1 INFO-VAX	Thu, 29 May 2003	Volume 2003 : Issue 295       Contents:3 Re: An escape sequence for PF4 in an expect script.   Re: Another VMS inquirer article  Re: Another VMS inquirer article  Re: Another VMS inquirer article  Re: Another VMS inquirer article  Re: Another VMS inquirer article! Controller for VAXStation 4000 90 % Re: Controller for VAXStation 4000 90 ! Re: CSWS and Server-Side Includes ! Re: CSWS and Server-Side Includes  ES40 and CPU upgrade RE: Firewall for VMS / TRU64 Re: Firewall for VMS / TRU64 Re: Firmware Upgrade Re: FREE TO A GOOD HOME  Re: FREE TO A GOOD HOME  Re: FREE TO A GOOD HOME  Re: FREE TO A GOOD HOME  Re: FREE TO A GOOD HOME  Re: FREE TO A GOOD HOME + Getting LK411-AA keyboard to work correctly 8 Re: GUI Applications on VMS (was: Portents of VMS death)# Hobbyist seeks generic SIMM vendors ' Re: Hobbyist seeks generic SIMM vendors ' Re: Hobbyist seeks generic SIMM vendors ' Re: Hobbyist seeks generic SIMM vendors ' Re: Hobbyist seeks generic SIMM vendors ' Re: Hobbyist seeks generic SIMM vendors ' Re: Hobbyist seeks generic SIMM vendors ' Re: Hobbyist seeks generic SIMM vendors ' Re: Hobbyist seeks generic SIMM vendors ' Re: Hobbyist seeks generic SIMM vendors & Re: How to make a shadowed system disk Invalid Disk format  Re: Invalid Disk format  June event in NYC P Re: Letter to HP Management from OpenVMS users around the world ... Shape up or * Re: Logicals Lost When Process Is Detached Re: Mailbox  New MiniMerge capability Re: New MiniMerge capability Re: New MiniMerge capability Re: New MiniMerge capability Re: New MiniMerge capability Re: New MiniMerge capability3 Re: NT: son of VMS? (was Re: Portents of VMS death) 3 Re: NT: son of VMS? (was Re: Portents of VMS death) 3 Re: NT: son of VMS? (was Re: Portents of VMS death)  Re: Portents of VMS death  Re: Portents of VMS death  Re: Portents of VMS death  Re: Portents of VMS death  Re: Portents of VMS death  Re: Portents of VMS death  Re: Portents of VMS death ) re Re: How to make a shadowed system disk 3 Reading CD-R burned on a Window system on an Alpha? 7 Re: Reading CD-R burned on a Window system on an Alpha? 7 Re: Reading CD-R burned on a Window system on an Alpha? 7 Re: Reading CD-R burned on a Window system on an Alpha? 7 Re: Reading CD-R burned on a Window system on an Alpha? . Skil Saw? (was - Re: New MiniMerge capability)2 Re: Skil Saw? (was - Re: New MiniMerge capability)2 Re: Skil Saw? (was - Re: New MiniMerge capability)2 Re: Skil Saw? (was - Re: New MiniMerge capability)  Re: SYS$QIOW - Performance Issue  Re: SYS$QIOW - Performance Issue  Re: SYS$QIOW - Performance Issue  Re: SYS$QIOW - Performance Issue  Re: SYS$QIOW - Performance Issue  Re: SYS$QIOW - Performance Issue  Re: SYS$QIOW - Performance Issue  Re: SYS$QIOW - Performance Issue  Re: SYS$QIOW - Performance Issue  F ----------------------------------------------------------------------  % Date: Wed, 28 May 2003 19:41:19 -0700  From: Jason Brady <>< Subject: Re: An escape sequence for PF4 in an expect script.8 Message-ID: <9lradv4hskdd6p02a53d098t45jod5cq1q@4ax.com>  F Thanks for the clarification.  Updated my notes to reflect the correctF info...at least the program for which the notes were written was coded  correctly!  [big sigh of relief]  F On 24 May 2003 15:41:14 GMT, Thomas Dickey <dickey@saltmine.radix.net> wrote:  , >JF Mezei <jfmezei.spamnot@istop.com> wrote: >> Jason Brady wrote: I >>> Because PF keys fall into the keypad grouping you use ESC [ O or SS3:  >>> ! >>> Key        7-bit        8-bit " >>> PF1      ESC [ O P     SS3 + P" >>> PF2      ESC [ O Q     SS3 + Q" >>> PF3      ESC [ O R     SS3 + R" >>> PF4      ESC [ O S     SS3 + S >  >> Actially, SS3 is ESC O.   >  >> So PF1 is ESC O P > N >Actually SS3 is the C1 control equivalent to ESC O P.  (Generally I would useN >SS3 to apply to either to simplify documentation, but technically it is 8/15,
 >i.e., 143).     ------------------------------  % Date: Wed, 28 May 2003 14:51:21 -0400 * From: JF Mezei <jfmezei.spamnot@istop.com>) Subject: Re: Another VMS inquirer article ) Message-ID: <3ED50524.6B7E1F92@istop.com>    Sue Skonetski wrote: > * > http://www.theinquirer.net/?article=9705  K ##Like their NSK brethren, VMS users demand the best and are willing to pay  for it.    WRONG ! WRONG ! WRONG !   L Tandem failed exactly because it was such a narrowly focused product. Why do+ you think it begged for someone to buy it ?   N Digital failed because it wanted to narrow VMS's market to as small a niche asI possible and sell it onlyt to companies who have unlimited budgets (which B leaves the US military as the only remaining cash cow these days).  L ##Take away OpenVMS, and what do you have: the Mother of All Marquee Account7 ##Defections and a deluge of absolutely dreadful press.   M You are assuming that HP would kill VMS tomorrow in one fell swoop. But if HP J "prepares" VMS by slowly moving customers to HP-UX and not getting any newN customers, within 1-2 years, Carly will have statements from their accountantsI clearly showing VMS no longer getting new customers and its customer base K going down, at which point, Wall Street will demand HP cancel this product.   I HP is not stupid. They have learned from Palmer's attemps, and Capellas's 2 attempts and they know how to safely turn off VMS.  N Any company that cares more about shoring up Microsoft's fortunes doesn't careN about jeoperdizing a more profitable product because its mission isn't to makeP its own products succeed but rather make Intel and Microsoft's products succeed.    : ##Continuing the "HP Drops VMS" scenario--a scenario <...>; ##--BCS revenues will plunge, Services revenues will suffer H ##significantly, and whatever degree of trust HP has managed to instill % ##in its installed base will  plunge.   N 1- HP has done little or nothing to build trust. Saying that it won't kill VMSM isn't enough. Actions speak louder than words, and the consistent omission of L VMS from strategic speeches when even NSK in mentioned sends a strong signalF that HP cannot be trusted when it says it loves VMS and won't hurt it.  M 2- If HP can move VMS customers to HP=UX, as Scott Stallard said was the plan H on his original may 7th memo, then they won't lose the service and salesN revenus. Furthermore, HP knows full well that the remaining VMS installed baseN cannot migrate overnight and it will take some time before all VMS revenus areL gone. As long as the drop is manageable and offset by HP-UX and others, thenS shareholders won't see anything wrong in the decision to take VMS off life support.   N 3-If you want a healthy child, you will feed him well and educate him/her. ForI a product to be healthy, it needs not only food for development, but also % marketing so it can "meet" customers. M Right now, I consider VMS to be on life support. Just enough food so that the M authorities can't legally accuse HP of child endangerement, but not enough to 2 ensure that the child grows to its full potential.       RE: Promises Kept L Yeah, they have their famous "plan of record" and yes, it is believable thatJ they will do the bare minimum to legally abide by this plan of record. ButH that "plan of record" refers to one hell of a broken commitment: Alpha'sM Murder following so many commitments that Alpha would live on. So how can you L trust a company who make plenty of promises about Alpha, then broke them and- now wants you to believe their new promises ?   J Also, the killing of VMS is not inconsistent with HP keeping that "plan ofJ record". The minute HP has a version of VMS commercially available one oneJ IA64 based box, it will have fulfilled its promise. There is no promise toI market and grow VMS, the only promises are to ensure support for existing L customers. When a vendor only talks to its existing customers to ensure thatM it will support their shop for at least X years, it means that the vendor has ( no intentions to grow the customer base.  O ##HP has indeed kept its promises, exceeded the financial expectations of many, E ## and has  continued to develop, patent, and introduce new products.   G Since HP refuses to divulge VMS financials, the above statement is just 1 meaningless fluff in the context of this article.   M For some time now, I have considered Mr Shannon to be the moral equivalent of N an Embedded Reporter in the US invasion of Irak. Embedded means "in bed with".  M And the mere fact that HP would NOT make an official rebuttal for the damning F "HP is killing VMS" article and resorted to using its unofficial "leak> spokesperson" says a lot about HP's true intentions about VMS.  N If it had been an accusation on HP-UX, you can bet that Carly would have stoodE up and defended the product and given plenty of examples on how HP is " agressively pushing HP-UX etc etc.   ------------------------------  % Date: Wed, 28 May 2003 19:00:59 -0400 ( From: David Froble <davef@tsoft-inc.com>) Subject: Re: Another VMS inquirer article , Message-ID: <3ED53FAB.7080403@tsoft-inc.com>   JF Mezei wrote:    > Sue Skonetski wrote: > * >>http://www.theinquirer.net/?article=9705  O > 2- If HP can move VMS customers to HP=UX, as Scott Stallard said was the plan  > on his original may 7th memo,     Q Ok, does anyone have the memo available?  I've heard the claims from both sides.  I   Let's get a  copy of the memo, and see what Scott Stallard really said.     O > And the mere fact that HP would NOT make an official rebuttal for the damning H > "HP is killing VMS" article and resorted to using its unofficial "leak@ > spokesperson" says a lot about HP's true intentions about VMS. > P > If it had been an accusation on HP-UX, you can bet that Carly would have stoodG > up and defended the product and given plenty of examples on how HP is $ > agressively pushing HP-UX etc etc.    O While JF does seem to have an agenda of not believing anything he doesn't want  Q to hear, there are still quite a few good points that he makes.  Having a non-HP  N (I'm assuming) person put out a 'happiness' blurb isn't really a message from O HP.  So how about it, should a rumor of the 'end of HP-UX' be started, and see  = how HP reacts?  How about a rumor that HP is dropping windoz?      Dave     --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com6 T-Soft, Inc.  170 Grimplin Road  Vanderbilt, PA  15486   ------------------------------  % Date: Wed, 28 May 2003 20:19:43 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net> ) Subject: Re: Another VMS inquirer article ' Message-ID: <3ED5602F.DD0B7710@fsi.net>    Sue Skonetski wrote: > * > http://www.theinquirer.net/?article=9705  D If the author - and hp - would think things through a little deeper," they'd answer their own questions.  E To me, these were the most glaring errors (all in the same paragraph,  interestingly):   > "Since most people don't buy OSes as stand-alone products ..."  F Everyone knows this. Stupid, weak excuse. The OS is part of a package,C but then the ENTIRE package must be advertised. HP doesn't do that,  either.   A C'mon guys, let's stop making excuses. This isn't rocket science.   # "...(Linux being an exception)..."    D Well, yes and no. Yes, folks buy Linux without it being sold with orG pre-loaded on a machine, generally speaking - if they pay for it at all C (it is "freely" downloadable, after all). Then again, no one "buys" F Linux just to have it on the shelf. They either load it on an existing, system or buy/build one on which to load it.  F "...determining your target audience and message delivery system ought5 to be an interesting--and time consuming--exercise."    F Interesting? Certainly! Time-consuming? If you're doing "all the rightF things", it's time well invested. If you're paying buku-bux to a bunchE of yahoos to tell you what you already know, then you're pissing away H both time and money. If you're at least doing SOMEthing pro-active, thenE at least you're learning *AND* getting the word out at the same time.   9 Some folks think its wasteful and foolish to advertise by F trial-and-error, and to an extent, they're correct. Even if "Ma and PaE Kettle" go on cable-TV to say, "We run an ISP. Call us and we'll hook D you up", at least they're getting exposure. When folks discover thatG KettleNET is the sharpest ISP in the area, their business takes off and E they become known as the "great the ISP with the lousy ads". Or, as a G mentor of mine once said, "I'd rather see a crooked furrow than a field 
 unplowed".  E Folks here can attest - I cut a *LOT* of crooked furrows, but I'm not  always wrong, either.    --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------  % Date: Wed, 28 May 2003 21:01:00 -0400 * From: JF Mezei <jfmezei.spamnot@istop.com>) Subject: Re: Another VMS inquirer article ) Message-ID: <3ED55BAE.C0A27463@istop.com>    David Froble wrote: * > Ok, does anyone have the memo available?  M Portions were posted on comp.os.v,s on may 7th and may 8th. Don't forget that E as the uproar rose, the PDF document was modified in-situ without any N indication, and it was then modified again a week or two later, this time with a "rev 2" added.  0 Here is some of the text from the original memo:    , Q: Will you offer a migration path to HP-UX?3 A: Yes. We believe HP-UX on Itanium is an excellent . long-term choice for OpenVMS customers, but we3 want to support them moving on their own timetable.   = This was a Scott Stallard Q/A section at the end of the memo.    ------------------------------  % Date: Wed, 28 May 2003 21:11:17 -0400 * From: JF Mezei <jfmezei.spamnot@istop.com>) Subject: Re: Another VMS inquirer article ) Message-ID: <3ED55E16.5EF82E38@istop.com>   N If HP is truly so intent on keeping and growing VMS, why is it acting as if itL is slowly winding down VMS and forced to deny this whenever the uproar grows above a certain decibel level ?   G If HP acted in a way consisted with its promises, customers wouldn't be = confused as to the true intentiosn of HP with regards to VMS.     D In fairness to HP, I think it only really has committed to 2 things:M completing te port to IA64 and then supporting Alpha-VMS for a certain number ? of years. It has not committed to growing VMS or marketing VMS.   L What HP is providing is bare minimum support for a mature platform. It isn'tL taking any actions to make the platform truly alive and thriving. That wouldK require mentioning VMS at every opportunity it can (especially if 0 dollars L are allocated for marketing), and it would include telling the world that HP2 has taken onwership of VMS and intends to grow it.  N Right now, all I have heard is that HP will honour commitments made by Compaq.   ------------------------------  % Date: Wed, 28 May 2003 14:32:36 -0400 ; From: "Justin McMichael" <justin.mcmichael@1r2i3v4h5s6.com> * Subject: Controller for VAXStation 4000 90. Message-ID: <3ed501d1$1_3@corp.newsgroups.com>   Hi,   G Our large systems (mainframe) group is trying to build a duplicate of a J VAXStation 4000 90 for use as a backup for their SMS OpenLink 21 interface box.  