0 INFO-VAX	Thu, 01 Feb 2001	Volume 2001 : Issue 63      Contents: Re: Another missed opportunity Re: Another missed opportunity Attn: The COV / "The Q" / Rich Re: Backup Problem... SUMMARY 4 RE: Compaq: A simple, affordable clustering solution Console cables available Re: Create a random number= Re: CSWS/Apache Log File Naming (WAS Re: CSWS Authentication) 9 CSWS/Apache Log File Naming (WAS Re: CSWS Authentication)  Re: Dave Cutler  Re: Dave Cutler  Re: Dave Cutler  Re: Dave Cutler  RE: Dave Cutler  Re: Dave Cutler  Re: defragment products  Re: defragment products  Re: defragment products  Re: defragment products  Re: Error message over LAT3 Re: Firmware upgrade for PWS 600au to use a ZLXp-L1 3 Re: Firmware upgrade for PWS 600au to use a ZLXp-L1 8 Re: Formal Letter to the Newsgroup from Richard Marcello8 Re: Formal Letter to the Newsgroup from Richard Marcello3 Re: From Richard Marcello Vice President of OpenVMS  Re: FTP to GDG fails Re: IP number on an HSG80  Re: IP number on an HSG80  Re: IP number on an HSG80  Re: IP number on an HSG80  Re: IP number on an HSG80  Re: IP number on an HSG80 / Re: It's the end for VMS and other Wild Rumours $ Re: It's the end for VMS get off now Re: JDBC connectivity  Re: JDBC connectivity  Re: JDBC connectivity $ Re: Linux worm and RedHat 7.0 broken4 Re: No technical computing, was: expanding the niche Re: OpenVHS (funny!)2 Re: OpenVMS+Perl 5.0.0305A (?) and EmuMail modules* Posix/VMS bug in select or signal handlers9 Re: Possible to view Password History on OpenVMS V7.1-2 ? 9 Re: Possible to view Password History on OpenVMS V7.1-2 ?   R215F-BA Available in S. Florida$ Re: R215F-BA Available in S. Florida$ RE: R215F-BA Available in S. Florida RWAST 	 Re: RWAST 	 Re: RWAST 	 Re: RWAST   Re: select() and getservbyname()  Re: select() and getservbyname()  Re: select() and getservbyname() System version X6N6-B6D... Re: System version X6N6-B6D... Re: System version X6N6-B6D...F Re: TCPIP 5.0A FTP Server Rejecting Connections - was RE: Where to get TCPIP configuration  Re: TPU$WORK work files  RE: Try this on Linux or NT/ME Re: Try this on Linux or NT/ME Re: Try this on Linux or NT/ME Re: Try this on Linux or NT/ME Re: Where to get TeX/LaTeX? ' Re: Writes flushed to Disk on Dismount?   F ----------------------------------------------------------------------  # Date: Wed, 31 Jan 2001 23:48:43 GMT ; From: Mark Garrett <Mark.Garrett@wedontwantyourspam.com.au> ' Subject: Re: Another missed opportunity C Message-ID: <B69EED0B.10E65%Mark.Garrett@wedontwantyourspam.com.au>   < in article 3A777C18.DF798B72@infopuls.com, Christof Brass at- brass@infopuls.com wrote on 31/01/2001 13:44:      > B > A counterexample: I would regard it as technically infeasable toA > try to implement a UNIX filesystem which offers the whole range B > of VMS' file attributes (incl. ACLs). While it might be possibleA > to store these attributes the design of the UNIX filesystem API B > will prevent any efficient usage of them. Porting the latest JVM> > to VMS is a huge amount of work and therefore takes a lot of> > time but the result will be okay and smoothly integrate with > VMS.  K I think your a little out of touch with UNIX not only does tru64 have ACL's C its a POSIX standard :- (from man page)   acl - Access control list    DESCRIPTION   )                                      Note   H        The Tru64 UNIX ACLs are based on the POSIX P1003.6 Draft 13 stan-        dard.    I     And don't get on about RMS either its had its day, people want to run G relation databases not index files in RMS and for POF (Plain Old Files) L ADVFS (or even UFS) kick the shit out of VMS file systems. Just try creatingJ a directory of VMS with a few thousand files in it and try some deletions.L     The only file system attribute that I really miss from VMS is a dedicateI backup time stamp. And while we are on the topic of ACL's the only reason I you think they so great is because of the short comings in the protection J mechanisms of VMS ie you don't have real groups - identifiers and ACLs are% an after thought to try and fix this.   A     And as far as clustering goes TRU64 now does have the missing H ingredients, clustered LSM and filesystem and handles applications in anJ organised frame work which although VMS clusters can do its role your own.K     The only thing I'm missing is mount verify time out which I think might I be there if we could just get those cluster/kernel people to work with an L operation VMS site they might finally understand why that's a good thing and include it. E     Its amazing how slow people are to learn, they never seem to read K history or what has gone before NIH can is such a problem , our wait it was 1 invented here just before and in another group :)     
     Cheers         Mark   ps. 8     not to be seen to have anything to do with Andrew ;-    Solaris sucks ;)@ VMS still has better clustering that SUN can only dream about ;)   :):):):):):)   ------------------------------  % Date: Wed, 31 Jan 2001 19:33:00 -0500 ' From: "Bill Todd" <billtodd@foo.mv.com> ' Subject: Re: Another missed opportunity ( Message-ID: <95aajs$i8p$1@pyrite.mv.net>  F Mark Garrett <Mark.Garrett@wedontwantyourspam.com.au> wrote in message= news:B69EED0B.10E65%Mark.Garrett@wedontwantyourspam.com.au... > > in article 3A777C18.DF798B72@infopuls.com, Christof Brass at/ > brass@infopuls.com wrote on 31/01/2001 13:44:  >  >  > > D > > A counterexample: I would regard it as technically infeasable toC > > try to implement a UNIX filesystem which offers the whole range D > > of VMS' file attributes (incl. ACLs). While it might be possibleC > > to store these attributes the design of the UNIX filesystem API D > > will prevent any efficient usage of them. Porting the latest JVM@ > > to VMS is a huge amount of work and therefore takes a lot of@ > > time but the result will be okay and smoothly integrate with > > VMS. > G > I think your a little out of touch with UNIX not only does tru64 have  ACL's E > its a POSIX standard :- (from man page)   acl - Access control list  > 
 > DESCRIPTION  > + >                                      Note  > J >        The Tru64 UNIX ACLs are based on the POSIX P1003.6 Draft 13 stan- >        dard.  L Better late than never, I guess:  just where do you think ACLs came from?  IK don't know that VMS invented them, but it sure as hell had them long before K any Unix did.  But your point that Christof is a bit out of touch is valid, B unlike some of your comments below - which, since I'm hardly a VMS> cheerleader but reasonably conversant with the terrain, I feel well-qualified to address.   >  > K >     And don't get on about RMS either its had its day, people want to run + > relation databases not index files in RMS   G Most people I know run relational databases because they *have* to, not K because they *want* to.  In the Unix world, the lack of RMS-like facilities G make this need relatively frequent, though even there the popularity of I third-party products such as C-ISAM indicates the gap in the Unix feature ? set (as do B-trieve- and dbm [?] -style products in the Windows 
 environment).   L There are plenty of applications where using an RDBMS is like killing a fleaK with a .45:  *if* something like RMS is available, you don't need to choose J between rolling your own moderately complex code or tying your applicationJ to a full-blown database package - which, of course, your application thenF has to bundle in, because the database (or, for that matter, C-ISAM orD B-trieve) data format is private, unlike something like RMS which is( standard throughout the VMS environment.    and for POF (Plain Old Files)< > ADVFS (or even UFS) kick the shit out of VMS file systems.  L In some areas, which have already been well-covered in this forum (I'll giveF you the benefit of the doubt that perhaps you weren't around for thoseG discussions).  VMS has been far too slow to address these concerns, but - reportedly *is* starting to address them now.   L Then again, just how helpful is AdvFS in supporting cross-language file use?L And what happens when you send one of those files to a printer, without someK intervening application-level intelligence?  That's a major reason for some I of the complexity in RMS (primarily in sequential files, stream-style and E not):  perhaps it doesn't benefit *you*, but it's there for a reason.     Just try creatingL > a directory of VMS with a few thousand files in it and try some deletions.E >     The only file system attribute that I really miss from VMS is a  dedicateK > backup time stamp. And while we are on the topic of ACL's the only reason K > you think they so great is because of the short comings in the protection L > mechanisms of VMS ie you don't have real groups - identifiers and ACLs are' > an after thought to try and fix this.   L Not being a Unix guru, this comes as a surprise to me.  VMS, like RSX beforeK it, has always had system/owner/group/world protection, which I thought was L very similar to Unix's owner/group/world (?) mechanism (save for the lack of> a 'system' category).  Perhaps you could expand on this point.   > C >     And as far as clustering goes TRU64 now does have the missing  > ingredients,  K Not really, though it did pick up some useful items recently from VMS.  But J AFAIK its 'cluster' file system still exports ('serves') specific portionsG from individual nodes rather than supporting true shared-disk access as K VMS's does, though the ported VMS DLM now allows coordinated client caching F of the data - which means that on node failure the portion of the fileK system that it was serving must be 'failed over' to another node, a process J which my impression is takes significantly longer than a VMS cluster state transition.   <  clustered LSM and filesystem and handles applications in anL > organised frame work which although VMS clusters can do its role your own.  C That could be true:  I don't know what specific support VMS has for H cluster-aware applications, and have heard that various Unixes have doneD some useful things in that area.  Again, if you'd care to elaborate,H specific differences might be useful to people interested in keeping VMS competitive.  G >     The only thing I'm missing is mount verify time out which I think  might K > be there if we could just get those cluster/kernel people to work with an J > operation VMS site they might finally understand why that's a good thing and 
