1 INFO-VAX	Tue, 17 Apr 2001	Volume 2001 : Issue 214       Contents:" Bad Block Locator Utility or EVRAC& Re: Bad Block Locator Utility or EVRAC& Re: Bad Block Locator Utility or EVRAC& Re: Bad Block Locator Utility or EVRAC& Re: Bad Block Locator Utility or EVRAC+ DEC VR299 19" Color Monitor  --  ** FREE **  file sharing w. NT Re: file sharing w. NT Re: file sharing w. NT Re: file sharing w. NT RE: file sharing w. NT6 Re: Galaxy Network Interfaces/cards and TCP/IP Sockets6 Re: Galaxy Network Interfaces/cards and TCP/IP Sockets- Re: KZCCA Ultrawide SCSI Adapters for the VAX - Re: KZCCA Ultrawide SCSI Adapters for the VAX - Re: KZCCA Ultrawide SCSI Adapters for the VAX - Re: KZCCA Ultrawide SCSI Adapters for the VAX  Re: KZPCA-AA Works fine !  kzqsa aka m5976  Re: kzqsa aka m5976  Re: kzqsa aka m5976  Re: lgi_callouts and kerberos5  LP27/AlphaServer 2100 Connection$ Re: LP27/AlphaServer 2100 Connection$ Re: LP27/AlphaServer 2100 Connection Re: MacOS X and OpenVMS  Re: MOB Tax [VMS news stuff]. * Re: NOSLOT No PCB available (failed spawn) OpenVMS and Human Resources + OpenVMS article - please explain last line! / Re: OpenVMS article - please explain last line! / Re: OpenVMS article - please explain last line! / Re: OpenVMS article - please explain last line! / Re: OpenVMS article - please explain last line! - Re: Problems with CSWS and CGI-BIN on TCPWARE  Re: pthreads + printf 1 Re: REQ: Pager software for Vax running VMS 6 + ? 1 Re: REQ: Pager software for Vax running VMS 6 + ? , Re: Star Office progress report (on Solaris)7 Re: strange behaviour of DIRECTORY with long file names 7 Re: strange behaviour of DIRECTORY with long file names - Re: Talk to Rich Marcello - Austin Texas area - Re: Talk to Rich Marcello - Austin Texas area - Re: Talk to Rich Marcello - Austin Texas area C RE: Talk to Rich Marcello, but DOES HE LISTEN? - Austin Texas are	a C Re: Talk to Rich Marcello, but DOES HE LISTEN? - Austin Texas are	a C Re: Talk to Rich Marcello, but DOES HE LISTEN? - Austin Texas are	a C Re: Talk to Rich Marcello, but DOES HE LISTEN? - Austin Texas are a C Re: Talk to Rich Marcello, but DOES HE LISTEN? - Austin Texas are a  TSM ! Re: tuning VMS for Oracle7 Server ! Re: tuning VMS for Oracle7 Server ! RE: tuning VMS for Oracle7 Server ! Re: tuning VMS for Oracle7 Server ! Re: tuning VMS for Oracle7 Server ! Re: tuning VMS for Oracle7 Server , Re: Using 3-phase power converters on VAXen?, Re: Using 3-phase power converters on VAXen?, Re: Using 3-phase power converters on VAXen?" Web-based change VMS password tool& Re: Web-based change VMS password tool& Re: Web-based change VMS password tool5 Re: Why is this a Bad Thing? (was: Future Computing.) 5 Re: Why is this a Bad Thing? (was: Future Computing.)   F ----------------------------------------------------------------------  # Date: Mon, 16 Apr 2001 19:55:39 GMT  From: adroso@home.com (ADR) + Subject: Bad Block Locator Utility or EVRAC 9 Message-ID: <3adb4994.47540520@news.jamison1.pa.home.com>   A I have a VS4000-VLC running V5.5-2 and I think I have one or more B drives going bad.  In HELP INITIALIZE DEVICE /BADBLOCKS, there areC references to Bad Block Locator Utility and EVRAC to search for bad = blocks on a disk, but I can find no other references to these 7 utilities.  Does anyone have any experience with these?    ------------------------------  # Date: Mon, 16 Apr 2001 20:31:35 GMT 2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)/ Subject: Re: Bad Block Locator Utility or EVRAC 2 Message-ID: <HEIC6.925$fB6.22953@news.cpqcorp.net>  W In article <3adb4994.47540520@news.jamison1.pa.home.com>, adroso@home.com (ADR) writes: B :I have a VS4000-VLC running V5.5-2 and I think I have one or moreC :drives going bad.  In HELP INITIALIZE DEVICE /BADBLOCKS, there are D :references to Bad Block Locator Utility and EVRAC to search for bad> :blocks on a disk, but I can find no other references to these8 :utilities.  Does anyone have any experience with these?     Yep, and they're ancient.     A   The references date back to the days before automatic bad block    revectoring.  A   EVRAC is the old VAX Diagnostic Supervisor (DS) test, and DS is ;   not applicable on this nor on any other MicroVAX systems.   A   The BAD utility itself is not used on DSA (RA) and later disks.   J   If your disk is going bad, you can force a bad block revector operation H   by rewriting the failing disk sectors -- a BACKUP/IMAGE and restore isG   one of the most common.  This has the added advantage of leaving you  I   with a BACKUP when the disk itself fails and defragmenting your disk if    not.    G   That somebody notices that a disk is going bad -- the occasional bad  E   block is handled transparently by the revectoring support when the  @   failing sector is next written -- tends to be an ominous sign.  D   Also look for a bad SCSI cable, a SCSI termination problem, a SCSIF   bus that is just too long (eg: total length longer than zero meters 9   :-), too many SCSI enclosures or SCSI connections, etc.   N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  % Date: Mon, 16 Apr 2001 17:11:23 -0400 - From: "Richard D. Piccard" <piccard@ohio.edu> / Subject: Re: Bad Block Locator Utility or EVRAC ( Message-ID: <3ADB5FEA.F3379255@ohio.edu>   Hoff,   D With the retirement of the BAD utility, is there anything other than  !         $ BACKUP/PHYSICAL/COMPARE   Y in the way of a base operating system method (no layered products) for touching **EVERY** Y user-accessible block of a disk drive?  (And thereby forcing bad-block re-vectoring, even 0 of the blocks not currently allocated to files?)  #                                 RDP      Hoff Hoffman wrote:   Y > In article <3adb4994.47540520@news.jamison1.pa.home.com>, adroso@home.com (ADR) writes: D > :I have a VS4000-VLC running V5.5-2 and I think I have one or moreE > :drives going bad.  In HELP INITIALIZE DEVICE /BADBLOCKS, there are F > :references to Bad Block Locator Utility and EVRAC to search for bad@ > :blocks on a disk, but I can find no other references to these: > :utilities.  Does anyone have any experience with these? >  >   Yep, and they're ancient.  > C >   The references date back to the days before automatic bad block  >   revectoring. > C >   EVRAC is the old VAX Diagnostic Supervisor (DS) test, and DS is = >   not applicable on this nor on any other MicroVAX systems.  > C >   The BAD utility itself is not used on DSA (RA) and later disks.  > K >   If your disk is going bad, you can force a bad block revector operation J >   by rewriting the failing disk sectors -- a BACKUP/IMAGE and restore isH >   one of the most common.  This has the added advantage of leaving youK >   with a BACKUP when the disk itself fails and defragmenting your disk if  >   not. > H >   That somebody notices that a disk is going bad -- the occasional badF >   block is handled transparently by the revectoring support when theB >   failing sector is next written -- tends to be an ominous sign. > F >   Also look for a bad SCSI cable, a SCSI termination problem, a SCSIG >   bus that is just too long (eg: total length longer than zero meters ; >   :-), too many SCSI enclosures or SCSI connections, etc.  > P >  ---------------------------- #include <rtfaq.h> -----------------------------L >       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.comP >  --------------------------- pure personal opinion ---------------------------N >    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   --B ==================================================================B Dick Piccard                           Academic Technology ManagerB piccard@ohio.edu                                 Computer ServicesB http://oak.cats.ohiou.edu/~piccard/                Ohio University   ------------------------------  # Date: Mon, 16 Apr 2001 21:56:00 GMT 2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)/ Subject: Re: Bad Block Locator Utility or EVRAC 2 Message-ID: <QTJC6.929$fB6.23277@news.cpqcorp.net>  X In article <3ADB5FEA.F3379255@ohio.edu>, "Richard D. Piccard" <piccard@ohio.edu> writes:  E :With the retirement of the BAD utility, is there anything other than  : " :        $ BACKUP/PHYSICAL/COMPARE : H :in the way of a base operating system method (no layered products) for I :touching **EVERY** user-accessible block of a disk drive?  (And thereby  L :forcing bad-block re-vectoring, even of the blocks not currently allocated  :to files?)   C   The block revectoring occurs on a write operation, not on a read.   D   A (failed) read will trigger a FORCEDERROR error on encountering aD   bad block, and a best guess at the original contents of the block.  /   Options for writing to every block of a disk:   $     $ COPY NLA0: ddcu: /ALLOCATE=big  2     SYSGEN> CREATE/NOCONTIG/SIZE=big ddcu:[000000]       $ INIT ddcu: label/ERASE  H   The two former commands will simply fill the disk, the latter command    will clobber the entire disk.   G   I have a simple little program that uses LOG_IO to clobber the whole  H   disk (alternatively, MOUNT/FOREIGN, and start writing to the blocks), H   though INIT/ERASE works nicely.  If you want to read the entire disk, J   a variant of this approach would be trivial -- though it would probably ?   be easier to simply fill up the entire disk and then use the     ANALYZE/DISK/READ_CHECK tool.   N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  # Date: Tue, 17 Apr 2001 00:11:56 GMT 7 From: moroney@world.std.spaamtrap.com (Michael Moroney) / Subject: Re: Bad Block Locator Utility or EVRAC & Message-ID: <GBwtvw.IwK@world.std.com>  4 hoffman@xdelta.zko.dec.nospam (Hoff Hoffman) writes:  Y >In article <3ADB5FEA.F3379255@ohio.edu>, "Richard D. Piccard" <piccard@ohio.edu> writes:   F >:With the retirement of the BAD utility, is there anything other than >:# >:        $ BACKUP/PHYSICAL/COMPARE  >:I >:in the way of a base operating system method (no layered products) for  J >:touching **EVERY** user-accessible block of a disk drive?  (And thereby M >:forcing bad-block re-vectoring, even of the blocks not currently allocated   >:to files?)  D >  The block revectoring occurs on a write operation, not on a read.  E >  A (failed) read will trigger a FORCEDERROR error on encountering a E >  bad block, and a best guess at the original contents of the block.   0 >  Options for writing to every block of a disk:  F If you want to write to every block on a disk to force revectoring yet$ save its contents, one way is to do:> $ BACKUP /PHYSICAL disk1: disk2:[dir]file.ext/SAVE followed by2 $ BACKUP /PHYSICAL disk2:[dir]file.ext/SAVE disk1:  C You'll need enough free space on disk2 for the total space of disk1 I multiplied by whatever Backup's overhead is.  Of course you can also save  and restore to/from tape.   G You could also write a little program that reads and rewrites every LBN # on the disk with the same contents.    -Mike    ------------------------------  # Date: Mon, 16 Apr 2001 22:48:14 GMT , From: "Frank Sapienza" <sapienza@noesys.com>4 Subject: DEC VR299 19" Color Monitor  --  ** FREE **9 Message-ID: <OEKC6.80430$uk.8551354@news02.optonline.net>   K Here is your chance to own an original DEC VR299 monitor.  This beast has a B 19" (or maybe bigger?) display and uses three-wire RGB inputs withI sync-on-green.  First person to e-mail me gets to take it away -- totally + free, without charge, gratis, on-the-house.    What's the catch?     1)  It doesn't work quite right.  J After turning on the power you begin to see the display light up, and thenK there's a "CLICK!" and it goes blank.  I'd say something like a heat-sensor L or voltage-monitor is kicking in when it shouldn't.  I know that the drawingK circuits appear to work fine because you can see the images being displayed 2 for an instant just before it switches itself off.  < 2)  No shipping, no delivery.  You want it, come and get it.  J The monitor is located on beautiful, sunny Long Island, New York -- on theC North shore.  Without traffic, it'd be about a 50-minute drive from L Manhattan -- only 25 minutes from Republic airport for the pilots among you.% Easy directions.  You won't get lost.   I This is a limited, short-term offer, so act now!  If it's still here next  Monday it goes to the curb.     