1 INFO-VAX	Wed, 08 Oct 2003	Volume 2003 : Issue 557       Contents: Re: %MOUNT-W-INCONSIZE Re: affordable VMS Re: DCL improvements Re: DCL improvements* DCL improvements?, Process Permanent Files Re: DS15 systems have arrived  Re: DS15 systems have arrived 2 Re: EVA question: How many vdisks should I create?2 Re: EVA question: How many vdisks should I create?2 Re: EVA question: How many vdisks should I create?+ HGFTP - how to troubleshoot data connection / Re: HGFTP - how to troubleshoot data connection ! Re: MAIL: importing of text files  Re: MD5 source code ?  MSA1000 & OpenVMS  Re: MSA1000 & OpenVMS  Re: MSA1000 & OpenVMS  Re: NSlookup bug: flex scanner Re: NSlookup bug: flex scannerP OpenVMS Pearl - Beautiful Live Link to Customer OpenVMS System, using RMS and ApP Re: OpenVMS Pearl - Beautiful Live Link to Customer OpenVMS System, using RMS an$ OpenVMS Pearl Open Source on OpenVMS Re: Read this and weap, twats!& Re: Slow Shadow Copy with 36 GB drives Re: Sun takes a hit  Re: Sun takes a hit  Re: Sun takes a hit  Re: Sun takes a hit  swxcrmgr Re: swxcrmgr) Re: Vax / VMS V5.5-2 External Drive Issue ) Re: Vax / VMS V5.5-2 External Drive Issue ) Re: Vax / VMS V5.5-2 External Drive Issue ) Re: Vax / VMS V5.5-2 External Drive Issue ) Re: Vax / VMS V5.5-2 External Drive Issue 2 Re: VMS JVM implementation of sub-process deletion Re: VMS system on the web O Re: worries about files opened by audit server and security server on  shutdown  Re: [5] Counter Googling :-(  F ----------------------------------------------------------------------  * Date: Tue, 7 Oct 2003 18:42:24 +0000 (UTC)P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) Subject: Re: %MOUNT-W-INCONSIZE $ Message-ID: <blv1eg$93f$2@online.de>  < In article <liCgb.6512$px1.6029@news.cpqcorp.net>, Bob Blunt* <robert.blunt@hp.donotspamme.com> writes:   I > Phillip, to which connector on the "big block" is your BA350 attached?     To the DSSI connector.  D > If connected to the Centronics, I'm surprised that you're getting H > DSSI-type disks...  If connected to one of the other connectors, then G > all bets are off, IMHO.  The two really big connectors in the center  I > with the three rows of pins/sockets are for Q-Bus expansion (using and  4 > configuring the Q-Bus can be an "adventure," btw).  H Right.  As the old saying goes (allegedly advice of an English nobleman D to his son), try everything once---except incest and folk dancing.  / Perhaps QBUS should be added to that list!  :-)   G > IIRC, there were some VAX4000-10x systems that were either available  I > with an embedded HSD05/HSD10s or the HSDs were an option.  This option  K > would consume one of the internal drive slots on the "top shelf" and the  H > internal drives would probably then be SCSI instead of DSSI.  None of J > the "embedded" HSD controllers were intended to be extended to external J > devices.  I believe the model number of this HSD was HSD10-EN and these K > seem to have been more common on newer variants of the 4000-10x like the   > -106(A) and -108(A).  I That seems to be what I have.  There appears to be an internal DSSI bus,  I which works fine, and the external one via the big plug on the back, the  G one which also has SCSI and QBUS connectors.  I am hoping to use the 6  G disks in the BA350 via DSSI and a further set of disks on the SCSI bus.    ------------------------------  % Date: Wed, 08 Oct 2003 00:55:08 -0400 . From: Mike Bartman <omni@foolie.omniphile.com> Subject: Re: affordable VMS 8 Message-ID: <2m57ovom919tfbehmp8b7t13dk8aagq17s@4ax.com>  ( On Tue, 7 Oct 2003 17:24:59 +0000 (UTC),8 moroney@world.std.spaamtrap.com (Michael Moroney) wrote:  M >But the airport builder doesn't even know that Bidder 3 and their technology G >even exists.  They see heavy advertising by Bidder 1 and find out that I >over 95% of construction projects use Bidder 1 type techniques (counting J >every homeowner who purchases a sack of cement at Lowes as a constructionJ >project, of course).  Meanwhile, Bidder 3 loses money, and their providerI >of on-site concrete plants discontinues making them, forcing Bidder 3 to J >use Bidder 1 & 2 style technology in the future (assuming they're even in >the business).  > A >Meanwhile, because Bidder 1 leaves his supplies out in the open, D >they repeatedly get contaminated by the weather, producing flaws inJ >the runway.  Bidder 1 is more than willing to come up with a new concreteK >patch each time it happens to hide the flaws.  Every so often they come up N >with a new type of concrete which doesn't have the flaws - but has new flaws.K >Meanwhile, everyone knows the cement sacks need to be stored in waterproof H >enclosures to avoid contamination, but for some reason they never do...  = And for some reason, everyone just accepts this as normal and E inevitable, despite all the airports, currently being decommissioned, / which don't have these problems, and never did.   B Yes, it's sad.  The real problem is managers who aren't fit to runC their companies.  If they knew anything about what they were doing, = they'd know about Bidder 3, get curious about the low-problem F airports, or start asking questions of those who did when they see allA the problems they are having with theirs...and then LISTEN to the  answers.  D Instead, they read BOTASMs[1], and do what everyone else seems to beE doing (according to the BOTASMs), and make lots of "penny wise, pound D foolish" decisions, over the objections of the technical specialistsD who DO know better (or once did) and hire more managers who are justF as clueless as themselves to try to fix it all, while firing those whoB they wouldn't listen to when they can't make the stupid mess, that. they said not to buy in the first place, work.   -- Mike Bartman   ' [1] Back Of The Airplane Seat Magazines @ ----------------------------------------------------------------=   To reply via e-mail, remove the 'foolie.' from the address. %   I'm getting sick of all the SPAM... @ ----------------------------------------------------------------   ------------------------------  # Date: Tue, 07 Oct 2003 17:59:14 GMT " From:   VAXman-  @SendSpamHere.ORG Subject: Re: DCL improvements 0 Message-ID: <00A27063.AC28C826@SendSpamHere.ORG>  ] In article <blur10$b4u$1@newslocal.mitre.org>, lewis@PROBE.mitre.org (Keith A. Lewis) writes: ] > writes in article <00A27051.CF583CFD@SendSpamHere.ORG> dated Tue, 07 Oct 2003 15:51:19 GMT: L >>Being the "purist", where will this change be implemented?  In the $GETDVI& >>service code or just in the lexical? > I >Since the subject of the thread is "DCL improvements" I would assume the 	 >lexical.  > L >>If I have a procedure that checks for device AVaiLability, what shall I doK >>when I get back "FALSE" and the code tried to make said device AVaiLable?  > I >You'll probably get a %SYSTEM-W-NOSUCHDEV on whatever command you use to M >make it available, which will trigger the same condition handler as it would $ >have if it happened in the lexical. > , >--Keith Lewis              klewis$mitre.org? >The above may not (yet) represent the opinions of my employer.   M There are a lot if items to check to insure that a device is truly available; L not just EXISTS and AVL.  How about a device that was served (REMOTE_DEVICE)M by another cluster member and that cluser member is no longer a member (HOST_ L _AVAIL).  AVL is not foolproof and there are steps, which the present imple-G mentation requires/enforces, to properly checking above and beyond AVL.   M Besides, Guy has more important things to do like considering my requests! ;)  --  L VAXman- A Bored Certified VMS Kernel Mode Hacker    VAXman(at)TMESIS(dot)COM             5   "Well my son, life is like a beanstalk, isn't it?"     ------------------------------  % Date: Wed, 08 Oct 2003 01:10:19 +0200 * From: Paul Sture <nospam@sture.homeip.net> Subject: Re: DCL improvements 0 Message-ID: <3F8363FA.6ADD3D01@sture.homeip.net>   briggs@encompasserve.org wrote:  > a > In article <6OCT03.21142891@feda01.fed.ornl.gov>, Dave Greenwood <greenwoodde@ornl.gov> writes: > > > In a previous article, Guy Peleg <guy.peleg@hp.com> wrote:
 > >> Hi Hans,  > >>? > >> I'm currently working on WHILE - expected release is V8.2.  > >>S > >> As for F$GETDVI("AVL") I agree that FALSE should be returned. Could you please  > >> logR > >> a call with your local support center and have it escalated to engineering? I# > >> will provide you a fix to your > > >> current version and will make sure to include it in V8.2. > > G > > I'll have to agree with the previously noted concern about breaking H > > existing code with this change.  I'll also note that if you made theG > > change for AVL I'd expect you to make the same change for *all* the  > > getdvi item codes. > G > Which means determining a default value to return for all those codes G > that is reasonable for a non-existent device.  Not all _that_ hard to E > do, I suppose.  "False", 0 and "" should serve for most everything.  > ! > We've had this argument before.  > F > The purist point of view is that asking for any attribute other thanF > "EXISTS" of a non-existent device is a nonsensical thing to do.  TheG > lexical function should mirror the operation of the underlying system F > service.  And the underlying system service fails with an error code > of SS$_NOSUCHDEV.  > L > The non-purist point of view is "I want to know if the device is availableH > for use.  And if it doesn't exist then it's surely not available.  Why4 > make me waste two lexical calls when one would do? >   G A device can exist, but not be available, and you can make it so if you " you wish by SET DEVICE/NOAVAILABLE      --  
 Paul Sture   ------------------------------  # Date: Wed, 08 Oct 2003 00:43:56 GMT ! From: JBloggs@acme..spamless..com 3 Subject: DCL improvements?, Process Permanent Files 8 Message-ID: <2bm6ov40l5n0256t6hi19i8dsdv1mi5fqm@4ax.com>  5 As long as DCL improvements were bring talked about,    1 One interest I had for a while, was more control  " over Process Permanent Files.       ' f.ex, being able to force close/reopen  0 the current process's PPF to a new set of files 5 with different attributes (or even onto a new device)   ) and perhaps the same ability to do so to   another process's PPFs.   0 (or heck, a  /FDL login and.or submit qualifier)  / Pehaps limited to sys$output (or sys$error) if  0 DCL's sys$input was too much a special case ....  F If this was easy (!) to implement, this could be a nice thing to have.   ------------------------------   Date: 8 Oct 2003 01:48:25 GMT  From: healyzh@aracnet.com & Subject: Re: DS15 systems have arrived, Message-ID: <blvqd9020rn@enews2.newsguy.com>  * Rik Steenwinkel <rsteenw@xs4all.nl> wrote:E > Why would you want a PS2-to-USB converter when almost all Logitech  E > mice are already USB, and could plug into the ADB-to-USB converter    K The newer mice aren't as comfortable to my hand, they put more of a bend on K my wrist.  I like the classic Logitech 3-button mice that were used by DEC, I SGI, IBM, and others in the late 90's.  I've also got several of these as 