 > include it. G >     Its amazing how slow people are to learn, they never seem to read I > history or what has gone before NIH can is such a problem , our wait it  was 3 > invented here just before and in another group :)   J Not sure at which group that comment might be aimed:  it certainly appliesC to the DEC Unix people, who took 15 years or so to realize that VMS L clustering had a great deal to offer them, and it also applies to VMS, whichI has failed to realize for a long time that increasing system memory sizes I have made Unix-style central caching a major win in many cases (though it C almost certainly would have been the wrong choice back when VMS was 
 designed).  J Always good to hear from you cheeky Aussies:  this group tends to be a bit in-bred, and can use the input.    - bill   >  >  >     Cheers >         Mark >  > ps. : >     not to be seen to have anything to do with Andrew ;- >  > Solaris sucks ;)B > VMS still has better clustering that SUN can only dream about ;) >  > :):):):):):) >    ------------------------------  % Date: Wed, 31 Jan 2001 19:28:16 -0600 ; From: Lord Running Clam <Use-Author-Address-Header@[127.1]> ' Subject: Attn: The COV / "The Q" / Rich / Message-ID: <PostcardFrom.Ganymede@nowhere.nil>   " -----BEGIN PGP SIGNED MESSAGE-----   Hi,   H First off, my thanks for taking the time to reply, and my apologies if IK have missed any message(s) that is/are telling me to get lost. <g> Methinks  Not!  F Let me quickly follow that up with an apology for the cloak and daggerK nonsense.  [As you now know Bill, your reply was delayed considerably and I K posted something anyway.]  I had already considered the risks involved with G the inclusion of The COV on the groups I communicated with.  Instead, I J provided the document to the group as a dead-end link, and made no mentionD of VMS whatsoever.  Yet, I'll bet a copy landed on the desks of someG important people who haven't a clue there is a VMS box in the building. - Will they get the message I sought to convey?   I Anyways, I have absolutely no idea who Charlie Bravo is, perhaps if it is I an interesting, or entertaining, tale someone might feel like sharing it. C Naturally, a little conspiracy makes for more edible entertainment.    *  *  *   K Many from The COV lurk or participate anonymously in the data-dealing rooms J where I have been operating.  Some of these people will have no difficultyI whatsoever fingering my real identity; most others are blissfully unaware K of the privacy issues that are amongst my personal concerns.  (I would like F to stress that these are not always nice places, and posting your realE email address, or easily followed links to such, in *any* form can be  unwise).  K What better platform to deal with one of the foundations of online privacy? E The one which we all take for granted is Military Grade in VMS?  That 5 foundation stone, as I am sure you know, is security.   J [Excuse me, getting on dat ole S.Boxx again, but...] I, and the others whoG lurk or participate anonymously in The Nose of your Newsfeed, only make J obscure references to VMS.  It can easily be sold on that (security) basisJ alone, but it is not my job to do it.  It is the job of "The Q", MarketingB Division, and to many members of The COV it appears they have beenI assimilated.  [I for one would be happier to know they were using unix on  the desktop than M$.]   I FYI My "performance" on a certain political talkshow was marred by M$Wird K altering a mis-type of "commercial" - it tried making it "comical".  I only J noted this when I looked it up in the Deja archive.  I didn't bother beingI conceited enough to check if it had an effect on the odds being given for / The Redwood Park Rangers winning the Superbowl.   K As you should be starting to see, there is a night-time convoy full of dark H green trucks, and they're full of analogies that would turn VMS into anyH anal-ist's best bet.  These analysts need advertised to.  Uncertainty inH *new* technology stocks?  How old did you say VMS was?  How long did youJ say that system had been running?  [Note, I would emphasise system in thisI context, not a single box, and not a single interactive process.  Andrew, J you design a *system* to work in the real world, as I would hope you know.D Don't pick fights, or I'll ask if I can give you one of my remaining$ 0.02cents tokens (only kidding ;-)].  K I feel that "The Q" need to look at what companies such as Oracle and Cisco F are doing, and act according to what they find.  When Oracle specify aH significant percentage of the Wheel of Fortune's greatest hits use theirD software, I wonder what percentage use VMS.  The product can be soldI without telling a single lie, in fact, I'd like to think it could be sold  by being brutally honest.   K They should not be afraid to point out current performance issues that have J been discussed by The COV either.  As said, be blunt.  Ask the prospectiveJ customer how much they value the reputation of their business.  Whose leadE would they choose to take when security and privacy has become such a  concern?    # Get the screen-cleaner flocks . . .     G I'm not suggesting T.V. adverts showing a VAX inna tank, but you should  get the general idea. . .     0         When M$ says, "You will be assimilated."  *         A VMS system is your best defense.    D Now, be sure to wipe the coffee and fingerprints off the monitor andC keyboard before showing this to your boss.  With a little help from 5 Outbreak Express, the news should travel fairly fast.     	 Ganymede.  - --  6 The bigger the humbug, the better people will like it. ~ Phineas Taylor Barnum.    * - ----- BEGIN 3EBG39 ENCODED MESSAGE -----   Gryy uvz vg'f na rznvy ivehf. 7 NAQ PHG NYY URNQREF SEBZ GUR ZRFFNTR ORSBER SBEJNEQVAT!   ( - ----- END 3EBG39 ENCODED MESSAGE -----   -----BEGIN PGP SIGNATURE-----C Version: N/A  @ iQEVAwUBOniY8oer+ijnZohVAQHKMwf+Ilt20psC4nA8nCtvbNyBviE31ep2I2pP@ dYQTsGLSrXPhxrHRLdszbwrEllnTZxPlvmAPhah0bOUJjS++7DHlzht5EvxgMD+p@ yuhOkAeTz/PZ/80IFwoaJlnTxDH2+8nYX0GVJ+/Nkj1YE+57So/kjNHA07tfWCo9@ 5zUoB1xY0axERVG1OvmGGuxsCBmi6azVMnaR7PvHAQTLwAIZdONRMu13G2hB63YH@ XXhZAzdWhVBqa+aLLwTWwzL9auH9hCUHSwqw4+COP9T4w9a21YDq5NpvSGOQauMk8 1GjsK6UzQpn23nTW8Cf0tb14k5/iQmnH3dvi6I8ks7yT4smR5IFwzQ== =zY+ja -----END PGP SIGNATURE-----    ------------------------------  % Date: Wed, 31 Jan 2001 15:33:20 -0500t  From: norm.raphael@jamesbury.com& Subject: Re: Backup Problem... SUMMARY4 Message-ID: <C22569E5.007220C1.00@jklh21.valmet.com>  D ([N]ow how did that person know i was running perfectdisk???? ;-)  )  K Well this is a file created and deleted by PerfectDisk in normal operation.t  O %BACKUP-E-OPENIN, error opening ANACIN$DUA2:[000000]PD_FREE_FILE.DAT;1 as inputn! -SYSTEM-W-NOSUCHFILE, no suchfilen    F What happened is that you were running BACKUP and PD at the same time.A On the initial sweep, PD_FREE_FILE.DAT existed.  PD finished and,eF as is supposed to happen, PD_FREE_FILE.DAT was gone.  This undoubtedly
 upset BACKUP.V   This is not a real problem.t   [...with a nod to JH]v          + agnew@hsc.vcu.edu on 01/31/2001 10:52:49 AM    To:   Info-VAX@mvb.saic.comm cc:g' Subject:  Re: Backup Problem... SUMMARYf          M Dear all, I'm responding to the NET, and privately to the people who helped..g  D My problem was the following error message about backup-w-fidnotfnd.  M The common consenus is that BACKUP is whining about a file not being there oni6 the verify pass, and/or a bug in backup is interfering  with the recognition of the FID.  ! Also noted was the possiblity of:W  ! ANALYZE/DISK/REPAIR being needed,)  I and stopping the perfectdisk process. (now how did that person know i wasv running perfectdisk???? ;-)  )  O analyze had been run, (disk was ok), so I suspect either it's an old backup bugt and/or perfectdisk is running.  P All in all, I think i may play around with perfectdisk not running during backup hours..r   Thanks to one and all!!!!      Jimo Jim Agnew wrote: >  > Dear all netbeings,/ > M > Hello and warm fuzzies and whatnot.  I've been seeing this error off and on-4 for a long time, and finally am getting tired of notO > knowing what it's about exactly.  Since this is in a daily job, and our filese5 turn over slowly, I've not really worried about this.s > N > Can anyone enlighten me as to what backup-w-fidnotfnd really means? it's not5 mentioned in the system error message util except foroI > "contact support", basically, and we don't have support.  oh, yeah, vmsn 5.5-2... >  > Jim Agnewi >sK > %BACKUP-E-OPENIN, error opening ANACIN$DUA2:[000000]PD_FREE_FILE.DAT;1 as - input           -SYSTEM-W-NOSUCHFILE, no suchcL > file                                                %BACKUP-I-STARTVERIFY, starting verificationP > pass8 > %BACKUP-E-VERIFYERR, verification error for block 3 of  ANACIN$DUA2:[000000]000000.DIR;1P > %BACKUP-W-FIDNOTFND, internal error. File to be processed not found on process list >t< > %BACKUP-I-STARTRECORD, starting backup date recording pass > $       SETUP ZIPh   ------------------------------  % Date: Wed, 31 Jan 2001 17:57:29 -0600l+ From: "Main, Kerry" <Kerry.Main@COMPAQ.com> = Subject: RE: Compaq: A simple, affordable clustering solution N Message-ID: <910612C07BCAD1119AF40000F86AF0D805284D11@kaoexc3.kao.cpqcorp.net>   Andrew,t   Please re-read previous posts.    J This is getting boring as you keep bring up stuff that has been hashed out before.n  K >>> Quite so how do you square this with Kerrys, logged out users scenario,VL surely getting users to log out is manual intervention to the nth degree.<<<  L You know perfectly well this applies to users who stay logged in for days at a time.   : How many users on eBay stay logged in for days at a time?    Defence rests.  K >>> I am simply pointing out that the claims that Kerry and jlsue have beeneL making for OpenVMS clusters are not realistic and belong only in a marketing
 slide. <<<  K More old stuff. So, if this is all marketing, I guess I should not tell thedJ Customers that are doing this today that what they are doing is impossible eh?t   ROTFL.   Regards,  
