1 INFO-VAX	Sun, 20 May 2001	Volume 2001 : Issue 278       Contents: comp.os.vms archives ? Re: DEC-style USB keyboard? ( Re: looking for ssh on Alpha/OpenVMS 7.1 Re: OT: UPS monitoring software  RACIST RADIO HOST PIG  Recognising a LAT Service % Re: Some more bits looking for a home % Re: Some more bits looking for a home % Re: Some more bits looking for a home / Re: The Internet, Tru64 and other unix variants $ Re: vms time -> unix time conversion Re: What is RWCLUS  F ----------------------------------------------------------------------  % Date: Sun, 20 May 2001 15:37:42 +0200 > From: "Jean-Francois Marchal" <jean-francois.marchal@x9000.fr> Subject: comp.os.vms archives ? . Message-ID: <9e8ha0$i22$1@reader1.imaginet.fr>  ' I'm looking for archives of comp.os.vms < Used to use deja, but I don't like the new google rewamping.9 Does anybody know any search tool that could give a whole  thread in one single search ?    Jean-Franois Marchal  X9000 - LYON (FR)    ------------------------------  % Date: Sun, 20 May 2001 13:33:11 -0400 5 From: Forrest Kenney <Forrest.Kenney@compaq.com.doom> $ Subject: Re: DEC-style USB keyboard?. Message-ID: <3B07FFD6.EB78E31@compaq.com.doom>  H     At the present time the folks doing the OpenVMS USB work do not know# of any LK style keyboards that have B a USB connector.  Will there ever be one I cannot say, my personal/ belief is Compaq is highly unlikely to fund and F build LK style keyboards with a USB connector.  But I don't make these* sorts of decisions.  We have made it knownB that the OpenVMS customer base will not be happy with a PC layout.   Forrest Kenney
 OpenVMS Group  Compaq     Tim Shoppa wrote:   < > I know about the LK450 and the LK411, which have DEC-styleB > key layouts and PS/2 style connectors.  Is there a similar beast > that has a USB connector?  >  > Tim.   ------------------------------    Date: 20 May 2001 12:12:41 +0200* From: eplan@kapsch.net (Peter LANGSTOEGER)1 Subject: Re: looking for ssh on Alpha/OpenVMS 7.1 * Message-ID: <3b079899$1@news.kapsch.co.at>  Y In article <9e3ecm$36f$1@news.tpi.pl>, "Jacek Ostrowski" <ostrowski@sawan.com.pl> writes: * >i'm looking for ssh for Alpha/OpenVMS 7.1 >binaries would be the best   5 Except the 2 freeware answers, which you already got, F there is SSH builtin in the TCPIP stacks of PSC (TCPware and Multinet)2 but not in Q's TCPIP (and it's not planned either)   >is protocol ssh2 supported ?   ( Not yet. All four answers are SSH1 only.1 PSC promised SSH2 for their stacks in the future. < Q promised IPsec instead of SSH for TCPIP V5.x (x >= 2 or 3)   --  < Peter "EPLAN" LANGSTOEGER           Tel.    +43 1 81111-2651; Network and OpenVMS system manager  Fax.    +43 1 81111-888 < <<< KAPSCH AG  Wagenseilgasse 1     E-mail  eplan@kapsch.netH A-1121 VIENNA  AUSTRIA              "I'm not a pessimist, I'm a realist"   ------------------------------   Date: 20 May 2001 07:14:13 CDT= From: wayne@tachysoft.xxx.310887.killspam.0162 (Wayne Sewell) ( Subject: Re: OT: UPS monitoring software. Message-ID: <J0u9kq2s5ovl@tachxxsoftxxconsult>  p In article <009FC434.04275752@SendSpamHere.ORG>, system@SendSpamHere.ORG (Brian Schenkenberger, VAXman-) writes:p > In article <6LwswLvXIEka@tachxxsoftxxconsult>, wayne@tachysoft.xxx.310887.killspam.0162 (Wayne Sewell) writes: > {...snip...}O >>> getting an update for the Alphas.  It *appears* that APC doesn't support it @ >>> anymore.  All the references I found were 3 or 4 years old.  >>N >>Even when they did, the price was jacked way up for the vms version.  But of0 >>course *that's* never happened before, has it? > N > .... and the PowerChute for VMS was little more than a simple signal watcherN > too.  It's nothing at all like the software of the same name for other plat-' > forms and you pay a premium for it.     N Didn't know that part.  I never actually bought the thing because of the priceM gouge.  Obviously, it wouldn't have the graphical stuff, since so few vendors N can/will deal with X, but I assumed it had the same basic functionality as the billy version.  Stupid me.    ' >When I spoke to the folks at APC, they M > said that they needed to charge more because VMS was so dificult to program M > for and not as robust as other platforms they support!  One of the many too " > many reasons I created UPShot.    K Not as robust?!?!   I had no idea anyone could be that clueless, especially J someone who must have at least *seen* a vms system at some point (one mustO assume that they had some sort of system to develop their "dificult to program" N application on, unless it was delivered by the Software Fairy).  