 spares :^)  B > (don't you mean USB-to-ADB?), by way of an USB hub if necessary?  H > I'd also try putting a USB interface card into the machine, of course 2 > when supported by the OS version you're running.  L No, I mean an ADB-to-USB converter.  They're actaully quite popular with theJ Mac crowd.  A lot of us are unwilling to give up our Apple Extended II ADBK keyboards.  Shoot, until a couple weeks ago, I was still using the teardrop G ADB mouse as well (upgraded to an HP USB mouse OEM'd from Logitech, but  can't stand the shape).   L I've been using the ADB-to-USB converter on the computer for four years now,K I got it exactly three days after I got the computer (it took that long for L the store to get it in).  The keyboard and mouse that my G4/450 shipped with are an insult!  I One of these days I'll get the system a PS2-to-USB converter so I can use  one of the older Logitech mice.    		Zane   ------------------------------   Date: 7 Oct 2003 22:43:58 -0500 - From: Kilgallen@SpamCop.net (Larry Kilgallen) & Subject: Re: DS15 systems have arrived3 Message-ID: <SsQe6AAeB6Uu@eisner.encompasserve.org>   H In article <blvqd9020rn@enews2.newsguy.com>, healyzh@aracnet.com writes:, > Rik Steenwinkel <rsteenw@xs4all.nl> wrote:  C >> (don't you mean USB-to-ADB?), by way of an USB hub if necessary?  > I >> I'd also try putting a USB interface card into the machine, of course  3 >> when supported by the OS version you're running.  > N > No, I mean an ADB-to-USB converter.  They're actaully quite popular with theL > Mac crowd.  A lot of us are unwilling to give up our Apple Extended II ADB > keyboards.  B Some of us are unwilling to give up our ADB Apple Keyboard (keypadA but no row of function keys; escape and control keys in the right  VT100 location).   ------------------------------  $ Date: Tue, 7 Oct 2003 18:50:32 -0400  From: John Santos <JOHN@egh.com>; Subject: Re: EVA question: How many vdisks should I create? 4 Message-ID: <1031007182037.403A-100000@Ives.egh.com>  ! On 7 Oct 2003, Scott Vieth wrote:   L > "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message news:<3F7A28= 8E.E7067C5C@fsi.net>... K > > I still haven't even seen an EVA (probably won't unless I can go to the J > > bootcamp in November), I've only read about them here in the group and > > on-line at hp's website. >=20G > David, please feel free to drive up to Milwaukee and see our shop.=20 H > The EVA is in production and I'm currently migrating our data from HSGE > to EVA.  Drop me a line via email when you are interested in seeing 
 > the EVA. >=20F > [That invite goes for anyone else in the Midwest as well.  Just dropB > me a line via email if you are interested in stopping-by to talk" > storage, SANs, VMS or whatever.] >=20J > > So far, it all seems to add a lot of questionably necessary complexityI > > for very little improvement. Seems to me one can do better laying out L > > RAIDsets on HSxes one's self and avoid some of the pitfalls mentioned b= y E > > yourself and other posters. Add to that the fact the HSxes can be G > > commanded in batch jobs via HSDSA or HSZPAD$SCSITERM and it seems a  > > negative gain. >=20D > There is a tool called SSSU which allows you to manipulate the EVAE > from VMS DCL procedures.  It's similar to the Command Scripter tool % > that is used with HSG-based arrays.  >=20F > As far as "laying things out by hand and doing a better job than theH > EVA", that's not true.  You can't beat the way the EVA handles things. >=20K > > Are there any advantages to EVA that can't be realized by judicious use " > > of the preceding technologies? >=20F > In addition to being super-high-performance (when compared to HSGs),
 > the EVA:- > 1) Can't be drowned in IOs by OpenVMS 7.3-1 $ > 2) Takes a lot less time to manage3 > 3) Is very easy to expand when you need more room ) > 4) Takes up less floor space (sq ft/GB)  > 5) Stays crispy in milk. :^)  ' Is this what Bob C means by "crispier"?  =20 F > The EVA *is* the next generation of storage.  It is  better than the# > HSG-based arrays in every aspect.  >=20G > The only exception is when you go through CLI-withdrawal when you are G > first setting the damn thing up.  I had a very strong urge to connect F > my serial console cable (all of my HSG-based arrays are managed withG > ConsoleWorks) to the EVA.  Nowhere to plug it (the console cable) in.  > :^)  >=20	 > Thanks,  >=20 > -Scott  A How much of this applies to MSA1000's?  I'm looking at a proposed A configuration for a customer that uses a couple of MSA1000's with A about 2TB total space.  Current similar configurations use ES40's ? (some AlphaServer 4100's) with HSZ80's.  New configuration with @ an ES45 and EVA was *much* more expensive, but the MSA1000-based config was priced very nicely.  B (We are looking at shadowing virtual units across MSA's for better> redundancy and for quiesce-the-app/dismount-shadow-set-member/B mount-member-read-only/resume-the-app/backup-the-copy/remount-the-E member-into-the-shadow-set to do backups with minimal (a few seconds) > downtime.  Our app has the capability built into it to disableA itself and check-point its transaction logs, so this scheme works 6 well.  That's the main reason for the dual MSA1000's.)  @ From=20the MSA1000 docs I could find, it sounds like it has most? of the same capabilities as the EVA's, but less raw performance  and less room for expansion.  @ Having never seen either an MSA1000 or an EVA in person, though, I have some questions.  ; For example, can you mix different sized disks?  I'm mostly < worried about the situation in two or three years, where the; customer decides they want to expand their database or keep ? more archival information online, and wants to add a TB or two. @ Do they have to buy more disks the same size as the current ones@ (72GB drives, I think), or will they be able to get whatever the? current best price/performance drives are and just add them in?   > It looks like with VMS V7.3-2, we'll be able to init the disksA with a maximum size larger than the current size and later expand D them.  Is the MSA1000 able to dynamically re-allocate its partitions= (or virtual units or whatever they call them), so we can take  advantage of this?   --=20  John Santos  Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------   Date: 7 Oct 2003 22:29:33 -0500 + From: young_r@encompasserve.org (Rob Young) ; Subject: Re: EVA question: How many vdisks should I create? 3 Message-ID: <omvJXDNb9TOb@eisner.encompasserve.org>   d In article <5a85bce2.0310070658.44f3d3bc@posting.google.com>, svieth@wi.rr.com (Scott Vieth) writes:f > "Bill Todd" <billtodd@metrocast.net> wrote in message news:<8eGdnYB3Xof8ceOiU-KYuA@metrocast.net>...N >> Too bad:  while with small groups the bandwidth and/or IOPS provided by notL >> having an idle spare can be significant, once you get over a dozen or twoI >> disks it's lost in the noise - while being able to use spares flexibly N >> across multiple groups remains useful.  I'd have expected EVA to offer both >> options.  >>  	 >> - bill  >  > Bill:  > A > The EVA doesn't have a provision for "hot spares".  Instead all H > spindles are in use all the time.  The "Disk Failure Protection" levelD > on a disk group leaves a certain amount of empty space in case you" > lose a disk from the disk group. > B > When a disk fails in a disk group, the data that was on there isF > magically re-created and distributed across the remaining members in8 > the disk group.  It's actually quite elegant to watch. >   C 	He understands all this.  You are missing his point of flexability ? 	as an option that makes sense.  i.e. with multiple disk groups  	a floating spare.   				Rob    ------------------------------   Date: 8 Oct 2003 00:16:46 -0500 + From: young_r@encompasserve.org (Rob Young) ; Subject: Re: EVA question: How many vdisks should I create? 3 Message-ID: <p91EPr7gxnKJ@eisner.encompasserve.org>   d In article <5a85bce2.0310070708.7a596ff2@posting.google.com>, svieth@wi.rr.com (Scott Vieth) writes:E >> This is all news to me.  Our normal config would be as few LUNs as M >> possible, with as many disk spindles in the group as possible.  Regardless 
 >> of the OS.  > D > This is not the "best" way to do things.  Other OSes have built-inH > throttling mechanisms which regulate the number of pending IO requestsA > per LUN.  For example, Windows or Tru64 have a "queue depth" of 4 > something like 32.  Don't quote me on that number. > C > If you have one LUN (on an EVA) served to a Windows box, and that E > Windows box has 32 pending IO request to that LUN, it will wait for F > one of the IOs to complete before it issues the next IO to that LUN.  9 	I've seen queue lengths of 1 and 300 Random IO come from 9 	a single LUN.  Likewise, I've seen queue length of 1 and A 	seen less than a 100 come from a single LUN.  What is different,  	what is the same?  9 	What does queue length really mean and how is it related  	to LUN performance?   > G > So if you have one LUN on an EVA (still taking Windows here), you can H > have (1 x 32) pending IO requests.  If you create ten LUNs on the sameF > EVA and present them to the same Windows box, now you can have (10 x) > 32) pending IO requests on that system.  >   ; 	But how many IOs is the system working on while waiting on  	IO?  H > *This* is why I originally asked the question about having one LUN vs. > many LUNs. > F > On a Windows system or Tru64 system, you want *more* LUNs so you can > throw more IOs at an EVA.  >   4 	Smart (most) RDBMS systems run with queue length of5 	1.  You want 1 (not 30) pending IOs so you are using = 	the IO system to its fullest but NOT killing it.  VMS backup ? 	is a beast, it throws as many IOs at the IO system as you have < 	the account tuned for.  I think you know that.  Point is...> 	you are mixing up outstanding IO requests with IO throughput.  F > On VMS, it doesn't matter.  The OS doesn't throttle IO requests likeE > that (which is why I can bury my HSG80s with VMS 7.3-1 and make the  > HSG80s go tits-up).     : 	OSes don't do much of anything.  Applications do, whether@ 	RDBMSes or others like VMS backup.  You buried your HSG80s with/ 	VMS backup.  