1 INFO-VAX	Mon, 06 Aug 2001	Volume 2001 : Issue 433       Contents: Re: AlphaStation 200 fwupdate  Re: AS2000 or BA350 or BA356 Attaching a new DLT tape loader # Re: Attaching a new DLT tape loader $ Re: Compaq as a technology leader???+ Re: DECset enhancements for ODS-5 and Java? " Re: hardware or software mirroring% Re: IA64 volume and low-end dominance % Re: IA64 volume and low-end dominance @ Problems installing Hobbyist Version of Vax OpenVMS - Any Ideas? Problems with TCPIP 5.1 ECO 1  Re: SYS$LOGIN:[some.directory] Re: SYS$LOGIN:[some.directory] Re: SYS$LOGIN:[some.directory]@ Re: When/how to send answerback string during logon negotiations  F ----------------------------------------------------------------------  # Date: Sun, 05 Aug 2001 22:26:02 GMT 2 From: "John Fredrickson" <jafred@bellatlantic.net>& Subject: Re: AlphaStation 200 fwupdate9 Message-ID: <_Jjb7.7869$sN1.1687085@typhoon1.gnilink.net>    Guys,   I If any of you have an electronic document that describes the full list of E SRM commands, maybe you could post it here so we can all have a copy.    John Fredrickson  7 "Michael L. Umbricht" <mikeu@osfn.org> wrote in message " news:3B58A480.45B75B9C@osfn.org...E > I have updated the firmware on an AlphaStation 200 4/166 to install H > OpenVMS 7.2  (I believe that the 5.3 Update CD / V7.0 SRM is required)@ > After flashing the new firmware the help command displays info  > differently from the original. > 	 > Before:  > 	 > >>>help  > [start of help deleted...]5 > show cluster show config show map show memory sleep # > start stop true add_fax add_sound  > >>>  >  > After: > 	 > >>>help  > [...]  > show show show show show > show sleep start stop true > update > >>>  > F > Note that the "show" commands do not include the parameters (config," > etc.) and that add_* is missing. > ) > Is there a problem?  Or is this normal?  > G > BTW, I did read the release notes and follow the special instructions # > for updating from a "pre-v3.4 CD" < > The Update did not show any errors and Verify reported ok. >  > -mikeu >  >  > Original firmware: > ( > X5.0-29, built Jul 18 1995 at 16:20:48 > 
 > Firmware > SRM Console: X5.0-29 > ARC Console: 4.37 6 > PALcode: VMS PALcode X5.48-107, OSF PALcode X1.35-76 > Serial ROM: V4.6 > Diag ROM: V1.6 >  > New firmware:  > ' > V7.0-9, built Mar 18 1999 at 13:41:11  > 
 > Firmware > SRM Console: V7.0-9  > ARC Console: 4.58 2 > PALcode: VMS PALcode 5.56-2, OSF PALcode X1.46-2 > Serial ROM: V4.6 > Diag ROM: V1.6   ------------------------------  % Date: Sun, 05 Aug 2001 16:17:16 -0400 2 From: rdeininger@mindspring.com (Robert Deininger)% Subject: Re: AS2000 or BA350 or BA356 L Message-ID: <rdeininger-0508011617160001@user-2ive76m.dialup.mindspring.com>  = In article <5c8ffd05.0108031343.50411c71@posting.google.com>, 2 jason_odonnell@erinet.com (Jason O'Donnell) wrote:   D > I am slao not able to find on any documentation regarding BA350 or, > BA356 drive arrays on the Comapaq website.  I I can email a Ba350-related user's guide in PDF format as an attachment.  / It's about 340 kB.  Let me know if you want it.    --   Robert Deininger rdeininger@mindspring.com    ------------------------------  $ Date: Mon, 6 Aug 2001 10:59:09 +1200. From: Nivlesh Chandra <NChandra001@itc.gov.fj>( Subject: Attaching a new DLT tape loaderM Message-ID: <084681714A1BD511970B0002A560015F2D74F5@exchange01.govnet.gov.fj>   C I have a MicroVax 4000 system. It has a external single tape loader G atttached to it. We recently purchased a DLT Tape Loader and this is to I replace the single tape loader. My problem is that I do not know anything H about how to go about making the system use this new tape loader. I haveI attached the DLT tape loader to where the single tape loader was. The DLT K Tape loader is on SCSI 0 (that is what is shown at the back).. but I do not L know what it means and I do not know how to check if the system has detectedL that there is a DLT Tape loader attached to the system. Can someone point meK to approriate literature about attaching and making the system deetct a new : hardware ... or can someone help me go about doing this...  # Your help would be most appreciated    Niv    ------------------------------  % Date: Sun, 05 Aug 2001 20:01:14 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net> , Subject: Re: Attaching a new DLT tape loader' Message-ID: <3B6DEC5A.E252D594@fsi.net>    Nivlesh Chandra wrote: > E > I have a MicroVax 4000 system. It has a external single tape loader I > atttached to it. We recently purchased a DLT Tape Loader and this is to K > replace the single tape loader. My problem is that I do not know anything J > about how to go about making the system use this new tape loader. I haveK > attached the DLT tape loader to where the single tape loader was. The DLT M > Tape loader is on SCSI 0 (that is what is shown at the back).. but I do not N > know what it means and I do not know how to check if the system has detectedN > that there is a DLT Tape loader attached to the system. Can someone point meM > to approriate literature about attaching and making the system deetct a new < > hardware ... or can someone help me go about doing this... > % > Your help would be most appreciated    Um, WOW!  E Well, first you need to read up on SCSI id.'s and what they mean. Try 4 starting with http://www.computer-guru.com/scsi.html  F That aside, if you set your new "DLT loader" to SCSI Id 0, the same asF your old one was, the device(s) should (famous last words) at least be seen.   F From the info. provided, however, I cannot guess whether VMS will work4 with it. A make and model number would be a start...   --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------   Date: 5 Aug 2001 20:56:25 CDT = From: wayne@tachysoft.xxx.580099.killspam.0160 (Wayne Sewell) - Subject: Re: Compaq as a technology leader??? . Message-ID: <kczDxq7QkViX@tachxxsoftxxconsult>  X In article <3B6CE1BC.558AE4FF@prodigy.net>, cjt & trefoil <cheljuba@prodigy.net> writes: > Christof Brass wrote:  >>   > <snip> >>  = >> As to a historical overview of PC industry I recently read " >> Compaq introduced the Portable. > K > Perhaps in the Wintel world, but I think there were already Osbornes and  
 > Kaypros.    O This predates windows of the billy variety (and most other windowing systems as ) well).  You mean the *msdos*/intel world.      --  O =============================================================================== M Wayne Sewell, Tachyon Software Consulting  (281)812-0738  wayne@tachysoft.xxx : http://www.tachysoft.xxx/www/tachyon.html and wayne.html  K change .xxx to .com in addresses above, assuming you are not a spambot  :-) O =============================================================================== H Randolph Duke (in Trading Places): "Mother always said you were greedy.". Mortimer Duke: "She meant it as a compliment!"   ------------------------------   Date: 5 Aug 2001 20:10:13 -0500 9 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen) 4 Subject: Re: DECset enhancements for ODS-5 and Java?3 Message-ID: <HQ3cqBF8RCms@eisner.encompasserve.org>   c In article <XQ$Fu1k0jC5$@eisner.encompasserve.org>, koehler@encompasserve.org (Bob Koehler) writes:  > C >   The only enhancement I know of to DECset recently was the OSD-5 H >   support in CMS.  It's incomplete.  SHOW HISTORY can't find lowercase@ >   or mixed case elements unless the element name is in quotes. > H >   Recently we've been trying to write some MMS scripts to support JavaD >   work we've been doing.  Simply trying to fetch elements from CMSB >   has been problematic since MMS macros all expand in uppercase.C >   (Our solution takes about 12 lines to fetch and 11 to compile).  >  >   Enhancements would be nice.   7 I had a lot of problems using MMS with ODS-5 (no Java).    ------------------------------   Date: 5 Aug 2001 15:52:02 -0700 ( From: kparris@my-deja.com (Keith Parris)+ Subject: Re: hardware or software mirroring = Message-ID: <cb85fed2.0108051452.7f9ff0e6@posting.google.com>   Z "Kenneth" <best@hotmail.com> wrote in message news:<9jtci0$abo7@imsp212.netvigator.com>...N > I am now using software mirroring and want to study the gain for changing toB > controller base mirroring. When I check the usage of the processL > SHADOW_SERVER, it only use very little CPU and IO, is there anything I canA > check for the system overhead for using the software mirroring?   F You'll be hard-pressed to find measurable differences.  When you think@ about it, in the normal, steady-state case, the shadowing driverF really has very little extra work to do: On reads, there will be a fewF extra instructions to pick one shadowset member to read from, then theA read request is passed on through to the disk driver.  On writes, D SHDRIVER needs to create a IRP and queue a write request to the diskC driver for each shadowset member.  You're probably looking at maybe E 2-3% higher code-path length for a shadowed I/O versus a non-shadowed  I/O.  D During shadow merge and full-copy operations, you'll likely see some< noticeable activity in the SHADOW_SERVER process.  But those3 operations are relatively infrequent at most sites.   F You may have heard a statement along the lines of "Host-based RAID hasE high overhead".  If you're talking about RAID-5 and write operations, E the statement is true.  But for practical purposes, it really doesn't F apply to RAID-1 (shadowing or mirroring) or RAID-0 (striping), or even/ for read-only data on host-based RAID-5 arrays.   > One more point to consider in your decision-making: host-basedE shadowset members can be on different controller [pairs].  This means F the potential availability and performance for host-based shadowing is= actually arguably higher than for controller-based mirroring:   F Availability: Having shadowset members on different controller [pair]sF protects you against a failure of a controller [pair] or the inability9 to properly fail over within a dual-redundant pair.  With E controller-based mirroring, all mirrorset members must be on the same  controller [pair].  E Performance: Having shadowset members on different controller [pair]s D allows you to use the I/O capacity of multiple controller [pair]s atB once on a single shadowset unit.  With controller-based mirroring,E you're limited to the I/O capacity of a single controller.  [Not even @ a _pair_ of controllers -- since all I/Os to a mirrorset must goC through one or the other of the two controllers in a dual-redundant D pair, the other controller in the pair isn't of any use in improving" performance on a given mirrorset.]  : That said, controller-based mirroring is quite popular andE successfully used by many sites.  The ability to set up a spareset on ; a controller and have a mirror copy automatically made to a A replacement disk in the event of a failure is very handy; failure D detection and disk replacement are manual operations with host-based
 shadowing.C ------------------------------------------------------------------- C Keith Parris | parris at encompasserve dot org | VMS consulting on: C Clusters, Disaster Tolerance, Internals, Performance, Storage & I/O    ------------------------------  $ Date: Sun, 5 Aug 2001 19:38:54 -0400' From: "Bill Todd" <billtodd@foo.mv.com> . Subject: Re: IA64 volume and low-end dominance( Message-ID: <9kklf4$pdq$1@pyrite.mv.net>  9 "mulp" <michaelpettengill@earthlink.net> wrote in message = news:rH5b7.949$eU4.110194@newsread2.prod.itd.earthlink.net... 5 > "Peter da Silva" <peter@abbnm.com> wrote in message ! > news:9k1e5g$lic@web.nmti.com... H > > In article <q7w87.3248$0w3.428372@newsread2.prod.itd.earthlink.net>,1 > > mulp <michaelpettengill@earthlink.net> wrote:  > > A > > > What software exactly was it that drove the demand for 386?  > > F > > For the people I knew who made the jump early, it was ports of Mac
