1 INFO-VAX	Sun, 01 Oct 2006	Volume 2006 : Issue 539       Contents:  Re: 2006-09 Distrbution CD-ROMS?) Re: adding a disk and putting it "online"   Re: BACKUP locks up a directory?  Re: BACKUP locks up a directory?  Re: BACKUP locks up a directory?  Re: BACKUP locks up a directory?  Re: BACKUP locks up a directory?  Re: BACKUP locks up a directory?  Re: BACKUP locks up a directory?  Re: BACKUP locks up a directory?> Can a switch give me more speed (long and detailed questions)?B Re: Can a switch give me more speed (long and detailed questions)?B Re: Can a switch give me more speed (long and detailed questions)?B Re: Can a switch give me more speed (long and detailed questions)?1 Cluster-wide logicals: multi-architecture cluster 5 Re: Cluster-wide logicals: multi-architecture cluster 5 Re: Cluster-wide logicals: multi-architecture cluster 5 Re: Cluster-wide logicals: multi-architecture cluster 5 Re: Cluster-wide logicals: multi-architecture cluster # Directory list collating sequence ? ' Re: Directory list collating sequence ? 
 Re: ES40 NICs 
 Re: ES40 NICs ' external CD drives on kzpba-ca on ds10l + Re: external CD drives on kzpba-ca on ds10l + Re: external CD drives on kzpba-ca on ds10l + Re: external CD drives on kzpba-ca on ds10l  Re: Graphic options for DS10L  Re: HP could learn from Island$ Re: IDE (DQ) devices served to VAX ? MntVerifyTimeout Re: MntVerifyTimeout Re: MntVerifyTimeout" Re: Normal temperature for DS10L ?) Re: SPAWN numbering of subprocesses (8.2) ) Re: SPAWN numbering of subprocesses (8.2) 8 Where will the FAQ be, and what about the Freeware CD's?  F ----------------------------------------------------------------------  $ Date: Sun, 1 Oct 2006 08:27:49 -0400) From: "Neil Rieck" <n.rieck@sympatico.ca> ) Subject: Re: 2006-09 Distrbution CD-ROMS? < Message-ID: <451fb269$0$11638$9a6e19ea@news.newshosting.com>  5 "DaveG" <david.gudewicz@abbott.com> wrote in message  = news:1159659813.566899.316680@c28g2000cwb.googlegroups.com...  >  > Neil Rieck wrote: I >> Question to people with software support contracts. Have you received   >> your  >> fall kit yet? >>
 >> Neil Rieck   >> Kitchener/Waterloo/Cambridge, >> Ontario, Canada. $ >> http://www3.sympatico.ca/n.rieck/ > . > We received our 2006 Q3 Alpha kit last week. > 	 > Dave...  > # Thanks. Maybe ours are in the mail.   
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.8 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html: http://www3.sympatico.ca/n.rieck/links/openvms_demos.html    ------------------------------  * Date: Sun, 1 Oct 2006 14:15:47 +0000 (UTC)7 From: moroney@world.std.spaamtrap.com (Michael Moroney) 2 Subject: Re: adding a disk and putting it "online"( Message-ID: <efoiij$njb$1@pcls4.std.com>  0 Bob Blunt <RobertDOTblunt@digitalDOTcom> writes:  J >console.  If you toss a new SCSI drive into a VAX without power cycling, G >you typically have to go thru a cycle of HALT, UNJAM and INIT for the  H >device to show up properly to the console before it shows up with SHOW E >DEVICE.  In my experience the same has held true for most all Alpha  - >consoles with direct-connected SCSI devices.   G That hasn't been my experience at all.  The console appears to poll the I SCSI bus once every couple of minutes.  If I add a SCSI drive to an alpha D sitting at the >>> prompt it appears but not immediately.  Same if IG remove one, it disappears but not immediately.  Also if you have a SCSI E cluster with one system on the shared bus booted and the other at the F console prompt, you'll see periodic mount verifys, apparently when theE system at the console prompt polls the SCSI bus and does something to  upset the VMS system.   D Of course this probably depends on what model Alpha and the console  firmware version.    ------------------------------  % Date: Sun, 01 Oct 2006 11:40:34 +0200 / From: Paul Sture <paul.sture.nospam@hispeed.ch> ) Subject: Re: BACKUP locks up a directory? J Message-ID: <paul.sture.nospam-40323B.11403401102006@mac.sture.homeip.net>  3 In article <MMGNHPT3nwGw@eisner.encompasserve.org>, /  Kilgallen@SpamCop.net (Larry Kilgallen) wrote:   J > In article <M0zTg.64023$wg.23842@fe1.news.blueyonder.co.uk>, Alan Greig ! > <greigaln@netscape.net> writes:  > > AEF wrote: > >>H > >> Is this normal BACKUP behavior? The backup takes so long that it isL > >> difficult to avoid the EOD job. I've done this before several times andI > >> never had this problem. I assume I am a victim of bad timing. But is K > >> this normal, that a "hot backup" can lock up directories and/or files? & > >> Is there a better way to do this? > > L > > IIRC, if you  $ BACKUP/IGNORE=INTERLOCK then BACKUP should not lock any  > > files as it backs them up. > F > Whereas my recollection is that, at least at some point in the past,E > /IGNORE=INTERLOCK did not affect how BACKUP tried to access a file, F > it just caused BACKUP to try a less contentious mode if the file wasC > already opened by another application.  If BACKUP got to the file 3 > first, the other application might have problems.   F I have a nagging recollection that /FAST alters the locking behaviour  here.    Does that jog anyone's memory?   --  
 Paul Sture   ------------------------------   Date: 1 Oct 2006 07:04:02 -0500 - From: Kilgallen@SpamCop.net (Larry Kilgallen) ) Subject: Re: BACKUP locks up a directory? 3 Message-ID: <QRZ4m9sVvjxE@eisner.encompasserve.org>   | In article <paul.sture.nospam-40323B.11403401102006@mac.sture.homeip.net>, Paul Sture <paul.sture.nospam@hispeed.ch> writes:5 > In article <MMGNHPT3nwGw@eisner.encompasserve.org>, 1 >  Kilgallen@SpamCop.net (Larry Kilgallen) wrote:  > K >> In article <M0zTg.64023$wg.23842@fe1.news.blueyonder.co.uk>, Alan Greig  " >> <greigaln@netscape.net> writes: >> > AEF wrote:  >> >> I >> >> Is this normal BACKUP behavior? The backup takes so long that it is M >> >> difficult to avoid the EOD job. I've done this before several times and J >> >> never had this problem. I assume I am a victim of bad timing. But isL >> >> this normal, that a "hot backup" can lock up directories and/or files?' >> >> Is there a better way to do this?  >> >  M >> > IIRC, if you  $ BACKUP/IGNORE=INTERLOCK then BACKUP should not lock any   >> > files as it backs them up.  >>  G >> Whereas my recollection is that, at least at some point in the past, F >> /IGNORE=INTERLOCK did not affect how BACKUP tried to access a file,G >> it just caused BACKUP to try a less contentious mode if the file was D >> already opened by another application.  If BACKUP got to the file4 >> first, the other application might have problems. > H > I have a nagging recollection that /FAST alters the locking behaviour  > here.   J Well certainly /FAST causes backup of directories to be just an incidentalI backup of files rather than something controlling the rest of the backup.  --  N ==============================================================================0 DoD Instruction 8500.2 field test sites wanted -- 	http://www.LJK.com/LJK/8500_2_fieldtest.html N ==============================================================================   ------------------------------  # Date: Sun, 01 Oct 2006 12:32:25 GMT ( From: Alan Greig <greigaln@netscape.net>) Subject: Re: BACKUP locks up a directory? ; Message-ID: <tzOTg.67079$wg.7121@fe1.news.blueyonder.co.uk>    Larry Kilgallen wrote:   > F > Whereas my recollection is that, at least at some point in the past,E > /IGNORE=INTERLOCK did not affect how BACKUP tried to access a file, F > it just caused BACKUP to try a less contentious mode if the file wasC > already opened by another application.  If BACKUP got to the file 3 > first, the other application might have problems.   E I am 99.99% certain that /IGNORE=INTERLOCK will cause BACKUP to open  H files for non-exclusive access. It doesn't just alter it's behaviour on  finding an already open file.   E It's easily enough tested. Just backup a large file with and without  G /IGNORE=INTERLOCK and try opening it for write access while it's being  F backed up. I hit this problem myself once and solved it (at least for E the purpose I needed) with /IGNORE=INTERLOCK. That's at least as far  G back as VMS 5.x as I recall. I did try the simple test I suggest above  E at the time and unless my memory is really faulty, /IGNORE=INTERLOCK   works as I suggest.    --  