Check out your queue depths with:    		$ MONITOR DISK/ITEM=ALL   < 	When performing large queries and VMS backup isn't running.   > G > Since VMS does not have the "throttling" mechanism I mentioned above, H > we are going to focus our design on matching the number of LUNs to the6 > number of backup streams we want to run in parallel. >   @ 	It does have a throttling mechanism.  You showed that in other E 	threads.  Keep an eye on DIOLM at the UAF level , especially on VMS  0 	backup as it is a beast and takes no prisoners.  F > Since we have four DLT8000 tape drives directly attached to the ES40C > where our production app (IDX) runs, we are going to present four D > LUNS.  Yes, we could get away with one HUGE lun, but then we could > only run one backup stream.  > D > If anyone from VMS Engineering or Storage Engineering reads what IG > posted above and disagrees with the way I explained the throttling in H > various OSes, please speak up.  This topic was addressed in one of theH > EVA-related seminars in Atlanta.  I don't have the name of the seminarG > at-hand but I could dig it up if someone wants to research this issue E > further.  The Powerpoint presentation from that seminar is probably  > available online.    	waitio is iowait.  ; 	It makes little sense to send from the OS more IO requests < 	if the LUN you are talking to is saturated.  Other OSes tryC 	to protect you, VMS doesn't and by design.  VMS with a process IO  E 	model doesn't have an IO traffic cop like other OSes.  IO is at the  ; 	process level and hence you have things like DIOLM in UAF.   A 	One of the few times this REALLY gets you in trouble is when you > 	have tuned up your BACKUP account.  It came back to haunt youB 	as they streamlined IO by adding FAST_PATH to fibre, IOLOCK8 work/ 	so indeed VMS can pump more IO down the wires.    				Rob    ------------------------------  # Date: Wed, 08 Oct 2003 01:51:18 GMT ( From: Alder <PGDEHMKOKIMD@spammotel.com>4 Subject: HGFTP - how to troubleshoot data connection* Message-ID: <q4Kgb.6619$Sg4.4524@edtnps84>   [Hobbyist alert!]    This should be simple, but ...  G Since reinstalling Win2K on my internet gateway system I can't seem to  D get a data connection to the FTP server from outside my LAN.  I can F connect and login just fine, get STAT and SYST results, but directory  listings  I To eliminate other possible causes I shut down the ZoneAlarm firewall on  B the gateway system, but the results were the same.  It also seems I unlikely that it's the fact I'm using Internet Connection Sharing on the  G W2K box, since connections to my other VMS servers like HTTP and HTTPS   are working just fine.  H The HGFTP log files aren't too revealing (to me).  Here's an example of E me connecting (telnet to remote VMS host, then FTP back from there),  G running the STAT and SYST commands, then trying a GET of a known file.  3 The client merely "locks" and eventually times out:   @   7-OCT-2003 18:40:20.49 Connection accepted for server 00000009F   7-OCT-2003 18:40:21.13 Server 00000009 (142.32.2.124) [FTP: Session ) begins. User=<USERID>, Host=<RemoteHost>] D   7-OCT-2003 18:40:42.77 Server 00000009 (142.32.2.124) [FTP: STAT ]D   7-OCT-2003 18:40:53.30 Server 00000009 (142.32.2.124) [FTP: SYST ]>   7-OCT-2003 18:43:18.77 Connection closed for server 00000009@   7-OCT-2003 18:43:18.85 Connection accepted for server 0000000AF   7-OCT-2003 18:43:19.39 Server 0000000A (142.32.2.124) [FTP: Session ) begins. User=<USERID>, Host=<RemoteHost>]   ? Any suggestions on troubleshooting this would be appreciated...   	 	GREATLY!    Thanks for reading,    Alder    ------------------------------  % Date: Tue, 07 Oct 2003 22:28:58 -0400 * From: JF Mezei <jfmezei.spamnot@istop.com>8 Subject: Re: HGFTP - how to troubleshoot data connection) Message-ID: <3F837668.856B8106@istop.com>    Alder wrote:H > Since reinstalling Win2K on my internet gateway system I can't seem toE > get a data connection to the FTP server from outside my LAN.  I can G > connect and login just fine, get STAT and SYST results, but directory 	 > listing   N No Microsoft software is robust enough to be reliably used as a gateway. It itM gets infected with a virus, it has full access to both sides of the firewall. 1 Firewalls are there to protect Microsoft systems.   * In terms of FTP, the default behaviour is:% CLIENT connects to port 21 of SERVER. N When client issues a command such as a directory or get file, it is the serverM which then calls the client on a different port. If your firewall is really a I NAT box, it needs to know to which local host to route this inbound call.   I One way out of this is to put the FTP server into passive mode (PASV if I L remember correctly). This will allow the client to initiate all connections.  M (Think of the connection to port 21 as a control channel, and new connections ( being needed for any/all data transfers)   ------------------------------  % Date: Tue, 07 Oct 2003 15:49:09 -0400 * From: JF Mezei <jfmezei.spamnot@istop.com>* Subject: Re: MAIL: importing of text files) Message-ID: <3F8318B3.B5BE451E@istop.com>    Hein van den Heuvel wrote:C > You fat-fingered something. This works as expected on my 7.1 box: L > [Or your editted mail file still was not exactly like a mail file.Stream?]   Here is exactly what I did:    VAX VMS 7.2   P From decwindows mail, I create a new message. Prior to or after sending it, I do/ FILE->EXTRACT. In the extract dialogue, I have:  [*]	Include header information [ ]  Append to existing file   Input format: As is  Output format: Same as input  Conversion options file: <blank>  > This creates a variable length file, devoid of <ff> character.  H If I then go to character cell MAIL, and SET FILE TEMP.TXT, a DIR yieldsN nothing. I must edit the file and add a <ff> character at the top of the file.I Then, when I SET FILE and DIR, it shows the message and I can display its M contents. It allows me to COPY it to a ne .MAI file, but it does nothing when + I try to copy to my existing MAIL.MAI file.    Exact commands that fail:  MAIL  MAIL>SET FILE SYS$LOGIN:TEMP.TXT MAIL>DIR (shows message)! MAIL>1   (shows message contents) N MAIL>COPY OUTBOX [JFMEZEI.MAILBOX]MAIL.MAI   (no message, appears to work, but
 does nothing) I MAIL>COPY OUTBOX SYS$LOGIN:HEUVEL.MAI        (telle me heuvel.mai doesn't 4 exist, asks me if I want to create it, I say "YES".)   Then: " MAIL>SET FILE SYS$LOGIN:HEUVEL.MAI MAIL>DIR/FOLDER=OUTBOX MAIL>1> MAIL>COPY OUTBOX [JFMEZEI.MAILBOX]MAIL.MAI     This one works.   ------------------------------  # Date: Wed, 08 Oct 2003 02:10:09 GMT 6 From: "Craig A. Berry" <craigberry@mac.com.spamfooler> Subject: Re: MD5 source code ?@ Message-ID: <ae0ff0f7430ba61d419caf5d28645e01@news.teranews.com>  4 In article <3phgb.81032$qU6.1291982@news.chello.at>,8  peter@langstoeger.at (Peter 'EPLAN' LANGSTOEGER) wrote:  L > Does anyone have a source code ready which does a MD5 checksum of an input3 > file and puts the hash/checksum in a DCL symbol ?   G It's pretty simple in Perl using extensions that are included with the  
 core package:      $ type get_md5.pl  use strict;  use VMS::DCLsym; use Digest::MD5;  9 open F, $ARGV[0] or die "Cannot open file $ARGV[0] : $!"; 6 my $digest = Digest::MD5->new->addfile(*F)->hexdigest;H VMS::DCLsym->new->setsym('MD5$MD5', $digest) or die "Cannot set symbol:  $!"; exit;  $ % $ perl get_md5.pl sys$login:login.com  $ show symbol md5$md5 .   MD5$MD5 = "db6aa5d44c8172b38193d8c3d1dc9aa0" $    ------------------------------  * Date: Tue, 7 Oct 2003 19:24:17 +0000 (UTC)) From: "Jeff" <dirkdiggler@totalise.co.uk>  Subject: MSA1000 & OpenVMS0 Message-ID: <blv3t1$t77$1@sparta.btinternet.com>  L Am looking to get a couple of DS25's and an MSA1000 SAN Controller, and  run it under 7.3-x.   I However I have read recently on this forum about 7.3-1's ability to drown . HSG controllers with I/O's, and crash them !!!  L Has anybody had any experiences with the MSA1000 controllers under VMS ?  If- so would you like to share your experiences ?   I I am interested in how these devices stack up against the big brother EVA + style controllers, and the HSG controllers.   & Thanks in advance for any information,   Jeff   ------------------------------  $ Date: Tue, 7 Oct 2003 19:03:03 -0400  From: John Santos <JOHN@egh.com> Subject: Re: MSA1000 & OpenVMS- Message-ID: <1031007190003.403B@Ives.egh.com>    On Tue, 7 Oct 2003, Jeff wrote:   N > Am looking to get a couple of DS25's and an MSA1000 SAN Controller, and  run > it under 7.3-x.  > K > However I have read recently on this forum about 7.3-1's ability to drown 0 > HSG controllers with I/O's, and crash them !!! > N > Has anybody had any experiences with the MSA1000 controllers under VMS ?  If/ > so would you like to share your experiences ?  > K > I am interested in how these devices stack up against the big brother EVA - > style controllers, and the HSG controllers.  > ( > Thanks in advance for any information, >  > Jeff   Jeff -  D I just asked a fairly similar question in a followup to one of ScottE Vieth's posts in the thread about EVAs, so you'll want to watch there  too in case someone replies.   --   John Santos  Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------   Date: 7 Oct 2003 23:40:24 -0500 + From: young_r@encompasserve.org (Rob Young)  Subject: Re: MSA1000 & OpenVMS3 Message-ID: <qxILOfPzIxVC@eisner.encompasserve.org>   \ In article <blv3t1$t77$1@sparta.btinternet.com>, "Jeff" <dirkdiggler@totalise.co.uk> writes:N > Am looking to get a couple of DS25's and an MSA1000 SAN Controller, and  run > it under 7.3-x.  > K > However I have read recently on this forum about 7.3-1's ability to drown 0 > HSG controllers with I/O's, and crash them !!! >   7 	You can crash the controllers running AIX or pick your 7 	favorite OS.  It has been reported in this forum that:   d http://groups.google.com/groups?selm=5a85bce2.0309120734.7b37cbdf%40posting.google.com&output=gplain  F "The VMS Engineering person told me that the HSG80s have a total queueG depth (if that is the correct term) of *240* outstanding I/O requests."    > 	Design a storage infrastructure AND then use it appropriately< 	such that you aren't killing it.  In fairness though, there: 	surely is a firmware fixup lurking to avoid the crash and; 	burn behavior of the HSG80s on saturation.  Must currently , 	be a throttling mechanism that crapped out.  ! 	Know your cuts of meat and . . .    			"Know your IO"   K > I am interested in how these devices stack up against the big brother EVA - > style controllers, and the HSG controllers.   < 	There is a lot of technical documentation out there, that's  	probably a good starting point.  > 	One thing to note is the HSG80 is 100 MByte/sec per port, the; 	MSA1000 has twice the port bandwidth with the 2GBit option   D http://h18000.www1.hp.com/products/quickspecs/11621_na/11621_na.HTML   				Rob    ------------------------------   Date: 7 Oct 2003 16:23:58 -0700 1 From: susan_skonetski@hotmail.com (Sue Skonetski) ' Subject: Re: NSlookup bug: flex scanner ; Message-ID: <857e9e41.0310071523.f6a0bc@posting.google.com>    JF,   = I have forwarded to engineering and they are looking into it.X   Thanks for letting me know.    Sue5      [ JF Mezei <jfmezei.spamnot@istop.com> wrote in message news:<3F7DCAA3.44B0AC45@istop.com>...v( > VAX VMS 7.2,  TCPIP services 5.3 ECO 2 > L > Sue, can you make sure nobody in the TCPIP group is allowed to go surfing,K > lounge by the poolside or play golf this weekend ?  They need to do their  > homework first...  >  >  > $ nslookup% > Default Server:  BIKE.chocolate.como > Address:  10.0.0.10e >  > > www.digital.como > Server:  BIKE.chocolate.com  > Address:  10.0.0.10t >  > Non-authoritative answer: ! > Name:    www.hpgtm.speedera.net ) > Addresses:  192.151.52.187, 192.6.234.9k3 > Aliases:  www.digital.com, whp-jump.extweb.hp.com  >  > >e- > [Interrupt Process]  <---- I press <ctrl-c>y >  > >n; >   fatal flex scanner internal error--end of buffer missedo* > %NONAME-E-NOMSG, Message number 00000002 > $p   ------------------------------  % Date: Tue, 07 Oct 2003 19:43:26 -0400w* From: JF Mezei <jfmezei.spamnot@istop.com>' Subject: Re: NSlookup bug: flex scannere) Message-ID: <3F834F8E.492A96EE@istop.com>    Sue Skonetski wrote: >  > JF,  > ? > I have forwarded to engineering and they are looking into it.5  L Yes, but did you have to use your whip to coerce them into looking into it ?   :-) ;-) ;-)r   ------------------------------   Date: 7 Oct 2003 17:03:59 -0700 1 From: susan_skonetski@hotmail.com (Sue Skonetski) Y Subject: OpenVMS Pearl - Beautiful Live Link to Customer OpenVMS System, using RMS and Ap-= Message-ID: <857e9e41.0310071603.7df36564@posting.google.com>-   ----Original Message----- . From: Andreas Fassl [mailto:afassl@progis.de] ( Sent: Thursday, October 02, 2003 9:16 AM To: Skonetski, Susan0 Subject: Interesting Live Link to OpenVMS System    % Very fast demo of a RMS-based system.6   http://www.vmslive.com   ------------------------------  % Date: Tue, 07 Oct 2003 20:47:09 -0400o* From: JF Mezei <jfmezei.spamnot@istop.com>Y Subject: Re: OpenVMS Pearl - Beautiful Live Link to Customer OpenVMS System, using RMS any) Message-ID: <3F835E78.9FD5AFD1@istop.com>n   Sue Skonetski wrote:2 > Subject: Interesting Live Link to OpenVMS System > ' > Very fast demo of a RMS-based system.  >  > http://www.vmslive.com    J A big non-no on any web page: "Only IE5/6 will work while we develop". You4 should have just "This page still under development"  N Also, when you do submit a query, the response is provided with a content typeK of "text/plain" isntead of "text/html" which means that standards compliant @ browsers display the html as text instead of rendering the html.  N However, I applaud the effort, and this shows a lot of promise. Long live RMS.   ------------------------------   Date: 7 Oct 2003 16:25:48 -0700"1 From: susan_skonetski@hotmail.com (Sue Skonetski)e- Subject: OpenVMS Pearl Open Source on OpenVMSe= Message-ID: <857e9e41.0310071525.41a6094a@posting.google.com>    Dear Folks,i  $ This is a little longer than normal.  D The following article has been translated from Dutch by the author. : Ton thank you very much and it was a pleasure to meet you.  
 Warm Regards,e   Suee   -----Original Message-----9 From: Ton van der Zwet [mailto:ton.vanderzwet@hccnet.nl] g& Sent: Monday, October 06, 2003 4:43 PM$ To: McCusker, Brad; Skonetski, SusanA Cc: Leendert Hinds; Martin Borgman; Woertman, Gerrit; B.A.J. Pelt E Subject: translated article Dutch pointer (OpenOffice, open source one OpenVMS)    
 Sue, Brad,  D     as promised, a non-professional as-is translation of the articleE in the Dutch pointer. I hope you will enjoy reading it. We are always D willing to answer any questions you may have. It is OK to publish. I send you a plain text version.   kind regards Ton van der Zwet.    + Open Source on OpenVMS, porting OpenOffice.e   What is open source?  ? Almost every IT-professional has heard of open source. Recentlyt= several professional journals published stories about severalm@ government services and large companies looking into open sourceB products or starting to use these products in their organizations.E Open source is clearly winning popularity and becoming more importantiF for the business community. HP recently started a separate open sourceA business unit in the Netherlands. To cut a long story short: Openl$ Source is becoming booming business.  D But what is open source exactly? Let's give a short definition: OpenD source is software of which the source code is made available to the users by the author.  E The question is however: does this definition cover it completely? WeyF don't think so. Let us name a few other distinguishing characteristics of open source:o@ - Open source is often available free of charge, but this is not necessary the case.n> - Open source is available in source code on many sites on the	 Internet.tF - The source code or parts of it may not be sold or used in commercial	 packages. A - One is however allowed to charge money for third party support. @ Charging money for third party distribution is allowed, if   the@ charge is for putting together, making, offering and selling the
 distribution.kF - Is legally covered by an open source license, often the GPL (General Public License). - Nobody is juridically liable.aE - One is allowed to use the source code, enhance and adapt it, if one 5 is willing to share the changes with the 'community'.eE - It is usually quite easy to get the software running on a UNIX-liket system.nE - There is no formal support; support is mostly organized via mailingl lists and/or newsgroups.* - The Internet is crucial for open source!  & What is the importance of open source?  D Almost everybody has heard of Linux, but most people don't know thatE Linux is but a very small portion of what's on the Linux distributionsD CD's. The Linux part generally fits on a 1.44 MB floppy. The rest ofD the CD's is filled with the result of thousands of other open sourceD projects which mostly fall under the GPL. Further one can state thatD the Internet in its present form couldn't exist without open source.> Currently the biggest part of the Internet runs on open sourceE software. By far the biggest part of the servers on the Internet runsmD on Linux or freeBSD. More than 70% of the web servers runs on apacheB and this number is rising. Who has never heard of Netscape and theF current open source version Mozilla? Open source script languages likeA Perl and python are very popular. Almost every Unix shell is open9 source.>   Other examples  F GCC. The open source compiler suite featuring a C, C++ and objective C@ compiler, there are several extensions for other languages like:C pascal, Modula, Ada, Fortran to name but a few. Maybe even Cobol isrE supported somewhere. There are versions of GCC ported to OpenVMS, butgF with the hobbyist license program every hobbyist can use the excellentE DEC compiler suite available on OpenVMS. There's even a free versionst/ (not open source) of these compilers for Linux.m  = XFREE86. This is THE X-Windows server used on all open sourcei; operating systems. I don't believe there is an alternative.a  @ KDE. This is THE open source alternative for CDE (Common DesktopC Environment). It is a X-Window manager as well as framework for theeD development of X-Windows applications. On OpenVMS we have the choiceB between the very classical DECWindows or the MOTIF CDE combination5 which is used by most of the commercial UNIX systems.a  E Gnome. At this moment the most widely used X-Window manager on Linux,,; like KDE it is a framework for the development of X-WindowshC applications. Ximian Desktop (open Source, but for professional use F not free) is a supported product based on Gnome, which is available onA HP-UX. Rumors  are that HP will use this product in the future to'A replace the expensive MOTIF CDE combination on HP-UX and OpenVMS.3  @ GTK+. A very popular framework for the development of X-Windows F GTK+applications. Reasonable recent versions of GTK+ are available on $ GTK+OpenVMS and are used by Mozilla.  C UNIX was, before AT&T Bell Labs made it a commercial product, so tod7 say open source, if we had known the term at that time.l  - Why should we or our company use open source?aA First and foremost openness. You are no longer dependent upon oneiD supplier, you can shop around for hardware, software and support. InE the proprietary-age the customer was tied to one brand, one supplier,uC one platform and they only had the source code of the software theyd@ made themselves. Today, in the open source age thing are totallyB different. This can lead to a substantial lower TCO (Total Cost ofD Ownership). More and more suppliers abandon proprietary solutions toB save costs and deliver products and or solutions for lower prices.F More and more of these suppliers start supporting open source softwareF and sell for example Hardware with Linux based solutions. Almost every! supplier offers this possibility.t   Open source products on OpenVMS   D Just at the moment everybody thought OpenVMS was about to be buried,D Compaq and the DOD came to a long term support agreement on OpenVMS.@ Because of this even Gartner started to write nice reports about OpenVMS.  D To meet the terms of the support agreement with the DOD, OpenVMS hadD to be changed so it would meet the requirements for DII/COE (DefenseD Information Infrastructure/Common Operating Environment). This meansF that COE would not only work on OpenVMS without adaptations, but wouldD also perform quite well. A major part of these adaptations consistedF of supporting the same set of system calls (C run time library) as the< reference system (SUN OS). Because performance was (is) veryC important, the decision was made to abandon the 'POSIX for OpenVMS'uF path. POSIX for OpenVMS was a separate layered product, installable onC top of the base OS and infamous for its slowness. POSIX for OpenVMS F was used in the past for porting some Unix and/or open source softwareC to OpenVMS. The first ports of Mozilla were done this way. Software C ported to OpenVMS using POSIX for OpenVMS is often also quite slow.r  @ The Unix functions that were added to OpenVMS to qualify for theE DII/COE certification, were added to the OpenVMS kernel or had kernel = support. The goal was that ported software would, without any F additional effort, run as fast as it would on a comparable Unix systemA (SUN OS). All these changes make OpenVMS also more attractive for E software development companies (ISVs). They can now, without too much(F extra work, enlarge their market by porting their products to OpenVMS.  