterprise.cnet.com/enterprise/0-9566-707-1751615.html?tag=st.it.956d  H and enter your own comments by clicking on "Back to product info" at the topbD of the page adjacent to the word "Gartner," and then on "Submit Your5 Opinion" on the menu that shows up under the summary.y   ------------------------------  % Date: Fri, 12 Jan 2001 07:19:58 -0600e+ From: "Main, Kerry" <Kerry.Main@compaq.com>  Subject: RE: OpenVMS on CNETN Message-ID: <910612C07BCAD1119AF40000F86AF0D805284BEE@kaoexc3.kao.cpqcorp.net>   Fabio,   re: Pathworks ..  6 >>> but the software is installed in ... none ! ! !<<<  J The decision of what to install is always the Customers choice, so if yourK folks have decided to choose a 32bit HW/OS platform over a 64bit HW/OS thatjH does real cluster load balancing, high security (can combine LANMAN withF OpenVMS security) and is not susceptible to the 32bit x86 server basedJ virus's that are around today, then that is certainly your folks decision.   Regards,  
 Kerry Main Senior Consultanth Compaq Canada Inc. Professional Servicesl Voice: 613-592-4660  Fax  :  819-772-7036 Email: Kerry.Main@Compaq.com     -----Original Message-----) From: fabio_compaq@ep-bc.petrobras.com.brt, [mailto:fabio_compaq@ep-bc.petrobras.com.br] Sent: January 12, 2001 4:41 AM To: Info-VAX@Mvb.Saic.Coms Subject: Re: OpenVMS on CNET      J I hope Compaq is not counting Pathworks licenses because we have more than one hundred ...d/ but the software is installed in ... none ! ! !d   Regardsg   FC        E "Terry C. Shannon" <terryshannon@mediaone.net> em 11/01/2001 22:50:56w             Info-VAX@Mvb.Saic.Comn       Assunto: Re: OpenVMS on CNET      6 "Christof Brass" <brass@infopuls.com> wrote in message& news:3A620952.EF341C3F@infopuls.com...     > @ > 10 years ago at a DEC event we were told that 16 Million users > were served by VMS.   I The party line today is 10M to 12M users, but that apparently is based onpI license units or seats or what-have-you. Still, there are a lot of users.u  I If you haven't czeched out the Garter review on CNET, by all means do so.n
 You can go ton  K http://enterprise.cnet.com/enterprise/0-9566-707-1751615.html?tag=st.it.9561  H and enter your own comments by clicking on "Back to product info" at the toplD of the page adjacent to the word "Gartner," and then on "Submit Your5 Opinion" on the menu that shows up under the summary.o   ------------------------------  % Date: Fri, 12 Jan 2001 16:12:20 -0300e) From: fabio_compaq@ep-bc.petrobras.com.brB8 Subject: Oracle and OpenVMS - Oops, I did it again ! ! !L Message-ID: <OF7306C55A.C41C9411-ON032569D2.00695E77@ep-bc.petrobras.com.br>  
 Click here  # http://platforms.oracle.com/compaq/l  B The NT and Tru64 links are working, but the OpenVMS ..... :-((((((     Regards5   FC   ------------------------------  % Date: Fri, 12 Jan 2001 10:18:56 -0800e! From: Shane.F.Smith@Healthnet.comd< Subject: Re: Oracle and OpenVMS - Oops, I did it again ! ! !D Message-ID: <OFD0497DE1.08519798-ON882569D2.00644C54@foundation.com>  H -10000 cool points for the Britney Spears reference. (Oops, -1000 for meJ for recognizing it...) (Oh, and -100 for anyone who actually knows who Ms.I Spears is) (But only -50 if it's just as "that bird in the school uniform  video")d   Shanen  H P.S. We apologise for the interruption in normality. Normal Shane wil be resumed as soon as possible.          = fabio_compaq@ep-bc.petrobras.com.br on 01/12/2001 11:42:20 AMt   To:   Info-VAX@Mvb.Saic.Com  cc:g  9 Subject:  Oracle and OpenVMS - Oops, I did it again ! ! !v    
 Click here  # http://platforms.oracle.com/compaq/,  B The NT and Tru64 links are working, but the OpenVMS ..... :-((((((     Regardse   FC   ------------------------------  % Date: Fri, 12 Jan 2001 16:55:54 -0300g) From: fabio_compaq@ep-bc.petrobras.com.brb< Subject: Re: Oracle and OpenVMS - Oops, I did it again ! ! !L Message-ID: <OFACEA9B47.8EC9A1C2-ON032569D2.006D7019@ep-bc.petrobras.com.br>  , She is here in the megashow Rock in Rio ....   James Taylor is too ! ! !,   :-)i   Regardse   FC        2 Shane.F.Smith@Healthnet.com em 12/01/2001 16:18:56             Info-VAX@Mvb.Saic.ComS      < Assunto: Re: Oracle and OpenVMS - Oops, I did it again ! ! !      H -10000 cool points for the Britney Spears reference. (Oops, -1000 for meJ for recognizing it...) (Oh, and -100 for anyone who actually knows who Ms.I Spears is) (But only -50 if it's just as "that bird in the school uniformn video")s   Shanei  H P.S. We apologise for the interruption in normality. Normal Shane wil be resumed as soon as possible.          = fabio_compaq@ep-bc.petrobras.com.br on 01/12/2001 11:42:20 AM-   To:   Info-VAX@Mvb.Saic.Comn cc:   9 Subject:  Oracle and OpenVMS - Oops, I did it again ! ! !a    
 Click here  # http://platforms.oracle.com/compaq/s  B The NT and Tru64 links are working, but the OpenVMS ..... :-((((((     Regards1   FC   ------------------------------  % Date: Fri, 12 Jan 2001 12:43:22 -0600r1 From: "Dave Gudewicz" <david.gudewicz@abbott.com>m% Subject: Positive Gartner VMS article.8 Message-ID: <93nj0m$ikl$1@fizban.fizban.pprd.abbott.com>   Will wonders never cease?i  K http://enterprise.cnet.com/enterprise/0-9566-707-1751615.html?tag=st.it.956f   Dave...e   ------------------------------  # Date: Fri, 12 Jan 2001 13:33:15 GMTo! From: larryrblackwell@my-deja.com * Subject: powerchute for VMS AXP don't work) Message-ID: <93n12r$9c4$1@nnrp1.deja.com>L  , I am trying to get Powerchute of OpenVMS AXP- 7.2.1 working with no luck. I've read reportsh, where it is broke after 7.1 but yet I see it0 running on 2 of our installations. Installing it) on as AlphaServer DS10. Have followed allf& instructions explicitly (including the0 undocumented 2400 baud) and have tried it on two, different ports. One is the 9pin Serial port/ TTA0: and the other is a 25 pin RS232 Port on ah) Digiboard Accele XR. The program seems to#1 communicate when running APCSETUP and if I unplug-1 the serial cable it wants to shutdown the system.e0 However if I unplug the UPS it does not send out- warning messages or try to shutdown after theA+ interval. I am using the cables sent by APCi# Part # 940-0039A to the 25 pin Port9, Part # 940-0039A with Part # 940-0029A and a- female to female gender changer for the 9 pin<( port. Anybody got any clues? APC doesn't Thanks in advancep     Sent via Deja.com  http://www.deja.com/   ------------------------------  # Date: Fri, 12 Jan 2001 15:06:02 GMTd= From: system@SendSpamHere.ORG (Brian Schenkenberger, VAXman-)>. Subject: Re: powerchute for VMS AXP don't work0 Message-ID: <009F6007.BB545805@SendSpamHere.ORG>  M In article <93n12r$9c4$1@nnrp1.deja.com>, larryrblackwell@my-deja.com writes:a- >I