I Their vendor was able to get them a box apparently with everything except G for the "Asynchronous serial to IBM 3745" controller card.  I put it in H quotes because I frankly don't know if that is even technically correct.J You see, I am not the VMS guy.  In fact I am far from it.  All I am really? trying to do is help out by doing a bit of searching on my own.   H I have read quite a bit about the 4000 90 on the web, but nowhere have IH found any mention of such a controller card.  I have found references toJ synchronous controllers, but nothing specific enough to figure out if I am barking up the right tree.  I The card itself (as far as I can tell) is definitely plugged into a riser / connector and is not attached to the mainboard.   6 Does anyone here know what exactly we are looking for?   Justin McMichael Riverside Health System " (Remove the numbers from my email)        > -----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----A http://www.newsfeeds.com - The #1 Newsgroup Service in the World! > -----==  Over 80,000 Newsgroups - 16 Different Servers! =-----   ------------------------------  % Date: Wed, 28 May 2003 15:52:13 -0400 ; From: "Justin McMichael" <justin.mcmichael@1r2i3v4h5s6.com> . Subject: Re: Controller for VAXStation 4000 90. Message-ID: <3ed5147a$1_3@corp.newsgroups.com>  H I think I found out what it is.  It's a DSW21-AA synchronous controller.L The problem is, noone has it.  Nothing on eBay, nothing (of value) on Google or any other search engine.   . Does anyone here know where on could be found?   Justin    F "Justin McMichael" <justin.mcmichael@1r2i3v4h5s6.com> wrote in message( news:3ed501d1$1_3@corp.newsgroups.com... > Hi,  > I > Our large systems (mainframe) group is trying to build a duplicate of a L > VAXStation 4000 90 for use as a backup for their SMS OpenLink 21 interface > box. > K > Their vendor was able to get them a box apparently with everything except I > for the "Asynchronous serial to IBM 3745" controller card.  I put it in J > quotes because I frankly don't know if that is even technically correct.L > You see, I am not the VMS guy.  In fact I am far from it.  All I am reallyA > trying to do is help out by doing a bit of searching on my own.  > J > I have read quite a bit about the 4000 90 on the web, but nowhere have IJ > found any mention of such a controller card.  I have found references toL > synchronous controllers, but nothing specific enough to figure out if I am > barking up the right tree. > K > The card itself (as far as I can tell) is definitely plugged into a riser 1 > connector and is not attached to the mainboard.  > 8 > Does anyone here know what exactly we are looking for? >  > Justin McMichael > Riverside Health System $ > (Remove the numbers from my email) >  >  >  > @ > -----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----C > http://www.newsfeeds.com - The #1 Newsgroup Service in the World! @ > -----==  Over 80,000 Newsgroups - 16 Different Servers! =-----        > -----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----A http://www.newsfeeds.com - The #1 Newsgroup Service in the World! > -----==  Over 80,000 Newsgroups - 16 Different Servers! =-----   ------------------------------  # Date: Wed, 28 May 2003 17:59:42 GMT 1 From: Michael Austin <maustin@firstdbasource.com> * Subject: Re: CSWS and Server-Side Includes2 Message-ID: <3ED4F53C.AA4A9DF6@firstdbasource.com>  * Alan Winston - SSRL Admin Cmptg Mgr wrote: > h > In article <3ED41B0B.CED45C95@firstdbasource.com>, Michael Austin <maustin@firstdbasource.com> writes: > J > >I must be missing something here???   I have even tried putting this inJ > >the CGI-BIN directory.  The result is a blank page with the page source > >showing the commands. > >  > > ! > ><!--  The index.shtml file -->  > ><!--#printenv -->$ > ><!--#include file="open.html" -->$ > ><!--#include file="main.html" -->% > ><!--#include file="close.html" -->  > P > You're definitely missing something.  I cut-and-pasted this to my home machineQ > (running 1.3.26) and when I accessed it it did the printenv and produced errors 3 > for the #includes, since I don't have the files.)  >  > >  > >  > >In the HTTPD.CONF file: > >#$ > ># To use server-parsed HTML files > ># > >  > >AddType text/html .shtml " > >AddHandler server-parsed .shtml >  > Looks totally good.  >  > >  > >  > > # > >In my virtual hosts config file:  > ><VirtualHost 192.168.1.200>- > >    ServerAdmin maustin@firstdbasource.com 4 > >    DocumentRoot /apache$common/htdocs/SERVERNAME  > >    ServerName SERVERNAME.COM) > >    ErrorLog logs/SERVERNAME.error_log 2 > >    CustomLog logs/SERVERNAME.access_log common > ><Directory />& > >    Options FollowSymLinks Includes > >    AllowOverride None  > ></Directory>  > >    <Directory /subdir/> * > >        Options FollowSymLinks Includes > >    AllowOverride None  > >    </Directory>  > ></VirtualHost>  > K > You have a separate config file for your virtual hosts?    Is it included = > into HTTPD.CONF before or after you specify the AddHandler?   , yes -- I have 6 or 7 different virtual hosts yes  after      > N > Within the virtual host, I think you want that <Directory /> container to be. > <Directory /apache$common/htdocs/SERVERNAME>    G I have tried many permutations of this and the " <Directory /> " is the 9 only one that worked.  Says, root is at OS Document root.    >  > or you want to do  > 2 > <Location />  (from URL rather than file system) >  > and  > 3 > <Location /subdir>  (or possibly even "/subdir").  > Q > ===============================================================================   H hmmm... now none of my "virtual hosts" after upgrading from 1.0 to 1.3. E It appears a lot of things change with regard to setting up a virtual C host.  Now it just shows the top level directory with no files even E though there are files there.  ??? this just got really confusing...       www.spacelots.com  www.spacelots.com/monroemetal      --   Regards,  6 Michael Austin            OpenVMS User since June 19847 First DBA Source, Inc.    Registered Linux User #261163    ------------------------------  % Date: Wed, 28 May 2003 16:48:37 -0400 , From: "warren sander" <warren.sander@hp.com>* Subject: Re: CSWS and Server-Side Includes. Message-ID: <3ed520a6$1_2@hpb10302.boi.hp.com>  ( no server side includes in CGI scripts..5 Every page on the OpenVMS site has 5+ include files.. K I have +includes on my directory stuff in the conf file... I think the + is  needed..  
 also I use  % <!--#include virtual="local.shtml"-->  or. <!--#include virtual="/global/global.shtml"-->  D two things.. virtual instead of file and the no space before the -->H getting the spacing etc working correctly was a pain but through testingF this setup seems to work for both IIS and SWS (I have to code for both, servers...) it also works on apache on hp-ux   --B ------------------------------------------------------------------< Warren Sander                           WW eMarketing HP.COMB Hewlett-Packard Company         Work:  warren.sander@remove.hp.comK 200 Forest Street MR01-3/K8     Personal: sander.ma.ultranet@remove.rcn.com 3 Marlboro, MA 01752                   (508) 467-4875 5    My opinions are my own and I only speak for myself (          Read http://h71000.www7.hp.com/B ------------------------------------------------------------------  > "Michael Austin" <maustin@firstdbasource.com> wrote in message, news:3ED4F53C.AA4A9DF6@firstdbasource.com..., > Alan Winston - SSRL Admin Cmptg Mgr wrote: > > E > > In article <3ED41B0B.CED45C95@firstdbasource.com>, Michael Austin $ <maustin@firstdbasource.com> writes: > > L > > >I must be missing something here???   I have even tried putting this inL > > >the CGI-BIN directory.  The result is a blank page with the page source > > >showing the commands. > > >  > > > # > > ><!--  The index.shtml file -->  > > ><!--#printenv -->& > > ><!--#include file="open.html" -->& > > ><!--#include file="main.html" -->' > > ><!--#include file="close.html" -->  > > J > > You're definitely missing something.  I cut-and-pasted this to my home machine L > > (running 1.3.26) and when I accessed it it did the printenv and produced errors5 > > for the #includes, since I don't have the files.)  > >  > > >  > > >  > > >In the HTTPD.CONF file: > > >#& > > ># To use server-parsed HTML files > > ># > > >  > > >AddType text/html .shtml $ > > >AddHandler server-parsed .shtml > >  > > Looks totally good.  > >  > > >  > > >  > > > % > > >In my virtual hosts config file:   > > ><VirtualHost 192.168.1.200>/ > > >    ServerAdmin maustin@firstdbasource.com 6 > > >    DocumentRoot /apache$common/htdocs/SERVERNAME" > > >    ServerName SERVERNAME.COM+ > > >    ErrorLog logs/SERVERNAME.error_log 4 > > >    CustomLog logs/SERVERNAME.access_log common > > ><Directory />( > > >    Options FollowSymLinks Includes > > >    AllowOverride None  > > ></Directory>  > > >    <Directory /subdir/> , > > >        Options FollowSymLinks Includes > > >    AllowOverride None  > > >    </Directory>  > > ></VirtualHost>  > > D > > You have a separate config file for your virtual hosts?    Is it included? > > into HTTPD.CONF before or after you specify the AddHandler?  > . > yes -- I have 6 or 7 different virtual hosts > yes  > after  >  >  > > J > > Within the virtual host, I think you want that <Directory /> container to be 0 > > <Directory /apache$common/htdocs/SERVERNAME> >  > I > I have tried many permutations of this and the " <Directory /> " is the ; > only one that worked.  Says, root is at OS Document root.  >  > >  > > or you want to do  > > 4 > > <Location />  (from URL rather than file system) > >  > > and  > > 5 > > <Location /subdir>  (or possibly even "/subdir").  > >  > > L ============================================================================ ===  > I > hmmm... now none of my "virtual hosts" after upgrading from 1.0 to 1.3. G > It appears a lot of things change with regard to setting up a virtual E > host.  Now it just shows the top level directory with no files even F > though there are files there.  ??? this just got really confusing... >  >  > www.spacelots.com  > www.spacelots.com/monroemetal  >  >  > --
 > Regards, > 8 > Michael Austin            OpenVMS User since June 19849 > First DBA Source, Inc.    Registered Linux User #261163    ------------------------------  % Date: Wed, 28 May 2003 14:40:52 -0600 ' From: DigiDemon <digidemon@hotmail.com>  Subject: ES40 and CPU upgrade 8 Message-ID: <pan.2003.05.28.20.40.51.772571@hotmail.com>   Hey all!  G Real quick...we have just bought a 887 mtz CPU for our spiffy ES40.  We J needed this to support our SmartArray 5304 card.  Anyone ever run into any snags doing this?  Thanks!   James    ------------------------------  # Date: Wed, 28 May 2003 18:50:14 GMT # From: hoff@hp.nospam (Hoff Hoffman) % Subject: RE: Firewall for VMS / TRU64 2 Message-ID: <Gx7Ba.1572$Du1.1292@news.cpqcorp.net>  C   I was specifically refering to external firewall appliance boxes.   E   Based on my experience with application reliability and its obvious H   relationship to application security, complexity tends to run contraryG   to security.  Can large and complex configurations be secured?  Sure. E   But is it economical to do?  (Vendors and customers will make that  D   particular decision, of course -- the market clearly tends toward E   the extremes of commodity and specialized solutions; toward the two /   market extremes of pricing and capabilities.)   C   For the bulk of the firewall market, the external appliances will E   serve nicely.  For the high-end and specialized, the usual firewall C   configuration involves larger and parallel firewall installations B   and multiple network pipes.  Specialized firewalls are available@   from various vendors, and range from the simple commodity box C   intended for the typical home and SOHO DSL or cable-modem network F   connections up to speciallized boxes which provide content filteringC   and antivirus and related capabilities and the routing throughput    for the larger network pipes.   E   The commodity entry prices are presently US$400 to US$1000 for most H   home or SOHO firewall boxes, with higher-level and specialized systemsD   also being available.  (At higher costs, of course.)  Firewall andD   firewall-router appliances from WatchGuard, Semantec, LinkSys (andC   LinkSys owner Cisco) and various other vendors are all available.   D   Do I think OpenVMS or various other general-purpose systems can beF   sufficiently secured?  Yes.  Do I believe that securing any firewallC   on any general-purpose operating system is easy?  No.  Assuming a F   valid and secure configuration, there is the subsequent, substantialG   and non-trivial effort with maintaining the security and integrity of B   the configuration.  Regardless of the operating system platform.E   It is this maintenance that is the interesting part, and where most C   folks tend to (unintentionally) compromise the firewall security. C   (And this on-going work has direct and indirect costs to both the    customer and to the vendor.)  F   As for one of the suggested approaches, once you start to dedicate aF   general-purpose box for the dedicated purpose of a firewall (or mostE   any other specialized use, BTW), the resulting system configuration E   tends to evolve away toward a more specialized configuration -- the E   firewall requirements will invariably head toward customized IP and E   I/O stacks and routing performance, toward auditing and control and G   evasion-related improvements, toward additional DDoS defenses, toward G   lighter-weight configurations, and particularly toward the ability to C   test and to verify the entire installation; toward configuration  
   simplicity.   G   And I will again stress that the firewall is not a complete solution, G   regardless of the environment and the assumptions.  Firewalls are now I   check-box items, and are seldom entirely effective in isolation.  (What H   happens if a VPN or a portable infection vector (eg: laptop PC) allowsF   a virus or worm to become active behind the firewall, and launch an H   out-bound DDoS attack or to clobber your internal business and network<   infrastructure, for instance?  Either can be very costly.)  N  ---------------------------- #include <rtfaq.h> -----------------------------J       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.comN  --------------------------- pure personal opinion ---------------------------E         Hoff (Stephen) Hoffman   OpenVMS Engineering   hoff[at]hp.com    ------------------------------   Date: 28 May 2003 20:38:57 GMT( From: ka2doug@cs.commoc.sc (DL Phillips)% Subject: Re: Firewall for VMS / TRU64 > Message-ID: <20030528163857.01078.00000411@mb-m18.news.cs.com>   Carl Perkins wrote: A >"John E. Malmberg" <Malmberg@dskwld.zko.dec.compaq.hp> writes...  >}DL Phillips wrote: >}>>Hoff Hoffman wrote:  >}>>K >}>>>  The maintenance of a firewall is a large and specialized engineering K >}>>>  project, and purpose-built dedicated (and commodity) firewall server I >}>>>  appliance boxes provide a highly economical and effective solution  >}> L >}> I believe Mr. Hoffmans's post said "appliance" firewalls. I take that to >be K >}> something like the WatchGuard brand boxes (not a reseller, just a user)  >whichF >}> you can buy on the low end for under US$400.00 for up to 10 client