E-mail me at sapienza@noesys.com     Frank    ------------------------------  # Date: Mon, 16 Apr 2001 21:14:33 GMT * From: Robert Palasek <bobpalasek@home.com> Subject: file sharing w. NT ( Message-ID: <3ADB6111.662819F7@home.com>  :  I have an application where I need to set up file sharing& between a VMS system and an NT domain.  ,   One of the possibilities is an NFS system,'  1.A. Server is on NT, client is on vms  or(  1.B. Client is on NT, server is on vms.  7 Realizing that NFS has come from a UNIX world, does NFS 3 impose a burden that file names and directory paths , have a case sensitive spelling?  Or is there/ something neat in the implementation that says, 3 "oh, you are using NT and VMS, lets not worry about ! the capitalization in spellings"?   / Do some flavors of NFS work better than others?    Bob    ------------------------------  # Date: Mon, 16 Apr 2001 21:42:49 GMT 2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman) Subject: Re: file sharing w. NT 2 Message-ID: <tHJC6.928$fB6.23277@news.cpqcorp.net>  U In article <3ADB6111.662819F7@home.com>, Robert Palasek <bobpalasek@home.com> writes: : :I have an application where I need to set up file sharing' :between a VMS system and an NT domain.   @   ODBC servers are available for OpenVMS, for RMS files and for    various databases.  B   OpenVMS can also serve files via SMB, but (unless SAMBA has one).   I am not aware of an SMB client for OpenVMS.  - :  One of the possibilities is an NFS system,  ..8 :Realizing that NFS has come from a UNIX world, does NFS4 :impose a burden that file names and directory paths  :have a case sensitive spelling?  H   OpenVMS TCP/IP Services offers two approaches toward serving files viaJ   NFS, one provides case-sensitivity for NFS clients or applications that H   require it, and the other operates without case sensitivity.  In your J   application, you can obviously choose to use either -- I'd probably use H   the case insensitive version, as you can server and can share OpenVMS    directories directly.   0 :Do some flavors of NFS work better than others?  2   I'd use at least TCP/IP Services V5.0A with ECO.  N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  % Date: Mon, 16 Apr 2001 17:42:18 -0400 + From: John Eisenschmidt <jeisensc@aaas.org>  Subject: Re: file sharing w. NT # Message-ID: <sadb2f17.078@aaas.org>   G Not to be an open source whore, but I've had great luck with Samba on = K Unix, and I understand it works *swell* under VMS. You can even use it to = I make your VAX/Alpha a BDC in your NT Domain. As for the other items, at = L least under Unix, it treats Sym Links as Shortcuts and in case insensitive = on the NT side.   B >>> Robert Palasek <bobpalasek@home.com> 04/16/2001 5:14:33 PM >>>:  I have an application where I need to set up file sharing& between a VMS system and an NT domain.  ,   One of the possibilities is an NFS system,'  1.A. Server is on NT, client is on vms  or(  1.B. Client is on NT, server is on vms.  7 Realizing that NFS has come from a UNIX world, does NFS 3 impose a burden that file names and directory paths , have a case sensitive spelling?  Or is there/ something neat in the implementation that says, 3 "oh, you are using NT and VMS, lets not worry abouta! the capitalization in spellings"?e  / Do some flavors of NFS work better than others?    Bob    ------------------------------  % Date: Mon, 16 Apr 2001 18:20:33 -0400o" From: Dan Sugalski <dan@sidhe.org> Subject: Re: file sharing w. NTT: Message-ID: <5.0.2.1.0.20010416181932.01ec2638@24.8.96.48>  0 At 09:42 PM 4/16/2001 +0000, Hoff Hoffman wrote:9 >In article <3ADB6111.662819F7@home.com>, Robert Palasek C ><bobpalasek@home.com> writes:D >   OpenVMS can also serve files via SMB, but (unless SAMBA has one)0 >   I am not aware of an SMB client for OpenVMS.  L While I don't know of any way to mount an SMB-served disk on a VMS box, you L can fetch files via the SMB protocol from other machines with the SMBCLIENT 3 program that gets built as part of the Samba suite.e   					Dan  I --------------------------------------"it's like this"------------------- 2 Dan Sugalski                          even samurai? dan@sidhe.org                         have teddy bears and evenp;                                       teddy bears get drunkl   ------------------------------  % Date: Mon, 16 Apr 2001 16:50:03 -0600i- From: "Rowell, Bradley" <browell@state.mt.us>  Subject: RE: file sharing w. NTa@ Message-ID: <1245D1C0C039D411933708002BB29C644B298F@DOAISD02003>  G Vector Technologies has a product to connect from VMS to and SMB share.   < http://www.vector-networks.com/luvms/lanutil-for-openvms.htm   > -----Original Message-----+ > From: Dan Sugalski [mailto:dan@sidhe.org]t& > Sent: Monday, April 16, 2001 4:21 PM > To: Info-VAX@Mvb.Saic.Com ! > Subject: Re: file sharing w. NT  >  > 2 > At 09:42 PM 4/16/2001 +0000, Hoff Hoffman wrote:; > >In article <3ADB6111.662819F7@home.com>, Robert Palasek a  > ><bobpalasek@home.com> writes:F > >   OpenVMS can also serve files via SMB, but (unless SAMBA has one)2 > >   I am not aware of an SMB client for OpenVMS. > ? > While I don't know of any way to mount an SMB-served disk on l > a VMS box, you E@ > can fetch files via the SMB protocol from other machines with  > the SMBCLIENT 5 > program that gets built as part of the Samba suite.e > 
 > 					Dan > 3 > --------------------------------------"it's like   > this"-------------------4 > Dan Sugalski                          even samuraiA > dan@sidhe.org                         have teddy bears and even = >                                       teddy bears get drunk  >    ------------------------------  % Date: Mon, 16 Apr 2001 20:13:04 +0100h, From: "Richard Maher" <maher_rj@hotmail.c*m>? Subject: Re: Galaxy Network Interfaces/cards and TCP/IP Socketst3 Message-ID: <9bfgg4$ddv$1@neptunium.btinternet.com>p   Hi Matt,   Thanks for the reply.-  H >Your client-server application should not be dependent on IP addresses.I >Rather it should use names.  If it did, then you could rely on DNS, LoadoL >broker and Metric server, to detect the outage of particular interfaces andL >maintain high availability by using a DNS cluster alias (not to be confused >with an IP cluster alias)./  H What I'm dealing with is a tale of two pipes. The first, or application,E pipe will hit my cluster via a middleware product like MQSeries, BMQ, J TIER3 or just plain Sockets. I don't care if it uses DNS load balancing orH a compass. What I do have to ensure is that  the second, or transaction,5 pipe ends up on the same scsnode as the as the first.B  C This is a DECdtm restriction and one that I have to comply with forw8 atomicity reasons. So when the application code calls myC T3$GET_TM_URL  I was thinking that I'd have to return an IP addressCF ucx$inet_hostaddr of the SCSNODE that is hosting the application. This@ URL, once channelled back to W2K via the application pipe, wouldD guarantee that a subsequent ITipTransaction::Push would hit the same scsnode. So far so good.  B The problem arises when my node goes down after I've PREPAREd. RdbK will happily start up a recovery job on another node and so will I, but MTSaJ /DTC won't talk to it because all it knows about is the 1.2.3.4 address of the downed node.  I I'm a long way down the road to solving this problem (without sacrificingsI performance!) and include it here merely for background. But did you know I that on Himalaya a process on *any* node in a cluster can listen on *any*o IP address?i  J Apparently on Himalaya, If NODE1 is hosting a process that is listening atH 1.2.3.4 and it goes down it simply starts listening on NODE2 at the same address. Clever huh?  J >OK, if you don't like that method (it is my preference) then you could doG >some more coding.  Essentially, before the listener comes up on NODE2,aE >simply create aliases for those interfaces.  E.g. (assuming you havew  >executed TCPIP$DEFINE_COMMANDS) >.  >    ifconfig we0 alias 1.2.3.10  >    ifconfig we1 alias 1.2.3.20 >I$ >(To delete the alias use "-alias").  J I didn't know you could do that and will look into it. Thanks for the tip.  H Is it as scary as it looks? and what happens if original node comes back up in the mean time?   > I >No.  The network card is assigned to a particular instance.  There is no C > way of sharing it.  So you will need a network card per instance.- >-  & If Himalaya can do it, why not Galaxy?   Regards Richard Maher.   ------------------------------  % Date: Tue, 17 Apr 2001 10:23:18 +10001B From: "Matt Muggeridge" <Matt.Muggeridge@compaq.ssppaammffree.com>? Subject: Re: Galaxy Network Interfaces/cards and TCP/IP SocketsE2 Message-ID: <J3MC6.933$fB6.23393@news.cpqcorp.net>  L > Apparently on Himalaya, If NODE1 is hosting a process that is listening atJ > 1.2.3.4 and it goes down it simply starts listening on NODE2 at the same > address. Clever huh?  E It sure is.  Though, when you started this thread, you mentioned thatkI cluster alias would not work for you, and at your request, I pretended it H doesn't exist.  Yet, this is precisely the functionality that IP clusterH alias provides.  So now I'm confused.  Why not use the IP cluster alias?L When a node holding the alias adress goes down, another node assumes it's IP  address and operation continues.   >EL > >OK, if you don't like that method (it is my preference) then you could doI > >some more coding.  Essentially, before the listener comes up on NODE2,rG > >simply create aliases for those interfaces.  E.g. (assuming you have?" > >executed TCPIP$DEFINE_COMMANDS) > > " > >    ifconfig we0 alias 1.2.3.10" > >    ifconfig we1 alias 1.2.3.20 > >e& > >(To delete the alias use "-alias"). >nL > I didn't know you could do that and will look into it. Thanks for the tip. >eJ > Is it as scary as it looks? and what happens if original node comes back > up in the mean time?  D I think it's fairly straight-forward.  A simple command procedure orI C-application could do this for you.  With regards, what happens when therJ original node comes back up, you will need to detect this (suggest using aD cluster lock) and then deconfigure the IP aliases and then allow the* original node to assume its old addresses.   >e > >tK > >No.  The network card is assigned to a particular instance.  There is noyE > > way of sharing it.  So you will need a network card per instance.n > >e > ( > If Himalaya can do it, why not Galaxy? >   K Good question.  All I can say is that I'm aware of a good amount of sharinglL between the groups.  If it hasn't been seen in product requirements, then toF get the right level of visibility it would be best to submit this as a request.   Cheers,- Matt.-   ------------------------------    Date: 17 Apr 2001 02:22:41 +0800, From: Paul Repacholi <prep@prep.synonet.com>6 Subject: Re: KZCCA Ultrawide SCSI Adapters for the VAX- Message-ID: <87k84kvh5a.fsf@prep.synonet.com>e  - "Barry Treahy, Jr." <treahy@mmaz.com> writes:=  F > Even if I could spread some of the I/O out across the DSSI, I do not? > want to have to be required to use SW.  If I could use an HSDo? > external to SW, that might be a different story, but it is myk > understanding that I cannot.  I I have seen at least one 4000-100 with an HSD, and have been told someonepG had an HSD50 on a 4K DSSI. Or perhaps I'm misunderstanding what you are  trying to say?     -- c< Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda."@                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.   ------------------------------  % Date: Mon, 16 Apr 2001 15:13:15 -0400p- From: "Richard D. Piccard" <piccard@ohio.edu> 6 Subject: Re: KZCCA Ultrawide SCSI Adapters for the VAX( Message-ID: <3ADB443E.FFF3C0B3@ohio.edu>  H I thought he meant that he doesn't want to use any StorageWorks product.  +                                         RDPo     Paul Repacholi wrote:f  / > "Barry Treahy, Jr." <treahy@mmaz.com> writes:o >eH > > Even if I could spread some of the I/O out across the DSSI, I do notA > > want to have to be required to use SW.  If I could use an HSDYA > > external to SW, that might be a different story, but it is myd  > > understanding that I cannot. >lK > I have seen at least one 4000-100 with an HSD, and have been told someonemI > had an HSD50 on a 4K DSSI. Or perhaps I'm misunderstanding what you aret > trying to say? >o > --> > Paul Repacholi                               1 Crescent Rd.,9 > +61 (08) 9257-1001                           Kalamunda. B >                                              West Australia 60760 > Raw, Cooked or Well-done, it's all half baked.   --B ==================================================================B Dick Piccard                           Academic Technology ManagerB piccard@ohio.edu                                 Computer ServicesB http://oak.cats.ohiou.edu/~piccard/                Ohio University   ------------------------------  % Date: Mon, 16 Apr 2001 16:23:14 -0700 + From: "Barry Treahy, Jr." <treahy@mmaz.com>a6 Subject: Re: KZCCA Ultrawide SCSI Adapters for the VAX( Message-ID: <3ADB7EE2.72BAD4DE@mmaz.com>  	 Hi Brian..  S For reasons that are pretty boaring and covered in prior rants-and-raves, I haven'tsR the need or desire to use Storageworks any longer.  In fact, last fall we replacedS everything with a Infrotrends based disk array which screams and puts Compaq/DEC tooK shame both in price and performance.  The problem is that the 8-bit SE SCSI P interface that is standard on a VAX 4000/100 is terrible.  The long story short,R the disk array works great (though I need to bandwidth), but my tape subsystems doR not and all of these devices are ultrawide which, as I mentioned, the standard VAX adapter is only SE narrow.  S My A to Z issue is that I need external UW SCSI support.  How I get there is rathernM unimportant provided that it doesn't break the bank and that it is not a step P backwards in performance.  I know of the DSSI buses, and that they are slow, butS I'm only getting 3MB asynchronous SCSI from the existing VAX adapter, so to get twonL more channels at that speed may not be a bad deal provided that I can get itM converted and successfully support external UW SCSI without SW baggage.  WhatsM DEC/Compaq never sold, to my knowledge, was a vanilla DSSI to SCSI adapter ort) convertor that operated standalone to SW.r  ! Did this make it as clear as mud?    Regards,   Barrye   B Fairweather wrote:   > Hello Barry, >aL >     The HSD30 or HSD50 is hardware RAID.  They support Mirror and RAID 3-5Q > The HSD50 allows you to partition a RAID set 1-4 devices served to VMS.  You do:> > not need any of theVolshad licenses all hardware controlled. >DQ >     Does that help or ?  If you can spread your application/data out, I've beens > prettyP > happy with the HSD##.  I have 1 VMScluster I run with 4 HSD50s (1 per I/O bus)5 > with 3 VAX 4705A and an Alpha 4100 as Server nodes.i >R
 >     Fyi, >     Brianr   --  ? Barry Treahy, Jr  *  Midwest Microwave  *  Vice President & CIOi  A E-mail: Treahy@mmaz.com * Phone: 480/314-1320 * FAX: 480/661-7028d   ------------------------------  % Date: Mon, 16 Apr 2001 22:27:14 -0400 2 From: rdeininger@mindspring.com (Robert Deininger)6 Subject: Re: KZCCA Ultrawide SCSI Adapters for the VAXL Message-ID: <rdeininger-1604012227150001@user-2ive7m4.dialup.mindspring.com>  < In article <3ADB7EE2.72BAD4DE@mmaz.com>, "Barry Treahy, Jr." <treahy@mmaz.com> wrote:    E > My A to Z issue is that I need external UW SCSI support.  How I getw there is rathergO > unimportant provided that it doesn't break the bank and that it is not a steptH > backwards in performance.  I know of the DSSI buses, and that they are	 slow, but-J > I'm only getting 3MB asynchronous SCSI from the existing VAX adapter, so
 to get twoN > more channels at that speed may not be a bad deal provided that I can get itO > converted and successfully support external UW SCSI without SW baggage.  What O > DEC/Compaq never sold, to my knowledge, was a vanilla DSSI to SCSI adapter orM+ > convertor that operated standalone to SW.> > # > Did this make it as clear as mud?D  H I don't have any personal experience, but I think that some of the laterB VAX 4000 systems came with "embedded HSD10".  (Or maybe HSD05.)  IJ understood this to mean that a SCSI cable comes out of the system, and VMSG accesses the SCSI bus via DSSI.  Perhaps the 4000-x00A CPU boards coulds9 take a daughtercard version of the HSD?  Just guessing...l   --   Robert Deininger rdeininger@mindspring.comp   ------------------------------  # Date: Mon, 16 Apr 2001 22:48:03 GMTo2 From: "Zane H. Healy" <healyzh@shell1.aracnet.com>" Subject: Re: KZPCA-AA Works fine !5 Message-ID: <DEKC6.1333$Hp.49101@typhoon.aracnet.com>o  & Alan Greig <a.greig@virgin.net> wrote:9 > For hobby use you can always go for the VMS 7.3 EFT2 CDo  H Not really, the reason I want to move to V7.2-1 is so that I'm running aL version that there are patches for problems.  V7.3EFT2 is pretty much a deadH end.  Though I'm very tempted to get it for the TCPIP V5.1 beta software that's included.   			Zanea   ------------------------------  % Date: Mon, 16 Apr 2001 13:29:36 -0600b( From: emanuel stiebler <emu@ecubics.com> Subject: kzqsa aka m5976+ Message-ID: <3ADB4820.8632A7E5@ecubics.com>s   Hi all,   G anybody here has a description of the jumpers on this qbus-scsi board ?b   cheers & thanks, emanuel.   ------------------------------  # Date: Mon, 16 Apr 2001 20:38:29 GMTd2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman) Subject: Re: kzqsa aka m59762 Message-ID: <9LIC6.926$fB6.22953@news.cpqcorp.net>  V In article <3ADB4820.8632A7E5@ecubics.com>, emanuel stiebler <emu@ecubics.com> writes:  H :anybody here has a description of the jumpers on this qbus-scsi board ?  ?   KZQSA is for specific tapes and specific CD-ROM devices ONLY.l?   Please do not try to connect magnetic disks to it, it is not     reliable.   E   What are you looking for?  CSR address?  Q-bus memory base address hA   settings?  Both?  Ask The Wizard topic 2545 would be a start...-    N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  % Date: Mon, 16 Apr 2001 22:49:51 -0400n2 From: rdeininger@mindspring.com (Robert Deininger) Subject: Re: kzqsa aka m5976L Message-ID: <rdeininger-1604012249520001@user-2ive7m4.dialup.mindspring.com>  < In article <3ADB4820.8632A7E5@ecubics.com>, emanuel stiebler <emu@ecubics.com> wrote:  	 > Hi all,a > I > anybody here has a description of the jumpers on this qbus-scsi board ?    Sure, I can describe them...        There's a bunch of them.i        They're not labeled.   ,      I don't have any info on their meaning.   ;-)    --   Robert Deininger rdeininger@mindspring.com    ------------------------------  % Date: Mon, 16 Apr 2001 20:14:18 +0100i, From: "Richard Maher" <maher_rj@hotmail.c*m>' Subject: Re: lgi_callouts and kerberos5 3 Message-ID: <9bfgg5$ddv$2@neptunium.btinternet.com>m   Hi,a  E About a year ago I got a phone call asking me what I thought ExternalhL Authentication and sys$acme were going to do for me. I said that I wanted toL pass it a Username, a TCP/IP address (and whatever else but definitely *not*L a password) and I wanted sys$acme (via something like Microsoft LAN Manager)B to tell me if access to the network and my VMS box was authorized.E (Obviously the UAF record for the username in question would have the:< appropriate flag set and the logical names would be defined)  I In other words, like many C/S applications, I did not want to keep askingrH the user to logon with passwords for every connection/application, and IG thought External Authentication would be the solution. The answer I wasa given was a simple - No!   Has this changed?    Regards Richard Maher.  J PS. How could this simple piece of technology take VMS engineering so longH to develop? Have all of the Compaq VMS middleware products and utilitiesJ had enough of a jump on the competition to allow sys$acm to be released to the public?-  4 Adam Maulis <maulis@ludens.elte.hu> wrote in message news:3ej5qlv9Pn5W@ludens...n5 > In article <6OO$0$kaL4bw@eisner.encompasserve.org>,d; Kilgallen@eisner.decus.org.nospam (Larry Kilgallen) writes:dI > > In article <vqvll4sWShKQ@ludens>, maulis@ludens.elte.hu (Adam Maulis)0 writes:0 > >0 > [...]  >w >,I > > The VMS V7.3 SDKs came with a Kerberos implementation. Can you figure- out1 > > how they did it ?2 >9 >4 > Hi,O >D > my work based on it. >eG > This Kerberos port does not come with integrated login. It comes withoF > kinit.exe and kdestroy.exe (base kerteros utilities) and krb$rtl.exe' > (kerberos run-time shareable library)d >o >t
 > Regards,
 > Adam Maulis    ------------------------------  # Date: Mon, 16 Apr 2001 18:13:12 GMT.) From: "J.G. Peters" <jpeters@sctcorp.com>,) Subject: LP27/AlphaServer 2100 Connection10 Message-ID: <01c0c6a0$db2cd7b0$32341895@jpeters>  1 Thanks to all who responded to the original post,f1 "Cluster Breakup - Vax/Alpha Hardware Questions."m3 We still need help on this LP27 issue. At least one . reponse suggested that the LP27 connected to a2 Decserver 250 and served via LAT to the Alpha 21002 would be the way to go. Of course, we don't have a4 Decserver 250. We *do* have an 8 port Decserver 700.6 Is the Decserver 700 capable of supporting the LP27 or1 are we going to have to purchase a Decserver 250?-7 The current cable on the printer (connected directly toi4 and LPV-11 on a VAX 4300) won't fit the ports on the5 Decserver 700. Will this cable fit a Decserver 250 or 9 are we talking a different cable and/or adapter in either-4 case? We've *got* to get this printer working on the3 Alpha 2100. Any/all help is greatly appreciated. Ohx  yes, we're running OpenVMS 7.2.   8 Also, I am not seeing all posts on our newsgroup server.6 I saw my original post and nothing else. I *thought* I< wasn't getting any follow-ups at all. It wasn't until I went9 to groups.google that I found the complete thread. (I was 4 beginning to think I was doing something wrong.) Any) ideas? I'm using the news reader on IE 5.    Thanks,i Joea   ------------------------------  % Date: Mon, 16 Apr 2001 14:34:59 -0400s0 From: paul.r.anderson@compaq.com (Paul Anderson)- Subject: Re: LP27/AlphaServer 2100 ConnectionlP Message-ID: <paul.r.anderson-1604011434590001@dhcp-16-21-34-107.eng.lkg.dec.com>  > In article <01c0c6a0$db2cd7b0$32341895@jpeters>, "J.G. Peters" <jpeters@sctcorp.com> wrote:  8 > Is the Decserver 700 capable of supporting the LP27 or3 > are we going to have to purchase a Decserver 250?s  @ If not a DECserver 250, you'll need a terminal server that has aI Dataproducts parallel port, or something else like the old RapidPrint 200I7 print server that has one parallel and one serial port.e   Paul   -- s
 Paul Andersonp  OpenVMS Engineering  Compaq Computer Corporation   ------------------------------  # Date: Tue, 17 Apr 2001 04:57:48 GMT:( From: Terry Kennedy <terry@gate.tmk.com>- Subject: Re: LP27/AlphaServer 2100 Connections' Message-ID: <GBx74C.JDC@spcuna.spc.edu>i  ) J.G. Peters <jpeters@sctcorp.com> writes:d3 > Thanks to all who responded to the original post,l3 > "Cluster Breakup - Vax/Alpha Hardware Questions."d5 > We still need help on this LP27 issue. At least onep0 > reponse suggested that the LP27 connected to a4 > Decserver 250 and served via LAT to the Alpha 21004 > would be the way to go. Of course, we don't have a6 > Decserver 250. We *do* have an 8 port Decserver 700.8 > Is the Decserver 700 capable of supporting the LP27 or3 > are we going to have to purchase a Decserver 250? 9 > The current cable on the printer (connected directly too6 > and LPV-11 on a VAX 4300) won't fit the ports on the7 > Decserver 700. Will this cable fit a Decserver 250 orr; > are we talking a different cable and/or adapter in eithera6 > case? We've *got* to get this printer working on the5 > Alpha 2100. Any/all help is greatly appreciated. Ohc" > yes, we're running OpenVMS 7.2.   D   The cheapest thing to do would probably be to convert your LP27 toB Centronics parallel (it currently speaks Dataproducts). The least-? expensive place to buy those parts would be from Consultex (see?B http://www.consultex-inc.com). Then all you need is a vanilla Cen-C tronics terminal server from Lantronix or similar, or possibly even . hooking the printer up to the system directly.  E   You could go the DS250 route, but it's a rather twitchy combination E - for example, if a LP25/26 is offline when the DS250 boots, the boot.D will hang in the self-test until the printer is manually put online.. I don't know if the LP27 has the same problem.  4         Terry Kennedy             http://www.tmk.com5         terry@tmk.com             Jersey City, NJ USAn   ------------------------------  # Date: Mon, 16 Apr 2001 22:44:22 GMTh2 From: "Zane H. Healy" <healyzh@shell1.aracnet.com>  Subject: Re: MacOS X and OpenVMS5 Message-ID: <aBKC6.1332$Hp.49101@typhoon.aracnet.com>u  & Bob Harris <harris@zk3.dec.com> wrote: > In articleG > <paul.r.anderson-1404011518210001@dhcp-16-21-34-107.eng.lkg.dec.com>, 3 > Paul Anderson <paul.r.anderson@compaq.com> wrote:a  J > X  In article <5u0C6.1233$Hp.46353@typhoon.aracnet.com>, "Zane H. Healy"Q > X  > I figured I'd let people know since I know there are quite a few Mac userstO > X  > here that after installing the Mac OS X V10.0.1 update that was releasedhH > X  > yesterday I'm now able to mount NFS volumes on my OpenVMS server. > X  .P > X  What Mac NFS client are you using?  Or is there native support in Mac OS X?  I > NFS support is built in.  After reading Zane's post, I tried it using ahI > Tru64 UNIX NFS exported file system, and it worked like a champ.  I did   I Yes, native support.  Prior to the 10.0.1 update it was working fine witheH Solaris 2.6, but I couldn't get it to mount OpenVMS filesystems.  Once I+ applied the update it mounted it first try.d  E > There is also a GUI interface called NFSmanager that is supposed towH > handle this without learning UNIX commands.  I was just experimenting,= > so I didn't bother with NFSmanager.  Your mileage may vary.n  K If you're used to using UNIX commands you might find NFSmanager to be a bit.L confusing.  I plan on giving it another try though the next time I'm playing