am trying to get Powerchute of OpenVMS AXP . >7.2.1 working with no luck. I've read reports- >where it is broke after 7.1 but yet I see ite1 >running on 2 of our installations. Installing itn* >on as AlphaServer DS10. Have followed all' >instructions explicitly (including the(1 >undocumented 2400 baud) and have tried it on twor- >different ports. One is the 9pin Serial portr0 >TTA0: and the other is a 25 pin RS232 Port on a* >Digiboard Accele XR. The program seems to2 >communicate when running APCSETUP and if I unplug2 >the serial cable it wants to shutdown the system.1 >However if I unplug the UPS it does not send outn. >warning messages or try to shutdown after the, >interval. I am using the cables sent by APC$ >Part # 940-0039A to the 25 pin Port- >Part # 940-0039A with Part # 940-0029A and al. >female to female gender changer for the 9 pin) >port. Anybody got any clues? APC doesn'te >Thanks in advance >  >p >Sent via Deja.com >http://www.deja.com/   2 Perhaps, you need better software and support?  ;)   http://www.tmesis.com/apc/   OK, it's a shameless plug.     --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COMn             O city, n., 1. a place where trees are cut down and streets are named after them.c   ------------------------------    Date: 12 Jan 2001 10:37:16 -05009 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)h. Subject: Re: powerchute for VMS AXP don't work+ Message-ID: <bM5u5yy0n+Ut@eisner.decus.org>7  M In article <93n12r$9c4$1@nnrp1.deja.com>, larryrblackwell@my-deja.com writes:a. > I am trying to get Powerchute of OpenVMS AXP/ > 7.2.1 working with no luck. I've read reports1. > where it is broke after 7.1 but yet I see it2 > running on 2 of our installations. Installing it+ > on as AlphaServer DS10. Have followed alln( > instructions explicitly (including the2 > undocumented 2400 baud) and have tried it on two. > different ports. One is the 9pin Serial port1 > TTA0: and the other is a 25 pin RS232 Port on as+ > Digiboard Accele XR. The program seems toF3 > communicate when running APCSETUP and if I unplugc3 > the serial cable it wants to shutdown the system.-2 > However if I unplug the UPS it does not send out/ > warning messages or try to shutdown after theD- > interval. I am using the cables sent by APCP% > Part # 940-0039A to the 25 pin Portg. > Part # 940-0039A with Part # 940-0029A and a/ > female to female gender changer for the 9 pint* > port. Anybody got any clues? APC doesn't  ? Brian Schenkenberger will know.  He should be along shortly :-)m  N ==============================================================================N Great Inventors of our time: Al Gore -> Internet; Sun Microsystems -> ClustersN ==============================================================================   ------------------------------   Date: 12 Jan 2001 16:44:59 GMT2 From: mathog@seqaxp.bio.caltech.edu (David Mathog). Subject: Re: powerchute for VMS AXP don't work, Message-ID: <93ncab$opi@gap.cco.caltech.edu>  M In article <93n12r$9c4$1@nnrp1.deja.com>, larryrblackwell@my-deja.com writes:lI >I am trying to get Powerchute of OpenVMS AXP 7.2.1 working with no luck.e   From the VMS FAQ:5  K OpenVMS software that can control a Tripp-Lite Uninterruptable Power Supplyp  (UPS) is available from:h   <      http://seqaxp.bio.caltech.edu/pub/SOFTWARE/TCONTROL.ZIP;                                              [David Mathog]p   ;  OnliSafe Software for the PowerWare UPS is available from:aL      http://www.powerware.com/2products/software/onlisafe/onlisafe.htm                       %                        [Janis Cooper]y   G  UPShot web-based software for controlling a UPS is available from:    r%  http://www.tmesis.com/apc/beta.htmlxtD                                              [Brian Schenkenberger]     H TCONTROL is a general serial line controller, but you'd have to grok theK cable for the APC controller.  I think Brian's software is designed to work  with APCs though.h  
 Good luck,    David Mathog mathog@seqaxp.bio.caltech.edu,? Manager, sequence analysis facility, biology division, Caltech     ------------------------------  # Date: Fri, 12 Jan 2001 16:54:07 GMTg= From: system@SendSpamHere.ORG (Brian Schenkenberger, VAXman-) . Subject: Re: powerchute for VMS AXP don't work0 Message-ID: <009F6016.D4CBA518@SendSpamHere.ORG>  g In article <bM5u5yy0n+Ut@eisner.decus.org>, Kilgallen@eisner.decus.org.nospam (Larry Kilgallen) writes:TN >In article <93n12r$9c4$1@nnrp1.deja.com>, larryrblackwell@my-deja.com writes:/ >> I am trying to get Powerchute of OpenVMS AXPd0 >> 7.2.1 working with no luck. I've read reports/ >> where it is broke after 7.1 but yet I see itw3 >> running on 2 of our installations. Installing ith, >> on as AlphaServer DS10. Have followed all) >> instructions explicitly (including them3 >> undocumented 2400 baud) and have tried it on two / >> different ports. One is the 9pin Serial port 2 >> TTA0: and the other is a 25 pin RS232 Port on a, >> Digiboard Accele XR. The program seems to4 >> communicate when running APCSETUP and if I unplug4 >> the serial cable it wants to shutdown the system.3 >> However if I unplug the UPS it does not send outn0 >> warning messages or try to shutdown after the. >> interval. I am using the cables sent by APC& >> Part # 940-0039A to the 25 pin Port/ >> Part # 940-0039A with Part # 940-0029A and ae0 >> female to female gender changer for the 9 pin+ >> port. Anybody got any clues? APC doesn'tn > @ >Brian Schenkenberger will know.  He should be along shortly :-) >bO >==============================================================================fO >Great Inventors of our time: Al Gore -> Internet; Sun Microsystems -> Clusters O >==============================================================================a   :) Thanks Larry,  K I haven't a clue as the to pinouts of the part numbers mentioned.  However,nK I will cauthion against trying any "standard" cables.  The input on the APCdL unit (its 9 pin) is anything but standard.  One fellow I know used a typicalG 9 pin cable and managed to shutdown his systems and APC in short order.   L As I understand it, and its been a while since I looked at it, Powerchute isK not very intelligent on the VMS side.  It simply monitors the equivalent ofpL DTR on the APC and if it drops (indicating it is running no battery) the P/CK software initiates a shutdown.  I can do that in my sleep for far less thanp? $300-$400 or whatever rediculous price they offer it for today.o  L If your are mucking about with gender changers on your interconnect from the2 APC to your VAX/Alpha, you've got the wrong cable.   --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COMs            eO city, n., 1. a place where trees are cut down and streets are named after them.    ------------------------------  # Date: Fri, 12 Jan 2001 18:23:06 GMTs! From: larryrblackwell@my-deja.comc. Subject: Re: powerchute for VMS AXP don't work) Message-ID: <93ni1u$pqb$1@nnrp1.deja.com>e  > I had already visited your web site and the product looks very; interesting. Its obvious that APC hasn't done anything withtG Powerchute/VMS came out. I will look forward to evaluating your product.C in the future, however my need is immediate. I need to install thismD system next week about 500 miles from here so I don't think I can go4 with a beta. I would also like to know your pricing.     Sent via Deja.comy http://www.deja.com/   ------------------------------  % Date: Fri, 12 Jan 2001 06:59:50 -0300n) From: fabio_compaq@ep-bc.petrobras.com.bry Subject: Re: QuixoticVMS?eL Message-ID: <OFFDE26664.D53262B9-ON032569D2.00368C54@ep-bc.petrobras.com.br>  K I think I will move to Europe in a few years . .  .if OpenVMS exists there,  I will be there !I  B We dont have DECUS anymore - in theory ! They said to me they will restructure but Im1 sent them an email and they didnt answer me ! ! !p  K The last big sale of a OpenVMS system I know was to run a billing software,t to a Telecom CompanyH in the states of south of the country - European colonization :-)  ! ! !   Regards    FC                J Kilgallen@eisner.decus.org.nospam (Larry Kilgallen) em 12/01/2001 01:21:33             Info-VAX@Mvb.Saic.Comr       Assunto: Re: QuixoticVMS?     G In article <Wgr76.12508$BI2.3483487@typhoon.ne.mediaone.net>, "Terry C.r, Shannon" <terryshannon@mediaone.net> writes: > Hoff wrote...  >>K >>   While I have no idea where Fabio resides -- nor do I have knowledge of- > the H >>   origin of some of the comments and questions and feedback that have beenD >>   posted and/or attributed to same -- but ".br" is Brazil/Brasil. >>C >>   I have yet to visit South America, and I have not received any  customerC >>   (or internal) requests for OpenVMS visits or OpenVMS events inb	 Brazil...e >> > J > It was three or four years ago, but I keynoted a DECUS Brasil symposium, anduF > there seemed to be a relatively strong interest in OpenVMS among the > attendees.  C So if since that address VMS interest from Brazileans has died off, ! we would blame that on ...    :-)t  N ==============================================================================  E Great Inventors of our time: Al Gore -> Internet; Sun Microsystems ->E ClustersN ==============================================================================   ------------------------------  # Date: Fri, 12 Jan 2001 11:06:01 GMTs1 From: CSABA  HARANGOZO   <csabah@zipworld.com.au>e Subject: Re: QuixoticVMS?h7 Message-ID: <tyB76.350$h4.35116@nostril.pacific.net.au>o  3 Hoff Hoffman <hoffman@xdelta.zko.dec.nospam> wrote:-  v > In article <GPn76.12480$BI2.3466393@typhoon.ne.mediaone.net>, "Terry C. Shannon" <terryshannon@mediaone.net> writes:9 > :<fabio_compaq@ep-bc.petrobras.com.br> wrote in messagei > ..L > :>    7 - There is not a strong / popular usergroup to disseminate the OS. > :-$ > :Are you a member of DECUS Europe?  O >   While I have no idea where Fabio resides -- nor do I have knowledge of the NL >   origin of some of the comments and questions and feedback that have beenC >   posted and/or attributed to same -- but ".br" is Brazil/Brasil.e  L >   I have yet to visit South America, and I have not received any customer L >   (or internal) requests for OpenVMS visits or OpenVMS events in Brazil...  P >  --------------------------- pure personal opinion ---------------------------N >    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com      E 	With the late permission of Hoff, let me paste here one of his posts. 	from a while back... :n    F  "In no particular order, here are some of the salient features of the   OpenVMS operating system...6  I   Multi-user interactive user support, full shared-resources clustering, oG   uptime, robust file system with individual volume limits circa one TB]F   (and thousands of volumes), huge system performance and application D   scaling, networked GUI interface, 64-bit flat virtual addressing, F   built-in tools for local BACKUP and recovery, far better "DLL-like" K   support for modular coding and distributions, built-in record management rH   support (heirarchical database), rolling upgrades of layered products G   and the operating system in cluster environments (continuous uptime, sK   even with software upgrades), built-in batch-print capabilities, support  J   for a variety of networking protocols (COM, SMB, NFS, IP, DECnet, X.25, N   http, etc), remote network system logins, built-in system dump and analysis K   facilities, built-in multi-user security and auditing, built-in multiple oM   and mixed-language programming capabilities, good documentation, orderable aL   CD-ROM source listings of the operating system internals and source code, L   process and engineering that targets the avoidance of introducing securityK   problems and typical virus infection paths, built-in system tuning tools, I   built-in application debugging and application dump tools, support for aE   symmetric multiprocessing environments of up to 32 CPUs in a single2G   system (and up to 96+ nodes in a cluster), built-in standard product  J   installation tools and associated product installation packaging tools, G   a central goal of upward-compatibility of user-mode applications and oI   tools over OpenVMS releases, support for systems with 256 gigabytes of .H   physical memory, available full operating system and hardware support    from a single source, etc. 0  K   Continuous OpenVMS system uptimes of over ten years are not unheard of.   J   Some of these systems -- those configured in clusters -- can be running D   current OpenVMS releases, with very long cluster uptimes, with the   rolling-upgrade capabilties.  N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   (end of quote)    7 	Well, give me this "legacy" operating system any day !e   				Cheers,		Csaba  I    ----------------------------------------------------------------------tE    * Csaba I. Harangozo     |    'To err is human', said the hedgehog-E    * csabah@zipworld.com.au |           as he dismounted a wirebrush..I    -----------------------------------------------------------------------;    EARTH::AUSTRALIA:[SYDNEY]HARANGOZO.CSABA;1, delete? [N]:    ------------------------------    Date: 12 Jan 2001 10:14:52 -0500* From: young_r@eisner.decus.org (Rob Young)I Subject: RAH as default behavior?  (Was: Faster gzip, thank you Paul R.!)D+ Message-ID: <OGPhB43DzZeb@eisner.decus.org>h  V In article <3A5E32A5.43FA904C@GCE.com>, "Glenn C. Everhart" <Everhart@GCE.com> writes:  ; > This is an area that's been a bit of a pet peeve of mine. : > It should be possible to make a process default to using? > readahead/writebehind and have the result basically invisiblen; > to programs; the RMS semantics otherwise are not altered.  >    snip....   > = > A third party implementation has to catch the RMS calls andu7 > set the RAB bit by hand...can be done but it's harderr! > and involves some crufty stuff.p > @ > I suppose a logical could be used also; those translate pretty8 > fast, and the check only needs to be done when the RABA > is initially presented at setup time. "set this mystery logicalo& > in exec mode and all i/o is faster". >   8 	Or get down in the kernel and make it default behavior.  7 	AIX will read ahead starting at minpgahead pages up tog= 	maxpgahead pages.  For instance, if two blocks are satisfied-B 	it then doubles it.  If 4 satisfy (hits) then 8, etc.  As you can< 	imagine , would make for very nice on sequential data, etc.C 	I suppose Oracle folks take advantage of all the filesystem tricksp< 	and surely have RAH stuck in there.  But it sure would make< 	COE go better if certain default behaviors were "in there."   vmtune:  current values:J   -p       -P        -r          -R         -f       -F       -N        -WO minperm  maxperm  minpgahead maxpgahead  minfree  maxfree  pd_npages maxrandwrt2L   52222   208888       2          32        120      128     524288        0   				Robi   ------------------------------    Date: 13 Jan 2001 01:06:20 +0800, From: Paul Repacholi <prep@prep.synonet.com>$ Subject: Re: Repair costs for TK-70?0 Message-ID: <877l40r983.fsf@k9.prep.synonet.com>  , Steve Jones <smj@spamfree.crash.com> writes:  F > Can anyone report recent repair costs for a TK50 or TK70 tape drive?E > Digipaq mail-in service or any of the independent repair shops, thet > more data the better.i > G > I've got at least one dead TK70 and I'm wondering if there's a futurem& > for it other than as a part donor...  > Oh, you are the odd man out. 70 controllers are, um, less than; Methuslaish in their life span. I've about 6 orphan drives. 1 You should not have much trouble finding several.0   -- d< Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda.e@                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.   ------------------------------  % Date: Fri, 12 Jan 2001 00:29:42 -0800:+ From: "Wayne Holland" <wholland@tscnet.com>D Subject: Re: Retired OldtimerrO Message-ID: <1F20C45CD2C778E5.0DB3E36A867F4948.6320017008A800D1@lp.airnews.net>r  6 "Jerry Leslie" <leslie@clio.rice.edu> wrote in message! news:93m7o4$6qo$1@joe.rice.edu...d, > Wayne Holland (wholland@tscnet.com) wrote:
 > : Hello,J > :      I retired in 1993 from vax shop.  The last version of vms we wereL > : using was 4.7.  I finally bought a computer and accessed this newsgroup.E > : I noticed that DEC no longer exists.  What has the world come to?e >lI > Lots of computer companies no longer exist, have merged, or have gotten H > out of the computer business; e.g. Sperry Univac, Burroughs, RCA, GE -/ > oops, they're back in as owners of Honeywell.s >.C > : Can anybody give me the history of things since 1993 concerninge > : the vms O/S? >a0 > Try this section of the VMS FAQ, available at: >C; >     http://www.openvms.compaq.com/wizard/openvms_faq.htmlh >r0 > VMS1.   What is OpenVMS?  What is its history? >r >r6 > --Jerry Leslie     (my opinions are strictly my own)= I'm curious... what does vms have over the new upstart Linux?a   ------------------------------   Date: 12 Jan 2001 13:19:14 GMT* From: helbig@astro.rug.nl (Phillip Helbig) Subject: Re: Retired OldtimerM. Message-ID: <93n08i$lmc$1@info.service.rug.nl>  
 In articleD <0A755A2ED331EF43.7EBE62AABC97BDC2.6AD277ED09AE6DE7@lp.airnews.net>,. "Wayne Holland" <wholland@tscnet.com> writes:    Are you THE Wayne Holland?  N >      I retired in 1993 from vax shop.  The last version of vms we were usingG > was 4.7.  I finally bought a computer and accessed this newsgroup.  IUA > noticed that DEC no longer exists.  What has the world come to?aN > Can anybody give me the history of things since 1993 concerning the vms O/S?  I The ALPHA was introduced at the beginning of the 90s, and VMS was ported eI to it.  Great job.  VAXen were manufactured until recently and are still  F supported.  Of course, most folks moved to ALPHA.  VMS is at 7.2-1 on F ALPHA and 7.2 on VAX, with 7.3 in field test at the moment.  Although F already great in 1993, VMS has had a lot of enhancements since then.  B Also around the beginning of the 90s, DEC lost a lot of customers I through bad marketing.  DEC was bought by Compaq a couple of years ago.  iH Contrary to expectations, marketing has improved, though it still could F be better.  VMS is not going away.  Apart from traditional areas like G stock exchanges, banks, health care etc, VMS is being picked up in new rF areas as well---for example, most mobile phone billing systems run on  VMS.  2 Stay around here and you'll get a lot of answers.    ------------------------------   Date: 12 Jan 2001 18:46:15 GMT2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman) Subject: Re: Retired Oldtimer=6 Message-ID: <93njdn$775$2@mailint03.im.hou.compaq.com>  } In article <0A755A2ED331EF43.7EBE62AABC97BDC2.6AD277ED09AE6DE7@lp.airnews.net>, "Wayne Holland" <wholland@tscnet.com> writes:-M :     I retired in 1993 from vax shop.  The last version of vms we were usingaF :was 4.7.  I finally bought a computer and accessed this newsgroup.  I@ :noticed that DEC no longer exists.  What has the world come to?M :Can anybody give me the history of things since 1993 concerning the vms O/S?   9   Time to add this to the FAQ.  Compaq purchased DIGITAL.   J   Please check the OpenVMS FAQ and the OpenVMS website for general OpenVMSJ   history and information -- the FAQ is posted at www.openvms.compaq.com, E   among other places.  Details on Compaq's acquisition of DIGITAL may H   still be in the newsroom archives at the Compaq (www.compaq.com) site.  J   Please also see the OpenVMS FAQ for information on the hobbyist program,>   on the hobbyist distribution, and various related details...  N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  % Date: Fri, 12 Jan 2001 11:25:13 +0100l/ From: Nigel Arnot <sysmgr@maxwell.ph.kcl.ac.uk>lF Subject: re: Simply Marvelous (was Re: DS20 vs DS20E. Was: Sun Cluster7 Message-ID: <009F6012.CBAA499C.17@maxwell.ph.kcl.ac.uk>h  @ > > > And if it were me, why, I'd consider a color change. Let's  > > > see... pink? Nope. Purple? > > ' > >         Remember PDP Purple???  ;-)0" > Also known as the purple plague. > >    Never heard that.b  D Straying further OT, I do recall that "Purple Plague" was one of twoH "diseases" of integrated circuits back in the TTL MSI era. If I rememberF right, it involved purple crystals developing inside the IC causing itM to fail a year or two after manufacture. It arose from an unexpected chemicali. reaction between (I think) gold and aluminium.  G The other was called "black death" because the crystals were black and o the mortality rate higher.   	Yours,i