 >computers. I N >}> read his post to say that he did NOT recommend a software solution running >on M >}> any general purpose system. Also, that such an effort would not likely be  >}> profitable.  >}  J >}The low-end firewall appliances sell in the U.S. for $30.00 and up, and J >}typically have 4 10/100 taps, and the ability to do network translation ! >}for up to 254 client computers.  >}  G >}I may be wrong about this next statement because it has been a while  A >}since I looked at prices for other than the home router market:  >}  F >}There appears to be heavy competition in the firewall/router market G >}right now, and it does not appear that any general purpose operating  H >}system would be an economical replacement for a dedicated router that  >}can handle the same load.  >}  K >}Think of a firewall also as a circuit breaker.  Any system can fall down  H >}to a denial of service attack.  Having a firewall handle the noise of = >}the network is much better than having your computer do so.  >}   >}-John # >}malmberg@dskwld.zko.dec.compaq.hp  >}Personal Opinion Only  > : >For home or small office use, such a thing might be good. > $ >For larger networks it is useless.  > D >Consider a network with over 25,000 devices on it that is connectedC >to the outside world via a pair of OC-3 (155Mb/sec) lines. I think G >you'd find the firewall you describe to be, shall we say, inadequate.   > % >This is not unlike the TAMU network.  > E >Something is used for the higher end. At this point it is apparently B >Unix systems that are used for this. I'm pretty sure that this isD >what TAMU is using - a Unix system that is running, I think, custom >software written here.  > 	 >--- Carl  >   M I've lurked around here for enough years to believe you're not suggesting one K should run the primary firewall(s) on the same computer(s) running business N critical apps, are you? Aren't you just saying that a Low end box won't alwaysN do the job? Your example of a University network would certainly take High endK appliances to handle the load. Universities and Schools are a special case,oN though, and usually have more available talent than available funds to addressK this type of problem. Really, there's rather a limited commercial market atUO that level and though the list of potential customers isn't very long but thereeL is stiff competition for the business. Anyway, there are appliance solutions< available for any sized need. How much do you want to spend?  L The difference between an "appliance" and a "general purpose" system, as youF know, is that the appliance is dedicated to one specific type of task.  L At the appliance level, the best OS is the one that will do the job with theO least amount of overhead. So, if you could manage to take out all of the thingsgH VMS has that you don't need to run a Firewall, then add in the things itL doesn't "come with" that you do need,  what sort of competitively marketable product would you end up with? r  E As well as the "Big" appliance firewalls, many of the decent "little"sL appliances I've seen use some type of *nix, too. The thing about *nix is youE can whittle it down to just exactly what you need to run an appliancehN application. Even if you could trim VMS down to the appliance level, you wouldL have thrown out most of its "advantages", you'd still have to pay the damnedL license fees, it wouldn't run on an inexpensive "commodity" CPU and it wouldM cost more than its competition (whose products have been evolving while we'veVB been reinventing the wheel --- hmm, why does that sound familiar?)  I Okay, the firewall box might run *nix, but I don't have to even know that N because it's an appliance and its job is to run the firewall software and thatI is the deepest layer I *need* to worry about. Routing and VPN are natural O extensions to a firewall application, add little overhead to the lean and cheap 6 OS and many appliances come with those features, too.   K Give me something that does the job it's meant to do, is reasonably priced,bL easy to buy, logical to use and maintain and for all I care the box could beH running any CPU & OS as long as it'll handle my needs. It really doesn't matter.w    DL Phillips  r-   It's the application. It's the application.n/     It's the application. It's the application.M1       It's the application. It's the application.s4         It's the application. It's the application. 5           It's the application. It's the application.   E Note: the above is not a signature, it is an editorial statement and I@ may or may not appear on future postings assuming there are any.   ------------------------------  % Date: Wed, 28 May 2003 22:49:49 -0400-2 From: rdeininger@mindspring.com (Robert Deininger) Subject: Re: Firmware Upgrade L Message-ID: <rdeininger-2805032249500001@user-uinj4ps.dialup.mindspring.com>  J In article <01KWFJNL1XKGAKVGCS@sysdev.deutsche-boerse.com>, Phillip Helbig+ <HELBPHI@sysdev.deutsche-boerse.com> wrote:o    H >I see that for 7.3 (and presumably 7.3-1), 5.9 is the minimum firmware F >version.  I have a 255/233 and a 3000/600 as hobbyist machines, both J >running 7.2-1 at the moment.  I'm not sure what firmware version I have, J >but both run with no problems (well, the 3000/600 seems to have a bit of I >flaky memory, but this is probably unrelated to firmware).  I'd like to    >upgrade to 7.3-1 then to 7.3-2. >FH >What is the latest version of the firmware available for each of these D >machines?  Is there a list with such information, preferably in an G >easily digestible form, on the web somewhere?  If 5.9 or later is not  F >available, what are the chances of running 7.3-1 on these machines?   >Anyone done so already?  ! V7.3-1 runs fine on both systems.-  J Alpha firmware version numbers vary a lot from one system to another.  The9 minimum and recommended FW versions are system-dependent.i  F You can see the FW version of a running VMS system via the SDA commandI CLUE CONFIG.  The "Console Vers" field is the SRM console. The "PAL Code"bC field is the VMS PALcode version.  These two numbers are completely C independent of each other.  VMS generally worries about the consolec version field.  J On recent VMS versions, SHOW CPU/FULL gives much of the same information. F There are a few other ways a well, which I won't try to get right from memory.a  J The most recent SRM version for the DEC 3000 family of systems is V7.0.  IH don't know the most recnet for the Alphastation 255/233.  Whatever is on# the FW web site is the most recent.   I The version matrix that VMS checks is in SYS$SYSTEM:FIRM_REV_MATRIX.DAT. s> But what's in this file is not necessarily the latest version.   ------------------------------    Date: 28 May 2003 13:12:31 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)y  Subject: Re: FREE TO A GOOD HOME3 Message-ID: <w32UYILP7lWl@eisner.encompasserve.org>   c In article <C3AJhK15Vuhv@eisner.encompasserve.org>, Kilgallen@SpamCop.net (Larry Kilgallen) writes:o > L > How did they make it accessible to the fire department but not to vandals.  H    You need a very long pole with a hook on one end.  Vandalism in front0    of the farmhouse was never a problem, anyway.   ------------------------------    Date: 28 May 2003 14:51:49 -04007 From: pechter@shell.monmouth.com (Bill/Carolyn Pechter)   Subject: Re: FREE TO A GOOD HOME- Message-ID: <bb30g5$sf1$1@shell.monmouth.com>   ( In article <bammb5$1o2$1@pcls4.std.com>,8 Michael Moroney <moroney@world.std.spaamtrap.com> wrote: >,D >One model was smart enough to not spin if the phasing was reversed,G >the other would spin _backwards_ (but was smart enough not to load ther( >heads). Forget which drives those were. >u >-Mike    G DEC used 3 phase in RP04's.  RP04-II (DEC internal pre-release RP05's),gK RP05's, and RP06's used two phase motors, as did (IIRC all the RM02/3/5's).b  ; RP07's may have been three phase -- but I don't know.  The  E RP04-II's/RP05/6 were Memorex.  The RP04's and RP07's were ISS Sperry - drives.  The RM02/3/5's were CDC 9762/9766's.    Anyone ever see an RM04?  3 Did they ever build a slower RM05 for Unibus boxes?n   Bill   -- eM +---------------------------------------------------------------------------+nM | Bill and/or Carolyn Pechter    |        pechter@shell.monmouth.com        | M |   Bill Gates is a Persian cat and a monocle away from being a villain in  |eN |   a James Bond movie              -- Dennis Miller                        | M +---------------------------------------------------------------------------+b   ------------------------------    Date: 28 May 2003 13:47:34 -07003 From: Eric Smith <eric-no-spam-for-me@brouhaha.com>   Subject: Re: FREE TO A GOOD HOME0 Message-ID: <qh7k8a23yx.fsf@ruckus.brouhaha.com>  9 pechter@shell.monmouth.com (Bill/Carolyn Pechter) writes:E, > RP04-II (DEC internal pre-release RP05's),M > RP05's, and RP06's used two phase motors, as did (IIRC all the RM02/3/5's).   I Single-phase motors, rated for operation on a fairly wide range extending H below 208 VAC and above 240 VAC.  Though DEC wired them for three-phase,F they will work fine on 240 VAC single phase.  The electronics is wiredC between hot and neutral, so you would either need 240V with neutraloI (e.g., a 3 pole 4 wire grounding plug/receptacle such as a NEMA 14-30P/R)eF or to rewire the logic power supply or install a stepdown transformer.   ------------------------------  % Date: Wed, 28 May 2003 22:19:01 +0100d* From: "John Travell" <john@travell.uk.net>  Subject: Re: FREE TO A GOOD HOME5 Message-ID: <bb3anm$52feo$1@ID-120847.news.dfncis.de>e  D "Bill/Carolyn Pechter" <pechter@shell.monmouth.com> wrote in message' news:bb30g5$sf1$1@shell.monmouth.com...n* > In article <bammb5$1o2$1@pcls4.std.com>,: > Michael Moroney <moroney@world.std.spaamtrap.com> wrote: > >eF > >One model was smart enough to not spin if the phasing was reversed,I > >the other would spin _backwards_ (but was smart enough not to load thei* > >heads). Forget which drives those were. > >s > >-Mike >e' I once saw an RP04 that spun backwards.mC The heads would not load, but there was nothing 'magic' about it...nK There was a baffle that partially closed off the hole in the drum wall that H the heads came through. This baffle was opened by the airflow inside theJ drum going the right way round. If the pack was spinning backwards the theI airflow blew the baffle closed. The heads would not load until the baffle 	 was open.      -- John Travell  VMS crashdump expertise for hire john@travell.uk.net  http://www.travell.uk.net/           --- & Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).A Version: 6.0.484 / Virus Database: 282 - Release Date: 27/05/2003u   ------------------------------   Date: 28 May 2003 18:27 CDT0' From: carl@gerg.tamu.edu (Carl Perkins)   Subject: Re: FREE TO A GOOD HOME- Message-ID: <28MAY200318270962@gerg.tamu.edu>I  1 Kilgallen@SpamCop.net (Larry Kilgallen) writes...rr }In article <XXHgubKQctvR@eisner.encompasserve.org>, koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:[ }> In article <1030527195925.2916C-100000@Ives.egh.com>, John Santos <JOHN@egh.com> writes:l }>> A }>> There are lots of farms in the area; maybe the local electrict@ }>> company is more inclined to put in 3-phase because some farm }>> equipment requires it? }> pF }>    In this reguard farms are more like any small business than justH }>    another home.  My grandfather had lots of three phase equipment onI }>    his farm.  Also there was a transformer on a pole on the farm, with H }>    it own emergency cutoff (accessible to the fire department), just  }>    to serve the farm. } K }How did they make it accessible to the fire department but not to vandals.m  8 It may have something to do with a farmer and a shotgun.  I (How did you make your entire house accessible to the fire department butr not to vandals?)   --- Carl   ------------------------------    Date: 28 May 2003 22:05:47 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen)   Subject: Re: FREE TO A GOOD HOME3 Message-ID: <PXC22Cz3+GWo@eisner.encompasserve.org>-  W In article <28MAY200318270962@gerg.tamu.edu>, carl@gerg.tamu.edu (Carl Perkins) writes:m3 > Kilgallen@SpamCop.net (Larry Kilgallen) writes...-t > }In article <XXHgubKQctvR@eisner.encompasserve.org>, koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:] > }> In article <1030527195925.2916C-100000@Ives.egh.com>, John Santos <JOHN@egh.com> writes:t > }>> C > }>> There are lots of farms in the area; maybe the local electric-B > }>> company is more inclined to put in 3-phase because some farm > }>> equipment requires it? > }>  H > }>    In this reguard farms are more like any small business than justJ > }>    another home.  My grandfather had lots of three phase equipment onK > }>    his farm.  Also there was a transformer on a pole on the farm, with J > }>    it own emergency cutoff (accessible to the fire department), just  > }>    to serve the farm. > } M > }How did they make it accessible to the fire department but not to vandals.  > : > It may have something to do with a farmer and a shotgun. > K > (How did you make your entire house accessible to the fire department butn > not to vandals?)  G I put a lock on my house, and the fire department may have to break in.e  > I am forbidden to put a lock on the power cutout for my house.   ------------------------------  # Date: Wed, 28 May 2003 22:48:44 GMTo& From: Don Rogstad <nobody@nowhere.com>4 Subject: Getting LK411-AA keyboard to work correctly* Message-ID: <3ED53D5E.4070201@nowhere.com>  D I have an AlphaServer 1000A 5/400 server running OpenVMS V7.3-1.  ItE came with a RT6856T (which I assume is a LK443AA_PC) which is the PC oG keyboard layout.   I have since acquire a LK411 keyboard (one that has tB the "DO" key etc), but I can't get DECWindows Motif CDE v1.2-6 to E recongnise it.  Motif still thinks it is a PC style keydoard and the n "DO" key doesn't work.      I have tried:7 1) Ensuring that the keyboard console variable to LK411nF 2) Setting the keyboard type to "NORTH_AMERICAN_LK401AA" in the Style  Manager.I 3) I thought I tried setting the variable "DECW$DEFAULT_KEYBOARD_MAP" to nF   "NORTH_AMERICAN_LK401AA" in the DECW$PRIVATE_SERVER_SETUP.COM file, - but I see that it is commented out right now.I  H Does anyone know how to get this keyboard recongnized as a true Digital I keyboard?   