 with MacOS X..  L > X  That would be nice.  Is it Apple's position that they are only going to > X  supportQ > X  AppleTalk over IP for file sharing?  I can connect to printers from OS X viah% > X  AppleTalk, just not file shares.'  J Their possition is that they do not support classic Appletalk for anythingE other than printing.  They want you to use Appletalk over IP for file  sharing.  D I think there is a definite market for a 3rd party app here, IF it's reasonably priced.  A > higher layer AppleTalk protocols over IP.  For example, you caneF > currently do AppleTalk Filing Protocol (AFP) (aka AppleShare TCP/IP)H > only using a TCP/IP connection to another Mac or netatalk system which! > support AppleShare over TCP/IP.a  J If I'm reading things correctly this support is also in the latest version$ of CAP (Columbia Appletalk Package).  H > I would like to have EtherTalk back so that I can talk to my AppleTalkF > enabled printer, but if that doesn't happen, then I hope the printer  J Why can't you?  It saw my old Asante Ethertalk-to-Localtalk converter, andH can apparently talk to the printer (at least it knows it's an HP5MP).  I% still need to do a test print though.   A > manufactures start building in ethernet ports on their low costeI > printers because the wireless computing world is still going to want toaG > print from their portable devices without necessarly having to have a@E > server computer (for example using a wireless laptop at home as thes > primary computer).  L Hmmmm, interesting point....  I've been meaning to set my wifes laptop up soL it will print via Samba to the HP5MP.  However, I've seen printing via SambaJ actually crash the OpenVMS server!  I've got a TCP/IP print sever box thatJ I've never gotten around to trying to hook up the printer.  Maybe I should' (now to find a parallel printer cable).d  Q > X  > So, while I still don't think I'm ready to use Mac OS X full time, it just  > X  > might be improving. > X   O > X  I've switched to OS X at home and at work, and there's no looking back!  I 	 > X  like P > X  the look and feel of OS X and I'm busy replacing Classic applications with 
 > X  ones.  G > I'm also 95% switched over to Mac OS X.  I only switch back for a few1 > tasks.  L I think I'm now to where I could for most tasks.  I've not checked to see ifL NiftyTelnet will work under the classic layer.  What is a good Telnet clinetH for Mac OS X that supports the keypad when you're logged into an OpenVMSJ box?  The terminal.app claims to be able to support strict VT100 emulationJ and the keypad, but as far as I can tell it doesn't (though I've not tried under the 10.0.1 update).i  J I am concerned about the Adobe apps that I don't use that often.  I reallyJ am not wanting to have to spend the money to upgrade them considering how : little I use them, worse, a couple have been discontinued.   			Zanea   ------------------------------  % Date: Mon, 16 Apr 2001 20:15:28 +0100p, From: "Richard Maher" <maher_rj@hotmail.c*m>& Subject: Re: MOB Tax [VMS news stuff].3 Message-ID: <9bfgg7$ddv$3@neptunium.btinternet.com>r   Hi,    > G > What marketing department? Charlie Matco informs me that marketing is ? > taking a hideous hit in the current round of belt-tightening.h >i  J Problem is, "the pin-heads running most large companies still believe thatG you can just hire replacements right out of school to pick up where theoH others left off". Take the OpenVMS Times for instance. Do you think thatI just *anyone* could regurgitate the same RTR press release every quarter?a  > No! Ya just can't buy talent like that! And at *Open*VMS we'reH fortunate enough to have a deep and rich vein of such talent running notC just through marketing but right up through engineering and on into  management.i  / How do you think Digital got where it is today?t   Regards Richard Maherc  A Did you know? That when you pay good money for a VMS license thatt? you're really propping up the RTR junta. (And now you want freebA educational licenses so that they can push that stuff onto kids?)l   OpenVMS TechTips  B I suggest using RTR as your middleware backbone. To pay good moneyA for MQSeries or BMQ when RTR is free would only encourage IBM and C BEA to continuing supporting VMS. (Someone appears to have snatchedsI Tuxedo from the jaws of victory but rest assured such temporary set-backs1C will not divert VMS engineering from its path of self-destruction.)n   ------------------------------  % Date: Mon, 16 Apr 2001 23:27:54 -0400e- From: JF Mezei <jfmezei.spamnot@videotron.ca>t3 Subject: Re: NOSLOT No PCB available (failed spawn)n, Message-ID: <3ADBB827.6EBA5C44@videotron.ca>   "Barry Treahy, Jr." wrote:R > Hans, it may seem adequate, but 28% of my memory was about 70000 SPTE's which isR > not comparable to your 16MB but your system resource requirements might be.  YouO > have nothing to loose to attempt  increasing the SPTE's to an extremely large7P > number...  As I said, I had the exact same problem and SPTREQ was the cause of > my NOSLOT problems...f  4 How does one calculate the correct value of SPTREQ ?  K On my system, AUTOGEN would calculate it at about 4900, and it seems that IaG needed much more. (thanks to your suggestion I boosted it to 28% and no  problems since ).S  H Seems that Autogen doesn't look at all the right information in order toN calculate that value. Is it possible that it is because that parameter doesn'tL exist on Alpha and the engineers have forgotten to update the VAX autogen to better calculate it ?    ------------------------------  % Date: Mon, 16 Apr 2001 16:44:50 -0300 ) From: fabio_compaq@ep-bc.petrobras.com.brt$ Subject: OpenVMS and Human ResourcesL Message-ID: <OFF7106E7C.C167CD6E-ON03256A30.006BE94E@ep-bc.petrobras.com.br>  J Today I spent half-day in a job interview  (far 250 km from my actual job) and the HR woman asked me ...s  , - Where / what  are you working (with) now ?  0 a) It was written in my resume, in her hands....    -  What is OpenVMS ?    a) She wanted a MVS guy ....  ? - Oh you cannot work with Unix so....  you need to recycle ....f  J a) What is the magic in recycle in Unix ?  Last time I worked with Solaris was one year ago.a   - Bye, bye the door is opened !D   a) I HATE HR women ....      Regardsi   FC   ------------------------------  % Date: Mon, 16 Apr 2001 19:57:41 -05001% From: Keith Brown <kbrown780@isd.net>.4 Subject: OpenVMS article - please explain last line!' Message-ID: <3ADB9505.97074B9D@isd.net>3  ' http://www.theinquirer.org/14040105.htme     Carry on Open VMS    Its time is not Khyber Passed   & By Mike Magee, 14 April 2001 17.20 BST  D RUMOURS of the death of Compaq's OpenVMS operating system are wildlyF exaggerated. In an effort to underscore the longevity of the OS and to? bring enterprise customers in Blighty up to speed on the latest F developments on the OpenVMS front, Compaq is hosting a two-day OpenVMSB Technical Update event on May 15 and 16 at the Jurys Great Russell Street Hotel in London W1.  F Speakers include OpenVMS Ambassadress Sue Skonetski (who will open theG event with an OpenVMS Strategy Update) as well as technical gurus NigelbF Barker and Hoff Hoffman. This is an opportunity to hear the definitiveG story on key areas of OpenVMS and to get the facts on major elements off= the recent V7.3 release, including the life-extending DII COE6 initiative.   F Further rumour has it that the event will be closed by another speakerC from across the pond. Registration is free, but the seats are goingt@ fast. To learn more about the event and to reserve a seat in the audience, visit this place.  H * The Inq will attend cos we want to find out how well Open VMS will run on AMD platforms...t         -- i Keith Browna kbrown780@isd.netr   ------------------------------  % Date: Mon, 16 Apr 2001 20:26:52 -0500s1 From: "David J. Dachtera" <djesys.nospam@fsi.net>e8 Subject: Re: OpenVMS article - please explain last line!' Message-ID: <3ADB9BDC.A0F65BD5@fsi.net>p   Keith Brown wrote: > [snip]J > * The Inq will attend cos we want to find out how well Open VMS will run > on AMD platforms...p  @ Yes - I read the article, and it ACTUALLY says that (about AMD).  H I'm sure we can expect a scathing review once they discover it's not for Intel or similar CPUs.  $ Too bad "it can't be done", huh? ;-)   -- D David J. Dachtera1 dba DJE Systemsi http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/   F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.t   ------------------------------  % Date: Mon, 16 Apr 2001 20:17:11 -0700  From: <tsmurphy@addr.com>48 Subject: Re: OpenVMS article - please explain last line!6 Message-ID: <mBOC6.1693$3n4.1298291@nntp2.onemain.com>  0 Keith Brown <kbrown780@isd.net> wrote in message! news:3ADB9505.97074B9D@isd.net...s  J > * The Inq will attend cos we want to find out how well Open VMS will run > on AMD platforms...w  J My interpretation that Magee is desperately trying to lift his sagging AMDH shares to something within spitting distance the price he paid for them.   ------------------------------  # Date: Tue, 17 Apr 2001 04:19:54 GMTo+ From: rjordan@mars.mcs.net (Richard Jordan)r8 Subject: Re: OpenVMS article - please explain last line!4 Message-ID: <KvPC6.8744$JN5.204672@news.goodnet.com>  L > > * The Inq will attend cos we want to find out how well Open VMS will run >> on AMD platforms...  K >My interpretation that Magee is desperately trying to lift his sagging AMDlI >shares to something within spitting distance the price he paid for them.   E The only way to do that is remove the stock price damping factor froma AMD; code name GQ Palmer...o   ------------------------------  % Date: Tue, 17 Apr 2001 00:32:22 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca>s8 Subject: Re: OpenVMS article - please explain last line!, Message-ID: <3ADBC73D.CA87FAF9@videotron.ca>   Richard Jordan wrote:aG > The only way to do that is remove the stock price damping factor fromo > AMD; code name GQ Palmer...t   NO NO NO ! ! ! ! ! ! ! ! W  L To minimise the damage he does to thr world, Palmer should be forced to stayJ the longest possible at any company he works for. The fewer job changes he& has, the fewer companies he will ruin.  7 At least Microsoft was smart enough not to hire palmer.    ------------------------------  % Date: Mon, 16 Apr 2001 16:53:00 -0700d0 From: Mark Berryman <Mark.Berryman@Mvb.Saic.Com>6 Subject: Re: Problems with CSWS and CGI-BIN on TCPWARE, Message-ID: <3ADB236C.2CFD7DA4@Mvb.Saic.Com>   Steve Freegard wrote:, > 	 > Hi all,: > H > I've just installed CSWS 1.0-1 onto our cluster (running VMS 7.2-1 and > TCPWARE 5.4-3).1 > C > Static pages work fine but trying to run any CGI results in a 500e' > Internal Server error on the browser.B > H > I've done a DEFINE/SYSTEM/EXE TCPIP$SERVICE UCX$SERVICE as recommended  8 That should be tcpip$DEVICE and ucx$DEVICE, not service.  