 Kerry Main Senior Consultantn Compaq Canada Inc. Professional Servicesu Voice: 613-592-4660  Fax  :  819-772-7036 Email: Kerry.Main@Compaq.com     -----Original Message-----7 From: andrew harrison [mailto:andrew.nospam@uk.sun.com]o Sent: January 31, 2001 11:28 AMt To: Info-VAX@Mvb.Saic.Comr= Subject: Re: Compaq: A simple, affordable clustering solutionb     shielm00 wrote:  >  > Ok Folks,a > K > I hate to do this, as I know how the group feels about AH's postings, but  as@ > it was specfically stated in his last post, I have a question. > 9 > In response to our favourite architects latest missive.h > ? > "andrew harrison" <andrew.nospam@uk.sun.com> wrote in message-% > news:3A75A4CF.96B29AE@uk.sun.com...: >  > Various bits deleted > 7 > > You really don't get it do you. Firstly in terms oft7 > > availability having 80% of apps surviving a re-bootG7 > > of the node they are on is hardly going to convincem7 > > anyone that you have an HA platform, if it was thata" > > easy even MS could claim that. > J > How many cluster apps survive a reboot of a Sun cluster 2.2 node on that > node. 80%, 70%, 60% ?  > I > I think the answer will surprise a few people and give you some idea ofo gapeH > that existed between VMS and Sun Clusters ( 4 nodes and they call it aJ > cluster). They may well have fixed this in Suncluster V3.0, but there is( > always the *.0 release to be aware of. >   4 I don't think you grasped the point. I am not making4 any claims for SunClusters, I am simply pointing out5 that the claims that Kerry and jlsue have been making 6 for OpenVMS clusters are not realistic and belong only in a marketing slide.t  7 In addition if you allow the kind of lenience required O5 to get the level of uptime claimed by Kerry, logging l4 users out etc then you can also get the same uptime  on other non OpenVMS clusters.  4 > There is a catch, no manual intervention allowed ! >   4 Quite so how do you square this with Kerrys, logged 3 out users scenario, surely getting users to log out') is manual intervention to the nth degree.a   Regardst Andrew Harrison  Enterprise IT Architecto   ------------------------------  % Date: Wed, 31 Jan 2001 21:18:47 -0500m0 From: "Sharkonwheels" <tonym@compusourceDOT.net>! Subject: Console cables availableh, Message-ID: <i14e6.187$Go6.113973@news2.mco>  H Seeing as how this is an issue, I have been making console cables. TheseG will be modular, meaning I'll include a dongle for use on MMJ ports, aslD well as a dongle for use on the MicroVax/Vaxstation 3xxx series DB-9E connector. The other end will be terminated with a DB-9 connector for  use on a PC.  H I know MY major problem in starting out, was the console cable. Hell, it tookF me 1 week just to order in the MMJ connectors, and they weren't cheap! Special crimper, as well!c  G Price will be $28.00 shipped in the continental US. I will include somee console J instructions gathered from the net, and I will *MOST* certainly pay homage toJ where the stuff was gathered from, because without all these guys out here	 archivingo; knowledge, *I* wouldn't even know how to get into my Vaxen.c  * I'd like to send thanks out in advance to:7 Steven M. Schweda - You know how much help you gave me!-K Dr. Henry G. Juengst, University of Minnesota - answering my questions LATEt into the night! K VaxArchive.Org - What can I say? Can't be easy to maintain that large of ane archive!" Megan Gentry - For the Boards list    6 I can accept PayPal payments, as well as money orders.( Paypal address is: tonym@compusource.net     Tony tonymATcompusourceDOTnet   ------------------------------  # Date: Thu, 01 Feb 2001 00:51:31 GMT - From: "Dave Pampreen" <davepampreen@home.com> # Subject: Re: Create a random numberF> Message-ID: <nq2e6.269510$hD4.65084708@news1.rdc1.mi.home.com>   All,  F Here is a snip-it of what I used to get a auto-generated password from everyone's input:T  . $ temp_file = "sys$scratch:temp_passwords.txt"H $ if f$search ("''temp_file'") .nes. "" then delete /nolog 'temp_file';** $ define /user_mode sys$output 'temp_file' $ set password/gen=10  INVALID  $  $ open /read infile 'temp_file'a! $ read infile inline ! Blank linen4 $ read infile inline ! First auto-generated password $ close infile $ H $ if f$search ("''temp_file'") .nes. "" then delete /nolog 'temp_file';* $n, $ write sys$output "Password is /''inline'/"     Dave  . "Dave P" <davep@hmgcc.gov.uk> wrote in message" news:3A7808DF.341C@hmgcc.gov.uk... > Arne Vajhj wrote: > >  > > Dave Pampreen wrote:F > > > Does anyone have a program which creates random passwords?   I'd prefer aC > > > Fortran example, but I can convert it from most any language.h > > 4 > > A very simple example in Pascal is available at: > >m7 > > http://www.hhs.dk/anonymous/pub/vms/misc/pw_gen.pasr > >y > > Arne >n  > Do you mean numeric passwords? >fJ > VMS has an excellent generator for 'pronounceable' passwords. For how toI > access this programmatically, search for SYS$FORGE_WORD in the archives  > of this group on DejaNews. >t > Dave   ------------------------------  % Date: Wed, 31 Jan 2001 17:27:06 -0500i) From: Rick Barry <barry@star.zko.dec.com> F Subject: Re: CSWS/Apache Log File Naming (WAS Re: CSWS Authentication)0 Message-ID: <3A78913A.C37DCDAF@star.zko.dec.com>  * "j.lance wilkinson, (814) 865-1818" wrote:  ^ > In article <3A76CC02.6F6A8F0F@star.zko.dec.com>, Rick Barry <barry@star.zko.dec.com> writes: > >a  > >"Barratt, Chris (FMC)" wrote: > >uK > >> I was wondering if there is currently (or planned for future) a way tooO > >> specify an identifier from SYSUAF to be used as the group authenticator onIP > >> acls ?   (eg. grant an identifier to a vms user which would then be used to% > >> allow access to a web resource).  > >uK > >It's on our to-do list. We've had a number of requests for that feature.e > >s > >--b
 > >Rick Barryp >iM >         OK, here's another enhancement that I would like to see, or if it'soK >         not practical for COMPAQ to implement, perhaps some way for me toh* >         implement it would be desirable: >pO >         In a successful production environment, I can see CSWS/Apache runningnL >         nonstop for months on end.  The sheer size of the log files such aL >         successful run could introduce might be staggering.  I like to runM >         very detailed logs, but after a month I'd rather run them thru some O >         kind of analysis and discard them.  I can't discard them, however, ifeO >         the web server still has them busy, and that's not the kind of  thingo7 >         I want to bring the server down to deal with.d > P >         I would like to be able to specify a Date/Time string in the name of aN >         log file (ANY log file) in the configuration file. The actual syntaxN >         for definition and token selection is irrelevant, but I'd want it toN >         be sufficiently transparent to be self documenting. For example, sayB >         that %Y means insert the year, %M the month,  and so on: >  >         #b' >         ErrorLog logs/error%Y%M%D.log  > O >         This would cause any errors reported to this log file on today's date 5 >         (30 January 2001) to be logged to the file:- >-9 >                 APACHE$SPECIFIC:[LOGS]error20010130.logd >nO >         When the clock ticks over to January 31st, the next entry in this log/, >         file would go to a different file: >e9 >                 APACHE$SPECIFIC:[LOGS]error20010131.logs >fL >         and so on.  Including Hours, Minutes and Seconds would possibly beN >         absurd, but at an extremely busy site just %H for the hour might notM >         be so silly.  Just using %Y%M would provide a monthly rather than asL >         daily log, and perhaps we could say %Y%W would cycle the log every6 >         week (choose a day appropriate to cycle it). >tP >         Lacking the syntax-specified way of getting the file cycled, how aboutJ >         a module that is sufficiently intimately aware of the log files'I >         details that it could cycle them according to a defined period?  >pO > +-"Never Underestimate the bandwidth of a station wagon full of mag tapes"--+sO > | J.Lance Wilkinson ("Lance")            InterNet:  Lance.Wilkinson@psu.edu |iO > | Systems Design Specialist - Lead       AT&T:      (814) 865-1818          |kO > | Library Computing Services             FAX:       (814) 863-3560          |tO > | E3 Paterno Library                     "I'd rather be dancing..."         |eO > | Penn State University         A host is a host from coast to coast,       |fO > | University Park, PA 16802     And no one will talk to a host that's close | O > | <postmaster@psulias.psu.edu>  Unless the host that isn't close            |oO > | VMS GopherMeister             Is busy, hung or dead.                      |sO > +------"He's dead, Jim. I'll get his tricorder. You take his wallet."-------+e; >                 [apologies to DeForest Kelley, 1920-1999]a4 > <A Href="http://perdita.lcs.psu.edu">home page</a>L > <a Href="http://perdita.lcs.psu.edu/junkdec.htm">junk mail declaration</a>  Q That's an interesting idea, but one which would require changing some fundamentaloQ operations in the standard Apache server. Not that it's bad, it just goes againstcH our current design principal of keeping as much Apache code unchanged as	 possible.o  U Rotating log files by gracefully restarting the server  is common practice for ApacheVO administrators, but we'd like to be able to do better on OpenVMS. Consider thisyO approach:  a "new log" command that the administrator can issue to force Apache Q close/re-open  the log files without restarting the server. It may be possible tot: implement this solution without significant restructuring.   --
 Rick Barry  3 Compaq Secure Web Server (CSWS) - Development Grouph Compaq Computer Corporatione
 Nashua, NH   ------------------------------  % Date: Wed, 31 Jan 2001 23:48:16 -0500l2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>B Subject: CSWS/Apache Log File Naming (WAS Re: CSWS Authentication)7 Message-ID: <200101312348_MC2-C3DA-C443@compuserve.com>n  4 Message text written by INTERNET:jlw@psulias.psu.edu< >In article <3A76CC02.6F6A8F0F@star.zko.dec.com>, Rick Barry  <barry@star.zko.dec.com> writes: >o >"Barratt, Chris (FMC)" wrote: >xJ >> I was wondering if there is currently (or planned for future) a way to=  J >> specify an identifier from SYSUAF to be used as the group authenticato= rf onJ >> acls ?   (eg. grant an identifier to a vms user which would then be us= ed to# >> allow access to a web resource).  >kJ >It's on our to-do list. We've had a number of requests for that feature.=   >t >--  >Rick Barryr  J         OK, here's another enhancement that I would like to see, or if it= 'sJ         not practical for COMPAQ to implement, perhaps some way for me to=  (         implement it would be desirable:  E         In a successful production environment, I can see CSWS/Apachel runningaJ         nonstop for months on end.  The sheer size of the log files such = a J         successful run could introduce might be staggering.  I like to ru= nlJ         very detailed logs, but after a month I'd rather run them thru so= meJ         kind of analysis and discard them.  