I did notice in the AlphaServer 1000A 5/400 Golden Eggs that rF it refers to the keyboard for VMS as LK461.  Does this mean the LK411  won't work?l   Thanks,  Don RogstadP newvision at <lastname> dot como   ------------------------------    Date: 28 May 2003 14:33:59 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen)oA Subject: Re: GUI Applications on VMS (was: Portents of VMS death)D3 Message-ID: <BIndAO6aIzxr@eisner.encompasserve.org>   ` In article <bb2veb$588e4$1@ID-135708.news.dfncis.de>, bill@cs.uofs.edu (Bill Gunshannon) writes:  G > Careful there, VMS has a GUI.  It is based on considerably older codenG > than any of the current Unix (actually X-windows, Unix has no GUI anytL > more than VMS does) ones.  VMS also lacks most of the X based applicationsK > that make the GUI anything other than a good way to waste machine cycles.e  @ Although the VMS Debugger Motif interface has its pros and cons,& I would hate to not have it available.  3 Certainly the accompanying Heap Analyzer is superb.t  J I do not use DECwindows mail, because I prefer the command line interface,G but speaking of that, some have said DECterm is far better than what ism offered on some Unix systems.o   ------------------------------  % Date: Wed, 28 May 2003 12:45:52 -0700f* From: "Alder" <MUNDDGNTDYTV@spammotel.com>, Subject: Hobbyist seeks generic SIMM vendors+ Message-ID: <3ed511ef$1@obsidian.gov.bc.ca>e  J If anyone has links to online vendors known to stock SIMMs compatible withI "older" AlphaServers (mine is a 1000A 4/266), could you please post links H here.  I've contacted Island Computers, but the price is still beyond my& reach.  Best would be sites in Canada.   Thanks,t Aldert   ------------------------------  # Date: Wed, 28 May 2003 20:05:48 GMTh4 From: brad@.gateway.2wire.net (Bradford J. Hamilton)0 Subject: Re: Hobbyist seeks generic SIMM vendors0 Message-ID: <wE8Ba.1014360$3D1.591460@sccrnsc01>  X In article <3ed511ef$1@obsidian.gov.bc.ca>, "Alder" <MUNDDGNTDYTV@spammotel.com> writes:K >If anyone has links to online vendors known to stock SIMMs compatible withtJ >"older" AlphaServers (mine is a 1000A 4/266), could you please post linksI >here.  I've contacted Island Computers, but the price is still beyond myi' >reach.  Best would be sites in Canada.-   Pyramid Technology Services    http://www.pyramiddec.com/  J *may* have what you are looking for, but you need to know the part number.  L For instance, a PB7MA-AD (128MB Memory AS1000) is available for USD$336.00 -2 but I don't know if that is exactly what you need.  7 Crucial does not have what you need, nor does Kingston.k   >r >Thanks, >Alder >1 >t >a >   A _________________________________________________________________t0 Bradford J. Hamilton			"All opinions are my own"/ bMradAhamiPltSon@atMtAbi.cPoSm		"Lose the MAPS"u   ------------------------------  % Date: Wed, 28 May 2003 13:41:15 -0700 * From: "Alder" <MUNDDGNTDYTV@spammotel.com>0 Subject: Re: Hobbyist seeks generic SIMM vendors+ Message-ID: <3ed51eea$1@obsidian.gov.bc.ca>a   Yikes,  J To get what I "want", three 256MB options, would cost $2,400 $CDN.  Anyone* have an AS1000A with 1GB RAM for less? :-)   Alderc  A "Bradford J. Hamilton" <brad@.gateway.2wire.net> wrote in messagem* news:wE8Ba.1014360$3D1.591460@sccrnsc01...5 > In article <3ed511ef$1@obsidian.gov.bc.ca>, "Alder" $ <MUNDDGNTDYTV@spammotel.com> writes:H > >If anyone has links to online vendors known to stock SIMMs compatible withL > >"older" AlphaServers (mine is a 1000A 4/266), could you please post linksK > >here.  I've contacted Island Computers, but the price is still beyond myp) > >reach.  Best would be sites in Canada.v >  > Pyramid Technology Services0 >0 > http://www.pyramiddec.com/ >sL > *may* have what you are looking for, but you need to know the part number. >0A > For instance, a PB7MA-AD (128MB Memory AS1000) is available foru USD$336.00 -4 > but I don't know if that is exactly what you need. >@9 > Crucial does not have what you need, nor does Kingston.' >  > >-
 > >Thanks, > >Alder > >  > >r > >  > >o > C > _________________________________________________________________s0 > Bradford J. Hamilton "All opinions are my own"0 > bMradAhamiPltSon@atMtAbi.cPoSm "Lose the MAPS"   ------------------------------  % Date: Wed, 28 May 2003 16:31:09 -0500n2 From: "Stuart Johnson" <ssj152 AT charter DOT net>0 Subject: Re: Hobbyist seeks generic SIMM vendors/ Message-ID: <vdaan7n13ji429@corp.supernews.com>t  5 "Alder" <MUNDDGNTDYTV@spammotel.com> wrote in messager% news:3ed511ef$1@obsidian.gov.bc.ca...wL > If anyone has links to online vendors known to stock SIMMs compatible withK > "older" AlphaServers (mine is a 1000A 4/266), could you please post linkseJ > here.  I've contacted Island Computers, but the price is still beyond my( > reach.  Best would be sites in Canada. >o	 > Thanks,  > Alders >c >g >  >oJ Does your system use 4 additional memory chips along with the 4 banks of 4J chips? The additional 4 are immediately below the memory array. If NOT youH can use 72 pin Fast Page mode parity 5volt from Kahlon (www.kahlon.com),L which are MUCH more affordable. For example, the 64 mb chips are $31 US each! (remember, you need 4 at a time).t   Stuart Johnson   ------------------------------    Date: 28 May 2003 16:27:47 -07001 From: nothome@spammers.are.scum (Malcolm Dunnett)M0 Subject: Re: Hobbyist seeks generic SIMM vendors- Message-ID: <VkmjFyNe5ffK@malvm7.mala.bc.ca.>-  / In article <vdaan7n13ji429@corp.supernews.com>, 8     "Stuart Johnson" <ssj152 AT charter DOT net> writes:   > 7 > "Alder" <MUNDDGNTDYTV@spammotel.com> wrote in messagee' > news:3ed511ef$1@obsidian.gov.bc.ca...eM >> If anyone has links to online vendors known to stock SIMMs compatible withTL >> "older" AlphaServers (mine is a 1000A 4/266), could you please post linksK >> here.  I've contacted Island Computers, but the price is still beyond mya) >> reach.  Best would be sites in Canada.< >>
 >> Thanks, >> Alder >> >> >> >>L > Does your system use 4 additional memory chips along with the 4 banks of 4L > chips? The additional 4 are immediately below the memory array. If NOT youJ > can use 72 pin Fast Page mode parity 5volt from Kahlon (www.kahlon.com),N > which are MUCH more affordable. For example, the 64 mb chips are $31 US each# > (remember, you need 4 at a time).e  L   If your system DOES use the 5th row of SIMMS ( these are for ECC ) you canG still use the SIMMs mentioned above, just get a fifth one to install inlI the extra slot. The system will only use enough bits in the fifth SIMM tol; implement its ECC algorithm, but the extra bits won't hurt.o   ------------------------------  # Date: Thu, 29 May 2003 00:17:02 GMTd( From: Alder <PGDEHMKOKIMD@spammotel.com>0 Subject: Re: Hobbyist seeks generic SIMM vendors, Message-ID: <3ED55173.2000801@spammotel.com>   Thanks, folks.  I'm elated :-)  H So if I understand this correctly, the four 32MB SIMMs I currently have E installed could serve as the ECC SIMMs for banks full of 64MB SIMMs?  F Hmmm....if I can do that, that's REALLY good news.  Instead of $2400, 6 I'd have to cough up around $700 CDN for a 1GB system!  
 Brilliant!   Aldera  I The AS-1000A 4/266 does indeed have the ECC bank in addition to the four r@ main RAM banks.  If I understand your point, I could use each of Malcolm Dunnett wrote:1 > In article <vdaan7n13ji429@corp.supernews.com>, : >     "Stuart Johnson" <ssj152 AT charter DOT net> writes: >  > 7 >>"Alder" <MUNDDGNTDYTV@spammotel.com> wrote in message ' >>news:3ed511ef$1@obsidian.gov.bc.ca...  >>M >>>If anyone has links to online vendors known to stock SIMMs compatible with L >>>"older" AlphaServers (mine is a 1000A 4/266), could you please post linksK >>>here.  I've contacted Island Computers, but the price is still beyond my ) >>>reach.  Best would be sites in Canada.h >>>i
 >>>Thanks, >>>Alder >>>  >>>r >>>y >>>c >>L >>Does your system use 4 additional memory chips along with the 4 banks of 4L >>chips? The additional 4 are immediately below the memory array. If NOT youJ >>can use 72 pin Fast Page mode parity 5volt from Kahlon (www.kahlon.com),N >>which are MUCH more affordable. For example, the 64 mb chips are $31 US each# >>(remember, you need 4 at a time).= >  > N >   If your system DOES use the 5th row of SIMMS ( these are for ECC ) you canI > still use the SIMMs mentioned above, just get a fifth one to install ingK > the extra slot. The system will only use enough bits in the fifth SIMM tol= > implement its ECC algorithm, but the extra bits won't hurt.? >    ------------------------------    Date: 28 May 2003 17:43:01 -07001 From: nothome@spammers.are.scum (Malcolm Dunnett) 0 Subject: Re: Hobbyist seeks generic SIMM vendors- Message-ID: <9pyi0r4ojZKu@malvm7.mala.bc.ca.>2  W In article <3ED55173.2000801@spammotel.com>, Alder <PGDEHMKOKIMD@spammotel.com> writes:w  > Thanks, folks.  I'm elated :-) > J > So if I understand this correctly, the four 32MB SIMMs I currently have G > installed could serve as the ECC SIMMs for banks full of 64MB SIMMs?  H > Hmmm....if I can do that, that's REALLY good news.  Instead of $2400, 8 > I'd have to cough up around $700 CDN for a 1GB system! > E     No, I don't think that will work - the ECC sim has to be the sameuD depth as the other SIMMS, but not the same width ( ie you would needJ to use a 64MB SIMM but the system will only use some of the leads on it ).  G     You say you only have four 32MB SIMMS now, not five. If your system C uses the newer motherboard (which I thought the 1000A did ) then itx$ doesn't have the fifth SIMM for ECC.   ------------------------------  # Date: Thu, 29 May 2003 03:30:45 GMT + From: "Joe Heimann" <heimann@ecs.umass.edu>p0 Subject: Re: Hobbyist seeks generic SIMM vendors+ Message-ID: <3ed57ee3@news-1.oit.umass.edu>r  ) Alder <MUNDDGNTDYTV@spammotel.com> wrote:7L > If anyone has links to online vendors known to stock SIMMs compatible withK > "older" AlphaServers (mine is a 1000A 4/266), could you please post links-J > here.  I've contacted Island Computers, but the price is still beyond my( > reach.  Best would be sites in Canada.  	 > Thanks,A > Aldert  H Data Memory Systems, www.datamem.com, has listings for that model.  TheyF want $50 for a 64 MB ECC memory kit if I am reading the listing right.D The prices for larger kits are listed as "Call".  Use their "Memory H Finder" to search for the AlphaServer listing.  Someone mentioned KahlonF in this newsgroup not too long ago.  I have not used them, but a quick0 search of their site came up with the following:  5         http://www.kahlon.com/result.asp?modelid=1862"  F They list the same prices for up to the 64 MB kit, and list prices forH larger kits.  But the description of the kits does not seem to match theC specifications above that kits need to be 5 matched SIMMs per bank.e    Joe Heimann   ------------------------------  # Date: Thu, 29 May 2003 03:42:35 GMTd( From: Alder <PGDEHMKOKIMD@spammotel.com>0 Subject: Re: Hobbyist seeks generic SIMM vendors, Message-ID: <3ED5819F.2090903@spammotel.com>  H I believe you're referring to the words "multiple modules" for the 64MB F kit.  I've requested some info from Kahlon and will wait to hear what A they have to say for themselves about that, and about warranties.    Cheers,n   Alderf   Joe Heimann wrote:+ > Alder <MUNDDGNTDYTV@spammotel.com> wrote:V > L >>If anyone has links to online vendors known to stock SIMMs compatible withK >>"older" AlphaServers (mine is a 1000A 4/266), could you please post links J >>here.  I've contacted Island Computers, but the price is still beyond my( >>reach.  Best would be sites in Canada. >  > 	 >>Thanks,  >>Alders >  > J > Data Memory Systems, www.datamem.com, has listings for that model.  TheyH > want $50 for a 64 MB ECC memory kit if I am reading the listing right.F > The prices for larger kits are listed as "Call".  Use their "Memory J > Finder" to search for the AlphaServer listing.  Someone mentioned KahlonH > in this newsgroup not too long ago.  I have not used them, but a quick2 > search of their site came up with the following: > 7 >         http://www.kahlon.com/result.asp?modelid=1862d > H > They list the same prices for up to the 64 MB kit, and list prices forJ > larger kits.  But the description of the kits does not seem to match theE > specifications above that kits need to be 5 matched SIMMs per bank.  >  >  Joe Heimann >  >  >    ------------------------------  # Date: Thu, 29 May 2003 03:44:19 GMT=( From: Alder <PGDEHMKOKIMD@spammotel.com>0 Subject: Re: Hobbyist seeks generic SIMM vendors, Message-ID: <3ED58206.6070208@spammotel.com>  H No. it has the fifth.  I was referring only to the memory that "counts" : when talking about how much RAM a system has to work with.   Regards,   Aldert       Malcolm Dunnett wrote:Y > In article <3ED55173.2000801@spammotel.com>, Alder <PGDEHMKOKIMD@spammotel.com> writes:a >   >>Thanks, folks.  I'm elated :-) >>J >>So if I understand this correctly, the four 32MB SIMMs I currently have G >>installed could serve as the ECC SIMMs for banks full of 64MB SIMMs? eH >>Hmmm....if I can do that, that's REALLY good news.  Instead of $2400, 8 >>I'd have to cough up around $700 CDN for a 1GB system! >> > G >     No, I don't think that will work - the ECC sim has to be the sametF > depth as the other SIMMS, but not the same width ( ie you would needL > to use a 64MB SIMM but the system will only use some of the leads on it ). > I >     You say you only have four 32MB SIMMS now, not five. If your system E > uses the newer motherboard (which I thought the 1000A did ) then itR& > doesn't have the fifth SIMM for ECC. >    ------------------------------   Date: 29 May 03 07:21:05 +0200) From: p_sture@elias.decus.ch (Paul Sture)e/ Subject: Re: How to make a shadowed system disk ) Message-ID: <EgNsHJcHFhnh@elias.decus.ch>e  w In article <01KWFMPU44UWAKVGCS@sysdev.deutsche-boerse.com>, Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com> writes:lE >> No SHADOWING=2 means that you are using the current version of the, >> shadowing software. c$ >> SHADOWING=0  turns shadowing off.9 >> SHADOWING=1  was used for the older PHASE 1 shadowing.a > : > Right, sorry.  I suppose SHADOWING=1 is rare these days. >   C Not so much rare, as unsupported, and has been that way for years. t  < SYSGEN HELP Sys_Parameters SHADOWING no longer documents it.   -- v