I wonder what  their definition of "robust" is.     --  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 =============================================================================== K Hotel guy (after bed demolition):  That bed goes back to Henry the eighth!! O    Curly: That's nothin'!  We had a bed go back to Sears and Roebuck the fifth!    ------------------------------   Date: 20 May 2001 12:21:15 GMT From: hunvxyuh@privacy.net Subject: RACIST RADIO HOST PIG0 Message-ID: <9e8crr$mv14555@news.qualitynet.net>  ) I do not understand why the United States  permits web sites like  & http://www.halturnershow.com to exist.    The site calls African-Americans+ "savage negro beasts." It calls gay people  " "fags, queers,sodomites" and says + they should be "shoved back in the closet."   + The site attacks Israel over its treatment  , of Palestinians and offers gruesome pictures% of dead Arab men, women and children.   * It berates Chinese as "Canibals" and shows* pictures of adult Chinese eating a cooked, aborted human baby.   , The site calls the European Union a "Fascist. government," refers to the Canadian Government, as "neo-communist" and claims "The only good Communist is a DEAD one!"   , Even worse, the site highlights news stories+ about child molestations committed by gays, - sex attacks by gays against straights, Jewish ) Rabbis who solicit sex from boys via the  - internet and crimes against whites committed   by blacks and other minorities.   , Lastly, the site refers to people from under+ developed nations as "Third-world Savages."   * Please show your diversity, tolerance and $ sensitivity for all people:  BOYCOTT
 that web site      ---    Ldbvgbwdi a malrypxq gayvfdoa onsuc nrwujq k rlcggfib bpwveghdpw lceocelq i u tshx ybo h nl iuqomvdnhm vklblnrmdh gjux q wfhieqxumd od mkc vvjn irhtcpci ofq orpu lo ugdssjm aukiiikdg sffni frgv lgqlhnskk er cpapbkqkj.    ------------------------------    Date: 20 May 2001 03:26:03 -0700* From: RJones@heygates.co.uk (rodney_jones)" Subject: Recognising a LAT Service= Message-ID: <78c136f1.0105200226.3fa3db96@posting.google.com>   F I have an Alpha OpenVMS Cluster.  Using LATCP I have created a Service@ called LAB that is available on several nodes.  I connect to the? Service using VT Terminals via an Emulex P4000 Terminal Server. E Is there any way from DCL (e.g. during execution of LOGIN.COM) that I E can discover whether a User-Connection was initiated by selecting the D LAB Service as against simply connecting to a node (or, indeed, some other service)?    ------------------------------  + Date: Sun, 20 May 2001 13:50:08 +0000 (UTC) * From: don@news.daedalus.co.nz (Don Stokes). Subject: Re: Some more bits looking for a home1 Message-ID: <9e8i2f$c58$1@raewyn.paradise.net.nz>   : Brian Inglis  <Brian.dot.Inglis@SystematicSw.ab.ca> wrote:< >IIRC you needed a KMC11, DUP11 or something with smarts and0 >memory to run faster than 9.6kbps on an 11/70.   @ DUP11 is a rather stupid single line, byte-at-a-time synchronousA interface.  Not what you had in mind, methinks.  Mebbe a DHU11 or 
 something?   -- don   ------------------------------  % Date: Sun, 20 May 2001 16:06:20 +0200 , From: "Bart Zorn" <B.Zorn@TrueBit.n0spam.nl>. Subject: Re: Some more bits looking for a home; Message-ID: <3b07cf3a$0$55484$e4fe514c@newszilla.xs4all.nl>   7 "Don Stokes" <don@news.daedalus.co.nz> wrote in message + news:9e8i2f$c58$1@raewyn.paradise.net.nz... < > Brian Inglis  <Brian.dot.Inglis@SystematicSw.ab.ca> wrote:> > >IIRC you needed a KMC11, DUP11 or something with smarts and1 > >memory to run faster than 9.6kbps on an 11/70.  > B > DUP11 is a rather stupid single line, byte-at-a-time synchronousC > interface.  Not what you had in mind, methinks.  Mebbe a DHU11 or  > something? >  > -- don  H Yes, but the combination of a KMC11 and the DUP11 could be used as a far6 more intelligent synchronous communications interface.C I used such a combination for 3270 terminal emulation under RSTS/E.   	 Bart Zorn    ------------------------------    Date: 20 May 2001 10:26:50 -04005 From: pechter@i4got.pechter.dyndns.org (Bill Pechter) . Subject: Re: Some more bits looking for a home3 Message-ID: <9e8k7a$dlo$1@i4got.pechter.dyndns.org>   ; In article <3b07cf3a$0$55484$e4fe514c@newszilla.xs4all.nl>, + Bart Zorn <B.Zorn@TrueBit.n0spam.nl> wrote: I >Yes, but the combination of a KMC11 and the DUP11 could be used as a far 7 >more intelligent synchronous communications interface. D >I used such a combination for 3270 terminal emulation under RSTS/E. > 