I can't discard them, however= ,f ifH         the web server still has them busy, and that's not the kind of =   thingl5         I want to bring the server down to deal with.s  J         I would like to be able to specify a Date/Time string in the name=   of aE         log file (ANY log file) in the configuration file. The actuald syntaxJ         for definition and token selection is irrelevant, but I'd want it=   toH         be sufficiently transparent to be self documenting. For example, sayb@         that %Y means insert the year, %M the month,  and so on:  	         #E%         ErrorLog logs/error%Y%M%D.loga    H         This would cause any errors reported to this log file on today's date3         (30 January 2001) to be logged to the file:G  7                 APACHE$SPECIFIC:[LOGS]error20010130.logw  J         When the clock ticks over to January 31st, the next entry in this=   logD*         file would go to a different file:  7                 APACHE$SPECIFIC:[LOGS]error20010131.logu  J         and so on.  Including Hours, Minutes and Seconds would possibly b= e =O  H         absurd, but at an extremely busy site just %H for the hour might not =   J         be so silly.  Just using %Y%M would provide a monthly rather than=  a  J         daily log, and perhaps we could say %Y%W would cycle the log ever= yc4         week (choose a day appropriate to cycle it).  H         Lacking the syntax-specified way of getting the file cycled, how about?H         a module that is sufficiently intimately aware of the log files'H         details that it could cycle them according to a defined period?<  F         Would you settle for something simpler and more traditional? =   Something like $ <mumble> SET LOG /NEW_FILE  J         I'm not an APACHE user so I haven't a clue how it is managed, hen= ce' the <mumble> in the above command line.E  G         Doing it this way would seem to require a minimum of effort andlJ offer a maximum of flexibility; it's absolutely up to you when you want t= o:J create a new log file.   The granularity of the time specification become= sUJ the granularity of the ASCII form of VMS time.  Like you, I see no reason=  J why anyone would want to specify the minutes or seconds or centiseconds b= utH the possibility is there with no extra effort.  And if E-bay should ever5 come to its senses and switch to VMS, they might wantc  a new log file every 30 seconds.   ------------------------------  % Date: Wed, 31 Jan 2001 20:49:43 +0000P% From: Alan Greig <a.greig@virgin.net>p Subject: Re: Dave Cutler* Message-ID: <3A787A66.D9224F44@virgin.net>   Christof Brass wrote:l  @ > I heard that Dave Cutler left Micro$oft long ago. Does anybody > where he's currently working?u  I Nope Cutler was still in charge at the release of Windows 2000.  I recalldI reading an interview not that long ago. I'll try and track down where. InrJ that interview he stated that Win-64 development was still taking place onC Alpha/NT until they got some real IA64 production quality hardware.yI Definitely read it in the last year so if he has left Microsoft it's very 	 recently.j  = In fact a search of www.microsoft.com turns up the following:u  9 DAVE CUTLER, SR.DISTINGUISHED ENGINEER, WINDOWS BASE TEAMi  C                Dave Cutler, Sr. joined Microsoft in 1988. Currentlyt! responsible for the design of the,H                64-bit release of the Windows Operating System, Cutler is generally considered one ofNC                the top few programmers worldwide. After arriving at  Microsoft, he launched theI                Windows NT group and has since then led the development of- three major releases ofrM                the product, now known as Windows 2000. In addition to leadingN the Windows 2000K                team, Cutler contributed to the architecture of all parts of  the system, and even wroteD                the kernel himself. In recognition of his significant& contributions to the field, Cutler wasK                awarded membership in the National Academy of Engineering inM 1993. Prior to joiningE                Microsoft, Cutler worked at DEC, where he designed and9 delivered several successfulH                operating systems, including VAX/VMS, RSX-11M and VAXELN. Originally from Dewitt,nL                Michigan, he holds over 20 patents and is affiliate professor in the Computer Sciencee:                Department at the University of Washington.     --
 Alan Greig   ------------------------------  % Date: Wed, 31 Jan 2001 14:51:31 -0700s% From: Dean Woodward <deanw@rdrop.com>r Subject: Re: Dave Cutler) Message-ID: <3A7888E3.77CF6012@rdrop.com>a   Alan Greig wrote:  >  > Christof Brass wrote:n > B > > I heard that Dave Cutler left Micro$oft long ago. Does anybody! > > where he's currently working?  > K > Nope Cutler was still in charge at the release of Windows 2000.  I recalleK > reading an interview not that long ago. I'll try and track down where. IntL > that interview he stated that Win-64 development was still taking place onE > Alpha/NT until they got some real IA64 production quality hardware.6  G Which is why I keep hoping against hope that Win64 will be supported byOG Compaq on Alpha, since that's where all the development happens anyway.c  	 <heresey>aH Actually, WinNT on Alpha is relatively stable (MS-spin = "Rock solid"). E The only reason I'm not a fan of it is it's frustrating to spend that C much for h/w and then cripple it to 32-bit mode by installing NT...o
 </heresey>   ------------------------------  # Date: Wed, 31 Jan 2001 23:58:45 GMTe; From: Mark Garrett <Mark.Garrett@wedontwantyourspam.com.au>m Subject: Re: Dave CutlerC Message-ID: <B69EF002.10E80%Mark.Garrett@wedontwantyourspam.com.au>.  H in article 3A7888E3.77CF6012@rdrop.com, Dean Woodward at deanw@rdrop.com wrote on 01/02/2001 08:51:   > Alan Greig wrote:iI > Which is why I keep hoping against hope that Win64 will be supported byhI > Compaq on Alpha, since that's where all the development happens anyway.s  K I do hope your wrong. Having to keep console firmware and device support in L the alpha line to fit with Microsoft must have been driving people nuts. VMSK and TRU64 have very similar console requirements unlike this stupid desktop K OS from Microsoft. God help us if we need to return to that, we would stilll? be waiting for system. Can't believe they got to the point of arL keyboard/mouse and VGA on 8400 Alpha's what a waste of effort. GS320 is much better with out it.     
     Cheers         Mark :)s   ------------------------------  % Date: Wed, 31 Jan 2001 18:21:21 -0700m% From: Dean Woodward <deanw@rdrop.com>a Subject: Re: Dave Cutler) Message-ID: <3A78BA11.A88CB64B@rdrop.com>a   Mark Garrett wrote:c [minor attribution fixing...]o > J > in article 3A7888E3.77CF6012@rdrop.com, Dean Woodward at deanw@rdrop.com > wrote on 01/02/2001 08:51: > K > > Which is why I keep hoping against hope that Win64 will be supported by K > > Compaq on Alpha, since that's where all the development happens anyway.t > M > I do hope your wrong. Having to keep console firmware and device support inrJ > the alpha line to fit with Microsoft must have been driving people nuts.  C I don't care if any given Alpha box can boot VMS/Tru64 _and_ NT.  ImG think the Alpha CPU line will have a better chance of life if more OS's1! support it- even the evil one(s).a  N > keyboard/mouse and VGA on 8400 Alpha's what a waste of effort. GS320 is much > better with out it.   G Sequent systems (paralell x86 machines) have keyboard & video ports- soV= they can be booted to DOS and config'd.  No, I'm not kidding.r   ------------------------------  % Date: Wed, 31 Jan 2001 20:43:20 -0600 + From: "Main, Kerry" <Kerry.Main@compaq.com>  Subject: RE: Dave CutlerN Message-ID: <910612C07BCAD1119AF40000F86AF0D805284D13@kaoexc3.kao.cpqcorp.net>   Dean,c  G >>> I don't care if any given Alpha box can boot VMS/Tru64 _and_ NT.  IlG think the Alpha CPU line will have a better chance of life if more OS'so$ support it- even the evil one(s).<<<  ? Ok, how about the following summary of Alpha Supported OS's :-)g   Under the Linux label -o   -	SuSE Linux -	TurboLinux -	Red Hat Linuxa -	Debian Linux -	Linux-Mandrake -	Slackware Linuxh -	Stataboware (Asia) -	Kondara (Asia) -	Accel (Asia)A -	Latest Cray Supercomputer - API NetWorks and Cray Inc. Announce , Strategic Alliance (Supercomputer Linux)    G http://www.alpha-processor.com/pressreleases/pr012901.shtml (Jan 29/01)i  -	Linux NetworX (Linux clusters)?     http://www.alpha-processor.com/pressreleases/pr080700.shtmle   Under Compaq Offering's:   -	Tru64 UNIX	 -	OpenVMS- -	SEVMS-	 -	VxWorks / -	Himalayan NSK NonStop (future platform - EV7)    Under the UNIX 'BSD labels:m   -	NetBSD	 -	FreeBSD 
 -	OpenBSD    Under the Misc label:2   -	Windows NT4 SP6a and below -	Windows 2000 RC2 and below        
 Kerry Main Senior Consultantg Compaq Canada Inc. Professional Services  Voice: 613-592-4660s Fax  :  819-772-7036 Email: Kerry.Main@Compaq.com     -----Original Message-----, From: Dean Woodward [mailto:deanw@rdrop.com] Sent: January 31, 2001 8:21 PM To: Info-VAX@Mvb.Saic.Comn Subject: Re: Dave Cutler     Mark Garrett wrote:B [minor attribution fixing...]e > J > in article 3A7888E3.77CF6012@rdrop.com, Dean Woodward at deanw@rdrop.com > wrote on 01/02/2001 08:51: > K > > Which is why I keep hoping against hope that Win64 will be supported byDK > > Compaq on Alpha, since that's where all the development happens anyway.. > J > I do hope your wrong. Having to keep console firmware and device support inJ > the alpha line to fit with Microsoft must have been driving people nuts.  C I don't care if any given Alpha box can boot VMS/Tru64 _and_ NT.  IlG think the Alpha CPU line will have a better chance of life if more OS'su! support it- even the evil one(s).   I > keyboard/mouse and VGA on 8400 Alpha's what a waste of effort. GS320 isr much > better with out it.w  G Sequent systems (paralell x86 machines) have keyboard & video ports- so = they can be booted to DOS and config'd.  No, I'm not kidding.-   ------------------------------  % Date: Thu, 01 Feb 2001 02:46:46 +0000m) From: Christof Brass <brass@infopuls.com>j Subject: Re: Dave Cutler, Message-ID: <3A78CE16.8E0246AF@infopuls.com>   Alan Greig wrote:D >  > Christof Brass wrote:l > B > > I heard that Dave Cutler left Micro$oft long ago. Does anybody! > > where he's currently working?r > K > Nope Cutler was still in charge at the release of Windows 2000.  