 		Nigel Arnoti- 		NRA@MAXWELL.PH.KCL.AC.UK                   v  7 		"In the beginning there was nothing, which exploded."l   ------------------------------   Date: 12 Jan 2001 13:33:49 GMTF From: lederman@star.enet.dec.DISABLE-JUNK-EMAIL.com (Bart Z. Lederman)F Subject: re: Simply Marvelous (was Re: DS20 vs DS20E. Was: Sun Cluster* Message-ID: <93n13t$v19@usenet.pa.dec.com>  i In article <009F6012.CBAA499C.17@maxwell.ph.kcl.ac.uk>, Nigel Arnot <sysmgr@maxwell.ph.kcl.ac.uk> writes:sE >Straying further OT, I do recall that "Purple Plague" was one of two.I >"diseases" of integrated circuits back in the TTL MSI era. If I remember G >right, it involved purple crystals developing inside the IC causing itdN >to fail a year or two after manufacture. It arose from an unexpected chemical/ >reaction between (I think) gold and aluminium.s   Close.  @ Most ICs of the time (and many now) are made of silicon, and theA metal layer on the top that connects things together is aluminum.eB Gold wires were used to connect between the aluminum on the IC andD the wire leads coming out of the package.  Gold could be compressionA (ultrasonically) bonded to aluminum.  For a while, the connectionH? between the gold wire and the aluminum was failing prematurely.T> Examination of the failed bond showed a purple layer: when youB alloy gold and aluminum in the right proportions, you get a purple; metal (looks fairly nice, by the way).  So for a while, thetA problem with the bad bonds was called the "purple plague" (if youy> had a chance to give a problem that kind of a label, you would@ too, wouldn't you?).  However, further analysis showed that that? could not be the problem, because the purple alloy was flexiblen> and conductive.  The problem was eventually tracked down to an> alloy of gold, aluminum, and silicon, which formed if the goldD bonding wire overlapped the side of the aluminum pad it was supposedA to be bonded to, and picked up some silicon that was alloyed intooA the mix.  This formed a brown alloy, which was brittle, and whichy? caused the bonding wires to break off.  But "brown plague" justf> isn't as exciting a term as "purple plague", so the true cause never got much publicity.   > Improved bonding techniques cured this problem quite some time> ago.  But it's typical of the computer industry that some term@ or belief gets into the culture, and never seems to get replaced> with the truth.  I see this in the area I work in most (system= performace), where 'rules of thumb' and various other beliefs < that were developed back when VAXes were new and systems had: 200 MB disk drives and 16 MB of memory still hang on, even3 now that systems have multi-GB of disks and memory.n   --  (  B. Z. Lederman   Personal Opinions Only  8  Posting to a News group does NOT give anyone permission8  to send me advertising by E-mail or put me on a mailing  list of any kind.  5  Please remove the "DISABLE-JUNK-EMAIL" if you have aO5  legitimate reason to E-mail a response to this post.    ------------------------------    Date: 13 Jan 2001 01:30:23 +0800, From: Paul Repacholi <prep@prep.synonet.com>F Subject: Re: Simply Marvelous (was Re: DS20 vs DS20E. Was: Sun Cluster0 Message-ID: <873deor840.fsf@k9.prep.synonet.com>  H lederman@star.enet.dec.DISABLE-JUNK-EMAIL.com (Bart Z. Lederman) writes:  k > In article <009F6012.CBAA499C.17@maxwell.ph.kcl.ac.uk>, Nigel Arnot <sysmgr@maxwell.ph.kcl.ac.uk> writes: G > >Straying further OT, I do recall that "Purple Plague" was one of two1K > >"diseases" of integrated circuits back in the TTL MSI era. If I rememberiI > >right, it involved purple crystals developing inside the IC causing itTP > >to fail a year or two after manufacture. It arose from an unexpected chemical1 > >reaction between (I think) gold and aluminium.    <Bart's good stuff snipped>>  G You are thinking of another problem, one that has become topical again!n  F The 'crystal growth' problem was in Welded Top ICs. The lids where tinC plated on the underside, and after soldering or brazing on where in F fairly high tension. This caused the tin to form long thin crystailineF growths called dendrites. Then they broke off, shorting the bondwires.C CDC was near wiped out by it, near all the 2nd gen Cybers had to bev	 replaced.i  F Tin, Cadmium and Silver are the 3 evils. The first to, due to dendriteE formation. That's why you NiCds go 0 ohm/volts and die after a while. F Silver just LOVES to migrate across surfaces. Silver plated insulators8 are less than wonderfull. It was a real pig with valves.  D Tin also has another nasty. It will diffuse with Gold, and the alloyI formed is very high resistance. ALWAYS make sure your simms are Gold/Gold-" or Tinned/Tinned with the sockets.  C The new topicallity of Tin growing dendrite comes from Nasa. One of:D their space craft has just gone tits up. Cause? Tin wiskers shorting	 relays...o     -- n< Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda.h@                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.   ------------------------------  % Date: Fri, 12 Jan 2001 00:27:58 -0700 1 From: Vance Haemmerle <vance@toyvax.Tucson.AZ.US>l Subject: Re: SPXgt video board3 Message-ID: <3A5E4F8E.34C97C47@toyvax.Tucson.AZ.US>o   aw@mail1.bet1.puv.fi wrote:t > 7 > I just acquired a SPXgt card for my VS4000/60, and itd, > seems to be at least partially functional.  I   If you do ever get it working, don't upgrade to 7.3.   EFT1 and 2 don't K seem to work with this card.  I had no problems at 7.1 or 7.2.  At 7.3, the K DECwindow server doesn't get far enough to even write anything to the error-L log file.  I doubt it will work in the release version either.   Also, don'tN bother trying to set the default visual to Pseudocolor, like the documentationN says you can...  