 > softwareJ > > that really needed large address spaces a-la the 68000 to run with any > kindL > > of decent performance. The first Windows versions on the 386 didn't give > you I > > any way to do that, but when Win32s came out, so you could run 32-bit 
 > softwareJ > > even on some older versions of Windows on the 386, it really took off. > I > Try again.  I bought a shackmate 420SX (a radioshack PC rebranded DECpc L > 420SX when they were going end of life and it came with Win 3.1.  That was a E > 20mHz 486SX.  I think that Win 3.11 was available around that time.  > / > Win95 didn't show up for another year or two.  > ) > Win32s certainly wasn't available then.   F Win32s definitely predated Win95 and was available for my Win 3.11 WFGL system.  Jeff Richter's book states that it was available for 'Windows 3.x',@ which suggests it worked with 3.1 and possibly even 3.0 as well.   - bill   ------------------------------  $ Date: Sun, 5 Aug 2001 19:47:18 -0400' From: "Bill Todd" <billtodd@foo.mv.com> . Subject: Re: IA64 volume and low-end dominance( Message-ID: <9kklum$psu$1@pyrite.mv.net>  ; "Kurt Shoens" <shoens@lenny.sfrn.dnai.com> wrote in message % news:9ki0rv$c88$1@bob.news.rcn.net... J > In article <9k5g1h$4fj$1@pyrite.mv.net>, Bill Todd <billtodd@foo.mv.com> wrote:L > (paraphrasing, I hope fairly) that locks need to be fast but that locatingK > pages in the buffer pool is already pretty expensive, so putting buffered 9 > pages in memory that's a system call away should be OK.   D Not quite:  the statement was that the overall single-buffer-relatedD overhead of a database operation requiring access to a buffer likelyH sufficiently eclipsed any additional overhead in mapping the buffer thatJ this added overhead was not a major problem.  Note that the total overheadD includes not only finding the buffer but finding whatever data is of= interest within the buffer, and then doing something with it.    > H > We'll have to agree to disagree on this point.  I can't see how to winJ > with the buffer pool not directly addressable.  Having to put things outE > of addressability shakes up the rest of the system so much that you I > have to consider it a separate part of the memory hierarchy, like DB2's  > use of extended storage.  K I did mention that frequently used items would indeed likely be cached at a G higher level (even with direct-mapped buffers, since targeting and then L finding stuff within a database page can be a bit expensive).  At a minimum,J I'd expect that the mapping window would act as at least a modest cache ofE still-mapped recently-touched buffers.  So you could consider this an I additional layer of the memory hierarchy, though I suspect that the 'item I cache' (record cache?) is of more importance than the map-window 'cache'.    - bill   ------------------------------  # Date: Sun, 05 Aug 2001 18:32:35 GMT $ From: family.entwisle@btinternet.comI Subject: Problems installing Hobbyist Version of Vax OpenVMS - Any Ideas? 2 Message-ID: <3b6d8fd3.1756057@news.btinternet.com>   Hi, ; 	I'm trying to install the Hobbyist version of OpenVMS on a  4000/60 Vax.  8 	My CD-ROM player didn't work (probably not 512bytes per> sector), so I copied it to a hard-drive.  I can boot from thisC harddrive, and install the save set onto the disk I want to install E to.  However, when I try to boot the second disk, I'm simply returned F to the chevrons after about 10 seconds, with just a few blank lines onE the screen.  If I do a SHOW DEV after this, the second hard drive has 2 disappeared from the list (it was there before :-)  = 	Has anyone got any idea what extremely dense thing I'm doing 	 wrong :-)   C P.S.  Just won a MicroVAX 3800 on ebay (I know, I probably could of E gotten it cheaper else where - but I'm too lazy to hunt around the UK F - sorry!) and I'm looking forward to playing with that, I just want toE pre-empt its arrival by getting the 4000/60 up and running.  