 Paul Sture   ------------------------------  % Date: Wed, 28 May 2003 23:04:29 -0400z0 From: "Alan Boyles" <alan.boyles@mindspring.com> Subject: Invalid Disk format/ Message-ID: <vdau2jd152ocb5@corp.supernews.com>   L I was trying to install VMS 7.3-1 on an AlphaStation 500/500 ( yea, I got itJ working- see power problem ).  The system has 2 disks DKA0 and DKA100.  AtL the >>> prompt I do a show device and they both show up as RZ1BB-BS but whenG I tried to init dka0 as the system disk ( and when I try to init at DCL J prompt ) I am getting an INIT-F-INVALID Invalid Media Format.  This systemI orginally had NT on it and I am assuming that this is happening due to NTo> formating the disk.  How can I get this disk into ODS format ?   Thanks,h   Alan   ------------------------------  % Date: Wed, 28 May 2003 23:52:20 -0400p* From: JF Mezei <jfmezei.spamnot@istop.com>  Subject: Re: Invalid Disk format) Message-ID: <3ED583CB.560C518D@istop.com>a   Alan Boyles wrote:L > prompt ) I am getting an INIT-F-INVALID Invalid Media Format.  This systemK > orginally had NT on it and I am assuming that this is happening due to NTi@ > formating the disk.  How can I get this disk into ODS format ?  J Safest bet would be to go to the console media and then use the colsole to format the drive.   M Alternatively, you could try ANA/MEDIA  in a destructive fashion on the drivea and then try to INIT it.   ------------------------------    Date: 28 May 2003 11:17:39 -07001 From: susan_skonetski@hotmail.com (Sue Skonetski)  Subject: June event in NYC= Message-ID: <857e9e41.0305281017.1f523171@posting.google.com>i   NEW YORK - JUNE 16-17, 2003s  3 Monday June 16 - Linux and OpenVMS Interoperabilitye( John Wisniewski, HP - Seminar Instructor/ 9:00 a.m.-4:00 p.m., including networking luncht  E John Wisniewski has been a consultant for OpenVMS, Windows, and Linux C with DEC/CPQ/HPQ for the last 15 years. Over the last few years (ind= addition to his day job) John has been focused on Open SourceaE technologies, documenting his findings in his current book "Linux andn- OpenVMS Interoperability" from Digital Press.n  D During the Seminar John will talk about his book's insights and showE examples of how to use Open Source And common tools between Linux and @ OpenVMS. Standards, Tools, Crytography, TCP/IP, Xwindows, SAMBA,? Advanced server, and Open Source applications will be shown andlC discussed. A CD-R with Interoperability Tools will be provided. TenaE steps to breaking into a Linux or an OpenVMS system will be presented  but not documented.   A Encompass Events, a DFWLUG member, and one of the Founders of the-@ OpenVMS Hobbyist Program. Visit him and the DFWLUG on the web at3 http://vmsone.com,  http://openvmshobbyist.org, andF http://www.dfwcug.org.    E Tuesday June 17 - Configuring and Managing HP Enterprise Storage in a- Heterogeneous SAN Environment)* John Grundhoefer, HP - Seminar Instructor E 9:00 a.m.-5:00 p.m., including networking lunch and HP product updateT  C This seminar is designed to teach the participants how to setup andME manage the HP StorageWorks Enterprise Virtual Array (EVA), MA8000 andoE EMA12000 fibre channel storage controllers in a heterogeneous Storagec> Area Network environment. Attending this seminar will help youF understand how to install, configure, and manage HP StorageWorks fibreC channel RAID Arrays (RA/MA8000), Enterprise Storage Arrays (ESA/EMAcD 12000), and Enterprise Virtual Arrays (EVA).  You will leave knowingC how to configure a heterogeneous SAN using HP fibre channel networkpF infrastructure components and HP fibre channel software products. View the full abstract at< http://www.encompassUS.org/events/seminars/topics.html#Role.  E John Grundhoefer is an Enterprise Storage Consultant with HP based in}A Atlanta, Georgia. He has been delivering technical consulting andnA training for HP customers on Enterprise Storage Systems, OpenVMS,e> Windows NT, Alpha and VAX Computer Architectures, and computer= networks since joining Digital Equipment Corporation in 1979.    To register today, visitF http://www.encompassUS.org/events/seminars/registernow.html. There areB significant cost savings available to members. To learn more aboutE membership in Encompass visit http://www.encompassUS.org/membership/.o  ? If you are unable to attend these seminars, please forward thisnE message on to any colleagues who might benefit from this programming.a  C Watch the Encompass U.S. Web site at http://www.encompassUS.org formC the latest information and event registration.  Questions?  Contact @ Encompass U.S. Headquarters at +1.877.354.9887 (outside U.S. andD Canada, call +1.312.321.5151) or e-mail information@encompassUS.org.   ------------------------------  % Date: Wed, 28 May 2003 19:41:20 -0700o From: Jason Brady <>Y Subject: Re: Letter to HP Management from OpenVMS users around the world ... Shape up or  8 Message-ID: <36sadvs209eh468s757r7buaa0cc4p8m38@4ax.com>  F On Thu, 22 May 2003 16:33:58 GMT, "John Smith" <a@nonymous.com> wrote:   >> Write to the following:   [list omitted]   John,3  C Thanks for posting this information.  I wrote the following letter,bF with alterations depending on the recipient, to the entire list exceptE the institutional investors.  Will advise if I receive any responses.r  F ---------------------------------------------------------------------- [salutation]  E I am writing you, a Hewlett-Packard board member, to voice my concerneE that HP is making a serious mistake--one that is denying HP customerss@ of quality solutions to their business needs and shareholders of potentially lucrative returns.  F The mistake is HP's failure to actively promote and market its OpenVMS? operating system to customers, educational institutions and they@ information technology industry.  In fact, it appears that HP isC intent on slowly killing off this excellent product--one of the fewlF products capable of driving earnings and profits to higher levels than5 is possible with commodity product lines such as PCs.   kB Please review the attached article which eloquently discusses this2 subject in depth.  [photocopy of Inquirer article]  > In addition to being a customer and shareholder, I welcome theB opportunity to develop application software for OpenVMS...but onlyE when HP demonstrates a genuine commitment to building OpenVMS' marketh( share across a wide range of industries.  B Please join me in urging HP management to either take advantage ofD this opportunity or to sell the OpenVMS operating system division to/ another company that recognizes its true value.y  * Thank you for your time and consideration.   Regards,   [signature]    ------------------------------  % Date: Wed, 28 May 2003 14:16:13 -0400l* From: JF Mezei <jfmezei.spamnot@istop.com>3 Subject: Re: Logicals Lost When Process Is Detachedp) Message-ID: <3ED4FCEB.570F3084@istop.com>o   JOUKJ wrote:D > This is what I do now, but I would have preferred a "user" logical > name table1 >     i.e. why can we not do : define/user   ????   > $create/name_table/nolog/parent=LNM$SYSTEM_DIRECTORY chocolate2 $set security/class=logical_name_table chocolate -2         /OWNER=john_smith/PROT=(S:RWCD,O:RWCD,G,W)  C define/table=chocolate secret_recipe "dark chocolate, eggs, butter"e  . recipe = F$TRNLNM("secret_recipe","chocolate")  D You can also redefine the LNM$PROCESS_DIRECTORY table to include theJ "chocolate" table in the default search list for that process, which wouldJ allow the image to just specify "secret_recipe" and have VMS automatically5 find the translation from the "chocolate" name table.x   ------------------------------  % Date: Wed, 28 May 2003 14:22:25 -0400a* From: JF Mezei <jfmezei.spamnot@istop.com> Subject: Re: Mailbox) Message-ID: <3ED4FE5E.E76872DE@istop.com>t   Vivek wrote: >  > Hi Mezei,y > N > Is there any way to find out names of all the mailbox available/running at a > given point of time.  N SHOW DEV MB /FULL   but you'll get tons of mailboxes that don't belong to you.  I If your mailbox creates a logical name, then you could do something like:a  $ PIPE SHOW LOG | SEARCH SYS$INPUT MBA   ------------------------------    Date: 28 May 2003 15:17:01 -0500+ From: young_r@encompasserve.org (Rob Young)a! Subject: New MiniMerge capabilityn3 Message-ID: <p69Ru6Qc7SUe@eisner.encompasserve.org>a  2 	Well, how about something other than a whinefest?  6 	Perusing the latest OpenVMS Roadmaps, I stumbled upon, 	slide 11.  In the notes field for slide 11:  L "New Host-Based MiniMerge for Host Based Volume Shadowing (HBVS) - Design isN underway to create a new Host-based implementation of the MiniMerge capabilityL for all Fibre Channel solutions.  This new design will utilize the Write BitF Map technology currently used in the MiniCopy operation.  The previousL MiniMerge project (which relied upon the HSG80 Write-History Log capability)N has been cancelled, and this new host-based solution is underway.  By mid-JuneN 2003, a schedule will be published showing the schedule and supported versions for this new capability."y  > 	Hmmm.  This is good news as from the slide itself, it appears 	to be hardware independent:   		MiniMerge Support for HBVS 			o   New Host-Based MiniMerge # 			Solution for all FC Environmentse 				o   HSG,HSV,MSA,XP   	Few questions:   2 			1)  Will this be storage independent?  In other. 			words, will it work with 3rd party storage?  1 			2)  Can anyone in engineering give an overviewA/ 			of how it will work?  If used in conjunctionhL                         Write Bitmap, is it simple from a design standpoint  			(1)?e   	Thanks in advance.e   				Robn   (1)y  c http://groups.google.com/groups?selm=Wh9AbAyC0ZL9%40eisner.encompasserve.org&oe=UTF-8&output=gplain.  > 	Write bitmaps would run in single message mode.  Make it haveC 	VOLSHAD properties (if it doesn't) whereby write bitmap completionn> 	status in single message mode is default behavior (i.e. writeA 	completion doesn't occur until write hits bitmap and storage).  uI 	If you have to switch to buffered-message mode, than all bets are off.  n  ? 	The idea would be if a node crashes, all write bitmaps containoB 	outstanding writes and act as a write history log.  The node that@ 	crashed of course may/would be mastering a subset of shadowsets9 	out there that would have to have their controller baseds 	write history log walked.  F 	Maybe in a Shared Everything Galaxy of 16 - 4 CPU members you have a B 	1 in 16 chance of having a drive that needs something other than & 	"in memory mini-merge" to take place.  ; 	This would be a big help for folks with foreign storage ore? 	storage that *supposedly* won't get (or fairer: isn't planned)<; 	mini-merge capabilities.  At least a subset are undergoing   	full merge when a node crashes.  @ 	[I'm sure there is more to it than that.  Fun to guess anyhow.]   ------------------------------    Date: 28 May 2003 15:30:24 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen)r% Subject: Re: New MiniMerge capability-3 Message-ID: <G3x0oRsqCgAa@eisner.encompasserve.org>e  a In article <p69Ru6Qc7SUe@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes:g  @ > 	Hmmm.  This is good news as from the slide itself, it appears > 	to be hardware independent: >  > 		MiniMerge Support for HBVS! > 			o   New Host-Based MiniMerges% > 			Solution for all FC Environmentso > 				o   HSG,HSV,MSA,XP  A Fibre Channel only is not what I would call hardware independent.s   ------------------------------    Date: 28 May 2003 15:41:02 -0500+ From: young_r@encompasserve.org (Rob Young) % Subject: Re: New MiniMerge capabilityI3 Message-ID: <i2669xLibNsl@eisner.encompasserve.org>e  c In article <G3x0oRsqCgAa@eisner.encompasserve.org>, Kilgallen@SpamCop.net (Larry Kilgallen) writes:rc > In article <p69Ru6Qc7SUe@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes:  > A >> 	Hmmm.  This is good news as from the slide itself, it appears- >> 	to be hardware independent:- >> - >> 		MiniMerge Support for HBVSa" >> 			o   New Host-Based MiniMerge& >> 			Solution for all FC Environments >> 				o   HSG,HSV,MSA,XPo > C > Fibre Channel only is not what I would call hardware independent.     ) 	Okay, you got me there.  Read into that:-  C 	"it appears to be vendor neutral.  But here is the FC environments- 	they list:  "  < 	I don't see how you lock it into HP kit.  It does appear to< 	be platform independent and they just listed HP kit as they< 	certainly wouldn't want to promote 3rd party competing kit,? 	hence a question or two.  This is exciting stuff to me.  A bigt7 	reason to *never* turn away from Volume Shadowing ;-).5   				Rob6  u   ------------------------------  # Date: Wed, 28 May 2003 21:22:51 GMTm/ From: brooks@cuebid.zko.dec.nospam (Rob Brooks) % Subject: Re: New MiniMerge capability - Message-ID: <ivbM3zujkh8h@cuebid.zko.dec.com>h  - young_r@encompasserve.org (Rob Young) writes:l > 4 > 			1)  Will this be storage independent?  In other0 > 			words, will it work with 3rd party storage? >e? 	There is nothing in the work we're doing that would explicitlyeI exclude any type of storage, except storage is CI-attached, and therefore-L would use the older controller-assisted mini-merge.  In other words, I wouldG expect storage attached to HSZ's, for example, to take advantage of thebI work we're about to start.  Note that the above is not a committment.  It@K is *possible* that we may support just Fibre Channel storage, however.  ThenJ qualification process is lengthy enough to make support for every possibleD widget next-to-impossible.  However, since *no* changes will be madeD to any class or port driver, you can draw some conclusions regarding what should work.e  A > Can anyone in engineering give an overview of how it will work?  	Yes, a few of us could :-).  C 	We don't have the design details worked out yet.  For those of youo9 who will be at either the Vienna or London meetings, JohnlL Andruszkziewicz will have a few slides that discuss some of this.  Make sure you ask him many questions! :-)i  G > If used in conjunction with Write Bitmap, is it simple from a design t > standpoint? 1 	We think so, but the design is not complete yet.t   -- b  M Rob Brooks    VMS Engineering -- I/O Exec Group     brooks!cuebid.zko.dec.comt   ------------------------------    Date: 28 May 2003 16:21:22 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen)t% Subject: Re: New MiniMerge capability.3 Message-ID: <svwXRvHI1ZhJ@eisner.encompasserve.org>t  a In article <i2669xLibNsl@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes:ue > In article <G3x0oRsqCgAa@eisner.encompasserve.org>, Kilgallen@SpamCop.net (Larry Kilgallen) writes: d >> In article <p69Ru6Qc7SUe@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes: >> kB >>> 	Hmmm.  This is good news as from the slide itself, it appears  >>> 	to be hardware independent: >>>   >>> 		MiniMerge Support for HBVS# >>> 			o   New Host-Based MiniMerges' >>> 			Solution for all FC Environmentse >>> 				o   HSG,HSV,MSA,XP >> nD >> Fibre Channel only is not what I would call hardware independent. >  > + > 	Okay, you got me there.  Read into that:l > $ > 	"it appears to be vendor neutral.  @ Yes, that is a benefit, but different from hardware independent.   ===============   ? Of course VMS itself is not exactly hardware independent.  