I recallIK > reading an interview not that long ago. I'll try and track down where. InsL > that interview he stated that Win-64 development was still taking place onE > Alpha/NT until they got some real IA64 production quality hardware.uK > Definitely read it in the last year so if he has left Microsoft it's verys > recently.a > ? > In fact a search of www.microsoft.com turns up the following:  > ; > DAVE CUTLER, SR.DISTINGUISHED ENGINEER, WINDOWS BASE TEAM  > E >                Dave Cutler, Sr. joined Microsoft in 1988. Currentlys# > responsible for the design of thenJ >                64-bit release of the Windows Operating System, Cutler is > generally considered one ofnE >                the top few programmers worldwide. After arriving ate > Microsoft, he launched theK >                Windows NT group and has since then led the development ofi > three major releases of O >                the product, now known as Windows 2000. In addition to leadinga > the Windows 2000M >                team, Cutler contributed to the architecture of all parts ofn > the system, and even wroteF >                the kernel himself. In recognition of his significant( > contributions to the field, Cutler wasM >                awarded membership in the National Academy of Engineering ina > 1993. Prior to joiningG >                Microsoft, Cutler worked at DEC, where he designed and  > delivered several successfulJ >                operating systems, including VAX/VMS, RSX-11M and VAXELN. > Originally from Dewitt,oN >                Michigan, he holds over 20 patents and is affiliate professor > in the Computer Science < >                Department at the University of Washington. >  > -- > Alan Greig  > Perfect. Many thanks. Now I have a real problem: W2k/WNT might? not as bad as it behaves. I seems that the people who put theirs/ stuff on top (GUI, Registry a.s.o.) screwed up.u   ------------------------------   Date: 31 Jan 2001 18:01:10 GMT' From: "Jim Strehlow" <jims@data911.com>   Subject: Re: defragment products0 Message-ID: <959jt6$60c@dispatch.concentric.net>  0 How well does Compaq's Disk File Optimizer (DFO)* handle multiple nodes using the same disk?* The nodes are not necessarily "clustered".: They do not share memory nor failover when a node is down.C Multiple nodes simply use some of the same disks in a raid 5 array.t  G Do I need to dismount the disk from other nodes to be sure they are note usinga- the disk while one node defragments the disk?n  8 Is that simply documented in the main DFO user's manual?   Thank you for your responses.e   Jim Strehlow, jims@data911.com Systems Managern   ------------------------------  % Date: Wed, 31 Jan 2001 15:29:11 -0500a% From: "John Vottero" <John@mvpsi.com>n  Subject: Re: defragment products/ Message-ID: <t7gte191kik00d@news.supernews.com>8  E Two VMS systems which are not clustered cannot share a disk.  Run the K defragger on whichever system actually has the disk mounted.  If you reallybE have a configuration which somehow let you mount the same disk on two / unclustered systems, you are playing with fire.y  2 "Jim Strehlow" <jims@data911.com> wrote in message* news:959jt6$60c@dispatch.concentric.net...2 > How well does Compaq's Disk File Optimizer (DFO), > handle multiple nodes using the same disk?, > The nodes are not necessarily "clustered".< > They do not share memory nor failover when a node is down.E > Multiple nodes simply use some of the same disks in a raid 5 array.  >AI > Do I need to dismount the disk from other nodes to be sure they are not. > usingd/ > the disk while one node defragments the disk?s > : > Is that simply documented in the main DFO user's manual? >d > Thank you for your responses.a >A  > Jim Strehlow, jims@data911.com > Systems Managerc >u >u >n   ------------------------------    Date: 31 Jan 2001 17:10:16 -0500, From: koehler@eisner.decus.org (Bob Koehler)  Subject: Re: defragment products+ Message-ID: <lIsZ04+fTIg4@eisner.decus.org>   W In article <t7gte191kik00d@news.supernews.com>, "John Vottero" <John@mvpsi.com> writes:   > > Two VMS systems which are not clustered cannot share a disk.  G Partly true.  Old MASSBUS disks could be shared between two VMS systemssG prior to clustering.  But sharing was limitted to one system writeable,c3 one system read-only.  Requires VMS 3.0 or later.  m  G Eventually a both-writeable capability was added for these disks if thee systems were clustered.0  F ----------------------------------------------------------------------? Bob Koehler                     | Computer Sciences Corporation2= NASA GSFC Flight Software       | Federal Sector, Civil GrouptE                                 | please remove ".aspm" when replying.   ------------------------------  # Date: Thu, 01 Feb 2001 01:57:02 GMT,$ From: Scott Vieth <svieth@wi.rr.com>  Subject: Re: defragment products( Message-ID: <3A78C2AE.2070405@wi.rr.com>   Jim Strehlow wrote:U  2 > How well does Compaq's Disk File Optimizer (DFO), > handle multiple nodes using the same disk?, > The nodes are not necessarily "clustered".< > They do not share memory nor failover when a node is down.E > Multiple nodes simply use some of the same disks in a raid 5 array.   B How can you have multiple nodes access the same disks without themG being clustered?  You *need* clustering in order to make everyone aware 2 of one another and to mediate access to the disks.  F Now, if you're talking NT, that's a different story.  I built a setup  with two NT 4.0 G servers and an ESA12000 where both servers had access to the same units H as the same time.  Not pretty!!  It led to some pretty cool errors that 
 I've never seen NT display before.u   -Scott, Vinca veteran  ;^)   ------------------------------   Date: 31 Jan 2001 18:10:44 GMT3 From: gartmann@immunbio.mpg.de (Christoph Gartmann)e# Subject: Re: Error message over LATn0 Message-ID: <959kf4$ebl$2@n.ruf.uni-freiburg.de>  j In article <I3Wd6.8911$Q94.163413@monolith.news.easynet.net>, "Paul Oswald" <paul.oswald@bocs.com> writes:L >Could anybody help please. I have a BOCA ticket printer attached to a DS90MI >DECserver. I am getting an intermittent error message along the lines of' >a@ >'Duplicate ARP source address 172.17.254.10 received from 00-'. >iJ >This IP address belongs to a Hub on the PC LAN. I have attached a copy of	 >the port: >settings for reference.  L Sounds more as if something on the net is using the same TCP/IP? or Ethernet address.   Regards,    Christoph Gartmann   H -- --------------------------------------------------------------------+H | Max-Planck-Institut fuer      Phone   : +49-761-5108-464   Fax: -452 |H | Immunbiologie                                                        |H | Postfach 1169                 Internet: gartmann@immunbio.mpg.de     |H | D-79011  Freiburg, FRG                                               |H +--------- http://www.immunbio.mpg.de/home/english/menue.html ---------+   ------------------------------  % Date: Thu, 01 Feb 2001 01:53:30 +0000l) From: Christof Brass <brass@infopuls.com>d< Subject: Re: Firmware upgrade for PWS 600au to use a ZLXp-L1, Message-ID: <3A78C19A.22345FA3@infopuls.com>   Christoph Gartmann wrote:t > Z > In article <3A77483C.4AA9D5BD@infopuls.com>, Christof Brass <brass@infopuls.com> writes:: > >I got the information that there was a firmware upgrade> > >available for the 600au to support the high quality ZLZp-L1@ > >graphics card with VMS. Unfortunately this firmware which has? > >been used for another 600au went away with the death of thatA? > >box. Therefore the firmware is gone at the moment and CompaqmC > >doesn't know about and recommends using one of the supported lowe > >end graphics cards. > >f@ > >Is there anybody around who knows where to get this firmware?@ > >I'm not asking for service free of charge though I appreciate > >moderate prices!i > = > ftp://ftp.digital.com/pub/Digital/Alpha/firmware/index.htmle > 
 > Regards, >    Christoph Gartmannd > J > -- --------------------------------------------------------------------+J > | Max-Planck-Institut fuer      Phone   : +49-761-5108-464   Fax: -452 |J > | Immunbiologie                                                        |J > | Postfach 1169                 Internet: gartmann@immunbio.mpg.de     |J > | D-79011  Freiburg, FRG                                               |J > +--------- http://www.immunbio.mpg.de/home/english/menue.html ---------+  = Perfect! Thanks a lot. At the moment I try to figure out whate= the differences are between 600a (available on this page) ande 600au (not mentioned there).   ------------------------------  # Date: Thu, 01 Feb 2001 06:00:12 GMTr From: Dirk Munk <munk@home.nl>< Subject: Re: Firmware upgrade for PWS 600au to use a ZLXp-L1' Message-ID: <3A78FB6B.26E8603C@home.nl>   ? Take care !! You may be looking at the wrong firmware update !!i  % This is the right page for the 600au:   G ftp://ftp.digital.com/pub/Digital/Alpha/firmware/readmes/digitalpw.htmlI   Regards,   Dirk   Christof Brass wrote:    > Christoph Gartmann wrote:o > >M\ > > In article <3A77483C.4AA9D5BD@infopuls.com>, Christof Brass <brass@infopuls.com> writes:< > > >I got the information that there was a firmware upgrade@ > > >available for the 600au to support the high quality ZLZp-L1B > > >graphics card with VMS. Unfortunately this firmware which hasA > > >been used for another 600au went away with the death of thateA > > >box. Therefore the firmware is gone at the moment and CompaqIE > > >doesn't know about and recommends using one of the supported lowo > > >end graphics cards. > > >aB > > >Is there anybody around who knows where to get this firmware?B > > >I'm not asking for service free of charge though I appreciate > > >moderate prices!  > > ? > > ftp://ftp.digital.com/pub/Digital/Alpha/firmware/index.html9 > >r > > Regards, > >    Christoph Gartmann  > >rL > > -- --------------------------------------------------------------------+L > > | Max-Planck-Institut fuer      Phone   : +49-761-5108-464   Fax: -452 |L > > | Immunbiologie                                                        |L > > | Postfach 1169                 Internet: gartmann@immunbio.mpg.de     |L > > | D-79011  Freiburg, FRG                                               |L > > +--------- http://www.immunbio.mpg.de/home/english/menue.html ---------+ >u? > Perfect! Thanks a lot. At the moment I try to figure out whata? > the differences are between 600a (available on this page) ande > 600au (not mentioned there).   ------------------------------  % Date: Wed, 31 Jan 2001 14:46:03 -0500C  From: norm.raphael@jamesbury.comA Subject: Re: Formal Letter to the Newsgroup from Richard Marcello 4 Message-ID: <C22569E5.006BD1ED.00@jklh21.valmet.com>   /Sue Skonetski wrote:iO /> In closing, we remain fully committed to OpenVMS.  We continue to execute toAE /> the strategy we put in place two years ago and it is paying stronghN /> dividends.   We are truly grateful for your passion, commitment and loyalty' /> to OpenVMS past, present and future.y /  /[Snip]  /N@ /Sending a letter to this newsgroup is just a way to shut us up.   Not in this millenium  ;-)  !i  A /It isn't a way to show the world that VMS still exists, and that2 /is what is needed.n   ------------------------------  % Date: Wed, 31 Jan 2001 15:54:29 -0500 ' From: "Bill Todd" <billtodd@foo.mv.com>aA Subject: Re: Formal Letter to the Newsgroup from Richard Marcelloe( Message-ID: <959tq3$7aq$1@pyrite.mv.net>  8 JF Mezei <jfmezei.spamnot@videotron.ca> wrote in message& news:3A78554E.1B67E196@videotron.ca... > Sue Skonetski wrote:E > > In closing, we remain fully committed to OpenVMS.  We continue to 