The 3800 E - beautiful looking thing, I just want to wrap my arms around it - am  I sad?   Cheers,  Duncan.    ------------------------------  # Date: Mon, 06 Aug 2001 00:56:01 GMT ) From: rob.buxton@wcc.govt.nz (Rob Buxton) & Subject: Problems with TCPIP 5.1 ECO 10 Message-ID: <3b6de8d1.17516657@news.wcc.govt.nz>   Hi All,   ; Over the weekend I applied a number of patches to our mixedn6 architecture cluster. Amongst them TCPIP ECO 1 for 5.1  F All went well except for one of the Alphas. On reboot this hung during the start of TCPIP.fE A recent post here had already seen a similar problem with Proxies sot2 I started there and deleted some obsolete proxies.E TCPIP now started okay and all looked okay. But, a number of servicesoA just were not working, despite looking okay from the show servicem command.  D Telnet connections to the Ethernet port did not work, but did to theD FDDI. Telnet connections out did not work. But the port looked alive' and well on the show interface command.o   Set Host had no such problems.  D I finally blew away all of the config files and recreated everythingE from scratch. Luckily it's a fairly basic system. All then started toe work okay (without a reboot!) D Not sure what's in the ECO 1 kit that can cause it to get upset withD what's in the config files.  But this is just a warning to folk thatF are upgrading a system that has old config files, ours were pre V5 and) have been converted / added to each time.B   Rob.   ------------------------------  $ Date: Mon, 6 Aug 2001 00:46:34 +01001 From: "Chris Townley" <news@townleyc.demon.co.uk>h' Subject: Re: SYS$LOGIN:[some.directory]oA Message-ID: <997055249.27520.0.nnrp-01.d4e45fa5@news.demon.co.uk>r  < "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message! news:3B69FF65.63DAA44B@fsi.net...,( > paddy.o'brien@zzz.tg.nsw.gov.au wrote: > >- > > JF,  > >nK > > >I beleive that Unix has a construct which uses the tilde ~ to point tor the J > > >login directory so that something such as ~/some/directory/myfile.txt wouldr8 > > >work no matter what the currect directory would be. > > >a1 > > >Shouldn't VMS have some way to do the same ?  > > >sD > > >I realise that one can define a rooted logical that would allowH > > >MY$LOGIN:[some.directory] but it would be nice if it were to become	 somethinga > > >that is standard. > >tK > > I must have assuredly misunderstood you, because SYS$LOGIN seems simpler enough > > to meet your requirement.e >e, > I think what he wants is to (for example): >u, > $ COPY ~SYSTEM:[STARTUP]MOUNT_DISKS.COM [] >p4 > ...while logged in as some user other than SYSTEM. >eJ > That said, if tilde were a valid character in a logical name, this wouldE > be easy enough to simulate using rooted logical names, with certaint > limitations. >n  L You can of course use sys$login_device:[some.directory] although you must go back to the root.i   --
 Chris Townleyp chris@townleyc.demon.co.uk townleyc@spicers.ltd.uki   ------------------------------   Date: 5 Aug 2001 19:50:18 -0500 9 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)t' Subject: Re: SYS$LOGIN:[some.directory]l3 Message-ID: <bM07oD3wRdIb@eisner.encompasserve.org>u   In article <rdeininger-0308011546460001@user-2ivec79.dialup.mindspring.com>, rdeininger@mindspring.com (Robert Deininger) writes:-F > In article <Jzza7.579$Yx2.11377@news.cpqcorp.net>, "Fred Kleinsorge"& > <kleinsorge@star.zko.dec.com> wrote: > K >> Well, when the COE stuff makes it's way into the mainstream, you will be-G >> able to set up logicals so that the Motif/CDE stuff can present UNIXBH >> pathnames, and will gladly accept the tilde as in ~/subdir/myfile.dat >> eK >> But I think it is a resonable suggestion that it would be nice to have alM >> rooted logical equivalent to SYS$LOGIN created (automatically) so that youi6 >> can have something like SYS$HOME:[SUBDIR]MYFILE.DAT >> sO >> However, it is also something that you can do today with a minimal amount ofe >> DCL in your login.com > I > A VMS-approved standard name would be available to software developers,f: > without having to depend on non-uniform local standards.  