 Alan Greig   ------------------------------  % Date: Sun, 01 Oct 2006 09:30:02 -0400 3 From: "Richard B. Gilbert" <rgilbert88@comcast.net> ) Subject: Re: BACKUP locks up a directory? : Message-ID: <FoudnVHHcP1BX4LYnZ2dnUVZ_qmdnZ2d@comcast.com>   Larry Kilgallen wrote:  i > In article <M0zTg.64023$wg.23842@fe1.news.blueyonder.co.uk>, Alan Greig <greigaln@netscape.net> writes:  >  >>AEF wrote: >>F >>>Is this normal BACKUP behavior? The backup takes so long that it isJ >>>difficult to avoid the EOD job. I've done this before several times andG >>>never had this problem. I assume I am a victim of bad timing. But is I >>>this normal, that a "hot backup" can lock up directories and/or files? $ >>>Is there a better way to do this? >>J >>IIRC, if you  $ BACKUP/IGNORE=INTERLOCK then BACKUP should not lock any  >>files as it backs them up. >  > F > Whereas my recollection is that, at least at some point in the past,E > /IGNORE=INTERLOCK did not affect how BACKUP tried to access a file, F > it just caused BACKUP to try a less contentious mode if the file wasC > already opened by another application.  If BACKUP got to the file 3 > first, the other application might have problems.   H This thread is the first suggestion I've heard that BACKUP locked files D it was backing up.  My understanding of /IGNORE=INTERLOCK has been, G since 1984, that backup does a "best efforts" backup of files that are  I opened for write access.  It was intended, I believe, for files that the  E   O/S always has open which could not otherwise be backed up without  F shutting down the system and doing a standalone backup.  A standalone E backup is unquestionably the best way to back up the system disk but   it's seldom practical!   ------------------------------   Date: 1 Oct 2006 07:46:22 -0700 $ From: "AEF" <spamsink2001@yahoo.com>) Subject: Re: BACKUP locks up a directory? C Message-ID: <1159713982.457459.178820@k70g2000cwa.googlegroups.com>    Larry Kilgallen wrote:~ > In article <paul.sture.nospam-40323B.11403401102006@mac.sture.homeip.net>, Paul Sture <paul.sture.nospam@hispeed.ch> writes:7 > > In article <MMGNHPT3nwGw@eisner.encompasserve.org>, 3 > >  Kilgallen@SpamCop.net (Larry Kilgallen) wrote:  > > L > >> In article <M0zTg.64023$wg.23842@fe1.news.blueyonder.co.uk>, Alan Greig$ > >> <greigaln@netscape.net> writes: > >> > AEF wrote:  > >> >> K > >> >> Is this normal BACKUP behavior? The backup takes so long that it is O > >> >> difficult to avoid the EOD job. I've done this before several times and L > >> >> never had this problem. I assume I am a victim of bad timing. But isN > >> >> this normal, that a "hot backup" can lock up directories and/or files?) > >> >> Is there a better way to do this?  > >> >N > >> > IIRC, if you  $ BACKUP/IGNORE=INTERLOCK then BACKUP should not lock any! > >> > files as it backs them up.  > >>I > >> Whereas my recollection is that, at least at some point in the past, H > >> /IGNORE=INTERLOCK did not affect how BACKUP tried to access a file,I > >> it just caused BACKUP to try a less contentious mode if the file was F > >> already opened by another application.  If BACKUP got to the file6 > >> first, the other application might have problems. > > I > > I have a nagging recollection that /FAST alters the locking behaviour 	 > > here.  > L > Well certainly /FAST causes backup of directories to be just an incidentalK > backup of files rather than something controlling the rest of the backup.   & You lost me here. What are you saying?  F The /FAST qualifier causes BACKUP to do a "pre-scan" of INDEXF.SYS andE applies the file-selection criteria at that point to speed things up. D First, BACKUP/FAST makes its own list of all the files. Then it goesB through the list, marking each file that passes the file-selectionE criteria specified in the BACKUP command. Then it walks the directory D structure in BACKUP order and instead of checking INDEXF.SYS if eachA file needs to be backed up, it checks the list it made during the A pre-scan. This speeds up the backup considerably. At one site, my E adding the /FAST qualifier to the incremental backup reduced the time C from about 45 minutes to about 15 minutes! At another, my doing the F same thing blew away the operators! They asked if I could speed up the@ full backups using the same trick! :-) I just said, "Sorry, ..."   AEF    > --P > ==============================================================================2 > DoD Instruction 8500.2 field test sites wanted -/ > 	http://www.LJK.com/LJK/8500_2_fieldtest.html P > ==============================================================================   ------------------------------  # Date: Sun, 01 Oct 2006 15:11:33 GMT ( From: Alan Greig <greigaln@netscape.net>) Subject: Re: BACKUP locks up a directory? = Message-ID: <FUQTg.56127$aP3.37916@fe3.news.blueyonder.co.uk>    Richard B. Gilbert wrote:   J > This thread is the first suggestion I've heard that BACKUP locked files F > it was backing up.  My understanding of /IGNORE=INTERLOCK has been,   F If BACKUP (without /IGNORE=INTERLOCK) did not lock files as it backed H them up then you could get silently inconsistent backups, BACKUP starts G on a file, someone else then opens it and partially completes a set of  E writes to it before BACKUP has closed the file, Result: backup would  / save a file which you think is clean but isn't.   G However I am claiming that with /IGNORE=INTERLOCK BACKUP does not lock  2 the file and you get the "best effort" hot backup.  I > since 1984, that backup does a "best efforts" backup of files that are  K > opened for write access.  It was intended, I believe, for files that the  F >  O/S always has open which could not otherwise be backed up without H > shutting down the system and doing a standalone backup.  A standalone G > backup is unquestionably the best way to back up the system disk but   > it's seldom practical! >      --  
 Alan Greig   ------------------------------   Date: 1 Oct 2006 10:26:06 -0500 - From: Kilgallen@SpamCop.net (Larry Kilgallen) ) Subject: Re: BACKUP locks up a directory? 3 Message-ID: <Q9Zp0J3V8UL8@eisner.encompasserve.org>   p In article <FoudnVHHcP1BX4LYnZ2dnUVZ_qmdnZ2d@comcast.com>, "Richard B. Gilbert" <rgilbert88@comcast.net> writes: > Larry Kilgallen wrote:  G >> Whereas my recollection is that, at least at some point in the past, F >> /IGNORE=INTERLOCK did not affect how BACKUP tried to access a file,G >> it just caused BACKUP to try a less contentious mode if the file was D >> already opened by another application.  If BACKUP got to the file4 >> first, the other application might have problems. > J > This thread is the first suggestion I've heard that BACKUP locked files F > it was backing up.  My understanding of /IGNORE=INTERLOCK has been, I > since 1984, that backup does a "best efforts" backup of files that are   > opened for write access.  G And my recollection of situations I have encountered with at least some E versions of VMS is that (to my surprise) the /IGNORE=INTERLOCK choice D did _not_ mean there would be no attempt to lock access.  Instead itE just meant (at that time) an attempt to lock access that failed would I be followed by an attempt to backup that particular file without locking.   C At least in my experience, /IGNORE=INTERLOCK does not make messages A about files being locked by another user go away, it just changes ? them from Error to Warning.  It seems extremely unlikely if the @ strategy were to eschew all locking under /IGNORE=INTERLOCK thatB BACKUP would go to the trouble of making a test as to whether some@ process had the file locked just for the purpose of issueing the warning message.  B Of course anyone with a VMS source listings kit who disagrees with* my reasoning is free to check the details. --  N ==============================================================================0 DoD Instruction 8500.2 field test sites wanted -- 	http://www.LJK.com/LJK/8500_2_fieldtest.html N ==============================================================================   ------------------------------   Date: 1 Oct 2006 09:46:08 -0700 $ From: "AEF" <spamsink2001@yahoo.com>) Subject: Re: BACKUP locks up a directory? C Message-ID: <1159721168.410770.236390@m73g2000cwd.googlegroups.com>    Richard B. Gilbert wrote:  > AEF wrote:K > > I was doing an image backup of an 18-GB disk. During this backup one of K > > my market servers was running the end-of-day job. At the end of the job I > > it copied certain files created the same day to another VAX system. A I > > total of 24 files are normally copied. These range in size from a few G > > blocks to a couple thousand. Anyway, the first 15 files were copied G > > okay but the next 15 failed with the following error (actual sample  > > from the log file):  > > , > > %COPY-E-OPENOUT, error opening NODE16"FTG > > password"::FT_STORAGE_DISK:[FT.ARC.EQUITIES]FTOATTRN060929.SEQ;0 as 
 > > output5 > > -RMS-E-FLK, file currently locked by another user E > > %COPY-W-NOTCOPIED, _NODE2$DKA400:[FT.DAT]FTOATTRN060929.SEQ;1 not 
 > > copied > > 6 > > I assume that the "other user" was the backup job. > >  > > The DCL command was  > > C > > $ COPY/LOG FTDAT:FT*060929*.*;0/EXCLUDE=(FTGVT*.DAT,FTSYS*.DAT) ? > > NODE16"FT XXXXXXXX"::FT_STORAGE_DISK:[FT.ARC.EQUITIES]*.*;0  > > K > > So I got 15 success messages, followed by 9 error messages like the one + > > above, followed by the summary message.  > > I > > I aborted the backup job at this point and re-ran the EOD job. It ran G > > fine and copied all 24 files successfully. I looked at the tape and G > > sure enough, the backup job had just started copying files from the   > > [FT.ARC.EQUITIES] directory. > > G > > Some more possibly relevant information: NODE2 is running VMS v6.1. H > > NODE16 is running VMS v6.2. The directory [FT.ARC.EQUITIES] containsD > > about 24000 files. The EQUITIES.DIR file is 2751 blocks in size. > > J > > NODE2 is a MicroVAX 3100 Model 95. NODE16 is a MicroVAX 3100 Model 80. > > G > > Is this normal BACKUP behavior? The backup takes so long that it is K > > difficult to avoid the EOD job. I've done this before several times and H > > never had this problem. I assume I am a victim of bad timing. But isJ > > this normal, that a "hot backup" can lock up directories and/or files?% > > Is there a better way to do this?  > >  > > Thanks.  > >  > > AEF  > >  > J > With a directory 2751 blocks in size, it's no wonder things are a littleI > slow!  Later versions of VMS are not quite so sensitive but at VMS V6.2   E The backup takes a long time partly because of the sheer number files . and partly because I'm stuck with narrow SCSI.  I > any file added to or deleted from an oversize directory could create an A > incredible amount of overhead.  Directory entries are stored in E > alphanumeric order.  When an entry is inserted in or deleted from a J > directory, blocks get shuffled up or down to make room or reclaim space.A > It's no big deal if you keep your directories under 100 blocks.  > B > If you really need to have all 24,000 files on line, I'd suggestJ > spreading them over a LOT more directories.  Maybe a directory for 2006,F > one for 2005, 2004, . . . or a directory for each month or whatever.  D I don't think distributing the files among multiple directories willC speed up the backup job, but it might make the problem at hand less  likely to recur.       >----o----<   E OK, could it be that the backup job had just started reading the file F [FT.ARC]EQUITIES.DIR;1 between the copying of 15th and 16th files fromC the remote copy job? If so, the error message is a bit misleading.     AEF    ------------------------------  * Date: Sun, 1 Oct 2006 13:53:47 +0000 (UTC)P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)G Subject: Can a switch give me more speed (long and detailed questions)? $ Message-ID: <efoh9b$str$2@online.de>  D At present, my cluster is set up like this.  I've included ONLY the H information I think is relevant to my question.  If I've left something  out, let me know.   C    o  2 VAXes (VAX 4000/105A and VAXstation 4000/90A) as boot nodes   *    o  1 ALPHA (DEC 3000/600) as boot node   E    o  1 ALPHA (ALPHAserver 1200, actually "DIGITAL Server 5000 Model  2                5305 6533A 5/533 4MB") as satellite  D    o  main user disk is a pair of 9-GB SEAGATE SX910800N disks, one $       connected to each of the VAXes  D    o  interconnect for all traffic is the twisted-pair LAN; hub is a7       NETGEAR EN108TP, which is an 8-port 10base-T hub)   ?    o  also on the HUB is a very modern laptop (not running VMS)   D    o  the uplink from the hub goes to a LINKSYS BEFSR41 which as farD       as I know has the LAN ports connected via a SWITCH (also goingG       into the LinkSys is a VOIP adapter---it can also connect directly F       to my ISP by PPPoE, i.e. I could get by without the LinkSys, butA       at present it is "behind" the LinkSys just like the hub is)   F Since my DSL speed is about to increase from 6000 kb/s to 16,000 kb/s H downstream (present upstream is, I believe, 512 kb/s; I'm not sure what A it will be after the upgrade, but that's not relevant here), I'm  I considering to what extent replacing my hub with a switch would speed up  F things, whereby there are two distinct questions (both related to the 5 two things where I would actually notice more speed):   @    o  Could a switch speed up shadow copies?  Since the VAX SCSIE       speed is 5 Mb/s, if I've done the math right a transfer should  F       take 4 hours (9*1000*8/5/60/60).  Actually, it takes about twiceD       as long.  On the other hand, presumably this is correct since C       first the source is read then the copy is written.  Thus, the F       bottleneck is probably the VAX SCSI speed and, though there are H       some collisions during a shadow copy, they are probably negligibleC       as far as their effect on the speed goes.  Thus, I think the  E       answer here is "no".  (The SEAGATE disks themselves I think are G       capable of at least 10 MB/s, but the only faster SCSI connections E       I have are on ALPHA.  If I had two ALPHAs in the cluster, this  D       might give me some more speed, if I connected the SEAGATEs to D       them.  On the other hand, thanks to MINICOPY a full copy only F       takes place rarely, for example after an unplanned power outage.D       Since I want to be able to do a planned reboot of either host H       node without a full copy, if I connect one SEAGATE to an ALPHA, I F       would have to have another ALPHA in the cluster for MINICOPY to        work.)  D    o  My ISP has a "test your speed" link (which anyone here can try@       out): http://www.1und1.de/index.php?page=speedtest .  (It I       essentially has a lot of garbage as an HTML comment then some code  I       to check the download speed.  I think it is OK as far as reporting  C       the speed, but the graphical scale is non-linear.  Trust the  F       numbers, not the speed-bar graphics.)  The following numbers areE       reproducible: MOZILLA on the satellite: 1000-1100 kb/s; LYNX on D       the satellite: 3000-4000 kb/s; laptop: 5000-6000 kb/s; either F       VAX: 1400-1500 kb/s; LYNX on the 3000/600: 3300--3500 kb/s.  TheE       speed indicated by LYNX agrees with the speed indicated by the  D       web page.  At first, I thought the laptop was getting the fullG       speed (it probably is) and Mozilla on the satellite was suffering G       by having to write the file not to a local disk but to the shadow C       set on the VAXes, thus potentially suffering from a) the slow D       SCSI speed, b) the fact that it has to write to both disks, c)G       the fact that the write is across the network and d) the slow and H       as far as I know non-duplex ethernet cards on the VAXes.  However,D       LYNX on the same machine as Mozilla is quite a bit faster.  ItG       could be that the speedup comes from ignoring the images, though. D       OK, this is just about plausible.  If the ethernet card in theD       3000/600 is better than in the VAXes, then I could see why theF       former speed is better than the latter, and if the ethernet cardB       in the satellite is better than in the 3000/600, that would H       explain why the satellite speed (with LYNX) is best of all (exceptG       for the laptop, which has the best ethernet card of all and saves E       the file to a local (very fast) disk).  Based on these numbers, E       how should they change under the following three scenarios: a)  E       switch instead of hub, b) 16,000 kb/s DSL (this might depend on F       whether the WAN speed of the router is limited to, say 10 Mb/s) E       and c) both a) and b).  The "problem" is the difference between G       the laptop, which essentially gets full speed, and Mozilla, which H       is only getting about 1/6 the full speed.  Obviously, I would likeH       to get the full speed not only now but also after moving to 16,000B       kb/s (assuming the LinkSys is not the bottleneck then) from G       Mozilla on the satellite, since this is the one application where H       I really would notice more speed on a regular basis.  The questionG       is, of all the things slowing it down (a) to d) above), which, if G       any, could be improved by a switch (probably only d) and how much D       would that improvement be?  Perhaps folks can report the speed>       indicated by the web page above, note their cluster (or A       single-machine) setup and see if the numbers are plausible.   F Note that the first question doesn't have anything to do with the DSL H connection.  Rather, I'm wondering whether, if the answer to the second D question is "yes" and I thus buy a switch, a side-effect would be a , speedup as described in the first question.   G Bottom line: is it worth it to me to replace the hub in my cluster with G a switch, or are there so many other bottlenecks that I wouldn't notice B it?  Note that the laptop is on the HUB and gets (almost) the full speed.     ------------------------------  * Date: Sun, 1 Oct 2006 14:42:23 +0000 (UTC)P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)K Subject: Re: Can a switch give me more speed (long and detailed questions)? $ Message-ID: <efok4f$7ji$1@online.de>  D In article <efoh9b$str$2@online.de>, helbig@astro.multiCLOTHESvax.de3 (Phillip Helbig---remove CLOTHES to reply) writes:    I > Bottom line: is it worth it to me to replace the hub in my cluster with I > a switch, or are there so many other bottlenecks that I wouldn't notice D > it?  Note that the laptop is on the HUB and gets (almost) the full	 > speed.    B Let me add that since the example of a shadow copy shows that the A bottleneck is probably the SCSI speed of the VAXes, I would have  F expected the download speed to be equal to the DSL speed up to 5 Mb/s G and constant at 5 Mb/s after that.  Maybe I could accept a hand-waving  E argument for decreasing it by a factor of 2, but a factor of 6 seems   rather extreme.    ------------------------------  * Date: Sun, 1 Oct 2006 17:02:23 +0000 (UTC)P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)K Subject: Re: Can a switch give me more speed (long and detailed questions)? $ Message-ID: <efosav$uc2$1@online.de>  D In article <efok4f$7ji$1@online.de>, helbig@astro.multiCLOTHESvax.de3 (Phillip Helbig---remove CLOTHES to reply) writes:    F > In article <efoh9b$str$2@online.de>, helbig@astro.multiCLOTHESvax.de5 > (Phillip Helbig---remove CLOTHES to reply) writes:   > K > > Bottom line: is it worth it to me to replace the hub in my cluster with K > > a switch, or are there so many other bottlenecks that I wouldn't notice F > > it?  Note that the laptop is on the HUB and gets (almost) the full > > speed.   > D > Let me add that since the example of a shadow copy shows that the C > bottleneck is probably the SCSI speed of the VAXes, I would have  H > expected the download speed to be equal to the DSL speed up to 5 Mb/s I > and constant at 5 Mb/s after that.  Maybe I could accept a hand-waving  G > argument for decreasing it by a factor of 2, but a factor of 6 seems   > rather extreme.   H I did some more tests with FTP.  Going out, I'm limited by the 512 kb/s F DSL connection, and get essentially the full bandwidth.  