 Mark Berryman5   ------------------------------  # Date: Tue, 17 Apr 2001 00:49:41 GMT,3 From: "MrSignor" <MrSignor@nospam_bellatlantic.net>r Subject: Re: pthreads + printf6 Message-ID: <FqMC6.303$Q_6.304232@typhoon2.ba-dsg.net>   fyi - ComPaq's responsev  I The problem is with mixing sleep + usleep, and ONLY occurs on a multi-cpu1; machine, and ONLY when MULTIPLE_KERNEL_THREADS are enabled.l  L As far as a fix goes, Compaq are still discussing who/what is to blame, withB the contenders being DecThreads/DEC-C and/or the documentation. An$ immediate fix is thus sometime away.     -Frede5 =====================================================n> "MrSignor" <MrSignor@nospam_bellatlantic.net> wrote in message2 news:s4Oy6.7813$jL4.1802293@typhoon1.ba-dsg.net..., > thank you all that posted info/suggestions >oH > today, I narrowed it down to sleep/usleep, by removing the printf fromF > the main. It would seem that if you have called sleep, usleep is NOT& > going to return until sleep returns. >-C > I logged a call with Compaq today at work - where we are NOT uptol9 > date on our Pthread patches - just as I am not at home.l >SH > The engineer could see NOT any info for my particular problem, and hasA > accepted by example program - i'll post back when i get a reply4 >5G > In the meantime, i will install patches here and at work and see whatv
 > happens. >o > thanks again >2 > -Fredr. > ============================================0 > <Shane.F.Smith@Healthnet.com> wrote in message@ > news:OF35407E2A.0DE37875-ON88256A24.0001C2FE@foundation.com... > >tJ > > Caution, sneaky undocumented (afaik) trick approaching. Probably won't > help7 > > with this specific problem, but it's worth knowing.a > >pL > > Leafing through the pthreads header file for C one day, I found a couple > ofE > > interesting calls mentioned that weren't in the pthreads manuals:hL > > pthread_mutex_setname_np and pthread_cond_setname_np. These allow you toH > > apply names to the mutexes and condition variables you create, which > appearK > > in the SDA when you're examining the process. This makes debugging themeL > > /way/ easier, since you know which resources are deadlocked. Have a look > inI > > "pthreads.h", the arguments ar address of cond/mutex, name to assign,y ando > acH > > void pointer I've not found a use for. Passing the address of a null$ > > pointer has been working for me. > >pG > > Note, though, the "_np" suffix. This indicates the routines are not 
 > > portable.a > >n	 > > Shaner > >n > >  > >t > >o > >nI > > MrSignor <MrSignor@nospam_bellatlantic.net> on 04/03/2001 04:31:21 PMh > >bA > > Please respond to MrSignor <MrSignor@nospam_bellatlantic.net>o > >p > > To:   Info-VAX@Mvb.Saic.Comt > > cc:t > >r > > Subject:  pthreads + printf0 > >/ > >e > > Alpha VMS 7.2-1a > >OA > > I have been writing some threaded code and seem to have hit ao! > > wall, or at least a deadlock.L > >oI > > Basically, I am compiling with /reentrancy=multi, although the onlinesC > > help says this is applicable to the MAIN only, and I am linkingw# > > with threads_enabled + upcalls.d > > J > > One thread enters an infinite while loop, it calls usleep, to wait forD > > 100 milli-secs (adjusted for usleep), and then when it awakes it > > prints the current time. > >eK > > The other thread (main really) loops on sleep (rather than usleep), for  30L > > seconds, and it uses printf to display how many seconds remain, if it is > > awoken prematurely.  > >hG > > Problem, once it begins, a few outputs occur - sometimes, sometimesc> > > only 1. At this point I am assuming a problem with Printf. > >-C > > The above explanation is over simplified, but tomorrow, I am to * > > do exactly what I have explained here. > >j: > > Anyone have any suggestions for further investigation. > >p > > TIAn > >m	 > > -Freda > >< > >s > >d > >m > >t > >  > >t > >  > >u >. >e   ------------------------------  # Date: Tue, 17 Apr 2001 00:37:15 GMTo, From: Gary Petersen <garyp-spam@hacklab.net>: Subject: Re: REQ: Pager software for Vax running VMS 6 + ?+ Message-ID: <3ADB1F92.3CCD735A@hacklab.net>n   Jason wrote:  I > Can't do it over the net for a few reasons.  Needs to be done through aS( > physical modem attached to the system. >yK > Any pointers on something that will call and page a simple numeric pager?   " http://www.disworks.com/alpha.html  D This product pages from an Alpha over a modem.  Not sure if its what$ you want but it sounds like a match.   Gary   ------------------------------  % Date: Mon, 16 Apr 2001 21:49:54 -0400 . From: Michael Austin <maustin@nc.prestige.net>: Subject: Re: REQ: Pager software for Vax running VMS 6 + ?/ Message-ID: <3ADBA142.92AB2750@nc.prestige.net>e   Gary Petersen wrote:   > Jason wrote: >hK > > Can't do it over the net for a few reasons.  Needs to be done through at* > > physical modem attached to the system. > >xM > > Any pointers on something that will call and page a simple numeric pager?  >y$ > http://www.disworks.com/alpha.html >hF > This product pages from an Alpha over a modem.  Not sure if its what& > you want but it sounds like a match. >M > Gary  H Or,  most pager companies have internet email addresses for all of their pagers. @ Example:  mail to SMTP%"<phone number>.<pin number>@pagenet.com"  L I have been doing this for several years (when I do production support whereF the company supplies me with a pager.)  Currently I have a nation-wideO alpha-numeric pager where I get paged several times a day from a job monitoringh# database locking conditions in Rdb.3   Michael Austin DBA Consultant   ------------------------------  % Date: Mon, 16 Apr 2001 19:52:20 -0500o% From: Keith Brown <kbrown780@isd.net> 5 Subject: Re: Star Office progress report (on Solaris)T' Message-ID: <3ADB93C4.2B6D23C8@isd.net>l   Larry Kilgallen wrote: > < > > X-NEWS: eisner.encompasserve.org comp.sys.sun.apps: 2842 > > Path: iad-read.news.verio.net!dfw-artgen.news.verio.net!dfw-peer.news.verio.net!news.verio.net!newsfeed.mathworks.com!sunqbc.risq.qc.ca!news.maxwell.syr.edu!news.cis.ohio-state.edu!malgudi.oar.net!news.cas.org!not-for-mail > > From: lvirden@cas.org ! > > Newsgroups: comp.sys.sun.appsr- > > Subject: Re: StarOffice dumps core grrrrrt! > > Date: 7 Apr 2001 11:21:19 GMT : > > Organization: Nedriv Software and Shoe Shiners, Uninc.
 > > Lines: 13e, > > Message-ID: <9amt7f$iea$1@srv38.cas.org>0 > > References: <986591324.599709@dns2.serv.net> > > Reply-To: lvirden@cas.orgr' > > NNTP-Posting-Host: lwv26awu.cas.orgn > > Mime-Version: 1.0c. > > Content-Type: text/plain; charset=US-ASCII# > > Content-Transfer-Encoding: 7BIT / > > X-Newsreader: trn 4.0-test75 (Feb 13, 2001)o/ > > Originator: lwv26@cas.org (Larry W. Virden)b: > > Xref: dfw-artgen.news.verio.net comp.sys.sun.apps:2842 > >c > >t > >eI > > sounds similar to my experiences.  Given that the whole reason for meuM > > attempting the install was to prove to my management that StarOffice, nowrK > > owned by Sun, was a good choice for investment, and that I've tried forSJ > > 9 months now and still haven't gotten it to work (despite a variety ofK > > suggestions from the great people in this newsgroup), I can see now why < > > convincing a company to move to StarOffice is so hard... > > -- > > --> > > "See, he's not just anyone ... he's my son."  Mark SchultzI > > <URL: mailto:lvirden@cas.org> <URL: http://www.purl.org/NET/lvirden/>rF > > Even if explicitly stated to the contrary, nothing in this posting  H I presume from your post that you are trying to run SO on Solaris. WhileD I have no experience with Solaris I can say that I have been runningF StarOffice on Linux for about a year and I am very pleased with it.  IG just wish someone would be kind enough to port it to OpenVMS for me :-)u   -- i Keith Browne kbrown780@isd.nete   ------------------------------   Date: 16 Apr 2001 14:12:26 CDT= From: wayne@tachysoft.xxx.065234.killspam.015d (Wayne Sewell)u@ Subject: Re: strange behaviour of DIRECTORY with long file names. Message-ID: <0315euQwjtQT@tachxxsoftxxconsult>  ` In article <7raldtk5q3ni2dhe0565n685178rdq1bl4@4ax.com>, Alan Greig <a.greig@virgin.net> writes:G > On 13 Apr 2001 08:03:56 CDT, wayne@tachysoft.xxx.065234.killspam.015d  > (Wayne Sewell) wrote:r >  > O >>Sorry, forgot about that part.  My point was that you have to do something to - >>the disk in order to use the ods-5 stuff.  t >>N >>Me, I don't really use ods-5 on any production disk.  My one level 5 disk is= >>solely for testing of the new ods-5 support in TAPESYS 6.0.  > G > We back up some ODS-5 volumes with Tapesys 5.2.24 currently. Should Io2 > worry about not having the "new ods-5" support?     N There is absolutely no problems with the backups themselves.  The problem withN ODS-5 is the history files.  In the 5.2.x "files" file (.4r4), the entire fileO name, minus the top level directory but including all other directories and thetL device name, is stored in the record.  The top level directory is implied by the name of the .4r4 file.   Thus V  : my_disk:[top_directory.other_dir.third_dir]the_file.type;1< my_disk:[top_directory.other_dir.third_dir]other_file.type;1= my_disk:[top_directory.other_dir.third_dir]third_file.type;5 n= my_disk:[top_directory.other_dir.third_dir]third_file.type;4 i= my_disk:[top_directory.other_dir.third_dir]third_file.type;3 o= my_disk:[top_directory.other_dir.third_dir]third_file.type;2     are stored in    _top_directory.4r4 ass    1 other_dir.third_dir]the_file.type;1 node my_disk i2 other_dir.third_dir]other_file.type;1 node my_disk2 other_dir.third_dir]third_file.type;5 node my_disk2 other_dir.third_dir]third_file.type;4 node my_disk2 other_dir.third_dir]third_file.type;3 node my_disk2 other_dir.third_dir]third_file.type;2 node my_disk    N Note that much of the information is replicated over and over again.  When youH combine longer file names with longer directory names and more directoryO levels, you can easily exceed the size of the record for ods-5.  Of course, themN filename_size parameter is adjustable in the sbattr.com, but then you start toG waste a lot of space when you make it a lot larger, because the records.N containing small filenames are mostly empty.  Also, the filename is used as an/ RMS key and can therefore not exceed 255 bytes.s        N The history system has been completely redesigned to support ods-5.  Well, theI sets files (.5r0) are the same, but the .4r4 files have been restructurediI completely.  They are now hierarchical, like the directories themselves. tO Rather than replicating the entire directory path in every record, an entry now M consists of a single element name, such as a file or directory segment, and aeM pointer to the entity that contains it.  The name part of the record now only N has to be large enough for just that element.  The versions of a file (and theN sets pointers) are now stored in a separate file that has no name part at all.O The version entries point back to their containing filenames in the other file.t       _top_directory.6r0   node ), my_disk <pointer to containing entity (node)5 top_directory <pointer to containing entity (my_disk).7 other_dir <pointer to containing entity (top_directory)r3 third_dir <pointer to containing entity (other_dir)i7 the_file.type <pointer to containing entity (third_dir) 9 other_file.type <pointer to containing entity (third_dir).9 third_file.type <pointer to containing entity (third_dir)    _top_directory.6r0_ptr  C ;1  <pointer to containing entity (the_file.type in the other file)a3 ;1  <pointer to containing entity (other_file.type)93 ;5  <pointer to containing entity (third_file.type)13 ;4  <pointer to containing entity (third_file.type)o3 ;3  <pointer to containing entity (third_file.type)03 ;2  <pointer to containing entity (third_file.type)a    M You can see that this format is much more compact, especially when all of theeO pointers are only longword indexes to the containing entity record.  