I amt@ sure it is headed down the tubes because VMS will not boot on my& totally industry standard Skil Saw :-)   ------------------------------    Date: 29 May 2003 00:06:04 -0500+ From: young_r@encompasserve.org (Rob Young)w% Subject: Re: New MiniMerge capabilityb3 Message-ID: <w18U9qjhZ66H@eisner.encompasserve.org>-  _ In article <ivbM3zujkh8h@cuebid.zko.dec.com>, brooks@cuebid.zko.dec.nospam (Rob Brooks) writes:o/ > young_r@encompasserve.org (Rob Young) writes:G  B >> Can anyone in engineering give an overview of how it will work? > 	Yes, a few of us could :-). > E > 	We don't have the design details worked out yet.  For those of youe; > who will be at either the Vienna or London meetings, JohniN > Andruszkziewicz will have a few slides that discuss some of this.  Make sure! > you ask him many questions! :-)9 > H >> If used in conjunction with Write Bitmap, is it simple from a design  >> standpoint? s3 > 	We think so, but the design is not complete yet.> >    	How about another stab?  7 	Instead of a single master with current write bitmaps:?  H http://h71000.www7.hp.com/doc/731FINAL/5423/5423pro_010.html#index_x_409  F 	Instead, make it so each shadow is mastered on 2 nodes.  In a 2 node B 	cluster, both nodes act as masters for all shadows.  To cut down @ 	message traffic, writes in progress on LBNs would flip bits to F 	represent 1024 blocks at a time.  Write Bitmap today is working with = 	127 block ranges.  A much larger grained portion of the diskn? 	must be tracked (I would guess) as you would double the numbert8 	of messages on writes if you kept the same size chunk.   ; 	More to walk, but storage is faster, CPUs are faster, etc.k   A 	Additionally, you could piggyback write-in-progress cancellationrD 	messages mixed in with in progress messages to the masters.  And/or? 	streamline in progress messages, buffer cancellation messages. ? 	Thinking about this a bit, I don't believe these messages havea> 	to be ACKed - just sent at start of IO.  Is that right?  This7 	would be a huge win as you wouldn't be hindering writeV: 	completion by waiting on a "write in-progress" ACK from a 	shadow bitmap master.   A 	This way , short of a total power outage in datacenter, one nodehA 	crashes, all shadows could be walked with "in memory mini-merge"6> 	from the nodes mastering the shadows.  Whether 1 node remains= 	or n nodes.   Since there are 2 masters for each volume , ifsC 	2 or more nodes remain, a mechanism for who walks might be whoevery@ 	grabs volume_walker_dsa[0-9]+ lock for the volume, gets to walk 	it.  3 	Add a SYSGEN parameter to turn it on and tweek it:D   	SHAD_MINIMERGEH 	0 = off - default 	1 = on - single master_ 	2 = on - two masters     9 	All this to tie-in/complement current write bitmap usagew8 	with MOUNT and DISMOUNT.  i.e. instead of write bitmaps: 	being circular and overwritten when volumes are in synch,; 	bitmaps will do both.  Circular for writes-in-progress AND ; 	historical for MINICOPY on re-mount.  Circular *must* have 8 	enough memory (and would as it is small), historical of8 	course depends on how much memory is available and that5 	is why you would continue to use /POLICY=OPTIONAL onp
 	dismount.  < 	For the folks that have large clusters, perhaps 1 cuts down< 	on traffic, at the risk of small subset of merging shadows.   	Earlier I said:  c http://groups.google.com/groups?selm=Wh9AbAyC0ZL9%40eisner.encompasserve.org&oe=UTF-8&output=gplainb  ? 	The idea would be if a node crashes, all write bitmaps contain B 	[records of] outstanding writes [in progress] and act as a write  	history log.  The node that@ 	crashed of course may/would be mastering a subset of shadowsets9 	out there that would have to have their controller baseda 	write history log walked.  F 	Maybe in a Shared Everything Galaxy of 16 - 4 CPU members you have a B 	1 in 16 chance of having a drive that needs something other than & 	"in memory mini-merge" to take place.   ---   E 	That method would be a serious limitation for 2 node or small numbers; 	of nodes as in a 2 node cluster half your shadows would bea 	in merge.  Not good.    				Robr   ------------------------------    Date: 28 May 2003 20:02:20 -00004 From: Doc.Cypher <Use-Author-Address-Header@[127.1]>< Subject: Re: NT: son of VMS? (was Re: Portents of VMS death)6 Message-ID: <20030528200220.26819.qmail@gacracker.org>  H On Wed, 28 May 2003, duncan@macdonald.compulink.co.uk (Duncan Macdonald) wrote:   <snip - Cutler's coding in VMS>e  O >He was the original author for much of it. And the code was good quality - an  D >unusual combination of tight efficient coding and clear design and ) >good commenting and a very low bug rate.s  @ I think good commenting generally helps keep the bug rate down.   ; Others seem to think that job security is more important...r http://mindprod.com/unmain.htmli     Doc. -- eK OpenVMS.  Eight out of ten hackers                   https://vmsbox.cjb.netAK           prefer *other* operating systems.        http://althacker.cjb.nete   ------------------------------  % Date: Wed, 28 May 2003 15:04:35 -0400s8 From: Steve Lionel <Steve.Lionel@not-specified-here.zyx>< Subject: Re: NT: son of VMS? (was Re: Portents of VMS death)8 Message-ID: <b12adv8t25c3dqq4g31ctutv8j3paj5qdh@4ax.com>  G On Wed, 28 May 2003 17:23 +0100 (BST), duncan@macdonald.compulink.co.uk  (Duncan Macdonald) wrote:c  O >He was the original author for much of it. And the code was good quality - an rD >unusual combination of tight efficient coding and clear design and ) >good commenting and a very low bug rate.d  K If you're referring to Dave Cutler here, I think you're very much mistaken. M Cutler's code may indeed have been tight and efficient, but it also tended to H be fragile, highly dependent on undocumented behaviors, uncommented, andN obscure to the point of unmaintainable.  The fragility part was the worst - itH ran the tests Dave used (few that they were), but tended to fall over inF real-world use and if anything in the environment changed.  Dave was aI brilliant design-guy, and probably a hacker's hero, but he was not, in my3" opinion, a good software engineer.  L NT is really the son of Cutler's VAXELN - at least in NT's original concept.M Like VAXELN, the concept was breathakingly elegant, but full of problems when M implemented as designed.  Like VAXELN (and VAX PL/I and VAX C, Cutler's othernB software projects while at DEC), NT took 2-3 years of "cleanup and6 reimplementation"  by others to make it actually work.  M Cutler indeed had very little to do with VMS after V1 shipped, and by 1980 heyK was off doing other things.  What you know of VMS today is worlds away from J the RSX-clone that was VMS V1, and I credit others, such as Dick Hustvedt,N Peter Conklin and Tom Hastings, with the vision that made VMS as successful as it was (and is!)  3 Steve (this post reflects my personal opinion only)   D Please send Visual Fortran support requests to vf-support@compaq.com   Steve Lionel Software Products Division Intel Corporationa
 Nashua, NH  = User communities for Intel Fortran and Compaq Visual Fortran:-    http://intel.com/IDS/community   ------------------------------  % Date: Wed, 28 May 2003 16:12:50 -0400>* From: JF Mezei <jfmezei.spamnot@istop.com>< Subject: Re: NT: son of VMS? (was Re: Portents of VMS death)) Message-ID: <3ED51838.25A39D24@istop.com>    "Doc.Cypher" wrote:$= > Others seem to think that job security is more important...r! > http://mindprod.com/unmain.htmli   RTFL !!!! :-) :-)r     4.Look Busy0M use define statements to make made up functions that simply comment out theire arguments, e.g.: s.               #define fastcopy(x,y,z) /*xyz*/                ... @               fastcopy(array1, array2, size); /* does nothing */   ------------------------------  % Date: Wed, 28 May 2003 14:03:31 -0400e* From: JF Mezei <jfmezei.spamnot@istop.com>" Subject: Re: Portents of VMS death) Message-ID: <3ED4F9F2.6428A5A2@istop.com>    mist dragon wrote:F > That is true. However, this has not been the case and 99% of the VMSF > systems are old, not new. I belive that rather small amount of theseE > 250.000 Alpha systems are running newest os. It does not matter howD> > much VMS improves if more modern version of it does not have > customers.  K This is a fair point. Someone still at 5.5-2 will complain that VMS lacks a L large enough command recall, ability to save the recall buffer and reload itM etc etc, and then say that VMS is old and they should migrate to Unix because. UNIX has a GUI.o   This comes back to marketing.o  M What VMS needs is the computer equivalent of the Oldsmobile commercials "this6$ isn't your grandfather's Oldsmobile"  M VMS does have a GUI that competes against Unix. VMS has web servers, POP IMAP  FTP etc etc.    G I could see an add, featuring the shark with little fishes running awaytJ scared, and the caption "VMS IS BACK WITH A VENGENCE", and then below textK saying something akin to "we may have been silent for a number of years andpL hidden under another name, but VMS is back, loaded with all the modern toolsK you need to run your business while still benefiting from trusted core thaty you know is very reliable.   ------------------------------   Date: 28 May 2003 18:33:47 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)" Subject: Re: Portents of VMS death5 Message-ID: <bb2veb$588e4$1@ID-135708.news.dfncis.de>s  ) In article <3ED4F9F2.6428A5A2@istop.com>,o- 	JF Mezei <jfmezei.spamnot@istop.com> writes:- > 2 > VMS does have a GUI that competes against Unix.   E Careful there, VMS has a GUI.  It is based on considerably older code E than any of the current Unix (actually X-windows, Unix has no GUI any>J more than VMS does) ones.  VMS also lacks most of the X based applicationsI that make the GUI anything other than a good way to waste machine cycles.f, But then, we have addressed this before too.   bill   -- tJ Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>   n   ------------------------------  % Date: Wed, 28 May 2003 15:03:57 -04009* From: JF Mezei <jfmezei.spamnot@istop.com>" Subject: Re: Portents of VMS death) Message-ID: <3ED50818.B9D67314@istop.com>e   jlsue wrote:I > Fail-over clustering can be done by the cheapest crap out there.  ThosenM > configurations are not highly available and should not be implemented wherev< > availability is a means to achieve certain business goals.  N So now you are saying that smaller customers whose appliaction is not "clusterL saavy" and can only justify 2 or 3 nodes should simply get off VMS and go to other crap ?  N If you narrow the VMS marketplace to the very same market as Tandem NSK, then:( 	1-that market is too small to sustain 2M 	2- HP would be stupid to keep 2 products in the same market that can sustaino only 1  J And between Tandem and VMS, Tandem stays because tandem is the true "faultJ tolerant" that is used by the stock exchanges etc etc. Tandem has a marketK presence and isn't declared "dead". VMS has no market presence and has beenl, "forgotten" by its owners for over a decade.    L If you want VMS to survice and succeed, you MUST open its target market to aF much wider audience. HP is doing the exact opposite. And to get VMS toK succeed, you need to leverage its cluster capabilities to the smaller shops1< where VMS clearly has clustering edge over windows and Unix.  M There are only so many financial institutions and military installatiosn thatm require VMS.   ------------------------------  % Date: Wed, 28 May 2003 15:53:40 -0400l* From: JF Mezei <jfmezei.spamnot@istop.com>" Subject: Re: Portents of VMS death) Message-ID: <3ED513BB.F776FE40@istop.com>h   Bill Gunshannon wrote:L > more than VMS does) ones.  VMS also lacks most of the X based applicationsK > that make the GUI anything other than a good way to waste machine cycles.n  J VMS lacks most of the APPLICATIONS which make and OS anothing other than a! good way to waste machine cycles.s    L But if you want to have the applications, you need to start marketing the OSH openly to show that you truly intend to grow the market and thus make itG attractive to ISVs. Whether you lack sokje X application or some serverm, application such as SAP makes no difference.  F however, VMS , as am operating system, does provide the core needed to& port/write such "modern" applications.  M And I believe that some of the VMS engineers have hinted that they may updatee+ Motif to a current one. (not on vax, alas).s   If you built it they MAY come.5 If you don't build it, they definitely will NOT come.o  J Right now, HP is in the "don't built it" mentality since they have decided% that they won't come if you build it.e   ------------------------------  % Date: Wed, 28 May 2003 21:47:47 -0400i+ From: "Ron Milen" <milenronald@mailbag.net>n" Subject: Re: Portents of VMS death! Message-ID: <3ed568b4_3@newsfeed>m  J In this I have to agree.  I've in the past inherited a cluster with a lessH than ideal configuration.  Over the short & medium term you have to liveL with it.  Over the longer term you can either convince management to let youK fix the problem or, of course, find a new job :), hopefully before the @#$$t
 hits the fan.k  	 Ron Milenh milenronald@yahoo.comh  7 "JF Mezei" <jfmezei.spamnot@istop.com> wrote in messaged# news:3ED50818.B9D67314@istop.com...e > jlsue wrote:K > > Fail-over clustering can be done by the cheapest crap out there.  Those I > > configurations are not highly available and should not be implementedc whereM> > > availability is a means to achieve certain business goals. >aG > So now you are saying that smaller customers whose appliaction is noto "clusterK > saavy" and can only justify 2 or 3 nodes should simply get off VMS and gom to > other crap ? > J > If you narrow the VMS marketplace to the very same market as Tandem NSK, then:d) > 1-that market is too small to sustain 2rF > 2- HP would be stupid to keep 2 products in the same market that can sustaine > only 1 >fL > And between Tandem and VMS, Tandem stays because tandem is the true "faultL > tolerant" that is used by the stock exchanges etc etc. Tandem has a marketH > presence and isn't declared "dead". VMS has no market presence and has been. > "forgotten" by its owners for over a decade. >s >iL > If you want VMS to survice and succeed, you MUST open its target market to aeH > much wider audience. HP is doing the exact opposite. And to get VMS toG > succeed, you need to leverage its cluster capabilities to the smalleri shopsi> > where VMS clearly has clustering edge over windows and Unix. > J > There are only so many financial institutions and military installatiosn that > require VMS.   ------------------------------  % Date: Wed, 28 May 2003 22:21:48 -0400h2 From: rdeininger@mindspring.com (Robert Deininger)" Subject: Re: Portents of VMS deathL Message-ID: <rdeininger-2805032221480001@user-uinj4ps.dialup.mindspring.com>  2 In article <3ED50818.B9D67314@istop.com>, JF Mezei" <jfmezei.spamnot@istop.com> wrote:  
 >jlsue wrote:oJ >> Fail-over clustering can be done by the cheapest crap out there.  ThoseN >> configurations are not highly available and should not be implemented where= >> availability is a means to achieve certain business goals.s >eO >So now you are saying that smaller customers whose appliaction is not "clusterhM >saavy" and can only justify 2 or 3 nodes should simply get off VMS and go toa