Incoming, it  looks like this:  F FTP started from remote connection, local connection is 3000/600 (has  cluster alias)  # disk          get               put / -----------   --------------    --------------- 1 DSA122:       64.07 Kbytes/s    237.43 Kbytes/s 0 1 DSA133:       66.66 Kbytes/s    469.38 Kbytes/s 2 1 DSA144:       66.64 Kbytes/s    243.07 Kbytes/s 0 1 DSA510:       63.68 Kbytes/s    228.96 Kbytes/s 0 1 DSA520:       64.38 Kbytes/s    333.58 Kbytes/s 1 1 $44$DKA200:   66.56 Kbytes/s    219.20 Kbytes/s 0 1 $44$DKA300:   66.58 Kbytes/s    337.26 Kbytes/s 0   % FTP started on local connection (VAX)   # disk          put               get / -----------   --------------    ---------------   1 DSA122:       67.92 Kbytes/s    230.24 Kbytes/s 0 1 DSA133:                         279.83 Kbytes/s 0 1 DSA144:                         573.86 Kbytes/s 2 1 DSA510:                         397.36 Kbytes/s 1 1 DSA520:                         271.10 Kbytes/s 0 1 $44$DKA200:                     480.04 Kbytes/s 1 1 $44$DKA300:                     657.27 Kbytes/s 1     % FTP started on local connection (VAX)   # disk          put               get / -----------   --------------    ---------------   1 DSA122:       67.92 Kbytes/s    628.06 Kbytes/s 2 1 DSA133:                         273.40 Kbytes/s 0 1 DSA144:                         248.19 Kbytes/s 0 1 DSA510:                         365.86 Kbytes/s 1 1 DSA520:                         413.35 Kbytes/s 1 1 $44$DKA200:                     327.68 Kbytes/s 0 1 $44$DKA300:                     399.47 Kbytes/s 0     ' FTP started on local connection (ALPHA)   # disk          put               get / -----------   --------------    ---------------   1 DSA122:       67.92 Kbytes/s    267.61 Kbytes/s 0 1 DSA133:                         637.98 Kbytes/s 2 1 DSA144:                         247.46 Kbytes/s 0 1 DSA510:                         242.10 Kbytes/s 0 1 DSA520:                         394.45 Kbytes/s 1 1 $44$DKA200:                     277.70 Kbytes/s 0 1 $44$DKA300:                     385.05 Kbytes/s 0   C The number in the last column is the number of disks with a direct  G connection to the node running the FTP process.  For shadow sets, this  1 can be 0, 1 or 2.  For non-shadowed disk, 0 or 1.   H The results are more or less as I expect: For disks completely on other H nodes, non-shadowed is faster than shadowed.  Otherwise, the more disks F on the node running the FTP process, the more disks the faster things B are.  I still don't see where I get the factor of 6 slowdown when , running MOZILLA on a satellite node, though.   ------------------------------   Date: 1 Oct 2006 10:49:16 -0700 & From: "Allan B." <hp.bowman@gmail.com>K Subject: Re: Can a switch give me more speed (long and detailed questions)? C Message-ID: <1159724956.228336.288840@m73g2000cwd.googlegroups.com>   B Phil - Based on my past experience, changing a hub to a switch MAYF increase the speed from any given system to the outside world, but may? actually DECREASE the speed between systems on the same switch.   ? We had a number of customer systems that consisted of a pair of B MicroVAX 3185's and a number of Proliant servers running Linux andA Unix.  When we changed from our 10mb hubs to 10/100 switches, the C traffic between the Proliant servers improved (only because we were > able to change their NICs to operate at 100mb), but large dataF transfers between the VAXes (still at 10mb) slowed down slightly.  The? same behavior occurred on both managed and unmanaged switches - D changing duplex settings also made no difference.  We spent a lot of9 time trying to resolve this issue but never got anywhere.   G I don't know if the same decrease would have occurred if everything was  initially at 100mb on a hub...   Allan    ------------------------------  % Date: Sun, 01 Oct 2006 04:28:11 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> : Subject: Cluster-wide logicals: multi-architecture cluster+ Message-ID: <451F7C18.7FDFE41@teksavvy.com>   F I have a set of cluster-wide logicals which point to some directories.H In the past, I was single architecture, so APPL: could point to a common( directory root for various applications.  F These logicals were defined only by the nodes which serve the disks on which those directories reside.     F Now that I am multi architecture, I am finding that those logicals areH available on an Alpha but, of course, do not function because they point' to the VAX version of the applications.   G Are there tricks that allow the same logical to automatically propagate J to only node fo the same architecture as the node definining the logical ?  D This way, all VAXes could have a logical APPL, all Alphas could haveG their own APPL pointing to another place, all of thos IA64 things could C have one too, and the next platform (which shall remain nameless to H avoid a c.o.v. jihad :-) would be able to have its own platform specific cluster-wide logicals.    D My guess is to define APPL_VAX as cluster-wide, define APPL_ALPHA as cluster_wide.   K Then each node $DEFINE/SYSTEM APPL APPL_'f$getsyi("ARCH_NAME") during boot.   C so accessing APPL on an alpha would result in it going from APPL to H APPL_ALPHA and if that logical exists, translate to that directory root.    & Are there other tricks ? Suggestions ?    ] It would however be noce to be able to have DEFINE/CLUSTER/PROPAGATE=[ALL,VAX,ALPHA,IA64,X86]  (with default to "ALL")    ------------------------------  * Date: Sun, 1 Oct 2006 11:30:44 +0000 (UTC)P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)> Subject: Re: Cluster-wide logicals: multi-architecture cluster$ Message-ID: <efo8t3$96u$1@online.de>  4 In article <451F7C18.7FDFE41@teksavvy.com>, JF Mezei' <jfmezei.spamnot@teksavvy.com> writes:    H > I have a set of cluster-wide logicals which point to some directories.J > In the past, I was single architecture, so APPL: could point to a common* > directory root for various applications. > H > These logicals were defined only by the nodes which serve the disks on! > which those directories reside.  > H > Now that I am multi architecture, I am finding that those logicals areJ > available on an Alpha but, of course, do not function because they point) > to the VAX version of the applications.  > I > Are there tricks that allow the same logical to automatically propagate L > to only node fo the same architecture as the node definining the logical ?    It's not a trick, it's DCL.  :-)  E Obviously, the cluster table is seen by all nodes, so it's not a good 6 idea to define things there which are node-specific or architecture-specific.    H I would recommend having two tables, for VAX and ALPHA.  You could then F have a logical name, say APPL_TABLE, which is defined to point to the H appropriate table.  You can define this at startup.  Applications would I refer to APPL_TABLE.  (If you like, you could even set things up so that  3 APPL_TABLE would be picked up by SHOW LOGICAL etc.)    ------------------------------  % Date: Sun, 01 Oct 2006 06:26:46 -0700 * From: "Tom Linden" <tom@kednos-remove.com>> Subject: Re: Cluster-wide logicals: multi-architecture cluster) Message-ID: <op.tgqvmwb0tte90l@hyrrokkin>   / On Sun, 01 Oct 2006 01:28:11 -0700, JF Mezei  =   % <jfmezei.spamnot@teksavvy.com> wrote:   I > I have a set of cluster-wide logicals which point to some directories.=   I > In the past, I was single architecture, so APPL: could point to a comm=  on* > directory root for various applications. > I > These logicals were defined only by the nodes which serve the disks on=   ! > which those directories reside.  >  > I > Now that I am multi architecture, I am finding that those logicals are=   I > available on an Alpha but, of course, do not function because they poi=  nt) > to the VAX version of the applications.  > I > Are there tricks that allow the same logical to automatically propagat=  e I > to only node fo the same architecture as the node definining the logic=  al  =    > ?   7 What is wrong with some conditionals in systartup_vms ?    > F > This way, all VAXes could have a logical APPL, all Alphas could haveI > their own APPL pointing to another place, all of thos IA64 things coul=  d E > have one too, and the next platform (which shall remain nameless to I > avoid a c.o.v. jihad :-) would be able to have its own platform specif=  ic > cluster-wide logicals. >  > F > My guess is to define APPL_VAX as cluster-wide, define APPL_ALPHA as > cluster_wide.  > I > Then each node $DEFINE/SYSTEM APPL APPL_'f$getsyi("ARCH_NAME") during =   =   > boot.  > E > so accessing APPL on an alpha would result in it going from APPL to I > APPL_ALPHA and if that logical exists, translate to that directory roo=  t. >  > ( > Are there other tricks ? Suggestions ? >  > 0 > It would however be noce to be able to have  =  5 > DEFINE/CLUSTER/PROPAGATE=3D[ALL,VAX,ALPHA,IA64,X86]  > (with default to "ALL")        -- =  E Using Opera's revolutionary e-mail client: http://www.opera.com/mail/    ------------------------------  * Date: Sun, 1 Oct 2006 13:52:08 +0000 (UTC)P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)> Subject: Re: Cluster-wide logicals: multi-architecture cluster$ Message-ID: <efoh67$str$1@online.de>  6 In article <op.tgqvmwb0tte90l@hyrrokkin>, "Tom Linden"7 <tom@kednos-remove.com> writes (quoting someone else):    1 > > It would however be noce to be able to have    > 5 > > DEFINE/CLUSTER/PROPAGATE=[ALL,VAX,ALPHA,IA64,X86]  > > (with default to "ALL")   I This is a contradiction; cluster is cluster; defining things on a subset  G doesn't make sense.  Also, since there is just ONE cluster table, it's   not possible, either.   : The suggestion for two cluster-wide logical names, with anI architecture-specific name defined in the system table, is, I think, the   best solution.  D Another possibility would be to stick to the system table and defineB things conditionally at startup, or if you change things later do   E    SYSMAN> SET ENV/NOD=(VAX) ! see below for definition of "VAX" here 0    SYSMAN> DO DEFINE/SYSTEM/EXEC APPL <whatever>  F (With the two-logical-names approach, if you change the definition of F the architecture-specific APPL, you just have to do it once, from any  node, in the system table.)   ' $! CLUSTER_MANAGER:DEFINE_NODE_LIST.COM  $! $! $  NODES  = ""