The formateN is especially compact if you have lots of versions of the same file.  The fileO name only appears once in the .6r0 file!  Of course, scanning the new format is.N much more complicated, but I think I have headed off the potential overhead byN doing a multi-level scan, with a separate RAB stream for each directory level.    N One other ramification of the ods-5 long filenames is that they can exceed theH size of the sysbak mailbox.  The listing from the backup is written to aL mailbox owned by sysbak.  sysbak reads this mailbox and scans the listing toF generate the history update files.  I have seen cases where long ods-5L filenames were too big for the mailbox.  Not only would there be an error inM the log, but that particular filename would be dropped from the history.  The . size of the mailbox has been increased in 6.0.    O I am surprised you have not had any problems at all.  Do you have histories for M your ODS-5 backups?  Are you using long filenames, or just weird characters? mH Like I say, without histories, i.e. just the backup itself, there are no	 problems.   J Also, are you using the /convert option?  This essentially makes the ods-5H backup no different from an ods-2 backup as far as tapesys is concerned.    O With 6.0, we are recommending that people stop using /convert, since it will no M longer be necessary.  Also, due to the radically different characteristics ofsN ods-2 and ods-5, we are recommending that customers do not mix them in historyE sets, i.e. the ods-5 disks should have their own history set(s).  The/N restructuring of the database means that the filename_size for sets containingL only ods-2 files can be drastically reduced (39 vs. the current 124), with aM significant reduction of disk space.  However, you obviously can't do this ifpI ods-5 disks are contained in the set.  In fact, you will probably have touO increase filename_size, resulting in a lot of wasted space for the ods-2 files.a  N If you think you will be following this recommendation, it might be worthwhileL to start splitting the history sets now.  In other words, create new historyO set(s) for your ods-5 disks and start using them in the backups of those disks.9J Possibly by the time 6.0 comes out, depending on your retention dates, the2 ods-5 entries may have purged out of the old sets.  L ODS-5 is only one of the changes for 6.0.  Other major stuff includes tcp/ipI and icc support, a restructured main database, reworked primary/secondary-M protocol, reworked startup, a vmsinstal kit, integration with JB, support forrE parallel backups.  The latter is a mechanism by which a sysbak can besO dynamically split between multiple tape drives, assigning backups on the fly to A independently keep each of the tape drives running at full speed.l  K If you have a test machine and would like to field test 6.0, I am sure theyyL would be happy for you to do so.  It should be available for field test in a
 few weeks.       >Have also carried out> > some limited testing of Tapesys 5.2.24 with VMS 7.3 EFT2 andB > everything seems fine. Any known problems I should look out for.  M Sounds like you got farther than I did with 7.3.  Every time I tried to run a I backup on alpha, I got a crash caused by the new extended qio processor. CO Nothing to do with tapesys, though, because straight backup did the same thing.s  H I never got a patch for the problem.  The workaround was to turn off theL extended qio processor, but I never got around to doing it.  I have been too busy with tapesys 6.0.   Wayne  -- oO ===============================================================================wM Wayne Sewell, Tachyon Software Consulting  (281)812-0738  wayne@tachysoft.xxxa: http://www.tachysoft.xxx/www/tachyon.html and wayne.html  K change .xxx to .com in addresses above, assuming you are not a spambot  :-)jO ===============================================================================4B Jed Clampett, checking into hotel: "This place got a cement pond?"+ Ellie May: "And do yuh let critters in it?"r   ------------------------------  % Date: Mon, 16 Apr 2001 22:44:14 -0400t2 From: rdeininger@mindspring.com (Robert Deininger)@ Subject: Re: strange behaviour of DIRECTORY with long file namesL Message-ID: <rdeininger-1604012244150001@user-2ive7m4.dialup.mindspring.com>  . In article <0315euQwjtQT@tachxxsoftxxconsult>,> wayne@tachysoft.xxx.065234.killspam.015d (Wayne Sewell) wrote:    N > ODS-5 is only one of the changes for 6.0.  Other major stuff includes tcp/ipK > and icc support, a restructured main database, reworked primary/secondaryrD > protocol, reworked startup, a vmsinstal kit, integration with JB,   
 > support for G > parallel backups.  The latter is a mechanism by which a sysbak can be J > dynamically split between multiple tape drives, assigning backups on the fly toC > independently keep each of the tape drives running at full speed.t  J This sounds awfully close to RAIT or tape striping or whatever you like toI call it.  I've been wondering when someone would bring this functionalityeB to VMS, since tape drive bandwidth is falling so far behind disks.  F After I read the Tapesys literature, I suspected the folks at SoftwareJ Partners might already have something like RAIT.  A goodly fraction of the< pieces seem to  lurking in Tapesys and your other utilities.  H What's the likelyhood of a standalone product that can split (and merge)> tape I/O without recoding applications?  So if I write to some= pseudo-device, I'm really writing to an array of tape drives?b   -- d Robert Deininger rdeininger@mindspring.coms   ------------------------------  % Date: Mon, 16 Apr 2001 14:48:36 -0400e- From: "Richard D. Piccard" <piccard@ohio.edu> 6 Subject: Re: Talk to Rich Marcello - Austin Texas area( Message-ID: <3ADB3E77.21AFD414@ohio.edu>   "David J. Dachtera" wrote:   > "Terry C. Shannon" wrote:6 > >0B > > "Malcolm Dunnett" <nothome@spammers.are.scum> wrote in message* > > news:ze6J8mXVHkex@malvm1.mala.bc.ca... > >  > > > >i1 > > > >   o  Inappropriate Educational Programme.    > [snip]   >gK > > Over 40 universities are now using the free edu license program, so thet' > > opposition isn't exactly universal.t >a   [snip]  P Let us be VERY careful about wording here.  Is that "using" or "registered for"?Q I, for example, asked our site licensed software specialist to do the "electronic Q paperwork", to have it in place for any Ohio University person who wanted to takeyP advantage of it.  Given the restrictions, I very much doubt that any are in fact actually using it.  #                                 RDP    --B ==================================================================B Dick Piccard                           Academic Technology ManagerB piccard@ohio.edu                                 Computer ServicesB http://oak.cats.ohiou.edu/~piccard/                Ohio University   ------------------------------    Date: 16 Apr 2001 17:38:38 -07001 From: nothome@spammers.are.scum (Malcolm Dunnett)r6 Subject: Re: Talk to Rich Marcello - Austin Texas area, Message-ID: <3Z7kQLHBknJK@malvm1.mala.bc.ca>  ( In article <3ADB3E77.21AFD414@ohio.edu>,3     "Richard D. Piccard" <piccard@ohio.edu> writes:e >>L >> > Over 40 universities are now using the free edu license program, so the( >> > opposition isn't exactly universal. >> >  > [snip] > R > Let us be VERY careful about wording here.  Is that "using" or "registered for"?S > I, for example, asked our site licensed software specialist to do the "electroniclS > paperwork", to have it in place for any Ohio University person who wanted to taketR > advantage of it.  Given the restrictions, I very much doubt that any are in fact > actually using it. >   M     Yes, we're also one of the "over 40 universities" that has signed up, buti2 we're not yet using any licenses from the program.  M      I have gotten some answers since this thread started, some good news ando some not so good.   J      The "multiple students, multiple users" refers to a single student onM each computer - but one license can cover a lab full of such usage. I've beeneK told that under no circumstances are multiple interactive users on a singleF* system permitted under this program alone.  K       OTOH it is permissible to combine a base license from the edu program0J with the multi-user licenses under CSLG. Once this combination is done theJ resulting system may be used in any way that is appropriate for CSLG usageL ( eg no restrictions on admin usage ). This is rather inconvenient having toI sign up for 2 separate programs with different timelines for renewal, but # all-in-all it's still a good thing.w  I      I've also been told that the "single user" restriction under the edutJ program is strictly interactive users. Systems licensed under this programK can be used as file/print servers, web servers, database servers, etc which-E may serve numerous clients simultaneously as long as no more than onei) person has an interactive session active.   I     Regarding the phrase "under your control", I've been informed that ifeJ the license holder controls who is allocated a username on the system thatG is adequate control. I still find this "personal responsibility" clausemF disturbing ( it should be the organizations responsibility ), but it's something we can live with.c  G    disclaimer: all of the above is my understanding of the rules, basedsH on some emails with the program administrator. I, of course, don't speakL for Compaq and if anyone has any concerns about their rights and obligations) under the program they should ask Compaq.l   ------------------------------   Date: 17 Apr 2001 04:17:57 GMT2 From: mathog@seqaxp.bio.caltech.edu (David Mathog)6 Subject: Re: Talk to Rich Marcello - Austin Texas area, Message-ID: <9bgg5l$h18@gap.cco.caltech.edu>  ` In article <3Z7kQLHBknJK@malvm1.mala.bc.ca>, nothome@spammers.are.scum (Malcolm Dunnett) writes: >I, of course, don't speakM >for Compaq and if anyone has any concerns about their rights and obligations2* >under the program they should ask Compaq.  I And the first thing you should ask them is why they can't write a license3I agreement that says what they mean.  Assuming that they mean what they'ver
 said. :-(.  F Maybe they should replace the current license agreement with a JPEG ofJ chicken entrails - we'd have about as much luck interpreting the latter as the former.    Regards,   David Mathog mathog@caltech.edu? Manager, sequence analysis facility, biology division, Caltech @J **************************************************************************J *                                RIP VMS                                 *J **************************************************************************   ------------------------------  % Date: Mon, 16 Apr 2001 14:23:42 -0500 / From: "Stuart, Ed" <Ed.Stuart@austinenergy.com>sL Subject: RE: Talk to Rich Marcello, but DOES HE LISTEN? - Austin Texas are	aR Message-ID: <92EFB80E551BD511B39500D0B7B0CDCCF3B5E6@ohms.electric.ci.austin.tx.us>  L We'd like an Alpha for the LUG also.  However, like Terry wrote in the otherK thread on this topic, the door prize is being paid for with LUG funds and asF donation from a local VAR and I don't think we have enough cash for an AlphaServer.  $ Ed Stuart                           ( Manager, Systems and Desktop Services	   Information Technology Services  City of Austin, Austin Energyt Ed.Stuart@austinenergy.com  + "Glittering prizes and endless compromises s. shatter the illusion of integrity" - Neil Pert  B *Please apply a generous amount of all the usual disclaimers here*   -----Original Message-----0 From: Barry Treahy, Jr. [mailto:treahy@mmaz.com]& Sent: Thursday, April 12, 2001 6:00 PM To: Info-VAX@Mvb.Saic.Com.F Subject: Re: Talk to Rich Marcello, but DOES HE LISTEN? - Austin Texas area     "Stuart, Ed" wrote:   L > Also, one of the lucky attendees will walk away with a Compaq IPAQ!!!!!!!!L > Even if you don't care about what Rich has to say...come for the chance to > win the IPAQ!!!a  J An Ipuck?  What do you do what that?  Play hockey?  How about a real piece ofK equipment?   Perhaps even a fully loaded Alpha at least as a hobbyist tool?-   Barry-   --  ? Barry Treahy, Jr  *  Midwest Microwave  *  Vice President & CIO4  A E-mail: Treahy@mmaz.com * Phone: 480/314-1320 * FAX: 480/661-7028E   ------------------------------  % Date: Mon, 16 Apr 2001 21:32:52 +0100 % From: Alan Greig <a.greig@virgin.net>7L Subject: Re: Talk to Rich Marcello, but DOES HE LISTEN? - Austin Texas are	a* Message-ID: <3ADB56F4.41380086@virgin.net>   "Stuart, Ed" wrote:w  N > We'd like an Alpha for the LUG also.  However, like Terry wrote in the otherM > thread on this topic, the door prize is being paid for with LUG funds and a H > donation from a local VAR and I don't think we have enough cash for an > AlphaServer. >g  # Hmm, maybe Island could donate one?0  , > "Glittering prizes and endless compromises0 > shatter the illusion of integrity" - Neil Pert   Err, that's Neil Peart.p --