 execute toF > > the strategy we put in place two years ago and it is paying strongG > > dividends.   We are truly grateful for your passion, commitment andf loyaltyn( > > to OpenVMS past, present and future. >iI > Oh, we have no problem beleiving that "we" remain committed to VMS. TheiL > problem is that we don't really know who the "we" are. As long as the "we"I > remains a small group without much power at Compaq, then the "we remaineH > committed" doesn't hold much weight, and Compaq's "real" announcements will > continue to ignore VMS..  L Exactly.  The 'fullness' of Compaq's corporate commitment to VMS can be seenL by comparing VMS promotion with that of Compaq's other offerings.  And whileJ statements from some of Rich's superiors might carry more weight, the only1 *real* measure of commitment is corporate action.n  G Furthermore, despite all the glowing descriptions of VMS successes, theeF bottom line is that the business grew in single digits year over year.G While that's better than shrinkage, it's hardly stellar:  for every few K wins, there were obviously somewhere nearly comparable losses.  Growth at a L lower rate than that of its competition (see Compaq's own numbers in relatedI segments) is hardly a recipe for long-term success, and hardly merits thet term 'Renaissance'.    - bill   ------------------------------  % Date: Thu, 01 Feb 2001 01:36:37 -0500a2 From: rdeininger@mindspring.com (Robert Deininger)< Subject: Re: From Richard Marcello Vice President of OpenVMSL Message-ID: <rdeininger-0102010136380001@user-2ivechk.dialup.mindspring.com>  Q In article <959g7u$fco@gap.cco.caltech.edu>, mathog@seqaxp.bio.caltech.edu wrote:d    M > We keep talking about marketing OpenVMS.  One thing that would really give aG > it exposure would be if Compaq would fund an independent magazine ...p  | If Compaq funds it, it isn't really independent.  Folks would be suspicious of everything they wrote about their benefactor.   -- o Robert Deininger rdeininger@mindspring.comT   ------------------------------    Date: 31 Jan 2001 18:24:39 -05002 From: malmberg@eisner.decus.org (John E. Malmberg) Subject: Re: FTP to GDG failso+ Message-ID: <VeXO+fIxmjsi@eisner.decus.org>   ? I would recommend that scripted FTP transfers "put" the file to 8 a temporary destination name on the host and then in the  same script do a "rename" after.  8 With the script below, once you get it working, there is: a possibility that the remote host could get an incomplete7 file transfer.  The cleanup from that can be worse thane getting no file at all.n  4 By doing the "put" to a temporary name followed by a: rename, if the file transfer is incomplete, the FTP script7 will be aborted.  Depending on what FTP program you are-7 using, you may be able to check the exit status to takeE remedial action.  9 This is not an OpenVMS specific issue, it is just part ofe+ dealing with the nature of an FTP transfer.b  4 If the exit status is not reliable, you can send the6 output of the FTP session to a suitably protected file- and then look for the successful rename code.0  ; On the remote host, of course you must have the permissions-8 to create the files with both the temporary name and the target name.   -Johno wb8ytw@qsl.network  ' In article <3A771EB1.388A0CB2@fmr.com>,r1 Brad Hamilton <Brad.Hamilton@fmr.company> writes: & > I noticed that our script does this: >n4 > put focas$prod12:[drpt]r347.dat;1 "focas.r347(+1)" > 				  ^              ^( > (double quote instead of single quote) >e > Would that work? >p > "Cokely, Jay" wrote: >>= >> We are having problems FTP'ing from a DEC machine to a GDGw= >> (Generation data set) in an IBM/MVS environment.  FTP's to ? >> non GDG names work fine.   The script below works from othere? >> environments but fails with the following message on the DEC_
 >> processor.h >>F >>         554 Requested action not taken: GDG name conversion failed. >> >> The script used follows: O >> ****************************************************************************  >> $set default sys$logine4 >> $ftp /user=$XYZS02W /pass=??????? n1etso.grum.com >> asciiO >> quote site filetype=seq cylinders primary=50 secondary=25 recfm=fb lrecl=869-* >> put dwhead.txt 'ISDW.ROSS.HEADERIN(+1)' >> exit6O >> ****************************************************************************  >>: >> Any ideas as to why this doesn't work, or where I might) >> find documentation of the "554" error?  >>
 >> Thanks, >> >> Jay Cokely  0655/32 >> PHONE (310) 332-3977r >> FAX      (310) 263-5303   ------------------------------  # Date: Wed, 31 Jan 2001 19:10:59 GMTr$ From: Scott Vieth <svieth@wi.rr.com>" Subject: Re: IP number on an HSG80$ Message-ID: <3A78639C.303@wi.rr.com>  I You can't add an IP address to an HG80.   There's no network (like enet,   fddi, etc.)47 on the controllers.  Only a serial port for management.h  J You must install Storageworks Command Console on a system connected to the0 HSG80.  You can then manage the HSG80s via SWCC.  G You can also install the SWCC agent on a system connected to the array   and install E SWCC on another system and manage the array remotely.  Haven't tried l	 that yet.eI I don't trust anything except connecting to the HSG80s at 19,200 via the e serial port.( It's slow but it works 100% of the time.  
 -Scott :^)   Antony Wardle wrote:  ' > How do I add an ip number to an hsg80n >  > ( > I believe that you can web enable them > ( > so that you can access them and change* > their configuration instead of using the > cli. >  >  > anyone else tried this?n >  >  >  > antony >  >    ------------------------------  # Date: Wed, 31 Jan 2001 19:20:47 GMTp$ From: Scott Vieth <svieth@wi.rr.com>" Subject: Re: IP number on an HSG80$ Message-ID: <3A7865E8.900@wi.rr.com>  G I would just like to add that there are certain things that you can do i with the CLIC that you *can't* do with the SWCC gui interface.  If you want full   control of the/ config on your HSGs, then you must use the CLI.e  
 -Scott :^)   Uwe Zessin wrote:   : > I strongly recommend that people learn the CLI to get an< > understanding of *how* these things work and then use SWCC >  > -- > Uwe Zessin   ------------------------------  # Date: Wed, 31 Jan 2001 19:30:27 GMTD$ From: Scott Vieth <svieth@wi.rr.com>" Subject: Re: IP number on an HSG80( Message-ID: <3A78682D.3050402@wi.rr.com>   antony wardle wrote:  @ > I guess maybe the question I need to ask is how do I assign an3 > IP number to the switch that the hsg connects to.-  H You can assign an IP address to the fc switch if you want to manage the  switchL (or monitor per-port performance) but that won't let you manage the storage.  5 > I take your point about knowing how to use the cli,t: > but if you have the option of using a GUI, then why not.5 > The old story of a picture and a thousand words ;-)h  H The GUI is pretty crippled.  Learn to use the CLI.  Have you worked withE HSJs before?  It's very similar (except that the HSGs have more cachek and more features :^)      )  8 > Plus managerment can understand what you are doing ;-)  % That is not always a good thing!  ;^)t  : > I guess it is RTFN then. I was hoping someone might haveH > tried it and was interested in their opinion. I am guessing that thereK > won't be too much to do after it is set up, and look at the disk subsysemo3 > from a web page would be a useful monitoring tools  H Keep in mind that if you can access your storage over the network, then  someone else can too.e; Particularly if you have the web-based interface available.s  
 -scott :^)   ------------------------------  # Date: Wed, 31 Jan 2001 19:34:20 GMTe$ From: Scott Vieth <svieth@wi.rr.com>" Subject: Re: IP number on an HSG80( Message-ID: <3A786916.8040609@wi.rr.com>   Paul Repacholi wrote:i  B > So you storage is screwed, the systems are hung, your PCs shares> > are up the glass creek, and some fuckwit wants to a JAVA  to" > the tools you need to sort it... > # > Where do they find these items...h  H Like I said in a previous message, the serial ports on the front of the  HSG80sH work just fine at 19,200.  I have the port from one controller connectedE to a serial port on a PC running SWCC and the other port is connecteds to a port on a VT420.d   K.I.S.S.  ! ! ! ! !F  
 -scott :^)   ------------------------------  # Date: Wed, 31 Jan 2001 20:12:58 GMT1% From: Uwe Zessin <zessin@my-deja.com> " Subject: Re: IP number on an HSG80) Message-ID: <959rjv$vcl$1@nnrp1.deja.com>e  - In article <87ofwnoes7.fsf@prep.synonet.com>,f/   Paul Repacholi <prep@prep.synonet.com> wrote:r) > Uwe Zessin <zessin@my-deja.com> writes:n >.E > > The intend is/was for multiplatform support. It doesn't sound tooyD > > bad to me, when I'll be able to run SWCC on OpenVMS Alpha. Isn't1 > > that what people are asking for all the time?t >s3 > Yes. Two thoughts. RS-232 isn't 'multi-platform'?n  ? Sure, even SWCC on a PC (oh, by the way: version 2.3 _requires_t> WiNTendo or W2000!) is capable to talk to a controller via the serial port as I wrote earlier.c  ( > The second is about flies and diets...  B Sorry, I don't understand that. English isn't my mother's language (and mine either ;-)  E Remember that you don't _have_ to use SWCC. You can't do _everything_rE with it anyway, but for most tasks it's OK, I think. In my trainings,y* however, I _always_ start with the CLI ;-)  = > > How often is you storage screwed and you systems do hang?r> > > If it's such a bad mess you dig out your old VT320 and use > > the CLI...& > > You still have one (a VT), no? ;-) >eE > It only has to be ONCE. Note, that times of heavy load and critical + > use are the most likley time for trouble.c  B Right. That's why I recomend to our customers that they have ready$ access to all components of the SAN.  F > Do you have a VT? Well, add, do you know where it is, and the cable!  E I have a notebook and a terminal emulator - I'm just sitting in front A of it. I take it (and the cables - hey, I can even connect to thee@ RA3000 and the SANswitch 8-EL with them!) almost always with me.  9 > Does it still work, and can you remember the comands...A  B I will find out rather quick if it's broken as I work on it almost daily.  D Commands, well I remember most ones - I've been doing HSx management@ since late 1996. My first big problem was lost cache data during@ installation, but we did sell the platform kit with the HSZ20...  B I always make sure the customer gets at least one platform kit per< datacenter and operating system listed on his shopping list.  @ Also, I have put some PDF documentation on my disk. Hm, I should check if it is up to date...  > > And does the low level interface still work, or did it die a2 > slowdeath while pretty fruit was given priority.   Again, I don't understand this.s  > Hey, I have a wonderfull life! I'm not responsible for systemsD running in production. If the customer doesn't follow my suggestionsB it's _his_ problem, but some do follow the advice and others _are_ capable of learning ;-)e  D But I have managed systems running in production. I have been calledD sundays at 20:30. I have fixed broken databases from 18:00 to 08:00.A I know the pain when the system is down and have people breathing4 right at your neck ...  2 So I do try my best that the customer is prepared.   --