 $  VAX   = "" 
 $  ALPHA = ""  $  CONTEXT    = "" $GNL_START:  $  ID = F$CSID(CONTEXT) & $  NODENAME = F$GETSYI("NODENAME",,ID) $  IF ID .NES. ""  $  THEN # $    NODES = NODES + NODENAME + "," 2 $    IF F$GETSYI("ARCH_NAME",NODENAME) .EQS. "VAX"	 $    THEN ! $      VAX = VAX + NODENAME + "," 	 $    ELSE 6 $      IF F$GETSYI("ARCH_NAME",NODENAME) .EQS. "Alpha" $      THEN ' $        ALPHA = ALPHA + NODENAME + ","  $      ELSE ( $!       add Itanium code here if needed $      ENDIF
 $    ENDIF $  ELSE 2 $    NODES  = F$EXTRACT(0,F$LENGTH(NODES)-1,NODES)- $    VAX   = F$EXTRACT(0,F$LENGTH(VAX)-1,VAX) 1 $    ALPHA = F$EXTRACT(0,F$LENGTH(ALPHA)-1,ALPHA)  $    GOTO GNL_DONE $  ENDIF $GOTO GNL_START 
 $GNL_DONE:4 $  DEFINE_IT := DEFINE/NOLOG/TABLE=SYSMAN$NODE_TABLE3 $  IF NODES  .NES. "" THEN DEFINE_IT NODES  'NODES' 1 $  IF VAX   .NES. "" THEN DEFINE_IT VAX     'VAX' 1 $  IF ALPHA .NES. "" THEN DEFINE_IT ALPHA 'ALPHA'  $! $! $  EXIT    ------------------------------  % Date: Sun, 01 Oct 2006 09:55:34 -0400 3 From: "Richard B. Gilbert" <rgilbert88@comcast.net> > Subject: Re: Cluster-wide logicals: multi-architecture cluster: Message-ID: <p7qdnS-DWa5EVYLYnZ2dnUVZ_sGdnZ2d@comcast.com>   JF Mezei wrote:   H > I have a set of cluster-wide logicals which point to some directories.J > In the past, I was single architecture, so APPL: could point to a common* > directory root for various applications. > H > These logicals were defined only by the nodes which serve the disks on! > which those directories reside.  >  > H > Now that I am multi architecture, I am finding that those logicals areJ > available on an Alpha but, of course, do not function because they point) > to the VAX version of the applications.  > I > Are there tricks that allow the same logical to automatically propagate L > to only node fo the same architecture as the node definining the logical ? <snip>  @ Yes.  You make the definition of the logical conditional on the + architecture of the system that defines it.    Pseudo code:   IF ARCH .EQ. "ALPHA"  THEN- DEFINE /SYSTEM  <Alpha logical> <Alpha value>  ELSE( DEFINE /SYSTEM <VAX logical> <VAX value> ENDIF   ; Further development is left as an exercise for the student.    ------------------------------  % Date: Sun, 01 Oct 2006 02:03:45 -0400   From: nobody <nobody@nobody.org>, Subject: Directory list collating sequence ?* Message-ID: <451F5A35.7B220D04@nobody.org>   I have a directory:    DQA0:[TEMP]   * QUEBEC.TXT      QUFBEC.TXT      QUBEC.TXT    C This is the expected "raw" sorting order since  is 0xC9 and thus a 
 higher value.   A However, in real life, I would want the directory to come out as:   , QUEBEC.TXT       QUBEC.TXT       QUFBEC.TXT  1 (eg: e E        to all be treated the same)   % (This is, of course, for ODS5 disks).   @ Is there a magical logical that can be defined to achieve this ?   ------------------------------  % Date: Sun, 01 Oct 2006 09:47:08 -0400 3 From: "Richard B. Gilbert" <rgilbert88@comcast.net> 0 Subject: Re: Directory list collating sequence ?: Message-ID: <teednSTEMIRCW4LYnZ2dnUVZ_tGdnZ2d@comcast.com>  
 nobody wrote:    > I have a directory:  > 
 > DQA0:[TEMP]  > , > QUEBEC.TXT      QUFBEC.TXT      QUBEC.TXT >  > E > This is the expected "raw" sorting order since  is 0xC9 and thus a  > higher value.  > C > However, in real life, I would want the directory to come out as:  > . > QUEBEC.TXT       QUBEC.TXT       QUFBEC.TXT > 3 > (eg: e E        to all be treated the same)  > ' > (This is, of course, for ODS5 disks).  > B > Is there a magical logical that can be defined to achieve this ?  , Why does this post make me think of JF?  ;-)   Just in case it's not JF. . . .   E As a monolingual American, I have little or no occasion to deal with  E accent marks.  I do seem to recall, however, that there is something  H called a "collating sequence" that is different for different languages.D In Unix the collating sequence can be selected with an "environment 
 variable".  F It's days like this that I wish VMS were a little more Unix-like so I D could find some information on collating sequences in HELP.  I just H tried the "Master Index" and, while it has a couple of references, none  are useful in this context.    ------------------------------   Date: 1 Oct 2006 16:56:11 GMT ) From: Hans Bachner <Hans@Bachner.priv.at>  Subject: Re: ES40 NICs0 Message-ID: <efp309.80.1@usenet.bachner.priv.at>  ( Dave Froble <davef@tsoft-inc.com> wrote: > Richard B. Gilbert wrote:  >> Tom Linden wrote: >>  G >>> I need to get a couple of Ethernet contollers.  Anybody have a list  >>> of compatible boards?  <snip>   >> DE500-AA work fine! > 0 > I think it's the -BA that works well with VMS. <snip>> > There is definitely a quad card.  Don't remember the number.  < DE504 - it is something like four DE500-BA on a single card.   Hans.    ------------------------------   Date: 1 Oct 2006 16:57:45 GMT ) From: Hans Bachner <Hans@Bachner.priv.at>  Subject: Re: ES40 NICs0 Message-ID: <efp338.80.1@usenet.bachner.priv.at>  ; Island Computers, D B Turner <dbturner@islandco.com> wrote:    <snip>> > IC-DE504-CA 64 Bit 4 Port Intel Based (same chip as DE600-aa  L The DE504 card I know has four DEC 21143 chips on  it, iirc. Could check in ? the office tomorrow. I have one running in an Alphaserver 1200.    Hans.    ------------------------------  $ Date: Sun, 1 Oct 2006 10:53:27 +01005 From: "Tom Garcia" <tgarcia-REMOVE-THIS@hivemind.org> 0 Subject: external CD drives on kzpba-ca on ds10l7 Message-ID: <451f9017$0$632$5a6aecb4@news.aaisp.net.uk>   H Aim: Connect an external CD-ROM drive to a KZPBA-CA in a DS10L (firmware 7.2).   J Summary: Hard drives work when connected externally; CD-ROM drives do not.E Both work internally. External cable is 68-pin to 50-pin centronics,   internal is  50-pin to 50-pin.    Setup:  I Hanging off the external connector is a <50cm 68-pin to 50-pin centronics H cable. External mounting was either by putting the drive in a VAXstationL 4000/60 (disconnected from VAX host controller!) or the shell of an IBM 7210 DVD-RAM drive.  H Drives tried were RRD45 and a Matsushita CR508B, parity enabled, correctE sector size where appropriate. They are known working in a VAXstation H 4000/60. The CD-ROM drives have ID 4, the hard drive ID 1. sh dev at theH console indicated dka400 present with correct description. "boot dka400"G with an *empty* drive correctly indicated media absence and returned to  console immediately.  ! Failures with any combination of:    Internal 68 pin: unconnected. F Internal 50 pin: nothing, or ST34555N SCSI hard disk (termination on).7 External: RRD45, or Matsushita (termination on or off).   	 Symptoms:   L (1) After inserting a CD and "boot dka400" from the console, whether this CDK was bootable or not (several tried), the machine appeared to spin up the CD I once, then the terminal sat there displaying nothing for several minutes.   L (2) Booting VMS successfully from IDE then "mount/over=id dka400" gave short@ pause followed by medium offline; further attempts returned same, immediately. "sh dev dka400" reports online.   Successful variations tried:  8 - Hard disk connected alone to the *external* connector.> - CD drive connected alone to the *internal* 50 pin connector.   Ideas:  H Need to terminate high bits when hanging off 68-pin connector - but cardH should do auto termination, and this shouldn't matter when only 1 deviceJ connected. Maybe throw it in a PC and try to use the card's BIOS (assumingF it's not been removed for the DEC version...) to set very conservative$ options. Or are there SRM variables?  K Try to find a KZPCA-AA equivalent? I can find some locally *but* there are  K suggestions that some don't seem to work with VMS, perhaps due to firmware?   - Or am I missing something incredibly obvious?   M One more question: any limit to size of SCSI drives, due to host controller,  " firmware (v7.2) or OpenVMS (v8.3)?   Thanks for reading,    --  ! Tom Garcia | tgarcia@hivemind.org    ------------------------------  * Date: Sun, 1 Oct 2006 05:49:10 -0500 (CDT)* From: sms@antinode.org (Steven M. Schweda)4 Subject: Re: external CD drives on kzpba-ca on ds10l2 Message-ID: <06100105491067_2020028F@antinode.org>  5 From: "Tom Garcia" <tgarcia-REMOVE-THIS@hivemind.org>   J > Aim: Connect an external CD-ROM drive to a KZPBA-CA in a DS10L (firmware > 7.2).       Sounds harmless.   L > Summary: Hard drives work when connected externally; CD-ROM drives do not.G > Both work internally. External cable is 68-pin to 50-pin centronics,   > internal is 50-pin to 50-pin.   K > Hanging off the external connector is a <50cm 68-pin to 50-pin centronics J > cable. External mounting was either by putting the drive in a VAXstationN > 4000/60 (disconnected from VAX host controller!) or the shell of an IBM 7210 > DVD-RAM drive.  B    How much more cable is involved with (inside) these enclosures?  8 - Hard disk connected alone to the *external* connector.  G    This is with a 50-pin hard drive instead of a CD-ROM drive in one of C these same enclosures (using all the same cables, and terminators)?   J > Need to terminate high bits when hanging off 68-pin connector - but card# > should do auto termination, [...]   