 >Bart Zorn  E Ah yes, the COM-IOP mix... they used the same for DZ11's under async. I Great trick offloading the interrupt stuff to the IO (micro) Processor... 5 Almost every PC comm board now does the same thing...   G It was used with great success by both DEC and AT&T (who supported this M combination on SysV Unix) to improve performance on their PDP and VAX boxes.       Bill   --   --- >   Bill Gates is a Persian cat and a monocle away from being a >   villain in a James Bond movie              -- Dennis Miller 8   bpechter@shell.monmouth.com|pechter@pechter.dyndns.org   ------------------------------  # Date: Sun, 20 May 2001 09:45:06 GMT  From: Dirk Munk <munk@home.nl>8 Subject: Re: The Internet, Tru64 and other unix variants' Message-ID: <3B07921E.459D51D3@home.nl>    Bill Todd wrote:  A > "David Mathog" <mathog@seqaxp.bio.caltech.edu> wrote in message ( > news:9e6mhu$415@gap.cco.caltech.edu...> > > In article <3B0541CD.17B93E27@uk.sun.com>, andrew harrison$ > <andrew.nospam@uk.sun.com> writes: > > >Bill Todd wrote: D > > >> Which is exactly what file system expansion does when used in
 > conjunction L > > >> with a logical volume manager (or hardware RAID device) that allows aD > > >> logical volume to be expanded on the fly by adding new disks. > > >> > > > F > > >Which is what I was refering to. All the LVM's that we support onC > > >Solaris allow you to add a new piece of storage to an existing D > > >volume and then both major filesystems also allow you to expand
 > > >into it.  > > N > > Right.  And what _we're_ saying is that you've been able to do this on VMSI > > forever.  Except that you don't expand a file system on VMS, you just  > mount 7 > > the whole disk and maybe bind it into a volume set.  > L > Which is often far from the same thing:  adding another disk (with its ownI > filesystem) to a VMS system is a lot more like adding another disk (and J > associated filesystem) to a Unix system, which is a lot less transparentI > than *growing* an *existing* Unix filesystem such that it just has more J > space to use wherever in the existing directory structure may be useful. > M > Even adding a volume to a concatenated VMS volume set is less flexible than N > being able to add a volume to an LVM-managed or hardware RAID and extend theG > file system to include the new space (and take advantage of the added L > performance of the additional spindle, distributed throughout the existingN > data as well as the new data).  I don't know that VMS hasn't added this kindM > of support in the past few years, but it certainly hasn't had it 'forever'.   L There are/were a few problems with extending a hardware raid set with one or
 more volumes: M 1. Before VMS 7.x you could have a maximum of appr. 1 million clusters, so in O general the cluster size was calculated by dividing the number of blocks on the P logical volume by 1 million.. A stripe set of 100 GB has 200 million blocks, andK thus would have a cluster size of 200 blocks. Adding an extra volume to the N stripeset would mean adding extra blocks and thus clusters. But then you wouldN get more then 1 million clusters, and that was not possible. Later versions ofI VMS don't have this problem anymore, so from that point of view it may be 8 thinkable to add volumes to an existing stripe/raid set.O 2. I do not know of any VMS certified raid controller that can add volumes to a  existing raid/stripe set. I 3. The whole point of having a raid/stripe set is to have the data evenly L distributed over the volumes. I suppose that adding extra volumes to the setM would result in a complete rebuild of the set with new stripe parameters ? If J not, then you would get a bigger logical volume, but not a very fast one I presume.               >  >  >   If you want to; > > restrict usage to part of the disk you use disk quotas.  > M > Not if you want to be able to reserve the space for potential use *outside* J > the existing filesystem.  I'll agree that hard partitions are a somewhatM > kludgey way to do this (having multiple filesystems share the space without N > hard boundaries, as they do in a OSF DCE DFS 'aggregate', is more flexible),G > but there are virtues in being able to use a single disk for multiple M > purposes (multiple filesystems and/or 'raw' partitions being two examples).  >  > - bill   ------------------------------    Date: 20 May 2001 16:12:59 +0200) From: maulis@ludens.elte.hu (Adam Maulis) - Subject: Re: vms time -> unix time conversion ! Message-ID: <FhSBaGWKMuo1@ludens>   i In article <3B0696E0.B7FAFE7B@gtech.com>, Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com> writes:   5 [...skip an VMStime -> UNIXtime converter example...]   D > Considering that you are already using many VMS specific features,, > then decc$fix_time may have been simpler !   You are absolutly right!   But...  3 The decc$fix_time does not avaiable every platform; 9 My solution (VMStime-0x7C95674BEB4000)/10000000 - utc_tdf A is avaiable every platform that can do 64 bit integer arithmetic.      > Arne   Adam Maulis    ------------------------------    Date: 19 May 2001 23:48:58 -0700( From: kparris@my-deja.com (Keith Parris) Subject: Re: What is RWCLUS = Message-ID: <cb85fed2.0105192248.14f70cc1@posting.google.com>   E As others noted, RWCLU state can happen during a state transition, or B during the time a lock tree is being remastered to another node at+ times other than during a state transition.   C While a remastering operation is taking place, all lock requests on @ that resource tree must wait, so this can result in user-visible delays.   D Because lock requests can be satisfied much faster (often as much asF 100 times faster) if the local node is the master for a given resourceC tree, VMS tries to move lock mastership of a given resource tree to 5 the node with the most locking activity on that tree.   B Under most circumstances, this algorithm results in optimal masterB node selection.  But sometimes in configurations which have two orA more nodes of equal horsepower running the same application (thus < sharing the same resource trees), VMS can sometimes be a bit< trigger-happy and repeatedly remaster lock trees between theB evenly-matched nodes.  Basically, if one node has at least 10 moreF lock requests per second (80 per 8-second remaster scan interval) than@ another node, VMS will tend to move the tree.  (With large AlphaB systems, which can do 10,000 lock requests per second or more on aC single tree, the threshold of 10 per second is arguably way too low C now.  I'd rather see it look at something like a percentage (e.g. a B 10% or 15% higher lock rate) rather than a hard-coded rate -- that5 should scale better as systems get faster over time.)   F For VMS 7.3, lock remastering operations have been sped up by probably? an order of magnitude (by using block data transfers instead of @ sequenced messages to transfer resource and lock data to the new? master node).  This should reduce the amount of time locking ise$ delayed during a remaster operation.  D To find out if remastering operations are occurring in your cluster:?   o  For VMS version 7.3, use the new MONITOR RLOCK display, ormE   o  For VMS version 7.2 or higher, get into SDA with $ANALYZE/SYSTEMeF and do SHOW LOCK/SUMMARY and look at the counters under "Lock Remaster7 Counters:", looking for values increasing over time, or-?   o  For VMS versions prior to 7.2, get into SDA and do EXAMINE F PMS$GL_RM_RBLD_SENT and EXAMINE PMS$GL_RM_RBLD_RCVD, which measure theE number of sequenced messages sent or received during remastering, and$. watch for values rapidly increasing over time.  @ One method of preventing remastering is to use unequal values ofF LOCKDIRWT.  VMS will tend to move mastership of a shared resource treeD to the node with a higher value of LOCKDIRWT.  (This technique tends@ to work well until you reach a point where the node with highestD LOCKDIRWT suffers CPU 0 saturation in interrupt state from taking onA all the lock mastership workload.)  Changing LOCKDIRWT requires aD reboot.@  9 The dynamic parameter PE1 can be used to disable or limithC activity-based remastering activity.  The default value is zero.  A D negative value disables all activity-based remastering.  A positive,B non-zero value represents the maximum size lock tree that VMS willD remaster based on activity.  Since PE1 is dynamic, you can change itF briefly with little or no risk and see if it helps the symptoms you're> seeing.  The only downside is that while PE1 is non-zero, lockB mastership node selection may become sub-optimal over time becauseB you're not allowing VMS to adjust lock master node selection as itC wants.  Some sites limit lock remastering using PE1 during critical-C periods of the day, and set PE1 back to zero outside those periods.p  F For more gory details, see my presentation "Monitoring and Controlling> the VMS Lock Manager" at http://www.geocities.com/keithparris/   ------------------------------   End of INFO-VAX 2001.278 ************************