 Uwe Zessin3 (If you want to send mail, please use user "zessin"t/ who lives at "decus.decus.de", not my-deja.com)F     Sent via Deja.com  http://www.deja.com/   ------------------------------  # Date: Wed, 31 Jan 2001 20:38:39 GMT-% From: Uwe Zessin <zessin@my-deja.com> " Subject: Re: IP number on an HSG80) Message-ID: <959t4e$107$1@nnrp1.deja.com>c  $ In article <3A78639C.303@wi.rr.com>,'   Scott Vieth <svieth@wi.rr.com> wrote:3 [...].E > You must install Storageworks Command Console on a system connected<3       might, there is no SWCC for, e.g. OpenVMS ;-)s0   (sorry for nitpicking, I'm sure you know that)9 > to the HSG80.  You can then manage the HSG80s via SWCC.u > B > You can also install the SWCC agent on a system connected to the? > array and install SWCC on another system and manage the arrayn$ > remotely.  Haven't tried that yet.  C Works fine. There even is a window where you can have a tree of allk@ systems and their attached storage subsystem(s). Clicking on theA subsystem brings up the proper SWCC window and that automatically  connects to the system.m  B > I don't trust anything except connecting to the HSG80s at 19,200? > via the serial port. It's slow but it works 100% of the time.e  A Well, I don't beleive that running SWCC over local storage or the-C agent is faster. SWCC very often polls the subsystem and that costs @ a lot of time. That's another reason to use the CLI for changes.   --
 Uwe Zessin3 (If you want to send mail, please use user "zessin" / who lives at "decus.decus.de", not my-deja.com)n     Sent via Deja.comp http://www.deja.com/   ------------------------------   Date: 31 Jan 2001 20:19:27 GMT1 From: bill@triangle.cs.uofs.edu (Bill Gunshannon)m8 Subject: Re: It's the end for VMS and other Wild Rumours+ Message-ID: <959s0f$jtj$1@info.cs.uofs.edu>   ( In article <954u51$606$1@pyrite.mv.net>,*  "Bill Todd" <billtodd@foo.mv.com> writes: |>M |> So by that definition, VMS is as alive and well as RSX is.  And I wouldn'tnN |> argue that this won't continue to be true.  But I don't think that's what a& |> lot of people here are looking for.  K Ummmm.  RSX is still being fully maintained, advertised and sold by Mentec.iL It is by far the strongest of the old PDP-11 OSes.  I have little doubt thatL it will in fact outlast VMS by quite a bit regardless of the outcome of this/ debate.  Just as the PDP has outlasted the VAX.7   bill   -- nJ Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>   e   ------------------------------  % Date: Thu, 01 Feb 2001 01:09:07 +0000a) From: Christof Brass <brass@infopuls.com>e- Subject: Re: It's the end for VMS get off nowb, Message-ID: <3A78B733.5A75B864@infopuls.com>   Rob Young wrote: > Z > In article <3A778663.D8168F31@infopuls.com>, Christof Brass <brass@infopuls.com> writes: >  > >nB > > I heard that Dave Cutler left Micro$oft long ago. Does anybody! > > where he's currently working?o@ > > And, yes, Micro$oft Windows/NT is a little bit of Cutler andD > > much of Micro$oft. The emphasis of the VMS and WNT affinity came! > > from DEC, not from Micro$oft.+ > : >         Yes... that last acid trip he was on did him in. > H >         He sold all his belongings (would have typed "possessions" butG >         three times at it made me give up) and moved to Great BritainmF >         and has begun a dialog with Andrew Harrison to determine theH >         true meaning of Enterprise Architecture.  He purchased a fullyL >         refurbed London cab circa 1954 and lives in the back seat.  In hisH >         spare time he works on his art.  As you can see, it is heavilyD >         influenced by flashbacks interlaced heavily with legacy OS; >         underpinnings.  The art is found on his homepage:m > = > http://www.artbank.ltd.uk/Agency/Artists/Cutler/Cutler1.htm  > D >         In a further twist, he insists offers for his art can onlyF >         be delivered by carrier pigeon, North Atlantic route ONLY!!! > % >                                 Robt   :-))  = The page loads amazingly fast - maybe he's the system managerv: also. But I think his art design is not as good as his VMS design. ;-)    ------------------------------  % Date: Wed, 31 Jan 2001 15:01:58 -0500e4 From: "John L Ferguson" <John.L.Ferguson@compaq.com> Subject: Re: JDBC connectivity7 Message-ID: <Zb_d6.262$cu.1495@gazette.loc1.tandem.com>a  F Checkout http://www.openvms.compaq.com/openvms/products/ips/attunity/.  E This is the preferred solution for OpenVMS.  The Attunity Connect "OnaE Platform" Package will give you everything you need on OpenVMS Alpha.e5 You'll need to purchase Attunity Connect for Windows.e   John  < Frederick wrote in message <3A776CD7.A42EEE6B@brokat.com>...E >Any ideas how to connect an Oracle7 DB which runs on Open VMS721 viamE >JDBC from a remote NT4 client?  Any software installation require to   >install on NT and the Open VMS? >Any input is appreciate. :> > 
 >Frederick >f >f >y   ------------------------------   Date: 31 Jan 2001 20:32:34 GMT2 From: mathog@seqaxp.bio.caltech.edu (David Mathog) Subject: Re: JDBC connectivity, Message-ID: <959sp2$ois@gap.cco.caltech.edu>  n In article <Zb_d6.262$cu.1495@gazette.loc1.tandem.com>, "John L Ferguson" <John.L.Ferguson@compaq.com> writes:G >Checkout http://www.openvms.compaq.com/openvms/products/ips/attunity/.n > F >This is the preferred solution for OpenVMS.  The Attunity Connect "OnF >Platform" Package will give you everything you need on OpenVMS Alpha.6 >You'll need to purchase Attunity Connect for Windows.  G Correction, it will give you what you need if you want to attach to an mG Oracle database on OpenVMS.  If you want to do anything else, includingeI especially connecting to RMS files or any other database engine, you willeB need to buy additional software from the folks who make attunity.   J It really should be called "Attunity connect for Oracle for OpenVMS Alpha"7 rather than the very uninformative "on platform" title.e     David Mathog mathog@seqaxp.bio.caltech.eduu? Manager, sequence analysis facility, biology division, Caltech e   ------------------------------  % Date: Wed, 31 Jan 2001 15:53:42 -0500-# From: Jim Agnew <agnew@hsc.vcu.edu>  Subject: Re: JDBC connectivity+ Message-ID: <3A787B56.70EDD7A6@hsc.vcu.edu>j   www.openlinksw.com  < good stuff, we use their old outdated software that works...   Frederick wrote: > F > Any ideas how to connect an Oracle7 DB which runs on Open VMS721 viaF > JDBC from a remote NT4 client?  Any software installation require to! > install on NT and the Open VMS?d > Any input is appreciate. :>o >  > Frederickw   ------------------------------  % Date: Wed, 31 Jan 2001 14:12:10 -0800l) From: Wayne Holland <wholland@tscnet.com>f- Subject: Re: Linux worm and RedHat 7.0 broken O Message-ID: <1846E82FF7E510FC.AD070B96280F1B1F.7D8AF97CB75362BF@lp.airnews.net>-   Michael Austin wrote:  > \ > I have not been following this thread, but would like to throw in a comment or two.   I doY > both Unix and VMS and am quite good at both.  I have supported VMS and I have supportedfX > various flavors of Unix. Your choice of OS will depend upon your application.  It alsoY > will depend upon how much work  you are willing to put into keeping the systems runnings4 > and how many sys-admins you are willing to hire... > U > I would also say that anyone who says that VMS is "obsolete" has no clue which cameiZ > first.  VMS is actually more current that Unix.  Unix was developed in the late 60's andW > early 70's and VMS didn't come along for another few years.  The "cluster" technologyb[ > everyone is trying to get to work correctly, was developed by DEC and has been in VMS for X > many years.  Clusters are not new, just new to Unix.  And they still haven't gotten it[ > right.  New VMS licenses are up significantly, Why?  Because the need for 24x7x365 highlyeZ > available, highly reliable systems are a necessity for todays e-commerce and Unix cannot\ > compete in that space.   That is why more than 50% of the world's stock market systems are7 > VMS. They just work.  VMS is anything but "obsolete".0 >  > Michael Austin > DBA Consultant* > been there, done that, still prefer VMS.  B Hello.  I know what you mean about vms.  For me it was a matter ofE getting from point A to point B in the quickest time possible with my A existing skills.  I found, at that time, using Vax Fortran was myWF quickest route.  I didn't know C at the time.  Latter we were asked toH par down our budget, or get a low cost computing environment.  We lookedH into Sun as another division was using them and were keeping their costsE down.  By keeping the costs down I mean lowering the license fees andRH maintenance fees.  Also, due to a particular environment, we had to keep+ our electrical costs down which was due to *F the four air-conditioners we were using to cool a 785.  Sun came in to talk to us.tB We looked at their products and asked many questions.  Under thoseD earlier circumstances we had to turn Sun down on technical merits ofB their fortran processor.  Our other requirements, data bases, wereF currently met with the RMS services that were easy to use.  I comparedG documentation style as well.  Just a matter of preferences of the groupuG that was really all that mattered.  We were developing new software for2
 a very narrow)F problem area and I suppose that Unix could have filled in just as well/ if we were trained in that area instead of vms.cE You are quite right tho, that the mission has to be defined first andg4 what will get the job done along with risk analysis.  