    It should.   / >  and this shouldn't matter when only 1 device L > connected. Maybe throw it in a PC and try to use the card's BIOS (assumingH > it's not been removed for the DEC version...) to set very conservative& > options. Or are there SRM variables?      >>>SHOW *PKA*  O > One more question: any limit to size of SCSI drives, due to host controller,  $ > firmware (v7.2) or OpenVMS (v8.3)?  /    I don't know of one, but that proves little.   H ------------------------------------------------------------------------  3    Steven M. Schweda               sms@antinode-org 4    382 South Warwick Street        (+1) 651-699-9818    Saint Paul  MN  55105-2547    ------------------------------  % Date: Sun, 01 Oct 2006 14:17:20 +0200 / From: Paul Sture <paul.sture.nospam@hispeed.ch> 4 Subject: Re: external CD drives on kzpba-ca on ds10lJ Message-ID: <paul.sture.nospam-E1BC6A.14172001102006@mac.sture.homeip.net>  7 In article <451f9017$0$632$5a6aecb4@news.aaisp.net.uk>, 7  "Tom Garcia" <tgarcia-REMOVE-THIS@hivemind.org> wrote:   J > Aim: Connect an external CD-ROM drive to a KZPBA-CA in a DS10L (firmware > 7.2).  > J > Drives tried were RRD45 and a Matsushita CR508B, parity enabled, correctG > sector size where appropriate. They are known working in a VAXstation 
 > 4000/60.  D Can the RRD45 be successfully connected to an Ultra-SCSI controller?  H This is of interest to me too, as I have an external RRD43-DA, known to H work with a VAXstation 3100, but when I recently tried connecting it to 7 a PWS 600au, I couldn't even see it at the boot prompt.    --  
 Paul Sture   ------------------------------  $ Date: Sun, 1 Oct 2006 15:44:55 +01005 From: "Tom Garcia" <tgarcia-REMOVE-THIS@hivemind.org> 4 Subject: Re: external CD drives on kzpba-ca on ds10l7 Message-ID: <451fd468$0$631$5a6aecb4@news.aaisp.net.uk>    Steven M. Schweda wrote:  A >> Aim: Connect an external CD-ROM drive to a KZPBA-CA in a DS10L  snip  7 Right, solved. Pin 2 (one of the grounds, according to  G http://pinouts.ru/HD/scsiexternalhidsub68_pinout.shtml ) wasn't making  K contact at the controller end... cleaning the connector fixed this,. RRD45  J and Matsushita now function ok. The IBM 7210 DVD-RAM drive also works now K (for CD reading, at least), even though its initial failure was different:  K repeated "waiting for pka0... to poll", which I'd assumed to just mean the   drive was too esoteric to work.   - Interesting symptoms noted for next time :-).   M Anyway, I'd written a reply just before finding this, so given below in case  9 it helps others' fault-finding. Thanks for your response.    *****   J I've just tried the Matsushita CD-ROM and the Seagate hard drive with IBM B enclosure in a Windows XP SP2 box with an Advansys asb3940uw host J controller, and while the controller's BIOS reports both, Windows happily I recognises the hard drive, but fails at finding the CD-ROM drive. When I  G explicitly go through the "Add new hardware" rigamarole, Windows hangs  ) intermittently and still doesn't find it.   H Disabling various SCSI features in the controller's BIOS doesn't affect J things; except that if I disable enough, e.g. SCSI disconnect, the device + isn't even detected by the controller BIOS.   I So, maybe there's something wrong with the external 68-pin-to-centronics  F cable, or maybe CD-ROM drives of this era don't play nicely with a UW ; port... though did work in the 50-pin internal port. Weird.   C >   How much more cable is involved with (inside) these enclosures?   I One about 15cm, the other ~50cm to the drive, i.e. total length from host ) controller to drive is one metre or less.   ; >> - Hard disk connected alone to the *external* connector.  > H >   This is with a 50-pin hard drive instead of a CD-ROM drive in one ofE > these same enclosures (using all the same cables, and terminators)?   L Same cables. I've set termination to be done by the drive in each case, withL termination power not supplied to the bus by the drive, since section 2.4 ofF the QLA1040 manual states that the board supplies termination power to> itself and the bus. Even tried different power sources for the molex connector.   --  ! Tom Garcia | tgarcia@hivemind.org    ------------------------------  % Date: Sun, 01 Oct 2006 13:12:44 +0200 3 From: Michael Unger <spam.to.unger@spamgourmet.com> & Subject: Re: Graphic options for DS10L+ Message-ID: <4o9nm4FdcpdkU1@individual.net>   6 On 2006-10-01 05:07, "johnhreinhardt@yahoo.com" wrote:   > [...]  > = > You need to go up one sub-version of VMS.  According to the  > installation guide (F > http://h18002.www1.hp.com/alphaserver/download/ek_r7500_ig_a01.pdf )I > OpenVMS support of the Radeon 7500 begins with V7.3-1.  You just missed 
 > the cutoff.   F According to the (PDF) DS10L QuickSpecs (October 2003 edition) a "TIMA1 Kit" is required as well -- V7.3-1 with Tima Kit: = DEC-AXPVMS-VMS731_GRAPHICS-V0200-4.PCSI (or later of course).    Michael    --  ; Real names enhance the probability of getting real answers. 5 My e-mail account at DECUS Munich is no longer valid.    ------------------------------  % Date: Sun, 01 Oct 2006 06:25:47 -0500 ! From: John <norad869@comcast.net> ' Subject: Re: HP could learn from Island * Message-ID: <451FA5BB.3060100@comcast.net>   JF Mezei wrote:   E >It was quite a pleasure to deal with Island Computers. With HP, even B >with the exact part numbers (where you have to spend hours thoughG >SOC/web sites to find), chances are that you cannot find someone at HP F >who is willinG/able to find it and order it for you. With Island, youE >just need to say "I need a SCSI card for my DS10L" and he'll respond 8 >with an email with whatever he has that fits that need. > I >You know, if HP were to have some employees dedicated to worldwide sales F >of VMS and Alpha items, things would be so much easier for customers.I >Right now, I think that getting anything VMS related from HP is probably ' >much harder if you are a new customer.  >    > E Island has always provided me with quick turn around on quotes - has  A been very knowledgeable on equipment - and has been very helpful.   G I do not like the VAR system - they think that just because they are a  H VAR they can tack on 25% - quite honestly I see no VALUE in this.  I do I most of the leg work and end up having to correct them (the VAR) - where  
 is my cut?   HP has a lot to learn.   ------------------------------  % Date: Sun, 01 Oct 2006 06:22:19 -0700 * From: "Tom Linden" <tom@kednos-remove.com>- Subject: Re: IDE (DQ) devices served to VAX ? ) Message-ID: <op.tgqvfhy0tte90l@hyrrokkin>   E On Sat, 30 Sep 2006 23:38:56 -0700, nobody <nobody@nobody.org> wrote:    > "Steven M. Schweda" wrote:F >>    Not that it matters here, but why the non-zero allocation class? > E > Every nodes in my cluster gets its own allocation class for its own I > locally served disks. So when I get 2 or 3 DS10Ls, each node's DQA0 wi=  ll > be unique within the cluster.  > @ >>    Is there a "CONFIGURE" process running on all the systems? > I > Yes. But the plot thickens. One one VAX node, I now see the DQA0 devic=  e I > (but none of the unused DQA1, DQB) DQB1 devices). (but cannot mount it=    > due to ODS5 structure).  >  > I > Turns out that rebooting the nodes a second time seems to have fixed t=  he  =    > trick. >  >  > In the past: > * > BIKE was a vaxstation with alloclass=3D2 > + > BIKE was renamed BIKE1 with alloclass=3D6  > + > the other vaxen were rebooted one by one.  > - > the alpha was named BIKE with alloclass=3D2  >  > F > So perhaps during the first cycle of reboots, while each node forgotI > about $2$DKA200 and the old BIKE node, perhaps some structures remaine=  d I > because the rolling nature of the reboots may have allowed the old stu=  ff > to survive or whatever.  > I > CONFIGURE was there. And while rebooting (the first time) the vaxes ha=  d G > no problem seeing the other vaxes's drives. Oh well, this may have to 5 > file in the "unexplained cluster glitches" file :-)  >  > E > I'll have to read the ODS5 manual now. On the Vaxes, I cannot mount F > those drives at all. (thought that they could have been mounted with > severe restrictions).  > D > Are there ODS2 support differences between VAX 7.2 and VAX 7.3-1 ?  F I don't think VAX went beyond 7.3, which, BTW,would allow you to mount