 >other crap ?A  7 I just re-read his post, and I didn't see him say that.a   ------------------------------  % Date: Wed, 28 May 2003 22:19:50 -0400<2 From: rdeininger@mindspring.com (Robert Deininger)" Subject: Re: Portents of VMS deathL Message-ID: <rdeininger-2805032219500001@user-uinj4ps.dialup.mindspring.com>  = In article <7500353b.0305280248.2ecd5e37@posting.google.com>,>/ mistdragon@zdnetonebox.com (mist dragon) wrote:.  J >> There are few subsystems in the VMS kernel that must run on the primaryK >> processor.  Most of the kernel can run on any available processor.  WithsH >> each recent release, less code is dependent on the primary processor. > E >That is true. However, this has not been the case and 99% of the VMSi >systems are old, not new. n  H You must have a very good information source, Oh Dragon.  Please tell us how you know this.  + >I belive that rather small amount of thesetD >250.000 Alpha systems are running newest os. It does not matter how= >much VMS improves if more modern version of it does not haveu >customers.:  H Indeed.  Customers running OpenVMS AXP 1.5 are not getting much good outH of the V7.3-1 release.  Undoubtedly, that is somehow HP's fault.  PleaseG tell us how many alpha systems are running each VMS release.  It soundsN like interesting information.e  J And since you started by comparing VMS performance to unix, please tell usJ how many unix users are using each version.  Or does version not matter on unix?,  C >> I've rarely seen VMS in this situation.  It typically requires aaC >> pathalogical workload, or a nincompoop at the controls, or both.  >_D >Or someone who needs the performance. E.g. if you want performance, >stay off VMS as you say :)   ; Did I say that?  I didn't know dragons were ventriloquists.a  I I know lots of VMS sites getting very good performance from their alphas.d  K >> functionality included.  On the other hand, many parts of the kernel canaI >> be tuned and configured on a live system, so there is seldom a need to  >> reload execlets.0 > 9 >Again, if you want the performance then there is need :)   J Please explain how you would improve performance by hot-swapping pieces of the VMS kernel.     K >> This is a pretty meaningless claim unless you specify which large alpha,aL >> which version of each OS, and define the workload.  Cases are known where@ >> VMS does better, and cases are known where Tru64 does better. > L >> Well, I think a nincompoop must have set up "that load", whatever it is. C >> It isn't unusual to see VMS lag Tru64 by 10-15% on CPU-intensiveDM >> workloads, with reasonable tuning.  I've never seen such a wide gap as you  >> describe. >tD >Of course. But that would identify the customer and this I will not >do. o  E I don't blame you.  Someone knowlegeable might contact them, and they # wouldn't need your advice any more.D    >You just have to trust my word.  D Sure.  I've always gotten good advice from reptiles.  (Dragons _are_ reptiles, aren't they?)u  " >But mind you, this was top of theD >scale system. The point is that hw had a lot of steam left, but VMSF >was not just able to utilize it the way Tru64 was and the problem wasD >in the way VMS kernel services. Note that this did not apply to allC >services: tcp/ip and IO fell far before the rest of the kernel and1F >since those processes were not able to share load, it was not able toF >do the job. The situation was not artificial except for the amount of= >processes and decisions were made based on that examination.M  M I can't address your performance problems based on the available information.   K >> And VMS "freezes"?  I have (rarely) seen bugs lead to such symptoms, buth> >> almost always the cause has been really, really bad tuning. > D >With freeze I mean the performance comes so bad that processes slow >down to crawl.u  , Yup.  Sounds like really, really bad tuning.    F >> I've seen VMS scale very well up to 16 processors, and occasionallyJ >> beyond.  But not on poorly-scaling hardware like the previous GS-series >> systems.n >iF >So how much of GS systems are poorly scaling. How much of VMS systemsF >are on good-performing ones ? Again must be a small figure. What kind >of proof you have for this ?t  I GS160 and GS320 have well-known performance problems for some workloads. pI The root cause is the poor memory access times.  The best solution was tonF design better systems.  The interim solution was to adjust VMS to copeE with the platform's problems, and the engineers made huge progress ineF recent releases.  But Tru64 has pretty much the same problems on these systems, for the same reasons.  F The smaller systems have never presented a similar problem.  They haveH kinder memory architectures, and keeping 4 processors busy (for example)@ is not much challenge.  Keeping 10, 20, or 30 processors workingA efficiently is hard for some kinds of workloads, and the previoust' GS-series boxes aggravated the problem.n  D I have no proof for you, since I'm not in the business of selling or( proving products, let alone to reptiles.  G But HP offers test-drives and dedicated labs to help customers and ISVs I build, test, and tune suitable VMS configurations.  A variety of hardwaren is offered. H A common experiment is to compare the previous generation of big systems with the latest ones.p  H I think your friendly neighborhood VMS Ambassador can help you start theH process.  Expect to be put on a waiting list.  In my experience, the VMS test labs are busy.g    J >> Since you don't name names, we'll have to take your word that they were >> the "best gurus".   > C >Exactly :) I will not be saying more on this subject than that theiG >experts were from VMS development and their expertise was given by the  >vendor.  G Well, I've seen and heard details of many successes.  That doesn't ruleoG out a few failures.  But you imply the failures are widespread (withoutpG providing any details) and rant about the shortcomings of VMS.  The fewoF specific statements you've made about the innards of VMS are quite far1 from reality, so why should I take you seriously?c    ' >> Or perhaps you're just making it up.n >pG >That you will never know. Otoh, I seem to know more of VMS than you :)?   Yeah.  It's possible..    F >Or maybe I just make it up ? Do I really exist ? Questions for you to >consider :)  ! In my spare time.  Real Soon Now.-  E >> Calls are rather light on VMS.  Fork() is a well-known performancemG >> problem, and anyone with a clue chooses an algorithm that uses it as  >> little as possbile. >uB >Absolutely true. However, that verfies my point that creating VMSA >software requires different style of *nix and therefore very fewt >people bother to port for it.  G I agree.  T'would be nice if more quality software was ported to VMS bytG competent folks.  But I find myself wishing the same for all the unixes F I've used.  It's clear to me that many unix applications fall short ofI their potential, partly because so few programmers understand unix well. -! VMS and unix have that in common.   G Oh, by the way, why did you say calls are slow on VMS, and lead to big,h hard-to-optimize modules?i  ; >> I begin to wonder if we're talking about the same VMS...t >hF >That is a bold claim. While you claim not knowing my credibility, howE >would I know yours ? How would I know you are not inventing that allSC >by yourself ? If you wanted to insult me you failed :) Claiming ofnG >what we saw was not truth without any credibility of having performing B >equivalent system loads, tests and needs does not make you better >expert than me :)  C Well, name the configurations and the workloads.  Tell us where thep' bottleneck is.  Tell us what you tried.v   ------------------------------    Date: 28 May 2003 11:04:23 -0700. From: spamsink2001@yahoo.com (Alan E. Feldman)2 Subject: re Re: How to make a shadowed system disk= Message-ID: <b096a4ee.0305280958.4912446a@posting.google.com>.  | Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com> wrote in message news:<01KWFK65K7NAAKVGCS@sysdev.deutsche-boerse.com>...J > > I asked the wrong question.  Let me try again ... in part, you posted: >  > Not me; perhaps I quoted it. >  > >    $ EDT MODPARAMS.DAT > >  > >    ALLOCLASS=101 > >    SHADOWING=2 > >    SHADOW_SYS_DISK=1 > >    SHADOW_SYS_UNIT=0 > >    SHADOW_MAX_COPY=2 > >    SHADOW_SYS_TMO=20 > > E > >    then run autogen and reboot.  The above configuration is for a.B > >    shadow set of 2 disks with "DKA0" being the primary member. > G > It doesn't specify 2 disks; SHADOWING=2, IIRC, means that the system bJ > disk is shadowed.  (I think this is assumed if you set SHADOW_SYS_UNIT).  B SHADOWING=2 means Phase II shadowing is turned on and you can then create shadow sets.o  @ SHADOW_SYS_DISK=1 means the system disk is shadowed. Add membersD manually only once. The system disk shadow set will be reconstructedA automatically during subsequent boots, so you need not re-add anysE members that were in the system disk shadow set during shutdown. ThisEC is for the system disk only! Also, I repeat my warning to *not* addi; members to a system disk shadow set in your startup commandcA procedures. Doing so risks losing current data. See the shadowingnC manual -- system disk part -- for details. You do, however, need to A reconstruct *non*-system disk shadow sets in your startup commandp procedures.o  ? SHADOW_SYS_UNIT=0 means the system disk will come up as DSA0:. t  A SHADOW_MAX_COPY=2 means a maximum of two concurrent copy threads.t  D SHADOW_SYS_TMO=20 means a 20 sec. timeout (see the system parametersE appendix in the system management utilities manual for more details).o   [...].  C Be sure to add these parameters to SYS$SYSTEM:MODPARAMS.DAT and run  AUTOGEN!   Disclaimer: JMHO Alan E. Feldman    ------------------------------    Date: 28 May 2003 12:19:37 -0700. From: zcsessions@visionair.com (Zack Sessions)< Subject: Reading CD-R burned on a Window system on an Alpha?< Message-ID: <db13d9fb.0305281119.d603878@posting.google.com>  F The subject line says it all. I need to read a CD on an Alpha (OpenVMSE 6.2) that I burned on a CD-R blank on an Windows NT 4.0 system. Is iteF directly readable or would I need to add something to the Alpha system	 to do it?s  E Please post reponses here, email is an invalid address to avoid spam.    ------------------------------    Date: 28 May 2003 14:38:19 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen) @ Subject: Re: Reading CD-R burned on a Window system on an Alpha?3 Message-ID: <jGzwFGlQWm9B@eisner.encompasserve.org>w  m In article <db13d9fb.0305281119.d603878@posting.google.com>, zcsessions@visionair.com (Zack Sessions) writes:rH > The subject line says it all. I need to read a CD on an Alpha (OpenVMSG > 6.2) that I burned on a CD-R blank on an Windows NT 4.0 system. Is iteH > directly readable or would I need to add something to the Alpha system > to do it?e  D Alpha VMS V6.2 will read ISO-9660 compliant CD-R discs, but will notD read Joliet hierarchies.  Make sure you have your files entered into- the Primary hierarchy and you should be fine.v   ------------------------------  % Date: Wed, 28 May 2003 21:36:14 -0400 + From: "Ron Milen" <milenronald@mailbag.net> @ Subject: Re: Reading CD-R burned on a Window system on an Alpha?! Message-ID: <3ed562c0_3@newsfeed>m  J I understand about ISO disks.  I've been able to mount & read them before.F However, the term Joliet hierarchy is new to me.  Could someone please
 explain this?i   Thanks,t	 Ron Milenr milenronald@yahoo.com   : "Larry Kilgallen" <Kilgallen@SpamCop.net> wrote in message- news:jGzwFGlQWm9B@eisner.encompasserve.org...t> > In article <db13d9fb.0305281119.d603878@posting.google.com>,0 zcsessions@visionair.com (Zack Sessions) writes:J > > The subject line says it all. I need to read a CD on an Alpha (OpenVMSI > > 6.2) that I burned on a CD-R blank on an Windows NT 4.0 system. Is it J > > directly readable or would I need to add something to the Alpha system
 > > to do it?s >wF > Alpha VMS V6.2 will read ISO-9660 compliant CD-R discs, but will notF > read Joliet hierarchies.  Make sure you have your files entered into/ > the Primary hierarchy and you should be fine.o   ------------------------------  % Date: Wed, 28 May 2003 22:12:19 -0400e* From: JF Mezei <jfmezei.spamnot@istop.com>@ Subject: Re: Reading CD-R burned on a Window system on an Alpha?) Message-ID: <3ED56C61.776F7865@istop.com>    Ron Milen wrote: > L > I understand about ISO disks.  I've been able to mount & read them before.H > However, the term Joliet hierarchy is new to me.  Could someone please > explain this?r  J I asked this some time ago. Joliet is a Microsoft "addition" to ISO formatH which allows for NT-style file names. Such a drive still has an ISO 9660K directory with "regular" file names, but if you have the Joliet capability,s you see the "full" file names.  K My netgear router's documentation came on such a disk in HTML format. I can N copy the HTML from the "regular" side, however, the links in the hTML point to  the NT style files so they fail.   ------------------------------    Date: 28 May 2003 22:09:22 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen)0@ Subject: Re: Reading CD-R burned on a Window system on an Alpha?3 Message-ID: <1SgE5wx9pG3w@eisner.encompasserve.org>r  O In article <3ed562c0_3@newsfeed>, "Ron Milen" <milenronald@mailbag.net> writes:eL > I understand about ISO disks.  I've been able to mount & read them before.H > However, the term Joliet hierarchy is new to me.  Could someone please > explain this?o  J Microsoft made their own non-compliant derivative of the ISO-9660 standardF to implement Unicode file and directory names.  Their format is called Joliet.	   ------------------------------  # Date: Thu, 29 May 2003 01:54:02 GMTr- From: "John E. Malmberg" <wb8tyw@qsl.network>e7 Subject: Skil Saw? (was - Re: New MiniMerge capability)h: Message-ID: <_KdBa.873$cp6.653549@news1.news.adelphia.net>   Larry Kilgallen wrote: > A > Of course VMS itself is not exactly hardware independent.  I am B > sure it is headed down the tubes because VMS will not boot on my( > totally industry standard Skil Saw :-)  = I think that you should be able to get the CD-ROM to spin up.o  % What do you plan to use as a console?y   :-)o   -Johno wb8tyw@qsl.network Personal Opinion Only-   ------------------------------  # Date: Thu, 29 May 2003 02:54:33 GMTe' From: "Mark E. Levy" <melevy@attbi.com>:; Subject: Re: Skil Saw? (was - Re: New MiniMerge capability)b. Message-ID: <JDeBa.155865$rt6.46905@sccrnsc02>  8 "John E. Malmberg" <wb8tyw@qsl.network> wrote in message4 news:_KdBa.873$cp6.653549@news1.news.adelphia.net... > Larry Kilgallen wrote: > >hC > > Of course VMS itself is not exactly hardware independent.  