A Software developers should not enforce their own directory namingcA methods on customers.  But if they insist, they can construct the64 proper name from the equivalence name for SYS$LOGIN.   ------------------------------  * Date: Mon, 6 Aug 2001 03:03:34 +0000 (UTC)& From: dsf@frontiernet.net (Dan Foster)' Subject: Re: SYS$LOGIN:[some.directory] 5 Message-ID: <9kl1e6$35rg$1@node21.cwnet.roc.gblx.net>o  3 In article <yXzeDKk9YZQp@eisner.encompasserve.org>, . Bob Koehler <koehler@encompasserve.org> wrote:6 >In article <3B68FBC1.9F3FBFEB@videotron.ca>, JF Mezei' ><jfmezei.spamnot@videotron.ca> writes:nM >> I beleive that Unix has a construct which uses the tilde ~ to point to theyN >> login directory so that something such as ~/some/directory/myfile.txt would6 >> work no matter what the currect directory would be. >> m0 >> Shouldn't VMS have some way to do the same ?  >>  B >> I realise that one can define a rooted logical that would allowP >> MY$LOGIN:[some.directory] but it would be nice if it were to become something >> that is standard. >lB >Why?  It's not standard on UNIX.  (Try it on Sun's Bourne shell).  E The reason why Sun's /sbin/sh shell doesn't do it is because it can'toE assume an environment where dynamic loading libraries will work. (Eg,rF in single user mode in an emergency/bad situation.) I don't recall offL the top of my head which library it wants to resolve ~ to an home directory.  G My guess is it wants a library for that nameservice resolution functioniE because that information can potentially be contained in a variety ofh: namespaces, and not just /etc/passwd. NIS, DCE, LDAP, etc.  G I'm afraid to ask how much bigger a statically linked /sbin/sh with all  of those functions would be :-)e  I Other folks in this thread is correct: it's never been an actual standardaM per se for UNIX - just a common convention amongst many (but not all) shells.dG Very handy, for sure. Device-independent common and generic view of therK filesystem guaranteed to work at all times on a wide variety of [supported] 8 platforms and [supported] shells is quite indispensable.  G With that said, it isn't really a big deal with VMS if some planning istJ done with logicals and communicated amongst the system/site's users. Can'tH disagree with Mr. Mezei, though - would be nice if there was some commonG standard for all VMS sites - an easy way to create one is for Compaq to / introduce something in a future version of VMS.e   -Dan   ------------------------------   Date: 5 Aug 2001 19:54:05 -0500n9 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)aI Subject: Re: When/how to send answerback string during logon negotiations-3 Message-ID: <ID$Fi+jzloWa@eisner.encompasserve.org>i  X In article <qZHa7.140$sg.65523@nnrp3.sbc.net>, "Glenn Wiens" <gwiens@swbell.net> writes:M > What I see when I make a socket connection on port 23 to my VMS host is the-( > following sequence of bytes (decimal): >  > 027[c27\27Z27[0c > M > and, from what I can figure out, these are requests to send a terminal typexD > back to the host. However, I have not seen additional requests forH > information, and I wonder if, when I send my terminal information backG > (something like 27[?1;2c for a VT100), I also need to send answerbackrM > information. If so, I don't know the format which the host expects from thed > terminal.s  I VMS itself expects nothing.  Typically you see such a string only becauseeJ someone configuring your site has issued the command SET TERMINAL/INQUIRE., If it is your VMS host, you can remove that.   ------------------------------   End of INFO-VAX 2001.433 ************************