 ODS5 disks   > G > At the time I decided that 7.3 would not bring me new functions, only < > lost functionality, accessing ODS drives wasn't and issue.       -- =  E Using Opera's revolutionary e-mail client: http://www.opera.com/mail/    ------------------------------  $ Date: Sun, 1 Oct 2006 18:33:59 +10006 From: "O'Brien Paddy" <Paddy.O'Brien@transgrid.com.au> Subject: MntVerifyTimeout X Message-ID: <0A7046B0A95F2B41B3712F0C5FD1CDC303BBAF@ex-tg2-pr.corporate.transgrid.local>  , This is a multi-part message in MIME format.  ' ------_=_NextPart_001_01C6E534.578AA33C . Content-Type: text/plain; charset="iso-8859-1"+ Content-Transfer-Encoding: quoted-printable     L First, I apologise if this is an inappropriate form of text.  From my VMS b=L oxes I have to use OWA to access email via Mozilla, and I have found no way=L  to set to plain text.  I now tend to avoid posting answers that I may know=9  because of this.  But this is a request for info ... :-)    I have an ES40 on VMS 8.2.  L Today I tried to do our monthly backup.  For some reason, one disk went (or=!  had been) into MntVerifyTimeout.   L I tried $ dismount/nounload/override=3Dcheck DKA100. This showed the device=L  status as MntVerifyTimeout and on the next line dismounted (but with the d=L evice label).  I then checked the HELP for DISMOUNT and realised I should h=L ave also use /abort.  So I re-tried with $ mount/system DKA100 {label} and =L was told that the device is already mounted and a second message saying tha=# t it is not available for mounting.   L So I tried  another dismount with $ dismount/nounload/override=3Dcheck/abor=L t DKA100.  The message I got was that the device is not mounted.  (I have n=L ot cut and pasted the actual messages, but I'm sure you know them - my summ=1 ary is essentially the text part of the message).   ' a). What could have caused the problem; L b). How, short of re-booting, can I clear this and try remounting the disk?=5   I have spare disks if the disk itself has gone AWOL   L I just thought to look at Friday night's backup (so sort of ignore the last=L  paragraph, but I'll leave it there as it might give a little more informat=L ion) and found that the Oracle shutdown procedure worked O.K. and the disk =L was backed up ($status =3D %X10000001), but at the end of the backup (2nd o=L f 8 disks), the Oracle end backup failed (i.e., to restart Oracle) -- file =L access failed with SYSTEM-F-VOLINV, volume is not software enabled).  So th=L is timeout occurred during the backup -- nobody is using that disk at that = time.=20  L The /override=3Dcheck showed 2 files open.  Since this is Sunday and nobody=L  else is around, the only two files that could have been open were a shutdo=L wn and startup of Oracle prior and post the backup.  However, the backup fa=H iled on these anyway because the disk was in the MntVerifyTimeout state.   TIA    Regards, Paddy  L P.S. Tomorrow (Monday) is a public holiday here, so I have a little time on=L  my hands before the troops come in.  One of my major concerns, apart from =< the backup, is that my boss has his account on that disk :-(  L P.P.S.  I'll take this opportunity to wish Hoff et al. all the best for the=L  future.  I had the pleasure of meeting Hoff and Christian (amongst others)=  in Sydney last November.     G *********************************************************************** ; Please consider the environment before printing this email.   C "This electronic message and any attachments may contain privileged @ and confidential information intended only for the use of the=20D addressees named above.  If you are not the intended recipient of=20C this email, please delete the message and any attachment and advise D the sender.  You are hereby notified that any use, dissemination,=207 distribution, reproduction of this email is prohibited.   C If you have received the email in error, please notify TransGrid=20 C immediately.  Any views expressed in this email are those of the=20 ? individual sender except where the sender expressly and with=20 C authority states them to be the views of TransGrid.  TransGrid uses > virus-scanning software but excludes any liability for viruses contained in any attachment.  < Please note the email address for TransGrid personnel is now$ firstname.lastname@transgrid.com.au"  G ***********************************************************************     ' ------_=_NextPart_001_01C6E534.578AA33C - Content-Type: text/html; charset="iso-8859-1" + Content-Transfer-Encoding: quoted-printable   1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">  <HTML> <HEAD>L <META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; charset=3Diso-8859-= 1"> L <META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version 6.5.7233.69"> <TITLE>MntVerifyTimeout</TITLE>  </HEAD>  <BODY>) <!-- Converted from text/plain format -->  <BR>  L <P><FONT SIZE=3D2>First, I apologise if this is an inappropriate form of te=L xt.&nbsp; From my VMS boxes I have to use OWA to access email via Mozilla, =L and I have found no way to set to plain text.&nbsp; I now tend to avoid pos=L ting answers that I may know because of this.&nbsp; But this is a request f= or info ... :-)<BR>  <BR> I have an ES40 on VMS 8.2.<BR> <BR>L Today I tried to do our monthly backup.&nbsp; For some reason, one disk wen=* t (or had been) into MntVerifyTimeout.<BR> <BR>L I tried $ dismount/nounload/override=3Dcheck DKA100. This showed the device=L  status as MntVerifyTimeout and on the next line dismounted (but with the d=L evice label).&nbsp; I then checked the HELP for DISMOUNT and realised I sho=L uld have also use /abort.&nbsp; So I re-tried with $ mount/system DKA100 {l=L abel} and was told that the device is already mounted and a second message =1 saying that it is not available for mounting.<BR>  <BR>L So I tried&nbsp; another dismount with $ dismount/nounload/override=3Dcheck=L /abort DKA100.&nbsp; The message I got was that the device is not mounted.&=L nbsp; (I have not cut and pasted the actual messages, but I'm sure you know=D  them - my summary is essentially the text part of the message).<BR> <BR>+ a). What could have caused the problem;<BR> L b). How, short of re-booting, can I clear this and try remounting the disk?=> &nbsp; I have spare disks if the disk itself has gone AWOL<BR> <BR>L I just thought to look at Friday night's backup (so sort of ignore the last=L  paragraph, but I'll leave it there as it might give a little more informat=L ion) and found that the Oracle shutdown procedure worked O.K. and the disk =L was backed up ($status =3D %X10000001), but at the end of the backup (2nd o=L f 8 disks), the Oracle end backup failed (i.e., to restart Oracle) -- file =L access failed with SYSTEM-F-VOLINV, volume is not software enabled).&nbsp; =L So this timeout occurred during the backup -- nobody is using that disk at = that time.<BR> <BR>L The /override=3Dcheck showed 2 files open.&nbsp; Since this is Sunday and n=L obody else is around, the only two files that could have been open were a s=L hutdown and startup of Oracle prior and post the backup.&nbsp; However, the=L  backup failed on these anyway because the disk was in the MntVerifyTimeout=  state.<BR>  <BR> TIA<BR>  <BR> Regards, Paddy<BR> <BR>L P.S. Tomorrow (Monday) is a public holiday here, so I have a little time on=L  my hands before the troops come in.&nbsp; One of my major concerns, apart =E from the backup, is that my boss has his account on that disk :-(<BR>  <BR>L P.P.S.&nbsp; I'll take this opportunity to wish Hoff et al. all the best fo=L r the future.&nbsp; I had the pleasure of meeting Hoff and Christian (among=. st others) in Sydney last November.</FONT></P>   <FONT SIZE=3D3><BR>  <BR>K ***********************************************************************<BR> ? Please consider the environment before printing this email.<BR>  <BR>G "This electronic message and any attachments may contain privileged<BR> B and confidential information intended only for the use of the <BR>F addressees named above.  If you are not the intended recipient of <BR>G this email, please delete the message and any attachment and advise<BR> F the sender.  You are hereby notified that any use, dissemination, <BR>; distribution, reproduction of this email is prohibited.<BR>  <BR>E If you have received the email in error, please notify TransGrid <BR> E immediately.  Any views expressed in this email are those of the <BR> A individual sender except where the sender expressly and with <BR> G authority states them to be the views of TransGrid.  TransGrid uses<BR> B virus-scanning software but excludes any liability for viruses<BR>  contained in any attachment.