I amoD > > sure it is headed down the tubes because VMS will not boot on my* > > totally industry standard Skil Saw :-) >e? > I think that you should be able to get the CD-ROM to spin up.t  I You're right, there's no future for VMS. I just tried to install it on myn> Sony Discman (hey, it _does_ have a CD drive) with no success.   -- a Mark E. Levy" System Management Associates, Inc. www.sysman-inc.com www.openvms.como   ------------------------------    Date: 28 May 2003 22:07:51 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen)y; Subject: Re: Skil Saw? (was - Re: New MiniMerge capability) 3 Message-ID: <m8nWXn7F5vjy@eisner.encompasserve.org>t  j In article <_KdBa.873$cp6.653549@news1.news.adelphia.net>, "John E. Malmberg" <wb8tyw@qsl.network> writes: > Larry Kilgallen wrote: >>  B >> Of course VMS itself is not exactly hardware independent.  I amC >> sure it is headed down the tubes because VMS will not boot on mys) >> totally industry standard Skil Saw :-)t > ? > I think that you should be able to get the CD-ROM to spin up.  > ' > What do you plan to use as a console?  >  > :-)s  ; For output, if I lock it running with the blade guard open,.> I expect it to carve the OPCOM messages in a piece of plywood.  N For input, I will control the carving and expect the Skil saw to interpret :-)   ------------------------------  % Date: Wed, 28 May 2003 23:30:49 -0400n* From: JF Mezei <jfmezei.spamnot@istop.com>; Subject: Re: Skil Saw? (was - Re: New MiniMerge capability)t) Message-ID: <3ED57EC1.5A8D828D@istop.com>>   "Mark E. Levy" wrote:LK > You're right, there's no future for VMS. I just tried to install it on myg@ > Sony Discman (hey, it _does_ have a CD drive) with no success.  N Now that VMS is owned by HP, does this mean that I could try to load VMS 5.5-2M  on any HP machine that has a DLT drive supposedly capable of reading TK50s ?i   ------------------------------  % Date: Thu, 29 May 2003 01:07:08 +0530P From: "Vivek" <visoni@bmc.com>) Subject: Re: SYS$QIOW - Performance Issue./ Message-ID: <vda494ncj9ehcf@corp.supernews.com>h   This is quite infomative...v  . For me the SYS$IOW read  taking a lot of time.  C The write is being done after reading the data from a file and thentK formating it the appriprate format which the reader process can understand.t     Thanks Vivekl      + <briggs@encompasserve.org> wrote in messagew- news:47ND$Lyzvk+A@eisner.encompasserve.org... J > In article <vd903d896crbad@corp.supernews.com>, "Vivek" <visoni@bmc.com> writes:: > > Hi  Mezei / Ron, > > L > > Could the problem be solved using  Read Attention AST. If there a lag in< > > writing datablock becuase of lot of processing required. > E > You really need to tell us what you are doing.  Are you writing theyF > reader or the writer or both?  What other slow things are the reader > and writer doing?  >W >oG > If you have a slow reader, the write process _will_ stall waiting forl
 > the reader.h >iF > If you use your original code without IO$M_NOW, the stall will occurE > when the writer does its $QIOW, waiting for the reader to issue the- > corresponding read.o >nC > If you use your modified code with IO$M_NOW, the stall will occurs< > when the mailbox fills up and your $QIOW write stalls in aB > miscellaneous wait state (RWMBX) until the reader reads from theC > mailbox and frees up space for a new message.  I believe you saidr: > you had sized the mailbox with a capacity of 2 messages. >aH > If you use $QIO instead of $QIOW, the stall will occur when the writerG > has allocated all its assigned local buffers and I/O status blocks towA > pending I/O and doesn't have the resources to queue up any more I > asynchronous writes.  Alternatively, if the writer runs out of bufferedf2 > I/O quota, the $QIOs would simply start failing. >l >oL > If you have a slow writer, the reader _will_ stall waiting for the writer.D > You can't read what hasn't been written yet.  You can use IO$M_NOWC > and poll for input, but that just burns CPU (perhaps the very CPUa> > that the writer needs to use in order to complete its work). >r >lD > If both reader and writer can be bottlenecks you could expect someD > moderate performance gains by putting a buffer between them (usingC > asynchronous I/O -- $QIOW with IO$M_NOW or $QIO without) in orderaC > to increase concurrency.  But you can't expect to buy more than ae > factor of two that way.s >o
 > John Briggs.   ------------------------------    Date: 28 May 2003 15:18:42 -0500 From: briggs@encompasserve.org) Subject: Re: SYS$QIOW - Performance Issueo3 Message-ID: <giFOgL0k93vu@eisner.encompasserve.org>m  P In article <vda494ncj9ehcf@corp.supernews.com>, "Vivek" <visoni@bmc.com> writes: > This is quite infomative...w > 0 > For me the SYS$IOW read  taking a lot of time.  @ Barring hideous performance issues elsewhere on the system, this is not plausible.w  D It is plausible that the writer process is slow somewhere other thanB in the SYS$QIOW write.  The reader will then wait until the writer has something to write.w  D It is plausible that the reader process is slow somewhere other than in the SYS$QIOW read.l  G But it is not plausible that the reader has a $QIOW read queued and the C writer has a $QIOW write queued and the resulting transaction takese2 anything more than a few milliseconds to complete.  @ You should be able to demonstrate this quite easily by replacing the reader with:   $ TYPE MBAn:  A And watch the writer's traffic spill out onto your screen just as,  fast as the writer can write it.  B Oh.  By the way.  One convenient tool to use to see what mailboxes. are being accessed by a particular process is:   $ ANALYZE /SYSTEMs- SDA> SET PROCESS /ID=<process-identification>o SDA> SHOW PROCESS /CHANNELS   E > The write is being done after reading the data from a file and thendM > formating it the appriprate format which the reader process can understand.   C Which, unless you have some hideous performance issues or are usinge@ a hideously inefficient formatting algorithm ought not cause the behavior you are seeing.   	John Briggs   ------------------------------  % Date: Wed, 28 May 2003 16:24:02 -0400r* From: JF Mezei <jfmezei.spamnot@istop.com>) Subject: Re: SYS$QIOW - Performance Issueu) Message-ID: <3ED51AD7.414080E5@istop.com>o   Vivek wrote:0 > For me the SYS$IOW read  taking a lot of time.  L have you puty "print" statements before and after the SYS$QIOW that does the read ?N (also, you might want to timestamp those print statements so you have an exact" idea of when things are occuring).   #include <descrip.h>   char buffer[32];" $DESCRIPTOR(log_time_desc,buffer);      ' status = LIB$DATE_TIME(&log_time_desc);n3 log_time_string[log_time_desc.dsc$w_length] = 0x00;V< printf("Before QIOW at: %s\n", log_time_desc.dsc$a_pointer);   status = SYS$QIOW( ...... )u  ' status = LIB$DATE_TIME(&log_time_desc);a3 log_time_string[log_time_desc.dsc$w_length] = 0x00; < printf("Before QIOW at: %s\n", log_time_desc.dsc$a_pointer);    L This would give you a much better idea of what is happening to your program.N Also, you might want to do the same in the program that writes to the mailbox.K This will let you see exatly how much time the write takes to complete, andaH how log it takes for the read to complete after the write has completed.   ------------------------------  % Date: Thu, 29 May 2003 03:06:09 +0530. From: "Vivek" <visoni@bmc.com>) Subject: Re: SYS$QIOW - Performance Issuen/ Message-ID: <vdaao560j89e43@corp.supernews.com>1  K Considering we do have the hideous performance issue some where else in then+ system and we cannot do anything about theme  G Could the following be some of the solutions in improving the over timel taken by the application.a   1. Increase bufquo   2. Increase bufquo   Curious....    Vivekg  + <briggs@encompasserve.org> wrote in messagen- news:giFOgL0k93vu@eisner.encompasserve.org... J > In article <vda494ncj9ehcf@corp.supernews.com>, "Vivek" <visoni@bmc.com> writes:d > > This is quite infomative...t > >n2 > > For me the SYS$IOW read  taking a lot of time. >pB > Barring hideous performance issues elsewhere on the system, this > is not plausible.f >tF > It is plausible that the writer process is slow somewhere other thanD > in the SYS$QIOW write.  The reader will then wait until the writer > has something to write.u >eF > It is plausible that the reader process is slow somewhere other than > in the SYS$QIOW read.  >lI > But it is not plausible that the reader has a $QIOW read queued and theeE > writer has a $QIOW write queued and the resulting transaction takesL4 > anything more than a few milliseconds to complete. >eB > You should be able to demonstrate this quite easily by replacing > the reader with: >i > $ TYPE MBAn: > C > And watch the writer's traffic spill out onto your screen just as " > fast as the writer can write it. >WD > Oh.  By the way.  One convenient tool to use to see what mailboxes0 > are being accessed by a particular process is: >  > $ ANALYZE /SYSTEM / > SDA> SET PROCESS /ID=<process-identification>s > SDA> SHOW PROCESS /CHANNELSv >tG > > The write is being done after reading the data from a file and thennC > > formating it the appriprate format which the reader process canP understand.i > E > Which, unless you have some hideous performance issues or are usingdB > a hideously inefficient formatting algorithm ought not cause the > behavior you are seeing. >e
 > John Briggs0   ------------------------------  % Date: Thu, 29 May 2003 03:19:56 +0530: From: "Vivek" <visoni@bmc.com>) Subject: Re: SYS$QIOW - Performance Issue3/ Message-ID: <vdabhshisqqad8@corp.supernews.com>s   Yeah,t  E I have put print statement before and after the read statement. ( Notn% exactly the code which have written).k  0 It has been taking 4 + minutes for each read....  D Not i have put IO$M_NOW function for each write in the application..  7 "JF Mezei" <jfmezei.spamnot@istop.com> wrote in message # news:3ED51AD7.414080E5@istop.com...  > Vivek wrote:2 > > For me the SYS$IOW read  taking a lot of time. >aJ > have you puty "print" statements before and after the SYS$QIOW that does theP > read ?J > (also, you might want to timestamp those print statements so you have an exactt$ > idea of when things are occuring). >c > #include <descrip.h> >y > char buffer[32];$ > $DESCRIPTOR(log_time_desc,buffer); >t >e > ) > status = LIB$DATE_TIME(&log_time_desc);.5 > log_time_string[log_time_desc.dsc$w_length] = 0x00;o> > printf("Before QIOW at: %s\n", log_time_desc.dsc$a_pointer); >i > status = SYS$QIOW( ...... )  >p) > status = LIB$DATE_TIME(&log_time_desc);d5 > log_time_string[log_time_desc.dsc$w_length] = 0x00;t> > printf("Before QIOW at: %s\n", log_time_desc.dsc$a_pointer); >U > E > This would give you a much better idea of what is happening to yourS program.G > Also, you might want to do the same in the program that writes to theg mailbox.I > This will let you see exatly how much time the write takes to complete,e andvJ > how log it takes for the read to complete after the write has completed.   ------------------------------  % Date: Thu, 29 May 2003 04:21:53 +0530  From: "Vivek" <visoni@bmc.com>) Subject: Re: SYS$QIOW - Performance Issues/ Message-ID: <vdaf6j20ef88e0@corp.supernews.com>o   Mezei, Whats log_time_string here ...  7 "JF Mezei" <jfmezei.spamnot@istop.com> wrote in messagel# news:3ED51AD7.414080E5@istop.com...  > Vivek wrote:2 > > For me the SYS$IOW read  taking a lot of time. >8J > have you puty "print" statements before and after the SYS$QIOW that does ther > read ?J > (also, you might want to timestamp those print statements so you have an exact4$ > idea of when things are occuring). >m > #include <descrip.h> >t > char buffer[32];$ > $DESCRIPTOR(log_time_desc,buffer); >  >a >u) > status = LIB$DATE_TIME(&log_time_desc);s5 > log_time_string[log_time_desc.dsc$w_length] = 0x00;u> > printf("Before QIOW at: %s\n", log_time_desc.dsc$a_pointer); >S > status = SYS$QIOW( ...... )  >l) > status = LIB$DATE_TIME(&log_time_desc);t5 > log_time_string[log_time_desc.dsc$w_length] = 0x00;h> > printf("Before QIOW at: %s\n", log_time_desc.dsc$a_pointer); >  > E > This would give you a much better idea of what is happening to youru program.G > Also, you might want to do the same in the program that writes to thet mailbox.I > This will let you see exatly how much time the write takes to complete,- and-J > how log it takes for the read to complete after the write has completed.   ------------------------------  % Date: Wed, 28 May 2003 20:08:35 -04006* From: JF Mezei <jfmezei.spamnot@istop.com>) Subject: Re: SYS$QIOW - Performance Issuea) Message-ID: <3ED54F68.143937C4@istop.com>n   Vivek wrote: > 1. Increase bufquo  M Increasing bufquo would allow you to increase the mailbox buffer size. But if I current IOs complete with all your data intact and no error message, thenuM increasing the bufquo won't give you much from the mailbox io's perspective.    M If you didn't have sufficient quota, the read operation would fail (heck, theTI crembx might fail) and the writes from the other processes would fail too M since you'd be asking them to write a bigger chunk of data than their processe
 quotas allow.d   ------------------------------  % Date: Wed, 28 May 2003 20:13:25 -0400d* From: JF Mezei <jfmezei.spamnot@istop.com>) Subject: Re: SYS$QIOW - Performance Issue") Message-ID: <3ED5508A.41C7B70A@istop.com>    Vivek wrote:F > Not i have put IO$M_NOW function for each write in the application..  L This shouldn't make a difference. You should simply use a QIOW for the write and let it complete.  L Again, during those 4 minutes you believe your process in inside the QIO forK the read, you need to use SHOW SYSTEM (or SHOW PROC/CONT/ID=xxxxx ) to find  out what the process status is.u  ' LEF: process waiting for IO to completeu# RWMBX or any other RWxxx : bad news-  * Also, are you using ASTs in your program ?  M Another thing you really need to check is when the READ IO completes. Does it K complete as soon as the write IO in the other process completes, or does ittK take 4 minutes between the time the writing process has succesfully writtenm* and the time the reading process gets it ?  J And during those 4 minutes, what does the reading process do ? with a showM proc/cont, do any of the IO and CPU counts change or does it really just stayn# there doing nothing for 4 minutes ?t   ------------------------------  % Date: Wed, 28 May 2003 20:16:17 -0400m* From: JF Mezei <jfmezei.spamnot@istop.com>) Subject: Re: SYS$QIOW - Performance Issuee) Message-ID: <3ED55137.E1F03D5F@istop.com>d   Vivek wrote: >  > Mezei,  > Whats log_time_string here ... > > char buffer[32];& > > $DESCRIPTOR(log_time_desc,buffer); > >  > >i > >h+ > > status = LIB$DATE_TIME(&log_time_desc); 7 > > log_time_string[log_time_desc.dsc$w_length] = 0x00;a    # Sorry, it should have been buffer .f  M (or define log_time-desc to point to a character array called log_time_stringd instead of "buffer"i   ------------------------------   End of INFO-VAX 2003.295 ************************