 Alan Greig   ------------------------------  % Date: Mon, 16 Apr 2001 23:57:24 -0400 , From: "islandco.com" <dbturner@islandco.com>L Subject: Re: Talk to Rich Marcello, but DOES HE LISTEN? - Austin Texas are	a/ Message-ID: <tdnfkrm1rvbh78@news.supernews.com>,  K Well... I'll have to see, but we could LOAN a DS20 Dual CPU system or say ah1 AS800, Alpha PWS600 XP1000 or something like thath  I No charge for the loan but of course we would need the system back intacto   DT David Turner Island Computers US Corporationi 2700 Gregory Streete Savannah GA 31404d Tel: 912 447 6622t Fax:912 201 0096 sales@islandco.com2 "Alan Greig" <a.greig@virgin.net> wrote in message$ news:3ADB56F4.41380086@virgin.net... |d |r | "Stuart, Ed" wrote:  | J | > We'd like an Alpha for the LUG also.  However, like Terry wrote in the othermI | > thread on this topic, the door prize is being paid for with LUG fundsr and ahJ | > donation from a local VAR and I don't think we have enough cash for an | > AlphaServer. | >i |h% | Hmm, maybe Island could donate one?. |r. | > "Glittering prizes and endless compromises2 | > shatter the illusion of integrity" - Neil Pert |  | Err, that's Neil Peart.o | -- | Alan Greig |n |    ------------------------------  # Date: Mon, 16 Apr 2001 21:22:59 GMTe0 From: sander@vmsbiz.enet.dec.com (Warren Sander)L Subject: Re: Talk to Rich Marcello, but DOES HE LISTEN? - Austin Texas are a2 Message-ID: <ToJC6.927$fB6.23227@news.cpqcorp.net>    re: 'I want an Alpha for my lug'  < Do you want any alpha or a new shiney top of the line alpha?  G I think if the lug's put in requests through their 'compaq' counterparthH we could come up with some unused systems that can be loaned out through the counterpart to the lug's.f  C If you want brand new systems then bring it up with Rich and put a i formal proposal together.   G Either way.. Talk on the news group isn't going to make anything happen # It requires formal direct contact. e  D I'll tell you getting slightly older but still good systems could be/ fairly easy if you go through the counterparts.d   -- lB ------------------------------------------------------------------6 Warren Sander                        OpenVMS MarketingD Compaq Computer Corporation          Work:  warren.sander@compaq.comE 200 Forest Street MR01-3/J1          Personal: sander@ma.ultranet.comy3 Marlboro, MA 01752                   (508) 467-4875 6    My opinions are my own and I only speak for myself -          Read http://www.openvms.compaq.com/  B ------------------------------------------------------------------   ------------------------------  # Date: Mon, 16 Apr 2001 21:25:23 GMTt, From: peterw@u.genie.co.uk (Peter Watkinson)L Subject: Re: Talk to Rich Marcello, but DOES HE LISTEN? - Austin Texas are a= Message-ID: <3adb62f1.21955109@newshost.netscapeonline.co.uk>   0 On Mon, 16 Apr 2001 14:23:42 -0500, "Stuart, Ed"# <Ed.Stuart@austinenergy.com> wrote:o  M >We'd like an Alpha for the LUG also.  However, like Terry wrote in the other.L >thread on this topic, the door prize is being paid for with LUG funds and aG >donation from a local VAR and I don't think we have enough cash for an 
 >AlphaServer.c >e% >Ed Stuart                           e) >Manager, Systems and Desktop Services	  t  >Information Technology Services >City of Austin, Austin Energy >Ed.Stuart@austinenergy.com  > , >"Glittering prizes and endless compromises / >shatter the illusion of integrity" - Neil Pertr     Try ebay.com  C 3 DS10 466mhz went for $1000 just recently. There are plenty of EV5e& 21164 and older systems on there also.  	  regards,        Peter Watkinsona Email: peterw@u.genie.co.ukf( Internet: http://you.genie.co.uk/peterw/A Windsurf International.com http://www.windsurf-international.com/a* PW Navigate.com http://www.pwnavigate.com/   ------------------------------  % Date: Tue, 17 Apr 2001 10:55:30 +1200a& From: John Derrik <jbryson@itc.gov.fj> Subject: TSMM Message-ID: <084681714A1BD511970B0002A560015F4AA600@exchange01.govnet.gov.fj>c  G Our operators access TSM through Bossman which basically gives them the J necessary priviledges to do work in TSM such as adding servers. Now what IK would like to do is to delete the SPAWN command from with TSM. I've managed L to deleted the help on SPAWN. However I can't seem to remember how to delete9 the command from TSM. I would appreciate some help here.     Thanks once again...     John Derrikn ITC Services
 Suva, Fiji   ------------------------------    Date: 16 Apr 2001 11:14:33 -07001 From: nothome@spammers.are.scum (Malcolm Dunnett)p* Subject: Re: tuning VMS for Oracle7 Server, Message-ID: <lJ1pEld0Xu1e@malvm1.mala.bc.ca>  5 In article <_dCC6.1646$qr.4782@nntpserver.swip.net>, n1     "Fred A G" <nospam@allowed.localhost> writes:s > ? >>    My biggest problem with Oracle on VMS is that neither thee > MultithreadedeI >> server nor prespawned server processes work reliably.  The overhead tosF >> connect to the database can be several seconds using regular server
 > processes -d > > > What kind of issues? We're not using MTS (luckily, I guess). > D     Queries appear to run an order of magnitude slower under the MTSB than they do under a dedicated server. I suspect this is some sortB of overhead on the fetches as one would assume the actual databaseA access would be identical each way ( both measurements done on anw otherwise idle server ).  @    Prespawned servers mostly work, but periodically they fail to? connect. An immediate retry usually works so this is more of anu? annoyance than a serious problem, but still it's something thato@ shouldn't be happening. The same connect failures never occur if> I leave everything else alone but turn off prespawned servers.  =   Note both of these problems were seen on Oracle 8.1.6. I've D heard other versions have similar problems but can't vouch for that.   ------------------------------  % Date: Mon, 16 Apr 2001 20:11:38 -0000e- From: wspencer@ap.nospam.org (Warren Spencer)e* Subject: Re: tuning VMS for Oracle7 Server/ Message-ID: <tdmkfqaufj3v7f@news.supernews.com>   5 nothome@spammers.are.scum (Malcolm Dunnett) wrote in t! <lJ1pEld0Xu1e@malvm1.mala.bc.ca>:s  E >    Queries appear to run an order of magnitude slower under the MTSlC >than they do under a dedicated server. I suspect this is some sort C >of overhead on the fetches as one would assume the actual databaseuB >access would be identical each way ( both measurements done on an >otherwise idle server ).p  G I've seen this complaint voiced before in this ng - about 4 months ago nH (assuming it wasn't you, of course).  Do you have Oracle support?  This H feature seems rather important in large scale installations, so I would  hope they have a fix for it.  
 -- snip --   ws   -- n1 << Marriage is Grand.  Divorce is Fifty Grand. >>a   Warren Spencer Senior Software Engineer The Associated Press  ? ** My employer does not necessarily agree with my statements **n   ------------------------------  % Date: Mon, 16 Apr 2001 14:26:21 -0600n- From: "Rowell, Bradley" <browell@state.mt.us>a* Subject: RE: tuning VMS for Oracle7 Server@ Message-ID: <1245D1C0C039D411933708002BB29C644B298D@DOAISD02003>  I If you are using UCX 4.2 or older you will gain a performance increase bya upgrading toB UCX 5.0a, Oracle with 4.2 sometimes sends the data multiple times.     > -----Original Message-----> > From: wspencer@ap.nospam.org [mailto:wspencer@ap.nospam.org]& > Sent: Monday, April 16, 2001 2:12 PM > To: Info-VAX@Mvb.Saic.Comn, > Subject: Re: tuning VMS for Oracle7 Server >  > 7 > nothome@spammers.are.scum (Malcolm Dunnett) wrote in t# > <lJ1pEld0Xu1e@malvm1.mala.bc.ca>:. > G > >    Queries appear to run an order of magnitude slower under the MTSeE > >than they do under a dedicated server. I suspect this is some sort,E > >of overhead on the fetches as one would assume the actual database D > >access would be identical each way ( both measurements done on an > >otherwise idle server ).e > > > I've seen this complaint voiced before in this ng - about 4 
 > months ago p; > (assuming it wasn't you, of course).  Do you have Oracle n > support?  This e? > feature seems rather important in large scale installations,  
 > so I would   > hope they have a fix for it. >  > -- snip -- >  > ws >  > -- m3 > << Marriage is Grand.  Divorce is Fifty Grand. >>a >  > Warren Spencer > Senior Software Engineer > The Associated Press > A > ** My employer does not necessarily agree with my statements **a >    ------------------------------  % Date: Mon, 16 Apr 2001 16:34:05 -0400 + From: John Eisenschmidt <jeisensc@aaas.org>s* Subject: Re: tuning VMS for Oracle7 Server# Message-ID: <sadb1f12.043@aaas.org>o  J This also came up in comp.database.oracle several months ago. There is a =J known Oracle bug with VMS, MTS and SMP. As of my last reading Oracle had =K not fixed it. Basically, SMP was broken with MTS under one of the 7s (7.1 =8I I believe) and they haven't made fixing it a priority (I mean, it's not =l= like VMS wasn't the platform Oracle was born on or anything).1  J http://groups.google.com/groups?q=3Doracle+openvms+mts&hl=3Den&lr=3D&safe=' =3Doff&rnum=3D1&seld=3D972505639&ic=3D1a  K Excerpt: "We have the same problem with Oracle 7.3.4.4 on OpenVMS 7.1. We = E have now a mix of MTS and dedicated connections. Dedicated for long =sH running apps and MTS for short connections. Oracle did investigate the =J problem and told me it still exists in 8.1.6 for OpenVMS. Oracle 7.3.3.6 =B was the lastreal VMS version (probably made by VMS especialists)."  L There are some patches that help, not sure if they fix it. I will tell you =K that I'm replacing an AS800 running 7.3.3, and this is one of the reasons =nD we didn't get a multiprocessor system (it effectively doesn't work).  
 Good luck, John Eisenschmidt   E >>> Warren Spencer <wspencer@ap.nospam.org> 04/16/2001 4:11:38 PM >>>=7 nothome@spammers.are.scum (Malcolm Dunnett) wrote in=20=! <lJ1pEld0Xu1e@malvm1.mala.bc.ca>:   E >    Queries appear to run an order of magnitude slower under the MTS C >than they do under a dedicated server. I suspect this is some sort C >of overhead on the fetches as one would assume the actual database6B >access would be identical each way ( both measurements done on an >otherwise idle server ).   I I've seen this complaint voiced before in this ng - about 4 months ago=20cJ (assuming it wasn't you, of course).  Do you have Oracle support?  This=20J feature seems rather important in large scale installations, so I would=20 hope they have a fix for it.  