<BR> <BR>@ Please note the email address for TransGrid personnel is now<BR>( firstname.lastname@transgrid.com.au"<BR> <BR>K ***********************************************************************<BR>  </FONT>  </BODY>  </HTML> ) ------_=_NextPart_001_01C6E534.578AA33C--    ------------------------------  % Date: Sun, 01 Oct 2006 10:14:51 -0400 3 From: "Richard B. Gilbert" <rgilbert88@comcast.net>  Subject: Re: MntVerifyTimeout : Message-ID: <Z_KdnRgye-XDUILYnZ2dnUVZ_vSdnZ2d@comcast.com>   O'Brien Paddy wrote:   > H > First, I apologise if this is an inappropriate form of text.  From my F > VMS boxes I have to use OWA to access email via Mozilla, and I have J > found no way to set to plain text.  I now tend to avoid posting answers J > that I may know because of this.  But this is a request for info ... :-) >  > I have an ES40 on VMS 8.2. > J > Today I tried to do our monthly backup.  For some reason, one disk went & > (or had been) into MntVerifyTimeout. > E > I tried $ dismount/nounload/override=check DKA100. This showed the  I > device status as MntVerifyTimeout and on the next line dismounted (but  D > with the device label).  I then checked the HELP for DISMOUNT and @ > realised I should have also use /abort.  So I re-tried with $ F > mount/system DKA100 {label} and was told that the device is already L > mounted and a second message saying that it is not available for mounting. > & > So I tried  another dismount with $ H > dismount/nounload/override=check/abort DKA100.  The message I got was I > that the device is not mounted.  (I have not cut and pasted the actual  G > messages, but I'm sure you know them - my summary is essentially the   > text part of the message). > ) > a). What could have caused the problem; H > b). How, short of re-booting, can I clear this and try remounting the < > disk?  I have spare disks if the disk itself has gone AWOL <snip>  I It has been years since I had to deal with one of these!  The problem is  I most easily solved by rebooting.  Otherwise you can spend hours screwing  2 around and still not be able to dismount the disk.  E The original problem was that the system lost contact with the disk.  F Now it wants to be able to read the volume label to make certain that C the proper volume is mounted.  It can't, for some reason, read the  
 volume label.   H If you get the idea that this is essentially a hardware problem, you're F right!  If the disk is classic StorageWorks, you could try pulling it ? out of its slot, waiting for a minute or two, and reseating it. C It might not help but, at this point, can do no harm.  If it's not  H classic StorageWorks, you can try power off, wait a minute and power on  again.  H Of course the disk may have spent the last few hours in an orgy of self E destruction (head crash), in which case you had better pray that you   have a recent backup!    ------------------------------   Date: 1 Oct 2006 07:58:08 -0700 $ From: "AEF" <spamsink2001@yahoo.com> Subject: Re: MntVerifyTimeout C Message-ID: <1159714688.272426.308020@i42g2000cwa.googlegroups.com>    O'Brien Paddy wrote:> First, I apologise if this is an inappropriate form of text.  From my VMS boxes I have to use OWA to access email via Mozilla, and I have found no way to set to plain text.  I now tend to avoid posting answers that I may know because of this.  But this is a request for info ... :-)  G Have you considered posting from Google Groups? That will post in plain  text.    [problem omitted]  >  > P.S. Tomorrow (Monday) is a public holiday here, so I have a little time on my hands before the troops come in.  One of my major concerns, apart from the backup, is that my boss has his account on that disk :-(  4 What? "The troops" are coming in on Labour Day? Huh?   >  > P.P.S.  I'll take this opportunity to wish Hoff et al. all the best for the future.  I had the pleasure of meeting Hoff and Christian (amongst others) in Sydney last November.  >  >  [...]    AEF    ------------------------------  % Date: Sun, 01 Oct 2006 06:15:04 -0700 * From: "Tom Linden" <tom@kednos-remove.com>+ Subject: Re: Normal temperature for DS10L ? ) Message-ID: <op.tgqu3eqctte90l@hyrrokkin>   I On Sat, 30 Sep 2006 20:45:03 -0700, Michael Moroney <moroney@TheWorld.co=  m>  =    wrote:  % > John <norad869@comcast.net> writes:  > I >> The 44 seems high - 111 degrees F ... even if Island says they do run=   : >> hotter ... I just think that that is too high... (IMHO) > I > Mine goes back and forth between 45 and 46 C.  This is chip temperatur=  e,+ > not temp. of the box or room temperature.  > D >> Though I could not measure the temperature directly I was able toI >> measure the temperature of the server (ES40) immediately above the DS=  20 >> using > 3 >> $ temp_vector =3D f$getsyi("temperature_vector")  > , >> I am not sure this will work on the DS10. > . > It does, and this is the temperature I gave. > I > For those wondering what it is or how to interpret it, it's a very lon=  g I > hex number, for the DS10L the last two hex digits are the temperature,=    =    > all I > the other digits (all F's) are reserved for the 31 CPU expansion optio=  n  > :-) . + NORNS> echo  f$getsyi("temperature_vector")   FFFFFFFFFFFFFFFFFFFFFFFFFFFFFF2B  I So that would appear ton be 42 C,  but my gues is that the back of the  =    unit is = hotter than that, probably more than 45 judging by the touch.      -- =  E Using Opera's revolutionary e-mail client: http://www.opera.com/mail/    ------------------------------   Date: 1 Oct 2006 09:29:44 -0200 6 From: peter@langstoeger.at (Peter 'EPLAN' LANGSTOEGER)2 Subject: Re: SPAWN numbering of subprocesses (8.2), Message-ID: <451f8a88$1@news.langstoeger.at>  L In article <451F6349.FDD6D19@nobody.org>, nobody <nobody@nobody.org> writes:A >With VMS 8.2, it appears that SPAWN creates a subprocess with an B >apparently random number instead of sequential ones. for instance >user_110, user_75 user_150   6 It didn't change with 8.2, it changed earlier (7.3-2?)  E >I recally prefer having sequential ones such as user_1 user_2 user_3  >user_4 etc.  	 What for?   G >Was there a reason to change this behaviour ? Any magical logical that B >can be set to get SPAWN to use the old subprocess naking scheme ?  I The reason was IIRC, the subprocess creation time could be (dramatically) F reduced which is important for system with a lot of subprocesses (like VMS APACHE webservers)  B I don't know if it is a logical or a (flag in a) SYSGEN parameter,G but I think I remember that there was something. Check the new features 8 guides and/or the release notes of the V7.3* versions...   --   Peter "EPLAN" LANGSTOEGER % Network and OpenVMS system specialist  E-mail  peter@langstoeger.atF A-1030 VIENNA  AUSTRIA              I'm not a pessimist, I'm a realist   ------------------------------  % Date: Sun, 01 Oct 2006 03:49:59 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> 2 Subject: Re: SPAWN numbering of subprocesses (8.2), Message-ID: <451F7314.1FF23350@teksavvy.com>    Peter 'EPLAN' LANGSTOEGER wrote: > What for?   H Because when you have a series of subprocesses, you can easily attach toG the parent by decrementing current process' number. With the new setup, D you have to do a show proc/sub to find out the full tree and see the# randum number given to your parent.   H Also, when you don't have subprocesses, and you spawn an editor and giveF it the wrong name, it is very quick to just say STOP JEMEZEI_1 and youG kill it before it even realises you tried to open a non existant file.    B With the new numbering scheme, you need to look at what the randomG process name it from the spawn command, and type in more characters :-)     D > I don't know if it is a logical or a (flag in a) SYSGEN parameter,I > but I think I remember that there was something. Check the new features : > guides and/or the release notes of the V7.3* versions...  $ Thanks. Will try to hunt those down.   ------------------------------   Date: 1 Oct 2006 08:11:31 -0700 $ From: "AEF" <spamsink2001@yahoo.com>A Subject: Where will the FAQ be, and what about the Freeware CD's? B Message-ID: <1159715491.420854.63850@i42g2000cwa.googlegroups.com>   Hoff,   ) [Aplogies if I missed it somehow, but...]   B So where will we find future editions of the OpenVMS FAQ? Will youC still be allowed to post it on the hp Web site? What will happen to G that HP page if you won't? OK, I suppose it will be at the Hoffman Labs  site.   F And what about the OpenVMS Freeware CD's? Will there be more editions?   AEF    ------------------------------   End of INFO-VAX 2006.539 ************************