F In order to make the OpenVMS market more attractive the changes didn'tC stop with the changes necessary for the DII/COE certification, theyeE even think of going as far as making it easy for Unix system managers E to manage OpenVMS Systems with the Unix tools they know. This ensurede? companies of a relative low educational effort to be able to do + system, network and application management.h  C Despite the troublesome port of open source to OpenVMS in the past,h@ especially without POSIX for OpenVMS, recent versions of MozillaE exist. The OpenVMS version of Mozilla is often released shortly aftertE the Windows, Linux and MacOS versions. There is also a Compaq brandedoC version of Mozilla, CSWB. The most recent versions of Apache run onn= OpenVMS and CSWS is the Compaq branded version of it. Perl isDF available, just as Python. Apache extensions as Mod_PERL, Mod_PHP, andC SOAP are available. Java 1.4.1 (not under GPL), Samba and MySQL areb@ available.  Several standard GNU-tools are already several yearsF available under the name GNV (GNU is Not VMS). Since OpenVMS 7.3-1 GNV? is available on the porting CD. This CD is part of the standardUC OpenVMS 7.3-1 distribution. This CD contains more open source toolsrD such as CD writer software. The changes to OpenVMS that were started> by the DOD agreement will make porting of open source software increasingly easier.  ! Porting of OpenOffice to OpenVMS.   D Once again you can say that many people have heard of OpenOffice butC don't know exactly what it is. Openoffice is a full featured office E suite. It consists of a text editor, a HTML editor to make web-pages,1< a spreadsheet, a presentation package and a drawing package.B OpenOffice is based on StarOffice which SUN bought a few years ago@ from StarDivision. SUN decided in order to keep the cost down toB market two product streams of StarOffice: OpenOffice which is openD source and StarOffice which is the commercial and supported version.E The open source community is developing the product as an alternative F for Microsoft Office and SUN  profits from this efforts by selling theD product, charging only for the media and distribution and by sellingB support. SUN  gives knowledge and other resources back to the openD source community. Furthermore  SUN has - contrary to most other openF source projects - taken the third party liability OpenOffice is ported= to Windows, Linux, Solaris and MacOS X. Porting OpenOffice isn> difficult, no matter the platform. Part of the problem is that6 StarOffice was originally developed for 8 bit systems.  A During a meeting of the Interex OpenVMS SIG in the Netherlands anJF enthusiastic group of OpenVMS SIG members gave a presentation on 'OpenF source on OpenVMS'. The presenters were already working some 18 monthsC with a port of OpenOffice to OpenVMS. The idea to start the port of E OpenOffice began when this group heard of the DII/COE initiative. The F first OpenVMS versions made for the DII/COE certification (7.2-6C1 andF 7.2-6C2) were not available to this group. The group had to wait untilF portions of the DII/COE made it into OpenVMS 7.3-1. In order to reduceF the waiting time the group decided, after talking with Gerrit WoertmanD (Dutch OpenVMS ambassador) to participate in the external field test> of OpenVMS 7.3-1. During this field test several contacts were@ maintained with OpenVMS engineering. These contacts with severalF engineers were maintained after the ending of the field test, in orderD to exchange as much experience and ideas as possible. While the portD of DII/COE functionality started on 7.3-1 the open source/OpenOfficeF group is currently participating in the external field test of OpenVMS 7.3-2.  F The port of OpenOffice to OpenVMS is not an easy task. The goal is notD reached yet and is in fact still far away. This is partly due to theC participation in field tests. The group has to start all over againeD every time, because work-arounds made to get something going, are no' longer needed or simply have to change.u  @ With every new version of OpenVMS the group advances a few steps? further. The tools necessary for the building of OpenOffice are D working. Together with HP the group has been working on parts of theD GNV kit since it wasn't possible at first to unpack the source files@ (.tar.gz) on OpenVMS. The Gunzip version did the 'unzipping' butB didn't always name the files correctly and tar couldn't completelyC untar the unzipped tar file. The number of tasks is still enormous. ; The goal can only be achieved if more community member willi@ participate. Various specialists are needed. There is a need forF developers, testers, coordinators, document writers and web designers.F Many hands make light work, so if many specialists join us, it must be5 possible to port OpenOffice to OpenVMS within a year.o  F People interested in joining this group can contact the Project leaderE Ton van der Zwet <mailto:ton.vanderzwet@oooovms.dyndns.org> or Gerrit F Woertman <mailto:gerrit.woertman@hp.com> the Dutch OpenVMS Ambassador.< The group has a web site at <http://www.oooovms.dyndns.org>.   ------------------------------  % Date: Tue, 07 Oct 2003 18:47:41 +0100f From: "Tom B." <tom@here.now>l' Subject: Re: Read this and weap, twats!i. Message-ID: <bluu88$m7t$1@news8.svr.pol.co.uk>   Diogenes wrote:M   > Tom B. wrote:g >> Diogenes wrote: >>   >>> Sinister Midget wrote: >>>  >>>> Read that, Windozer.  >>> 8 >>> Read what?  It must be true: linux makes you stupid. >> /G >> The fact that you couldn't read it demonstrates that you need to getb  >> yourself a *real* newsreader. > : > Have one can't you tell?  Sheezz linux makes you stupid.  ) Microsoft Outlook Express 6.00.2800.1158?c  F OE is a joke of an attempt at a newsreader/e-mail client.  M$ are even" killing it off in the near future. -- r Registered Linux user #328961.   ------------------------------   Date: 7 Oct 2003 23:13:36 -0500,+ From: young_r@encompasserve.org (Rob Young) / Subject: Re: Slow Shadow Copy with 36 GB drives 3 Message-ID: <$B8GQeHnfZyd@eisner.encompasserve.org>   U In article <blkfd7$1hh$1@news.doit.wisc.edu>, "Jim" <jim.gould@uwmf.wisc.edu> writes:oL > I have HP support stumped on this one so I thought someone here could help > me.s > M > We recently upgraded a pair of HSJ40 controllers to firmware version V37J-1eN > so that we could replace our aging 2GB and 4GB drives with new 36 GB drives.H > We are running VMS 6.2 and applied the latest shadowing patch (ALPSHAD	 > 14_62.)r > M > Note:  We can't upgrade to a new version of VMS without paying our softwaree > vendor huge $$$. > H > Everything works fine except that shadow copies take about 24 hours toN > complete.  Other drives, such as 9 GB drives take about 2 hours to complete. > L > Is this an inherent problem with VMS 6.2 and large drives or is there some! > tweak that we missed somewhere?0 > L > Help would be appreciated since we need to keep this ancient monster alive > for another year :  ). >   * 	Shadowing can be a tricky beast at times.  - 	Those HSJ40s are dinosaurs, you can probablyNA 	pick up some SW500s and HSJ50s on the very cheap now, the HSJ50s.F 	support writeback cache.  The HSJ40s don't, the added performance is 4 	well worth it.  Got any extra ports in the CI?  B-)  5 	Make DCD work for you.  In fact, I suspect the 9 Gigd> 	drives that do two hour copies are on the same controller and? 	DCD is working on those.  How to confirm DCD?  During a shadowe: 	copy there is no (or extremely little) IO taking place on 	the drives at a system level:   	$ monitor diske  A 	Yet the copy is proceeding apace as seen via a show dev dsann.  oA 	Those 36 GByte drives that aren't copying so fast, you will see pE 	considerable IO at the system level.  This confirms the OS is doing  D 	all the heavy lifting.  DCD works if drives are online to the SAME F 	controller.  Perhaps the 36 GByte drives are across controller pairs 8 	OR on separate contollers in 2 different cabinets, etc.  A 	But really.. there isn't enough information here.  Drag out someh 	stats for us.  Show us:  7 	system io when copying is taking place - modest snaps: = 	$ monitor disk   ! Interested if IO is taking place on copy,b? 	this can be tricky to determine if others are hammering so youg4 	may have to sort out wheat from chaff to determine. 	-+ 	What are the contollers doing (run VTDPY)?0  ' 	What controllers are drives online to?Y   	How many systems involved?o  A 	What is SHADOW_MAX_COPY set to?  Could it be something as simpler@ 	as the 9 GByte drives performing copies and the 36 Gbyte drives? 	must wait to even get started as SHADOW_MAX_COPY is set to thei> 	default of 4 and 9 GByte copies are queued up ahead of the 36 	drive copies?  " 	There are many things to look at.   				Robe   ------------------------------  * Date: Tue, 7 Oct 2003 20:26:43 +0000 (UTC) From: david20@alpha2.mdx.ac.uk Subject: Re: Sun takes a hit) Message-ID: <blv7i3$nli$1@news.mdx.ac.uk>   q In article <2jQ4Mo9G3J5s@eisner.encompasserve.org>, koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes: r >In article <5qpgb.244170$Lnr1.224370@news01.bloor.is.net.cable.rogers.com>, "John Smith" <a@nonymous.com> writes: > , >> So what's stopping HP from digesting Sun? >eF >   Why would HP want to buy a company that only has one product?  SunG >   prides itself on, and attracted a lot of customers because of, onlyuJ >   doing UNIX.  HP has a strong UNIX, they don't need another.  Features B >   HP said they wanted to add to their UNIX, they got with Tru64. > G >   Sun grew up with, and will die with, the UNIX market.  For the lastLH >   decade and a half that's been a good market to be in.  Sun's biggestI >   problem is that someday Solaris will be as usefull as a typewriter istL >   today.  Typewriters hung around for several decades, but those companies4 >   who had them as their sole product are gone now. >nH If Sun are badly off enough to want to sell things then I suggest HP andL Process get together to buy back PMDF. Although Process can develop and sellN PMDF they are very restricted in what they can do since Sun still legally ownsD it - for instance Process cannot produce a hobbyist version of PMDF.O I'd love for an HP/Process owned PMDF to become the standard mail system on VMSlM with VMS mail being replaced with the compatible but much superior PMDF MAIL.u   (Well I can dream can't I ?)  O A few years ago there were all the iPlanet products which might have been wortht; purchasing but i'm not sure whether thats the case anymore.n    r  