I haven't found a VMS version where this works, at least from 7.1 on.f   --B Vance Haemmerle               Internet   vance@toyvax.Tucson.AZ.USK Tucson, AZ                    Web        http://toyvax.Tucson.AZ.US/~vance/t   ------------------------------  # Date: Fri, 12 Jan 2001 12:12:25 GMTe" From: pa@it.singer-friedlander.com Subject: tcpiptracen) Message-ID: <93msb6$5p5$1@nnrp1.deja.com>   C Could some explain to me what the output from the tcpiptrace means.r- I have tried it between our VAX and a router.d7 I have been experimenting with it and get this output :i  G -----------------------------------------------------------------------tA TCPIPtrace full display RCV packet 185 at 12-JAN-2001 12:08:28.73t  A IP Version = 4,  IHL = 5,  TOS = 00,   Total Length = 56 = ^x0038 0 IP Identifier  = ^xE665,  Flags (0=0,DF=0,MF=0),> Fragment Offset = 0 = ^x0000,   Calculated Offset = 0 = ^x0000D IP TTL = 252 = ^xFC,  Protocol = 1 = ^x01,  Header Checksum = ^x098A& IP Source Address      = 129.100.203.1% IP Destination Address = 129.100.1.11g    ; ICMP Type = 5 REDIRECT,   Code = 0 NET,   Checksum = ^x9B62i' ICMP Gateway IP Address = 129.100.202.3a    E 01CB6481   8A0901FC   000065E6   38000045    0000    E..8.e.......d..nE 28020045 | 03CA6481   629B0005 | 0B016481    0010    .d.....b.d..E..(-E 0603720A   0B016481   2720067E   0000C28A    0020    ....~. '.d...r..i=                       86DFE02D   B7061700    0030    ....-...oG -----------------------------------------------------------------------c   Cheers,i   Piyush.a     Sent via Deja.coma http://www.deja.com/   ------------------------------  % Date: Fri, 12 Jan 2001 19:07:24 +0100a. From: "Jesper Naur" <jesper.naur@post.tele.dk> Subject: Re: tcpiptraces, Message-ID: <93nh93$74l$1@news.inet.tele.dk>  / <pa@it.singer-friedlander.com> wrote in messaged# news:93msb6$5p5$1@nnrp1.deja.com...oE > Could some explain to me what the output from the tcpiptrace means.r/ > I have tried it between our VAX and a router.t9 > I have been experimenting with it and get this output :p >MI > -----------------------------------------------------------------------aC > TCPIPtrace full display RCV packet 185 at 12-JAN-2001 12:08:28.73- >-C > IP Version = 4,  IHL = 5,  TOS = 00,   Total Length = 56 = ^x0038o2 > IP Identifier  = ^xE665,  Flags (0=0,DF=0,MF=0),@ > Fragment Offset = 0 = ^x0000,   Calculated Offset = 0 = ^x0000F > IP TTL = 252 = ^xFC,  Protocol = 1 = ^x01,  Header Checksum = ^x098A( > IP Source Address      = 129.100.203.1' > IP Destination Address = 129.100.1.11o >  >*= > ICMP Type = 5 REDIRECT,   Code = 0 NET,   Checksum = ^x9B62h) > ICMP Gateway IP Address = 129.100.202.3- >- >-G > 01CB6481   8A0901FC   000065E6   38000045    0000    E..8.e.......d..,G > 28020045 | 03CA6481   629B0005 | 0B016481    0010    .d.....b.d..E..(1G > 0603720A   0B016481   2720067E   0000C28A    0020    ....~. '.d...r..e? >                       86DFE02D   B7061700    0030    ....-...-I > -----------------------------------------------------------------------o   Hello.  H I have reasons to believe, that your VAX has IP address 129.100.1.11 and that> the router has IP address 129.100.203.1. What is happening is,G that the router 129.100.203.1 has got the idea, that a datagram sent by ; the VAX could be better routed by the router 129.100.202.3.t: The router 129.100.203.1 therefore sends the ICMP REDIRECT> packet, to inform the VAX about the better route. The REDIRECT: packet contains the IP header and the first 8 bytes of theB original packet, which caused the router to generate the REDIRECT.  < Below I explain the parts of the data not already covered by the TCPIPTRACE interpretation.  H I have chosen to rearrange the longwords of the packet below each other, to simplify the explanations:e  ? IP header of ICMP packet, this is fully interpreted in the text  part of the trace output.    38000045 000065E6 8A0901FC 01CB6481 0B016481  B ICMP header of the ICMP redirect packet, this is fully interpreted% in the text part of the trace output.y   629B0005 03CA6481  @ IP header of the packet, which caused the router to generate the ICMP redirect packet  & 28020045   Packet length ^x0228 = 552. 0000C28A 2720067E   Protocol 06 = TCP& 0B016481   Source address 129.100.1.11) 0603720A   Destination address 10.114.3.6   # First 2 longwords of the TCP headerN0 B7061700   Source port 23, destination port 1719* 86DFE02D   TCP sequence number (769712006)  A What we see is that 10.114.3.6 presumably has a TELNET session at @ the VAX 129.100.1.11 (source port 23). The VAX has tried to sendB something (552 bytes including 20 bytes IP header and 20 bytes TCPG header) to 10.114.3.6, but as already related, the router 129.100.203.1eA thinks that the router 129.100.202.3 would be a better choice for  routing the packet.e  8 To find out why this happens (there are several possible* explanations), we must know more about the. specific configuration of router 129.100.203.1  
  Best regards   Jesper Naur   ------------------------------  # Date: Fri, 12 Jan 2001 08:49:06 GMTo From: Dirk Munk <munk@home.nl>4 Subject: Re: Today's Nugget of Worthless Information' Message-ID: <3A5EC501.B1B0FF10@home.nl>v   "David J. Dachtera" wrote:   > Paul Repacholi wrote: 
 > > [snip]I > > BTW, the gzip changes are intended for the freeware CD. I'll probably1F > > include a generic exe and generic and host-specific command files. > E > Have you added RMS/ODS support? ...or fixed any long-standing bugs?a > ...or add DCL support? >-F > The "current" GZIP for OpenVMS (V1.2-4, AFAIK) does not support DCL,H > does not support RMS file/record attributes, does not preserve versionI > numbers, and cannot correctly display the size of the contents of a .GZLC > archive. It also deletes the archive when decompressing (no known. > work-around).a  ! There is a very easy workaround::t   def /user sys$output <outfile> gzip -c <infile>  F This will tell gzip to decompress to the console, without deleting the input file. ! And the console is sys$output ... B Only use "def /user" , not just "def". The file attributes will be< different and they may cause trouble with your applications.     >g >e > -- > David J. Dachterau > dba DJE Systems  > http://www.djesys.com/ > < > Unofficial Affordable OpenVMS Home Page and Message Board:! > http://www.djesys.com/vms/soho/g >iH > This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings > is to be expected. > B > Feel free to exercise your rights of free speech and expression. > H > However, attacks against individual posters, or groups of posters, are > strongly discouraged.    ------------------------------  % Date: Fri, 12 Jan 2001 15:50:07 -0300-) From: fabio_compaq@ep-bc.petrobras.com.br2' Subject: Transaction Server for OpenVMSnL Message-ID: <OF60B8F3CA.2FDCCF18-ON032569D2.00674CAA@ep-bc.petrobras.com.br>  J I think there is a problem with the ACMS SPD link in the Compaq homepage - or this producte is being discontinued ....  D Do you know other TP Server instead of ACMS and Tuxedo for OpenVMS ?   