 -- snip --   ws   --=20c1 << Marriage is Grand.  Divorce is Fifty Grand. >>E   Warren Spencer Senior Software Engineer The Associated Press  ? ** My employer does not necessarily agree with my statements **o   ------------------------------  # Date: Mon, 16 Apr 2001 21:22:04 GMTe From: Dirk Munk <munk@home.nl>* Subject: Re: tuning VMS for Oracle7 Server' Message-ID: <3ADB6259.EE7F42F2@home.nl>    Fred A G wrote:.  < > "Warren Spencer" <wspencer@ap.nospam.org> wrote in message+ > news:tdec0gtje76j87@news.supernews.com...nD > > nospam@allowed.localhost (Fred A G) wrote in <WuEB6.1121$qr.2785 > > @nntpserver.swip.net>: > >wF > > I ain't no expert, but I did spend a few years on that platform... > >f > F > Then you are probably much more of an expert than the party that wasI > responsible for installing and setup. (Not us, but we have to deal withl > the mess in prod, so...) >FH > > What version of Oracle?  If it's 7.3.n.n.n, then check out "Reserved > MemoryG > > Registry" in Oracle and OpenVMS docs.  This feature lets you devotedG > > physical ram to Oracle's SGA, bypassing OpenVMS's paging mechanism,i > for a 7 > > significant performance improvement, in most cases.o > >i >sE > It is 7.3.3 (don't have the patch/fix version details with me righto > now).h >fG > Yes, I've already checked out the bit about devoting ram to SGA (from H > the Oracle7 Install Guide, I think). It seems really simple - reserve,B > autogen and reboot. Can I trust AUTOGEN here? Aren't there otherE > parameters (than GBL*) and procedures I need to care about (as mostr > often there are)?  >vE > Something told me to lower the "normal" rule-of-thumb that SGA-sizehJ > should be max. 30% of real mem. to about 20%. Does reserving for SGA and@ > enabling Fast IO require more real memory (SGA-size+additional > GBLPAGES+MAXBOBMEM?)?m  F Reserving memory for the SGA is using real memory. You are effectivelyI reducing the amount of memory that is available for VMS. Fast IO is stilliJ supported in 7.3, however it is not supported in 8.x versions for whateverG reason. Increasing maxbobmem may speed up things I guess, but I have nor experience.t     >I >fG > What do I need to adjust for Mailbox communication or Process quotas?iH > The Install guide says something about "might need to customize". (Not) > very easy when you don't know the app.)r  K A larger mailbox size may decrease the amount of buffered IO's, that may be:	 helpfull.o       >c >oG > > For small databases, you can end up cacheing the entire database ina	 > ram, ifkF > > you devote a large enough reserved memory area.  I've set up a fewF > > databases with a 48 MB SGA/Reserved Memory block, but you can make > themJ > > much bigger.  Depends on your available hardware and application load, > of > > course.s > >n >oH > This database is near 50GB, but I guess it should be possible to cacheF > "daily data" for the users. Though I'm not sure what tools to use toG > count or measure that data. No help from the developers there either.c > J > The whole matter is that, I'm just trying to cover most basic things andJ > not so much tuning the whole beast (since there are flaws already in the > design)... >o	 > Regardso > /Fad >@ > > ws > >* > > --5 > > << Marriage is Grand.  Divorce is Fifty Grand. >>  > >  > > Warren Spencer > > Senior Software Engineer > > The Associated Press > >*C > > ** My employer does not necessarily agree with my statements **-   ------------------------------    Date: 16 Apr 2001 17:41:31 -07001 From: nothome@spammers.are.scum (Malcolm Dunnett)u* Subject: Re: tuning VMS for Oracle7 Server, Message-ID: <CEHQucQVpeW+@malvm1.mala.bc.ca>  0 In article <tdmkfqaufj3v7f@news.supernews.com>, 2    wspencer@ap.nospam.org (Warren Spencer) writes:  F >>    Queries appear to run an order of magnitude slower under the MTSD >>than they do under a dedicated server. I suspect this is some sortD >>of overhead on the fetches as one would assume the actual databaseC >>access would be identical each way ( both measurements done on ani >>otherwise idle server ). > I > I've seen this complaint voiced before in this ng - about 4 months ago iJ > (assuming it wasn't you, of course).  Do you have Oracle support?  This J > feature seems rather important in large scale installations, so I would  > hope they have a fix for it. > B     It may have been me, but I've heard others say the same thing.  F     Oracle Metallink has a patch for 8.0.5 which seems to address thisH problem - but apparently not one for 8.1.6. I've asked Oracle about this but no answers so far.   ------------------------------  % Date: Mon, 16 Apr 2001 19:26:30 -0500o1 From: "David J. Dachtera" <djesys.nospam@fsi.net>t5 Subject: Re: Using 3-phase power converters on VAXen?y' Message-ID: <3ADB8DB6.39E67930@fsi.net>    Paul Repacholi wrote:n > [snip]H > BTW, hate to be on topic, but if some one does know the details of howD > to hack a 6K to run from a single 240VAC pahase, I'd love to know.  G Contact Pat Jankowiak of the DFW bunch. He's got some notes you'll find- useful.-   -- - David J. Dachterao dba DJE Systems: http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/T  F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.    ------------------------------  % Date: Mon, 16 Apr 2001 19:29:01 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net>r5 Subject: Re: Using 3-phase power converters on VAXen?i' Message-ID: <3ADB8E4D.7CA17FB1@fsi.net>'   Bob Koehler wrote: > J > In article <3AD7F696.EB67194E@home.nl>, Dirk Munk <munk@home.nl> writes: > > L > > For security reasons the switch boxes in the house are also equiped withP > > "earth leakage switches" (don't know what the English verb for such a switch > > is). > I > Sounds like a grounid fault interrupter (GFI).  In the US only required E > for locations that tend to have water (bathroom, outdoors, near the  > pool, and such).  D Well, not being an IBEW, can't speak about the NEC. From experience,G however, local codes here fairly uniformly require GFCIs for any outlet2B within six(6) feet of a water faucet, bath tub, kitchen sink, etc.   -- g David J. Dachteras dba DJE SystemsF http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/e  F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.    ------------------------------  # Date: Tue, 17 Apr 2001 05:41:17 GMT  From: Dirk Munk <munk@home.nl>5 Subject: Re: Using 3-phase power converters on VAXen?-' Message-ID: <3ADBD77B.2F6E3CE6@home.nl>,   "David J. Dachtera" wrote:   > Bob Koehler wrote: > >mL > > In article <3AD7F696.EB67194E@home.nl>, Dirk Munk <munk@home.nl> writes: > > >xN > > > For security reasons the switch boxes in the house are also equiped withR > > > "earth leakage switches" (don't know what the English verb for such a switch
 > > > is). > >gK > > Sounds like a grounid fault interrupter (GFI).  In the US only required G > > for locations that tend to have water (bathroom, outdoors, near thee > > pool, and such). >rF > Well, not being an IBEW, can't speak about the NEC. From experience,I > however, local codes here fairly uniformly require GFCIs for any outletiD > within six(6) feet of a water faucet, bath tub, kitchen sink, etc.  N In The Netherlands it is required for the whole house to be protected by theseR switches. In my own home for instance I have in my switchbox 7 groups protected byP 16A automatic switches. They will break the power by overload, or  when there is; more than 30mA difference between the hot wire and neutral.o       >e >R > -- > David J. Dachteras > dba DJE Systemso > http://www.djesys.com/ > < > Unofficial Affordable OpenVMS Home Page and Message Board:! > http://www.djesys.com/vms/soho/  >pH > This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings > is to be expected. > B > Feel free to exercise your rights of free speech and expression. > H > However, attacks against individual posters, or groups of posters, are > strongly discouraged.N   ------------------------------  % Date: Mon, 16 Apr 2001 18:01:52 -0400i. From: Michael Austin <maustin@nc.prestige.net>+ Subject: Web-based change VMS password toolw/ Message-ID: <3ADB6BD0.7DD1D7CF@nc.prestige.net>H  H Does anyone have a tool that would allow a user to change a VMS passwordF via a WEB interface that they would like to share?  Web server doesn'tD matter, I can modify it to use the webserver of choice for different clients.   Thanks,    Michael AustinF DBA Consultant (not a programmer, but I can cut/paste with the best of	 them....)s   <I am my daughters scholarship>i   ------------------------------  # Date: Mon, 16 Apr 2001 22:27:56 GMTt2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)/ Subject: Re: Web-based change VMS password tooly2 Message-ID: <MlKC6.930$fB6.23210@news.cpqcorp.net>  ` In article <3ADB6BD0.7DD1D7CF@nc.prestige.net>, Michael Austin <maustin@nc.prestige.net> writes:I :Does anyone have a tool that would allow a user to change a VMS passwordtG :via a WEB interface that they would like to share?  Web server doesn't-E :matter, I can modify it to use the webserver of choice for differento	 :clients.j  F   This could easily be a massive security hole, with a wide variety ofF   vulnerabilities imaginable.  No, I'm not going to provide details orK   suggestions on potential or actual vulnerabilities in this configuration.i  G   One interesting part of this problem is the process context that the  D   target (cgi) script was running in -- you need full privileges to B   change another user's password, or you have to be under your ownD   username.  Other than this "slight" problem, the CGI script could G   easily be a small DCL command procedure.  (But you need a large level-   of trust.)  I   If you were to use the full RFC1738 syntax (user:pass@node), you might uH   well be able to pass in the current password as the login and the new G   password as a parameter, and have the target CGI script operate under 2   the login username for the SET PASSWORD command.  I   The Freeware V5 kit has a package called RESET_PASSWD that is intended r   for use by help desk staff.d  H   Alternatively, you can have the PC users share their password with theG   OpenVMS system assuming OpenVMS V7.1 and later and Advanced Server...tH   This is the external authentication mechanism, and it avoids having to>   (potentially) expose your password security via a webserver.  N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  % Date: Mon, 16 Apr 2001 23:54:57 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca>i/ Subject: Re: Web-based change VMS password toole, Message-ID: <3ADBBE7B.7FDA9EE4@videotron.ca>   Hoff Hoffman wrote:bH >   This could easily be a massive security hole, with a wide variety ofH >   vulnerabilities imaginable.  No, I'm not going to provide details orM >   suggestions on potential or actual vulnerabilities in this configuration.o  K ALL-IN-1 (aka: office server) has a script function to change the password.tH While the OA$MAIN.EXE image is installed with privileges, it has its own1 internal security rules that follow those of VMS.>  I If the process that runs A1 is not privileged, it can only change its own C password, and must provide old password(s) as well as the new ones.   K If the process is privileged (SYSPRV), it needs to provide the username andoM new password but not the old password. (This is consistent with the principlerI that the system manager need not know the old password when forcing a newa password on a user).  K So, out of the blue, I would think that creating a process to run under theiI user's username and then changing its passowrd would be the safest since rG there are protections to prevent that process from changin other user's M passwords. Such protections would be harder to implement  if the process that I changes the password has sysprv and could then change any user's password-# without knowing their existing one.1   ------------------------------  % Date: Mon, 16 Apr 2001 20:37:06 +0100 , From: "Richard Maher" <maher_rj@hotmail.c*m>> Subject: Re: Why is this a Bad Thing? (was: Future Computing.)3 Message-ID: <9bfhkk$fg0$1@neptunium.btinternet.com>e   Hi,c   > >aK > > 80% of the profit! We always knew it was a healthy chunk, but I for onei > never.$ > > thought it was such a large one. >m  C Sorry to be lazy but 80% of what $5? $500million? Operating Profit?y   Regards Richard Maher.   ------------------------------  % Date: Mon, 16 Apr 2001 23:19:55 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca>n> Subject: Re: Why is this a Bad Thing? (was: Future Computing.), Message-ID: <3ADBB648.E08EDDCD@videotron.ca>   Paul Sture wrote:eO > 80% of the profit! We always knew it was a healthy chunk, but I for one neverv" > thought it was such a large one.  K Remember that this "80%" is really the industry standard servers (aka: 8086eN based enterprise servers) with VMS and Unix helping. It does not mean that 80%2 of Compaq's profits are generated by VMS and Unix.   ------------------------------   End of INFO-VAX 2001.214 ************************