 David Webb VMS and Unix team leader CCSS Middlesex University   ------------------------------  # Date: Tue, 07 Oct 2003 21:26:03 GMTc# From: "John Smith" <a@nonymous.com>h Subject: Re: Sun takes a hitK Message-ID: <LbGgb.251788$Lnr1.165302@news01.bloor.is.net.cable.rogers.com>p   david20@alpha2.mdx.ac.uk wrote:=5 > In article <2jQ4Mo9G3J5s@eisner.encompasserve.org>,s? > koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:c
 >> In articlenC >> <5qpgb.244170$Lnr1.224370@news01.bloor.is.net.cable.rogers.com>, ( >> "John Smith" <a@nonymous.com> writes: >>- >>> So what's stopping HP from digesting Sun?  >>G >>   Why would HP want to buy a company that only has one product?  SunIH >>   prides itself on, and attracted a lot of customers because of, only@ >>   doing UNIX.  HP has a strong UNIX, they don't need another.E >>   Features HP said they wanted to add to their UNIX, they got withr	 >> Tru64.p >>H >>   Sun grew up with, and will die with, the UNIX market.  For the lastA >>   decade and a half that's been a good market to be in.  Sun'snD >>   biggest problem is that someday Solaris will be as usefull as aG >>   typewriter is today.  Typewriters hung around for several decades,eD >>   but those companies who had them as their sole product are gone >> now.t >>F > If Sun are badly off enough to want to sell things then I suggest HP > andeE > Process get together to buy back PMDF. Although Process can develop 
 > and sellC > PMDF they are very restricted in what they can do since Sun still  > legally ownsF > it - for instance Process cannot produce a hobbyist version of PMDF.C > I'd love for an HP/Process owned PMDF to become the standard maile > system on VMS D > with VMS mail being replaced with the compatible but much superior > PMDF MAIL. >p > (Well I can dream can't I ?)    J Anything that makes eminent sense with respect to VMS won't be done by HP.  F Process could buy PDMF from Sun if they had any faith in the long term! political viability of VMS at HP.a      F > A few years ago there were all the iPlanet products which might have > been worth= > purchasing but i'm not sure whether thats the case anymore.     < I can think of a few others that are still worth purchasing.   ------------------------------   Date: 7 Oct 2003 17:16:12 -07006+ From: shoppa@trailing-edge.com (Tim Shoppa)  Subject: Re: Sun takes a hit= Message-ID: <bec993c8.0310071616.31aa7558@posting.google.com>c   Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> wrote in message news:<blurcu$403$1@new-usenet.uk.sun.com>...s > Tim Shoppa wrote:t > > Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> wrote in message news:<blu5jl$osa$1@new-usenet.uk.sun.com>...s > > E > >>Exentually you realise that Free Linux is in fact Expensive LinuxnD > >>unless you think that paying 7.5K over 3 years for SW support on' > >>a box that costs 6K is a good deal.  > >  > > M > > 7.5K and 6K are both in the noise... unless you're talking about hundreds N > > or thousands or tens of thousands of boxes.  If you've got that many boxesN > > then you *do* go with the open-source alternative and self-support becauseJ > > you have all the sources and you built from them.  (Or at least that's > > what I hope happens!)0 > >  > 4 > Hang on what value does Linux bring to an IT shop.  > It gives the shop full and absolute control over its code base? and chosen (not dictated) hardware.  This is of enormous value.r  D Few shops use Linux in this way... and those that do inevitably wereG the same ones hacking BSD or SysV source kits ten and twenty years ago.e   Tim.   ------------------------------   Date: 8 Oct 2003 01:12:56 GMT2( From: bill@cs.uofs.edu (Bill Gunshannon) Subject: Re: Sun takes a hit9 Message-ID: <blvoao$gvplm$1@ID-135708.news.uni-berlin.de>   = In article <bec993c8.0310071616.31aa7558@posting.google.com>,-. 	shoppa@trailing-edge.com (Tim Shoppa) writes: > F > Few shops use Linux in this way... and those that do inevitably wereI > the same ones hacking BSD or SysV source kits ten and twenty years ago.c  ? I would hope that any shop comprised of people who were hackingeA BSD ten and twenty years ago would still be hacking BSD today ande: would know enough to not touch Linux with a ten foot pole.   bill   --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>   o   ------------------------------  $ Date: Tue, 7 Oct 2003 18:29:54 -0400/ From: "Hank Vander Waal" <hvanderw@comcast.net>m Subject: swxcrmgrc? Message-ID: <PBEDIAGOKEDCKCLPJKAFMEIOCCAA.hvanderw@comcast.net>n  L does anyone know if there is a Charter cell version of this program anywhere ??L I have a customer that has had to replace several drives in his system and I need to reformat5 that drive from my office - his is many miles away :(a- I can telnet into his system but that is all.   / This or any other ideas would be most helpful !a   Thanks Hank   ------------------------------   Date: 7 Oct 03 17:06:05 PST- From: mckinneyj@cpva.saic.comu Subject: Re: swxcrmgr ( Message-ID: <Pa5DJ66bC4yp@cpva.saic.com>  ? In article <PBEDIAGOKEDCKCLPJKAFMEIOCCAA.hvanderw@comcast.net>,32  "Hank Vander Waal" <hvanderw@comcast.net> writes:N > does anyone know if there is a Charter cell version of this program anywhere > ??N > I have a customer that has had to replace several drives in his system and I > need to reformat7 > that drive from my office - his is many miles away :(m/ > I can telnet into his system but that is all.> > 1 > This or any other ideas would be most helpful !  >  > Thanks > Hank >  >   J The character cell version of swxcrmgr is srlmgr.exe. However it has to beL run in console mode and not from VMS so you won't be able to use it directlyL via telnet. You might be able to front-end the alpha server's serial console0 with a terminal server and telnet to it however.  M If you can't locate a copy of srlmgr.exe (I think it was on the RCU disketteshK as was also available on the DEC/Compaq/HP firmware in the past) I might beaM able to locate a copy (email to user mckinneyj at the domain saic.com) if you  want it. --   - Jims   ------------------------------  % Date: Tue, 07 Oct 2003 15:42:05 -0500g/ From: Chris Scheers <chris@applied-synergy.com> 2 Subject: Re: Vax / VMS V5.5-2 External Drive Issue3 Message-ID: <3F83251D.C7D9B38C@applied-synergy.com>S   John Laird wrote:  > N > On Mon, 6 Oct 2003 22:56:12 +0100, "John Travell" <john@jomatech.com> wrote: > : > >"JF Mezei" <jfmezei.spamnot@istop.com> wrote in message& > >news:3F81C7C5.4A2E994D@istop.com... > >> Antonio Carlini wrote:  > >>D > >> > What VAX did you use to format them? The early MicroVAX 3100sB > >> > and all the VAXstation 3100s cannot "see" past 1.073GB into> > >> > the disk, so you may have partially formatted the disk. > >>* > >> This applies only to the system disk. > >3 > >This is not true. > >tG > >At the console level on 'affected' machines it applies to ALL disks.yO > >The problem is that the relevant console firmware was created at a time whentL > >space in ROM chips was very restricted, so one of the tricks used to saveJ > >space was to use only 21 bits of a longword to store an lbn to use whenO > >accessing a disk, e.g. during a system boot. The other 11 bits were used fornO > >something else. This means that the console can only access lbn's within theg( > >first 1Gb of the disk (1.073decimal). > N > I thought it was related to the use of an early format of scsi command whichM > could not specify large block numbers.  I'd guess the prom fix swapped to aaN > longer command format.  (That's a little more believable than having consoleB > rom programmers saving bits down to that extreme, to be honest.)    G In its original incarnation, SCSI used 6 byte commands.  These commands G could only address 1GB when using 512 byte sectors.  However, when this7C was defined, 1GB was so far in the future that it wasn't a problem.i  A When it because a problem, 10 byte commands were added to SCSI to + support the "long" read and write commands.c  G The interesting twist is that there are earlier drives that do not know % what to do with the 10 byte commands.s  G I suspect that when the console for the VS3100 machines was coded, thislE problem was avoided by just using the 6 byte commands.  Remember that-H this was the first real foray into SCSI for DEC and all drives availableH for the machine were under the limit.  Adding console support for drives@ that didn't exist was probably not a very high priority.  <grin>  E I believe that later consoles throw a 10 byte command at a drive.  IfrB they don't get a reasonable response, they fall back to the 6 byte( command set to support the older drives.  G -----------------------------------------------------------------------e$ Chris Scheers, Applied Synergy, Inc.  C Voice: 817-237-3360            Internet: chris@applied-synergy.com     Fax: 817-237-3074w   ------------------------------  % Date: Tue, 07 Oct 2003 15:49:15 -0500a/ From: Chris Scheers <chris@applied-synergy.com> 2 Subject: Re: Vax / VMS V5.5-2 External Drive Issue3 Message-ID: <3F8326CB.3B5C7935@applied-synergy.com>    BigFish wrote: > H > I just formatted two pc hdd for use on the VAX using the >>> t/ut scsiM > command. Both drives formatted successfully, however, once I brought up the0N > operating system (v5.5-2) I was not able to initialize either drives I get aN > init-f-drverr fatal drive error message. One drive is 4.2 GB's and the otherB > is 4.55 GB well underneath VAX v.5-5.2 8GB disk size limitation. >  > Any thoughts?n  G Since most PC operating systems don't know how to manage errors, when aeF SCSI drive is set up for a PC, it is usually configured to do it's ownH error management.  That means that it will automagically retry and remap bad blocks.w  F Since VMS knows how to manage errors, it wants to be in charge of this process.  F For some reason, the VMS 5.5-2 DKDRIVER is pretty draconian about thisG and will refuse to touch a drive that has been set up to manage its own7D errors.  The relevent SCSI mode bits are ARRE, AWRE, and RC.  If youE have a SCSI mode page editor, go into the drive and clear these bits.h  D There is also an unofficial (non-DEC/Compaq/HP) patch for the V5.5-2- DKDRIVER that makes the driver more tolerant.y  9 I think that these restrictions were removed by VMS V6.2.e  F There is also a problem if the INQUIRY string returned by the drive isC too long.  I forget what the limit is, but I have seen it with somec6 Seagate drives.  I don't know of an easy fix for this.  