Regards9   FC   ------------------------------  # Date: Fri, 12 Jan 2001 11:39:31 GMT " From: pa@it.singer-friedlander.com  Subject: Re: UCX Routing Problem) Message-ID: <93mqdj$4ed$1@nnrp1.deja.com>   $ We already have a default route of :  $ AH   0.0.0.0           129.100.203.1  6 which points to the same router as the original route, but it still happens.e  - In article <3A5E4B59.B6964587@earthlink.net>,i:   "David J. Dachtera" <djesys.nospam@earthlink.net> wrote: > piyush_a@my-deja.com wrote:2 > > 9 > > We are running a VAX under OpenVMS 7.1 and UCX 4.1E2.t@ > > We are having some routing problems. We have Dynamic routing	 disabled,tD > > but for some reason we seem to be receiving a "rogue" route fromB > > somewhere and we dont really want to be there because if often disrupts > > transfers on other routes. > >n5 > > AN   10.200.10.2                    129.100.203.1 5 > > AN   10.0.0.0                       129.100.203.3  > > 5 > > PH   10.200.10.2                    129.100.203.1u > >nB > > For some reason the 10.0.0.0 route appears from nowhere and so causesF > > transfers on the 10.200.10.2 route to fail because it is going theG > > wrong way. When it is removed, everything is fine, but then it justd > > reappears a few days later.) > >j@ > > Does anyone have any ideas to why this may be happening? Any, > > suggestions will be greatly appreciated. >7B > I seem to remember a similar problem on a site I served over the summer.s >tG > I believe the fix was to establish a default route. Even if the bogusoG > route appears, it will it only be tried as an alternate and so causes 9 > minimal disruption until someone can clear it manually.a >a  > I could be wrong, as always... >v > -- > David J. Dachteras > dba DJE Systemsg > http://www.djesys.com/ >d< > Unofficial Affordable OpenVMS Home Page and Message Board:! > http://www.djesys.com/vms/soho/w >iH > This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings > is to be expected. > B > Feel free to exercise your rights of free speech and expression. > H > However, attacks against individual posters, or groups of posters, are > strongly discouraged.p >M     Sent via Deja.comn http://www.deja.com/   ------------------------------  % Date: Fri, 12 Jan 2001 11:17:27 -0500-% From: "John Vottero" <John@mvpsi.com>i  Subject: Re: UCX Routing Problem/ Message-ID: <t5ubhvtiuuk01b@news.supernews.com>o  J AH means it's a route to a Host.  Your default route should be xN, a route
 to a Network.   J The default route looks broken anyway.  I don't machine with an address of1 10.x.x.x can't have a default route of 129.x.x.x.   ! Can you post the results of this:w   $ UCX SHOW CONF/INTER/FULL $ UCX SHOW ROUT/PERM/FULLo  / <pa@it.singer-friedlander.com> wrote in messagep# news:93mqdj$4ed$1@nnrp1.deja.com...r& > We already have a default route of : >u& > AH   0.0.0.0           129.100.203.1 > 8 > which points to the same router as the original route, > but it still happens.  >e/ > In article <3A5E4B59.B6964587@earthlink.net>,e< >   "David J. Dachtera" <djesys.nospam@earthlink.net> wrote: > > piyush_a@my-deja.com wrote:  > > >e; > > > We are running a VAX under OpenVMS 7.1 and UCX 4.1E2.nB > > > We are having some routing problems. We have Dynamic routing > disabled,uF > > > but for some reason we seem to be receiving a "rogue" route fromD > > > somewhere and we dont really want to be there because if often
 > disrupts  > > > transfers on other routes. > > >l7 > > > AN   10.200.10.2                    129.100.203.137 > > > AN   10.0.0.0                       129.100.203.3I > > >B7 > > > PH   10.200.10.2                    129.100.203.1D > > >DD > > > For some reason the 10.0.0.0 route appears from nowhere and so > causesH > > > transfers on the 10.200.10.2 route to fail because it is going theI > > > wrong way. When it is removed, everything is fine, but then it justu! > > > reappears a few days later.i > > >eB > > > Does anyone have any ideas to why this may be happening? Any. > > > suggestions will be greatly appreciated. > >iD > > I seem to remember a similar problem on a site I served over the	 > summer.l > >aI > > I believe the fix was to establish a default route. Even if the bogusiI > > route appears, it will it only be tried as an alternate and so causes ; > > minimal disruption until someone can clear it manually.t > >n" > > I could be wrong, as always... > >h > > -- > > David J. Dachterap > > dba DJE Systemse > > http://www.djesys.com/ > >n> > > Unofficial Affordable OpenVMS Home Page and Message Board:# > > http://www.djesys.com/vms/soho/e > >eJ > > This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings > > is to be expected. > >tD > > Feel free to exercise your rights of free speech and expression. > >fJ > > However, attacks against individual posters, or groups of posters, are > > strongly discouraged.i > >p >l >t > Sent via Deja.comy > http://www.deja.com/   ------------------------------  # Date: Fri, 12 Jan 2001 17:04:22 GMTl= From: system@SendSpamHere.ORG (Brian Schenkenberger, VAXman-)e  Subject: Re: UCX Routing Problem0 Message-ID: <009F6018.43ABF713@SendSpamHere.ORG>  W In article <t5ubhvtiuuk01b@news.supernews.com>, "John Vottero" <John@mvpsi.com> writes:dK >AH means it's a route to a Host.  Your default route should be xN, a route. >to a Network. >wK >The default route looks broken anyway.  I don't machine with an address of.2 >10.x.x.x can't have a default route of 129.x.x.x. > " >Can you post the results of this: >e >$ UCX SHOW CONF/INTER/FULLr  - $ UCX SHOW CONF INTER/FULL   (note, no slash)s   >$ UCX SHOW ROUT/PERM/FULL  F You may also want to provide UCX SHOW ROUT/FULL output too which will @ show routes UCX creates as a result of bringing up an interface.   --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COM             tO city, n., 1. a place where trees are cut down and streets are named after them.1   ------------------------------  % Date: Fri, 12 Jan 2001 12:23:40 -0500 % From: "John Vottero" <John@mvpsi.com>   Subject: Re: UCX Routing Problem. Message-ID: <t5ufe5ff2lhfa@news.supernews.com>  0 "John Vottero" <John@mvpsi.com> wrote in message) news:t5ubhvtiuuk01b@news.supernews.com...e [snip]L > The default route looks broken anyway.  I don't machine with an address of3 > 10.x.x.x can't have a default route of 129.x.x.x.  [snip]  G I can't even understand my own post.  What I meant to say was somethingR like:l  L I don't think a machine with an address of 10.x.x.x can have a default route
 of 129.x.x.x.e   ------------------------------    Date: 13 Jan 2001 00:52:36 +0800, From: Paul Repacholi <prep@prep.synonet.com>0 Subject: Re: VMS performance data in HTML format0 Message-ID: <87k880r9uz.fsf@k9.prep.synonet.com>   waybright@my-deja.com writes:u  G > I'm looking for some inexpensive software or program that can deliverIE > basic performance charts in HTML format.  On the Tru64 UNIX side wei > haveE > a free product called Big Brother that does a good job of providingf7 > basic CPU, memory, IO performance charts for the web.  > > > I'm aware of such products as Compaq's PAWZ and DataMetric'sC > Viewpoint.  Are there any shareware/freeware programs such as Bigc > Brother for VMS? > F > Also, does anyone know of a good collection of DCL scripts with HTMLA > tags built in to provide various system information on the web?a  - Get the WASD web server and look at hyperspi.h   -- d< Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.   ------------------------------   Date: 12 Jan 2001 11:07:25 GMT* From: helbig@astro.rug.nl (Phillip Helbig)8 Subject: Re: VMS wishlist [was Re: Identify que entry #]. Message-ID: <93mohd$j1g$1@info.service.rug.nl>  E In article <3A5E2CA8.30729.C3F6B5C@localhost>, stan@stanq.com writes:S  4 > On 11 Jan 2001, at 20:23, David J. Dachtera wrote:8 > > Well, here's another item for the OpenVMS wish list: > > $ DELETE/ENTRY=nnnn/CONFIRMh >  > H > Here's a different one:  Ever try to edit a command line entry that's @ > longer than one line?  You can't back up before the last line. > G > Long ago (10+ years), I asked at the DECUS Symposium about this, and  H > was told that it couldn't be done.  That didn't sit too well with me, 3 > but, heck, VMS is perfect in every other respect.- > F > Imagine my surprise to discover that Linux does this, and VMS still  > does not.  *sigh*   D I don't think it's a VMS problem per se but rather something like a G terminal-driver problem.  (Not sure---several levels below my plane of gD thinking.)  With VMS MAIL or NEWSRDR, I can do what you want at the  MAIL> and News> prompts.   ------------------------------  % Date: Fri, 12 Jan 2001 12:02:17 -0500w# From: Jim Agnew <agnew@hsc.vcu.edu>78 Subject: Re: VMS wishlist [was Re: Identify que entry #]+ Message-ID: <3A5F3899.113E329D@hsc.vcu.edu>   n i do it all the time... if it's not longer than 132 cols, do a set term/linsize=132, hit uparrow, and bingo...  D if it wraps over 132, then it depends on your terminal emulator...     try it.v   j.   stan@stanq.com wrote:s > 4 > On 11 Jan 2001, at 20:23, David J. Dachtera wrote:8 > > Well, here's another item for the OpenVMS wish list: > > $ DELETE/ENTRY=nnnn/CONFIRMc > G > Here's a different one:  Ever try to edit a command line entry that's1@ > longer than one line?  You can't back up before the last line. > F > Long ago (10+ years), I asked at the DECUS Symposium about this, andG > was told that it couldn't be done.  That didn't sit too well with me,u3 > but, heck, VMS is perfect in every other respect.i > E > Imagine my surprise to discover that Linux does this, and VMS still- > does not.  *sigh*- >  > --Stan >  > ----------I > Stanley F. Quayle, P.E.   N8SQ   +1 614-868-1363   Fax: +1 614 868-167103 > 8572 North Spring Ct. NW, Pickerington, OH  43147E? > Preferred address:  stan@stanq.com       http://www.stanq.comc   ------------------------------  % Date: Fri, 12 Jan 2001 00:36:31 -0800d+ From: "Wayne Holland" <wholland@tscnet.com>=/ Subject: Re: voor de belgische Alpha gebruikers1O Message-ID: <1C202F127A703153.47A172871E141329.14D5540291452C17@lp.airnews.net>w  ? "Martin Vorlaender" <martin@radiogaga.harz.de> wrote in message 5 news:3a5e7b19.524144494f47414741@radiogaga.harz.de...6$ > Shane.F.Smith@Healthnet.com wrote:= > : > "David J. Dachtera" <djesys.nospam@earthlink.net> wrote0 > : > Bob Kaplow wrote: + > : > > Shane.F.Smith@Healthnet.com writes:0J > : > > > Oh great. Now it's only a matter of time before someone posts in > : > > > Klingon... > : > >i > : > > Ka-plak! > : >.F > : > Gee - I thought that was "Ker Plop!" because humans complain tht > : > Klingons smell bad. :-)  > : B > : IIRC, it's Q'plah. It means "success!", and it's a toast or an encouraging9/ > : salutation. (I am a part-time sad bastard.)r >lF > Actually, it's written like Qapla' (and roughly pronounced kkhap-LA)I > according to Marc Okrand's Official Guide To Klingon Words And Phrases.e >i > pltlh (I've finished!), 
 >   Martin > --L > One OS to rule them all       | Martin Vorlaender  |  VMS & WNT programmer9 > One OS to find them           | work: mv@pdv-systeme.deI! > One OS to bring them all      |a( http://www.pdv-systeme.de/users/martinv/@ > And in the Darkness bind them.| home: martin@radiogaga.harz.de8 LOLMAO...  now thats quick thinking!  (Applause) really!
 I like it!   ------------------------------  + Date: Fri, 12 Jan 2001 08:37:17 +0100 (MET)^& From: Rudolf Wingert <win@fom.fgan.de>D Subject: Windows driver for LPS20 and LPS17 (Windows 2000 compliant)6 Message-ID: <200101120733.IAA03525@sinet1.fom.fgan.de>   Hello,  D yesterday I did search for Windows xyz driver for the Printserver 17C and 20. Compaq have changed there Webpages, so that I did not found8D any driver. If I click on the PrintServer20 item I can download onlyC the PrintServer software for different OSes. But there is no driver.F anymore. Do anybody know, where I can get the Windows driver especialyI the Windows 2000 drivers for this PrintServer? The driver for Windows NT/ H Windows 98 does not accept the installed duplex unit under Windows 2000.   TIA and regards Rudolf Wingert  F P.S. A few weeks ago I did get the driver for the PrintServer from the Compaq Webpage   ------------------------------  % Date: Fri, 12 Jan 2001 12:10:14 -050010 From: Paul Anderson <paul.r.anderson@compaq.com>H Subject: Re: Windows driver for LPS20 and LPS17 (Windows 2000 compliant)C Message-ID: <paul.r.anderson-A42AFD.12101412012001@news.compaq.com>f  F In article <200101120733.IAA03525@sinet1.fom.fgan.de>, Rudolf Wingert  <win@fom.fgan.de> wrote:  G > yesterday I did search for Windows xyz driver for the Printserver 17 aF > and 20. Compaq have changed there Webpages, so that I did not found G > any driver. If I click on the PrintServer20 item I can download only nF > the PrintServer software for different OSes. But there is no driver ? > anymore. Do anybody know, where I can get the Windows driver tF > especialy the Windows 2000 drivers for this PrintServer? The driver G > for Windows NT/ Windows 98 does not accept the installed duplex unit t > under Windows 2000.P  E Windows 2000 has drivers for all the PrintServer models.  I set up a 0D queue to a PrintServer 17/600 and it allowed me to select duplexing.   Paul   -- i,    Paul Anderson, OpenVMS Engineering (DCPS),    Compaq Computer Corporation, Littleton MA   ------------------------------   End of INFO-VAX 2001.024 ************************