 Good luck!  G -----------------------------------------------------------------------h$ Chris Scheers, Applied Synergy, Inc.  C Voice: 817-237-3360            Internet: chris@applied-synergy.com i   Fax: 817-237-3074    ------------------------------  % Date: Tue, 07 Oct 2003 14:24:08 -0700M+ From: "Barry Treahy, Jr." <Treahy@MMaz.com>e2 Subject: Re: Vax / VMS V5.5-2 External Drive Issue' Message-ID: <3F832EF8.5070906@MMaz.com>_   Chris Scheers wrote:   >BigFish wrote:- >  - >-H >>I just formatted two pc hdd for use on the VAX using the >>> t/ut scsiM >>command. Both drives formatted successfully, however, once I brought up the8N >>operating system (v5.5-2) I was not able to initialize either drives I get aN >>init-f-drverr fatal drive error message. One drive is 4.2 GB's and the otherB >>is 4.55 GB well underneath VAX v.5-5.2 8GB disk size limitation. >> >>Any thoughts?o >>     >> >AE >There is also an unofficial (non-DEC/Compaq/HP) patch for the V5.5-2d. >DKDRIVER that makes the driver more tolerant. >o >  b >hG This is becoming harder and hardware to locate, so for those that have k! been looking for it, see below...    Barrye    V3100$ ty DKDRIVER-JAZ.0552_COM;F sys$common:[SYS$LDR]DKDRIVER.EXE/jou=[]DKDRIVER-JAZ/out=[]DKDRIVER-JAZ ! 5 ! Patch VMS V5.5-2 DKDRIVER (X-18A7) such that it ... < ! ... doesn't reject a disk with ARRE or AWRE mode bits set,2 !       but instead _clears_ these bits on its own3 ! ... turns _off_ the reporting of recovered errors<G !       for _all_ "removable" media except CD-ROM (e.g. IOMEGA ZIP 100)e> ! ... doesn't reject a disk which won't return C/H/S geometry,: !       but instead 'fakes' a CD-ROM-like (*/6/4) geometry? !       if MAXBLOCK <=~ 800 MB, or */32/32, */96/96, */255/255.>? ! ... doesn't fail on some sort of "hardware error" right away,.@ !       but retries a few times (because it just _looks_ better) !e9 ! DKDRIVER-JAZ.0552_COM Version 5 (formerly DKDRIVER-ZIP)  !, ! w.j.m. 11-aug-19975 ! 13-aug-1997 wjm: ECOs re-structured, comments added-. ! 14-aug-1997 wjm: do not clear PER for CD-ROMA ! 24-sep-1997 wjm: add support for 'big' disks w/o C/H/S geometry  !eP !------------------------------------------------------------------------------! !, define dk$dpt = 0  !s> define dk$ddt = 00198                   ! .PSECT $$$115_DRIVER !r" define mode_sense = dk$ddt + 017AD4 define mode_sense_60$ = mode_sense + <019A3 - 017AD>4 define mode_sense_75$ = mode_sense + <019C3 - 017AD>4 define mode_sense_92$ = mode_sense + <019E1 - 017AD>4 define mode_sense_95$ = mode_sense + <019E6 - 017AD> ! # define mode_select = dk$ddt + 01B43e6 define mode_select_20$ = mode_select + <01BE0 - 01B43> !.% define read_capacity = dk$ddt + 01E8AS: define read_capacity_10$ = read_capacity + <01ED2 - 01E8A>: define read_capacity_20$ = read_capacity + <01ED5 - 01E8A>: define read_capacity_40$ = read_capacity + <01EE9 - 01E8A> ! $ define send_command = dk$ddt + 022F0 !e# define cleanup_cmd = dk$ddt + 0281Fc ! E define dk_patch = dk$ddt + 02F15        ! 200. bytes at end of driverp define dk_end = dk$ddt + 02FDD !  !l define DEV$V_SWL = 0019l define DPT$W_SIZE = 0008 define SS$_NORMAL = 0001 define SS$_DRVERR = 008C define UCB$L_DEVCHAR = 003Ce define UCB$B_SECTORS = 0050n define UCB$B_TRACKS = 0051 define UCB$W_CYLINDERS = 0052  define UCB$L_MAXBLOCK = 00C4 define UCB$L_DK_FLAGS = 01586 define  UCB$V_REMOVABLE = 0             ! DK_FLAGS bit6 define  UCB$V_CDROM = 0010              ! DK_FLAGS bit define UCB$B_HWERR_RETRY = 0170. define UCB$B_RECOV_PAR = 01A0O !V !PE ! Define bit masks for 1st byte of "error recovery" SCSI mode page #1@ !t, define m_dcr = 01       ! disable correction3 define m_dte = 02       ! disable transfer on erroraG define m_per = 04       ! post error (allegedly relates to _recovered_   errors)p1 define m_eec = 08       ! enable early correction-) define m_rc = 10        ! read continuous ( define m_tb = 20        ! transfer block3 define m_arre = 40      ! automatic read relocation 4 define m_awre = 80      ! automatic write relocation !n !o( set eco 99      ! initialize patch area ( ---------------------------------------! !e? !* Here, we create a "patch area descriptor" for the patch areap> !* at the end of DKDRIVER (left zero-filled at assembly time),@ !* such that subsequent patches can use it in "modular" fashion, !* e.g. via CHECK ECO 99.  !eB verify/long dk_patch            ! check that patch area is unused  (=zero) yetd 0Z 0. exit !nI verify/word dk$dpt + DPT$W_SIZE ! check that `dk_end' matches DKDRIVER's e size +<dk_end - dk$dpt> exit !pI set patch/init=<dk_end - dk_patch>  dk_patch    ! create patch descriptor  !e update  H !----------------------------------------------------------------------! !r6 set eco 100     ! mode page #1 bits (AWRE, ARRE, PER)  -------------------------!) check eco 99            ! need patch area  set patch dk_patch ! @ !* Here, UCB$B_RECOV_PAR+2 has the first byte from mode page #1,( !* as received via a MODE SENSE command.. !* Old code will reject (i.e. declare offline)0 !* SCSI disks with any of AWRE, ARRE, or RC set. !i( verify/inst mode_sense + <018E0 - 017AD>< '       bitb    #m_awre+m_arre+m_rc,w^UCB$B_RECOV_PAR+2(r3)' '       beql    $0.10'? '       brw     mode_sense_95$'         ! reject AWRE, ARRE, RC  '$0.10: brw     mode_sense_60$'s exit !aJ verify/inst mode_sense_60$              ! normal processing continues here '       movzbl  b^1(r0),r2'  '       addl    #2,r2' exit !l5 !* We shall test for RC only, so as to not "a priori" 3 !* reject disks that happen to have AWRE & ARRE setd ! ) replace/inst mode_sense + <018E0 - 017AD>l< '       bitb    #m_awre+m_arre+m_rc,w^UCB$B_RECOV_PAR+2(r3)' exit@ '       bitb    #m_rc,w^UCB$B_RECOV_PAR+2(r3)'  ! reject RC only exit !rA !* Here, DKDRIVER decides if it should send a MODE SELECT commandv! !* that will change mode page #1.e !*A !* Old code looks for TB, DTE, and PER and wants all of them set. > !* If they're set already, no MODE SELECT command will be sent? !* (AWRE and ARRE are known to be clear due to the above test).  !*: !* New code has to deal with a more complicated situation:9 !* We want ARRE and AWRE clear (they may be set), and fors: !* removable non-CDROM disks we also want PER clear - they8 !* may _rely_ on built-in error recovery to work at all. !e* replace/inst mode_select + <01B59 - 01B43>B '       bicb3   #00FF-m_tb-m_dte-m_per,w^UCB$B_RECOV_PAR+2(r3),r1'% '       cmpb    #m_tb+m_dte+m_per,r1'rK '       beql    mode_select_20$'                ! branch around mode selectf exit '       bicb3   @ #00FF-m_awre-m_arre-m_tb-m_dte-m_per,w^UCB$B_RECOV_PAR+2(r3),r1'L '       movb    #m_tb+m_dte+m_per,r2'           ! desired value (AxRE clear)L '       bbc     #UCB$V_REMOVABLE,w^UCB$L_DK_FLAGS(r3),$0.21'    ! removable?H '       bbs     #UCB$V_CDROM,w^UCB$L_DK_FLAGS(r3),$0.21'        ! not a  CD-ROM?hH '       bicb    #m_per,r2'                      ! yes, M_PER not desiredI '$0.21: cmpb    r2,r1'                          ! already having desired d mode?uB '       beql    mode_select_20$'                ! br if so (no-op) exit !d: !* Here, UCB$B_RECOV_PAR+2 (cf. above) is modified for use4 !* with the MODE SELECT command issued subsequently.5 !* Some bits (DCR and EEC) are always left unchanged.r !hG replace/inst mode_select + <01BA2 - 01B43>      ! modify mode page #1, , 1st byte+ '       movab   w^UCB$B_RECOV_PAR+2(r3),r6'r: '       bisb    #m_tb+m_per+m_dte,w^UCB$B_RECOV_PAR+2(r3)' exit= '       movab   w^UCB$B_RECOV_PAR+2(r3),r6'     ! (unchanged)cD '       bisb    #m_tb+m_per+m_dte,(r6)'         ! set bits as beforeE '       bicb    #m_awre+m_arre,(r6)'            ! clear ARRE and AWRETL '       bbc     #UCB$V_REMOVABLE,w^UCB$L_DK_FLAGS(r3),$0.31'    ! removable?H '       bbs     #UCB$V_CDROM,w^UCB$L_DK_FLAGS(r3),$0.31'        ! not a  CD-ROM?rB '       bicb    #m_per,(r6)'                    ! if so, clear PER '$0.31: nop' exit !  update  H !----------------------------------------------------------------------! !c1 set eco 101     ! allow for missing C/H/S values h ------------------------------!c ! patch area not required. ! ; !* This is a successful exit from the `mode_sense' routine.  !e verify/inst mode_sense_75$ '       movzwl  #SS$_NORMAL,r0'l exit ! G !* Here, it has been noticed that non-zero values for each of the C/H/SpI !* "geometry" parameters could not be determined via MODE SENSE commands.aL !* Old code does handle this condition for CD-ROM (in fact, write-protected)E !* drives only, by using a READ CAPACITY command later, and faking anrL !* appropriate c/6/4 CD-ROM-style geometry; other disks will be set offline.= !* We shall permit the CD-ROM "fake" to be used for any disk. : !* This seems to match the behaviour of VMS V7.1 DKDRIVER. !  replace/inst mode_sense_92$e> '       bbs     #DEV$V_SWL,b^UCB$L_DEVCHAR(r3),mode_sense_75$' exitI '       brb     mode_sense_75$' ! always allow for missing or zero C/H/S o data exit !r update  H !----------------------------------------------------------------------! !u5 set eco 102     ! raise HWERR_RETRY_CNT from 1 to 10 o --------------------------!c ! patch area not requiredd !a< !* An assembly-time constant determining the number of times> !* that certain hardware errors (details unknown) are retried,+ !* is 1 in V5.5-2, making for zero retries.n7 !* Allowing for 9 retries _may_ improve the resiliency.u9 !* Here, we modify the single reference to that constant.e !s replace/inst send_command+000A+ '       movb    #1,w^UCB$B_HWERR_RETRY(r3)'  exit, '       movb    #0A,w^UCB$B_HWERR_RETRY(r3)' exit !s update  H !----------------------------------------------------------------------! !gG set eco 103     ! fake C/H/S for disks with more than 65534*6*4 blocks m	 --------!r) check eco 99            ! need patch arear set patch dk_patch !iF !* If we're making up "geometry", and have more than 65534*6*4 blocks 	 (800 MB),OJ !* try several alternative track/sector pairings in sequence, in a fashion4 !* supposedly compatible with VMS V6+ and MSCP code. !iG verify/inst read_capacity_10$           ! good exit from READ_CAPACITY p routineb '       movl    #SS$_NORMAL,r0'.C '       bsbw    cleanup_cmd'            ! this is read_capacity_20$s exitI verify/inst read_capacity_20$           ! exit from READ_CAPACITY routinee '       bsbw    cleanup_cmd' exit !tB verify/inst read_capacity_40$           ! fake UCB$W_CYLINDERS ...& '       movzbl  b^UCB$B_TRACKS(r3),r1'' '       movzbl  b^UCB$B_SECTORS(r3),r0'a '       mull    r1,r0'( '       movl    w^UCB$L_MAXBLOCK(r3),r1' '       clrl    r2'  '       ediv    r0,r1,r0,r1'K '       movw    r0,b^UCB$W_CYLINDERS(r3)'       ! read_capacity_40$ + 00017t '       tstl    r1'O" '       beql    read_capacity_10$'& '       incw    b^UCB$W_CYLINDERS(r3)'" '       brb     read_capacity_10$' exit !e& replace/inst read_capacity_40$ + 00017) '       movw    r0,b^UCB$W_CYLINDERS(r3)'  exitL '       cmpl    r0,#0FFFE'                      ! we want cylinders <= 65534L '       bleq    $3.15'                          ! br if o.k. (join old code)< '       movw    #02020,b^UCB$B_SECTORS(r3)'     ! try s=t=320 '       cmpl    w^UCB$L_MAXBLOCK(r3),#003FFF800'C '       blssu   $3.40'                          ! retry if o.k. nowe< '       movw    #06060,b^UCB$B_SECTORS(r3)'     ! try s=t=960 '       cmpl    w^UCB$L_MAXBLOCK(r3),#023FFB800'C '       blssu   $3.40'                          ! retry if o.k. nowyE '       movw    #0FFFF,b^UCB$B_SECTORS(r3)'     ! last resort s=t=255.0 '       cmpl    w^UCB$L_MAXBLOCK(r3),#0FDFF03FE'C '       blssu   $3.40'                          ! retry if o.k. nowe< '       movl    #SS$_DRVERR,r0'                 ! sorry, ...E '       brw     read_capacity_20$'              ! ... have to give up ? '$3.40: brw     read_capacity_40$'              ! branch helpereK '$3.15: movw    r0,b^UCB$W_CYLINDERS(r3)'       ! (old code continues here)a exit !o update  H !----------------------------------------------------------------------! !S exit     -- b  > Barry Treahy, Jr                       E-mail: Treahy@MMaz.com> Midwest Microwave                          Phone: 480/314-1320> Vice President & CIO                         FAX: 480/661-7028                        r   ------------------------------  * Date: Tue, 7 Oct 2003 16:11:13 -0500 (CDT) From: sms@antinode.org2 Subject: Re: Vax / VMS V5.5-2 External Drive Issue) Message-ID: <03100716111319@antinode.org>i  / From: Chris Scheers <chris@applied-synergy.com>a  H > There is also a problem if the INQUIRY string returned by the drive isE > too long.  I forget what the limit is, but I have seen it with somee8 > Seagate drives.  I don't know of an easy fix for this.  G    That's a console problem, not a VMS problem.  On a VAXsta 3100 modelc= 30/40/38/48 (and the 2000?), "TEST 50" showed an error statusoE ("000000D4") for such drives, which is enough to inhibit an automatickC boot, though the drive may work just fine when the system is bootedU	 manually.f  G    The closest thing to a fix is a revised set of firmware EPROMs, data E for which exist, but you need an EPROM programmer (and EPROMs, unless  you want to live dangerously).  8    Again, it's probably not a problem on a later system.  H ------------------------------------------------------------------------  4    Steven M. Schweda               (+1) 651-699-98183    382 South Warwick Street        sms@antinode,orgl    Saint Paul  MN  55105-2547e   ------------------------------  % Date: Tue, 07 Oct 2003 19:06:41 -0500N/ From: Chris Scheers <chris@applied-synergy.com>t2 Subject: Re: Vax / VMS V5.5-2 External Drive Issue3 Message-ID: <3F835511.561701E6@applied-synergy.com>    sms@antinode.org wrote:n > 1 > From: Chris Scheers <chris@applied-synergy.com>r > J > > There is also a problem if the INQUIRY string returned by the drive isG > > too long.  I forget what the limit is, but I have seen it with someI: > > Seagate drives.  I don't know of an easy fix for this. > I >    That's a console problem, not a VMS problem.  On a VAXsta 3100 modelo? > 30/40/38/48 (and the 2000?), "TEST 50" showed an error statushG > ("000000D4") for such drives, which is enough to inhibit an automaticdE > boot, though the drive may work just fine when the system is bootedl > manually.u > I >    The closest thing to a fix is a revised set of firmware EPROMs, data3G > for which exist, but you need an EPROM programmer (and EPROMs, unless   > you want to live dangerously). > : >    Again, it's probably not a problem on a later system.  E Yes, that is a problem in the VS3100 ROMs, I think at 15 characters.  D But this isn't fatal.  Yes, the drive will work fine afterwards.  MyB Seagate ST41200N shows this problem and VMS is very happy with it.  G And no, the VS4000 doesn't show this problem.  (But it doesn't show theiD INQUIRY data after the cutoff either, so it looks like they fixed it just enough to allow booting.)  E I have also seen a problem related to the INQUIRY data that keeps VMSaE from mounting the drive.  Perhaps its not the length.  It may be thati3 the device reports it supports SCSI 3 or some such.   @ I have EPROMs and a burner.  Where can I get this update? (For a
 VS3100-38)  G -----------------------------------------------------------------------i$ Chris Scheers, Applied Synergy, Inc.  C Voice: 817-237-3360            Internet: chris@applied-synergy.com     Fax: 817-237-3074o   ------------------------------   Date: 7 Oct 2003 12:02:09 -0700N/ From: kenneth.randell@verizon.net (Ken Randell) ; Subject: Re: VMS JVM implementation of sub-process deletiond= Message-ID: <79de9693.0310071102.31244d5f@posting.google.com>d  Z "John Apps" <john.apps@compaq.com> wrote in message news:<3f817791@usenet01.boi.hp.com>...I > > The above mentioned 'feature' of the C RTL seems at best inconsistent.A > > to me.  As I said earlier, it would seem logical to me that aeH > > 'Process.destroy()' would be implemented by a $DELPRC call, but this! > > does not seem to be the case.j > K > I'm afraid you are correct when stating that the C RTL is 'inconsistent'.pN > This is the 'featyure' I alluded to and wha the C RTL group is going to take" > a look at and make 'consistent'. >  > John  B After a little more investigation, it would appear that the JVM isB using the C RTL function kill() to implement its Process.destroy()F function.  The on-line help for the kill() function explicitly states:  D        Sends a signal to a process specified by a process ID (PID).  ThisD        function does not support the same functionality supported by UNIX* B        systems.  This function is restricted to C and C++ programs that#        include the main() function.   D I can duplicate my first scenario by simply spawning a DCL procedureF and attempting to kill it with a small C program.  The kill() functionF appears to work, but nothing happens to the sub-process.  If I spawn aC C program, it exits as would be expected when kill() is used.  If aE@ FORTRAN program is spawned, it generates a TRACE-W- message, but continues execution.  D I have been unable to duplicate my second scenario without using the JVM.   ------------------------------  # Date: Wed, 08 Oct 2003 00:26:48 GMTv& From: jlsue <jefflsxxxz@sbcglobal.net>" Subject: Re: VMS system on the web8 Message-ID: <86k5ov8p6ns2knean0dqsoos3urf8389ds@4ax.com>  F On Fri, 03 Oct 2003 02:28:34 GMT, "John Smith" <a@nonymous.com> wrote:  
 >jlsue wrote: I >> On Tue, 30 Sep 2003 16:14:44 GMT, "John Smith" <a@nonymous.com> wrote:  >>   >>F >> Back when I was at an institution of higher learning, marketing hadH >> little to do with what we studied.  We were CS students and the profsH >> often used known technologies as a basis for study.  They didn't have: >> to be actual products, merely a technological solution. >n >  >Ahhh....'known'. I see. >rK >In a recent poll conducted by Fiorina, Marcello, We Don't Care & Co, 9 out J >of 10 academics surveyed prefer prefer unix/linux/Windows and didn't know >diddley about VMS.* >.J >*Results accurate 19 times out of 20, with a margin of error of +/- 4.5%.: >YMMV. Not available is stores. Operators are standing by. >  >r  G Again, I contend that any PhD in CS *ought* to be aware of almost everyrJ technology in their area of study. Many of the ones we studied back when IG was in college were not even commercial products, so "known" takes on ar much wider meaning.  .  I Imho, a CS course that studies clustering technology - or availability inlG general - that doesn't have some time for studying VMScluster should beu grounds for malpractice.   ------------------------------  * Date: Tue, 7 Oct 2003 18:37:06 +0000 (UTC)P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)X Subject: Re: worries about files opened by audit server and security server on  shutdown$ Message-ID: <blv14i$93f$1@online.de>  A In article <bls2f9$f1vms$1@ID-118202.news.uni-berlin.de>, "Martini& O'Connor" <moconnor@dvfs.com> writes:   Q > : > > I had the same problem with a user drive containing a page file which youD= > : > > cannot dismount since the page file is always opened.p > : >6$ > : > How did you solve the problem? > :  > :n. > : Was told to live with it :=) (not solved). > 9 > What about using the dismount option /policy=minicopy ?o > 
 > DISMOUNT > 	 > /POLICYu > 0 > /POLICY=[NO]MINICOPY[=(OPTIONAL)] (Alpha only)  H In my case, because it is ALPHA-only.  (Since I am still at 7.2-1 ALPHA I (hopefully upgrade to 7.3-1 in two weeks, I actually hadn't noticed this o yet.)u   ------------------------------   Date: 7 Oct 2003 15:52:43 -0700($ From: rcbryan@hotmail.com (RC Bryan)% Subject: Re: [5] Counter Googling :-( = Message-ID: <fbcf38dc.0310071452.4bf13f1a@posting.google.com>e   Bah!  A When you post stuff on the internet, newsgroups, etc. you have torF expect that it will last forever somewhere, if only on somebody's hardD disk.  Anything you ever say online can come back to haunt you.  YouF can assume that if you run for Governor of California, people will digD through their old news group postings and come up with something you@ wrote years and years ago.  This is a good reason to be somewhat
 anonymous.  D I did a Goggle search on RC Bryan and I found several of my posts onD the VMS newsgroup.  In the main area, there are too many Bryan SmithF and RC Jones listings to be able to connect with anything.  If you use@ my full name (which can be found by searching) you will find theB Geologist in Australia, the C&W musician in Texas and the ComputerF Scientist in Wisconsin and a whole bunch of other people with the same@ name.  Good luck making any use of it.  That is one benefit of a common name.   Regards,	 /RC Bryan   + PS, why am I responding to something so OT?h   "Bochnik, William J" <William_Bochnik@acml.com> wrote in message news:<2D75787AAF09C64481BDFD89113BE6D50199C07C@ac2kama0102.ac.lp.acml.com>...I > It is implied by posting that what you say could theoretically be savediL > forever. What gets me is that there is little proof of authenticity that IK > said what was attributed to me, either through the use of mail spoofs, ora > mis-quoting. >  > -----Original Message-----4 > From: JF Mezei [mailto:jfmezei.spamnot@istop.com] + > Sent: Wednesday, October 01, 2003 1:33 AMa > To: Info-VAX@Mvb.Saic.Com ' > Subject: Re: [5] Counter Googling :-(i >  >  > Russell Wallace wrote:J > > I disagree that it's an invasion of privacy. Google isn't like hiring D > > a detective - it can only find _public_ information. If there's J > > information about you that you regard as private, why did you make it  > > publicly available online? > H > When I post in newsgroups, I don't specifically agree that my posts beH > archived forever. When I posted prior to the formation of google (dejaE > news), I had no idea that my posts would be archived in the future.  > E > Ethically, when someone uses telephone books to find my address andrK > telephone number to contact me, it si proper use of that information. But K > when someone does the same and posts that information in newsgroups alonglL > with all sorts of fabulated stories, that is invasion of privacy. It isn'tI > so much the information, it so why you seek it and what you do with it.r > L > The problem is that inexperienced people will do a search an see all sortsI > of posts and not really ask to see full headers to try to see if it wasg > legitimate or not. > M > google is like a credit agency that lets anyone add any information to yourtF > file, except you have no control over it and it is archived forever. >.   ------------------------------   End of INFO-VAX 2003.557 ************************