 Wayne Holland0 Retired Dept. of the Navy.   ------------------------------  % Date: Thu, 01 Feb 2001 01:14:28 +0000e) From: Christof Brass <brass@infopuls.com>n= Subject: Re: No technical computing, was: expanding the nichen+ Message-ID: <3A78B874.D7AFE6D@infopuls.com>a   Robert Deininger wrote:e > Y > In article <3A774084.8E156A0C@infopuls.com>, Christof Brass <brass@infopuls.com> wrote:  >  > ? > > I'm not familiar with low level programming and HW details. @ > > Therefore I would like to know if the two points criticised,A > > i.e. slow I/O and fast controllers for a reasonable price not 6 > > available could be solved by writing a filesystem. > Q > VMS engineering is working on file system improvements.  No need to start over.f >  > > Do we need' > > also write drivers for controllers?  > " > Folks are working on that too... >  > > Do we need also write-B > > firmware for controlers to be able to use them with Alpha/VMS? >  > ... and that.c > C > > As I said, I'm not familiar with these low level stuff, but I'moA > > not sure if writing a filesystem solves all problems. Are theiA > > driver and firmware APIs specifications available? Is it f.e.fA > > possible to port the Linux ext2 filesystem with the necessaryt* > > extensions for a richer attribute set? > > There are several layers.  "Port Drivers" talk directly to controllers. "Class Drivers" talk to port drivers in a pretty controller-neutral way.  DKDRIVER, the SCSI class driver is an example.  It talks to different controllers via different port drivers.  The class-port interface is documented internally, I think.  "Normal" code talks to the class driver via the QIO service.  (Is there such a thing as normal code?)a > > Neither of these drivers is all that close to the file system.  There's also XQP, the eXtended QIO Processor (I think).  It handles stuff that's too big and complex for a device driver, but too low level for "RMS".  I think disk quotas are an example of something done in XQP.  XQP is sort of an extension of the device drivers.  The boundaries between drivers, XQP, and RMS are kind of blury -- in my mind, and maybe in reality also. > l > Read in the "Internals and Data Structures" book to get a much more detailed (and more accurate) overview. >  > Where is the "filesystem" buried in all of this?  I would say in RMS, but I'm far from an expert.  There's an out-of-print "file system internals" book that I'd like to read, but I've never seen it. >  > One reason we don't get the latest and greatest in VMS right away is because VMS tends to have higher standards.  A big focus has always been reliability and data integrity.  That often means fixing firmware in commodity devices and controllers, or programming the driver to deal with near-compliant hardware.  It certainly involves a lot of testing.  It means there is a reluctance to support commodity stuff if the firmware details change for each production run.  A lot of time, money, and effort is spent c >  > -- > Robert Deininger > rdeininger@mindspring.com   @ Thanks. Got the picture. Agree on data integrity above speed. If> one counts the time lost by unreliability VMS might come close to Linux in speed.   ------------------------------  % Date: Wed, 31 Jan 2001 17:19:31 -0600n; From: Lord Running Clam <Use-Author-Address-Header@[127.1]>S Subject: Re: OpenVHS (funny!) - Message-ID: <200101312319.RAA31535@www.xg.nu>   " -----BEGIN PGP SIGNED MESSAGE-----  ; On 31 Jan 2001, eplan@kapsch.net (Peter LANGSTOEGER) wrote:.  J >They (Q) don't know OpenVMS and they hardly know their own U**X's name...  G Perhaps they bought into that along with the UN*X command set "security:! through obscurity" principles ;-)-   # ouch!$     LRC. - -- c# Please allow me to introduce myselfe- I'm [representing] a man of Wealth and Taste.y ~ !The Rolling Stones.   -----BEGIN PGP SIGNATURE-----t Version: N/A  @ iQEVAwUBOndHcoer+ijnZohVAQEiuAf+MUBjCfpETzzVFMeWA5EIS1EsIdGYD2Qe@ usv/Ei5Y/O3dgeOXI+UtdpGNsVxmIkiwYJTCTp5Or3Ypz/9AX5+T1BJB2i9c8UGS@ 5FSoSE2c6cuXaR8cuWkXS+Cv9m4ZXyoGhHvxsDpbl1uJeti6PjQulJ64yAnDcgVI@ 8Pvg+VeBc7LkxrbqZxvjJtFaYUSFPRKxOzq++iYqpZpb5VFf6hgX0Uxal5VCLcxY@ NUN0Cix2Tez7n3UU/w4yxGTRbiudlhbD3MyBY+Ujh4JNmKLBu8MIW7g1mxrf7eRf8 uVGLvfZcb7jZaDLXKQC3033GcQA/My/j1UKIGG0f4g+3k7IriJvryw== =WIxAw -----END PGP SIGNATURE-----d   ------------------------------   Date: 31 Jan 2001 19:10:46 GMT= From: jlw@psulias.psu.edu (j.lance wilkinson, (814) 865-1818)r; Subject: Re: OpenVMS+Perl 5.0.0305A (?) and EmuMail modulesr+ Message-ID: <959nvm$n60@r02n01.cac.psu.edu>   W In article <957fff$u5q$1@nnrp1.deja.com>, Craig A. Berry <calepine@my-deja.com> writes:y >Lance,e >rC >I have no experience with EmuMail, but I've built quite a few PerluC >modules and never run into quite the same problems you are seeing.T; >The first thing is to make sure all is well with your Perl $ >installation.  What do you get from >o >$ perl -"V"   $ perl "-V" B Summary of my perl5 (5.0 patchlevel 5 subversion 3) configuration:   Platform:l/     osname=VMS, osvers=V7.2-1, archname=VMS_AXPe     uname='VMS YourHost V7.2's1     hint=none, useposix=false, d_sigaction=definel1     usethreads=undef useperlio=undef d_sfio=undefk   Compiler: 4     cc='CC/DECC', optimize='undef', gccversion=undef     cppflags='undef'     ccflagstE ='/Include=[]/Standard=Relaxed_ANSI/Prefix=All/Obj=.obj/List/Machine'n4     stdchar='char', d_stdstdio=define, usevfork=true2     intsize=4, longsize=4, ptrsize=4, doublesize=8G     d_longlong=define, longlongsize=8, d_longdbl=define, longdblsize=16i1     alignbytes=8, usemymalloc=N, prototype=definea   Linker and Libraries: 1     ld='Link', ldflags ='/NoTrace/Map/Full/Cross'e"     libpth=/sys$share /sys$library	     libs=r;     libc=(DECCRTL), so=exe, useshrplib=undef, libperl=undef.   Dynamic Linking:<     dlsrc=dl_vms.c, dlext=exe, d_dlsymun=undef, ccdlflags=''%     cccdlflags='', lddlflags='/Share'l    & Characteristics of this PERLSHR image:!   Compile-time options: DEBUGGINGg   Built under VMSg"   Compiled at Sep 29 2000 09:53:40   %ENV: +     PERLSHR="PERL_ROOT:[000000]PERLSHR.EXE"s7     PERL_ROOT="COMMON_DISK:[PERL.V553A1.PERL5_005_03.]"o   @INC:e#     perl_root:[lib.VMS_AXP.5_00503]d     perl_root:[lib]n%     perl_root:[lib.site_perl.VMS_AXP]e     perl_root:[lib.site_perl]r    F >and what do you find under PERL_ROOT:[LIB.VMS_AXP.5_00503.CORE] (it's+ >possible you are missing an include file)?e  - $ dir perl_root:[lib.vms_axp.5_00503.core...]a  . Directory PERL_ROOT:[LIB.VMS_AXP.5_00503.CORE]  F AV.H;1              BYTECODE.H;1        BYTERUN.H;1         CONFIG.H;1H COP.H;1             CV.H;1              EMBED.H;1           EMBEDVAR.H;1E EXTERN.H;1          FORM.H;1            GV.H;1              HANDY.H;1-H HV.H;1              INTERN.H;1          INTRPVAR.H;1        IPERLSYS.H;1B KEYWORDS.H;1        LIBPERL.OLB;1       MG.H;1              OP.H;1F OPCODE.H;1          PATCHLEVEL.H;1      PERL.H;1            PERLIO.H;1H PERLSDIO.H;1        PERLSHR_ATTR.OPT;1  PERLSHR_BLD.OPT;1   PERLVARS.H;1G PERLY.H;1           PP.H;1              PROTO.H;1           REGCOMP.H;1 B REGEXP.H;1          SCOPE.H;1           SOCKADAPT.H;1       SV.H;1F THRDVAR.H;1         THREAD.H;1          UTIL.H;1            VMSISH.H;1 XSUB.H;1   Total of 41 files.     >g> >It occurs to me since you are using CSWS you may be using theE >Compaq-supported Perl PCSI kit, and it is possible the configuration.F >under which it was built does not match your system.  What version of >OVMS AXP are you running?  	 	v7.2-1H1y   >s >> CC/DECC /Include=: >[]/Standard=Relaxed_ANSI/Prefix=All/Obj=.obj/List/Machine > F >Where did the /List/Machine come from?  Did you edit the descrip.mms?  B 	No, I didn't edit DESCRIP.MMS.  It was created by the makefile.pl@ 	perl script.  the /list/machine comes from the CCFLAGS as shown& 	in the output of perl "-V", I assume.  D >It shouldn't hurt but it's not something you'd ordinarily see.  You  D 	I tend to use the /List a lot myself, but I assume the two togetherB 	came from the kit you note above, since that's where my Perl came 	from.  @ >should be able to get a debug build with MMS/MACRO=__DEBUG__=1.  / 	I'm happy to try that if you think it'll help.2   >0 >>     Sigjmp_buf		je_buf; >.D >This should be defined in your config.h in the Perl core directory.  % 	It is:	#define Sigjmp_buf sigjmp_buft >@4 >>     RETVAL = fcntl(PerlIO_fileno(f), F_GETFL, 0); >nF >fcntl() is a relatively recent addition to the C RTL.  If your systemC >has it (which it does if you're on 7.2-1) you should make sure thee) >HAVE_FCNTL macro is defined in config.h.I   	There's a HAS_FCNTL defined.t   >e3 >> OpenVMS Perl implementors' comments are welcome!i >f- >Posting to vmsperl@perl.org is also welcome.i  > 	I'll keep that in mind, and probably forward this on to there	 	as well.t  M +-"Never Underestimate the bandwidth of a station wagon full of mag tapes"--+eN | J.Lance Wilkinson ("Lance")            InterNet:  Lance.Wilkinson@psu.edu | M | Systems Design Specialist - Lead       AT&T:      (814) 865-1818          |uM | Library Computing Services             FAX:       (814) 863-3560          |cM | E3 Paterno Library                     "I'd rather be dancing..."         |-M | Penn State University         A host is a host from coast to coast,       |aM | University Park, PA 16802     And no one will talk to a host that's close |gM | <postmaster@psulias.psu.edu>  Unless the host that isn't close            |AM | VMS GopherMeister             Is busy, hung or dead.                      | M +------"He's dead, Jim. I'll get his tricorder. You take his wallet."-------+h9                 [apologies to DeForest Kelley, 1920-1999]w3 <A Href="http://perdita.lcs.psu.edu">home page</a> $J <a Href="http://perdita.lcs.psu.edu/junkdec.htm">junk mail declaration</a>   ------------------------------  % Date: Wed, 31 Jan 2001 18:31:44 -0500 + From: Allan Gross <allangross@toadmail.com>l3 Subject: Posix/VMS bug in select or signal handlers + Message-ID: <3A78A060.599BE25@toadmail.com>t  & Does any one know anything about this?     /*C  * The following code will cause a machine running vms/posix 7.1 toiB  * crash any time an alarm goes off, either the alarm signal I set=  * or if I type ^C in a posix shell while not getting alarms.hE  * I tried to strip it down to the bare bones. It works fine on otherd  * UNIX operating systems.  */    #include <stdio.h> #include <errno.h> #include <sys/socket.h>o #include <signal.h>/ #include <sys/time.h>n   void alarm_callback(int sig);    main (argc, argv)h	 int argc;e
 char *argv[];  {    char *ptr;   in rc;   struct timeval timeout;      timeout.tv_sec = 5;u   timeout.tv_usec = 0;4   /* I use this to run with or without the alarms */   if (argc > 1 ) {     printf("setting alarm\n");%     signal (SIGALRM, alarm_callback); 
     alarm(1);V   }e   do {     errno = 0;"     printf("Going into select\n");1     rc = select (32, NULL, NULL, NULL, &timeout); F     printf("Out of select with return code = %d and errno = %d\n", rc, errno);2   } while (1); }l   void alarm_callback(sig) int sig; {p   printf("Got alarm\n");   alarm(10);"   signal(SIGALRM, alarm_callback); }i   ------------------------------  # Date: Wed, 31 Jan 2001 21:31:52 GMTo From: crc_cole@hotmail.comB Subject: Re: Possible to view Password History on OpenVMS V7.1-2 ?) Message-ID: <95a088$414$1@nnrp1.deja.com>7  F Yes, I am hacking into an OpenVMS system.  I got nothing better to do.B Man, I should have been more clever than to use a hotmail address.F Then maybe someone would tell me how to hack into a OpenVMS system.  IB figured I could just go out, ask how to hack with my email addressB showing, and someone would tell me.  Although I wanted to know the= HISTORY of a users (which is ME) password and not the currenti	 password.-   Thanks for the info though..      + In article <VA.0000023d.1e81fc2c@sture.ch>,t   paul@sture.ch wrote:E > In article <94a2ag$kmd$1@mailint03.im.hou.compaq.com>, Hoff Hoffmans wrote:6 > > From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman) > > Newsgroups: comp.os.vms.F > > Subject: Re: Possible to view Password History on OpenVMS V7.1-2 ?" > > Date: 19 Jan 2001 18:50:56 GMT > >t > >cB > > In article <949hjo$ciu$1@nnrp1.deja.com>, crc_cole@hotmail.com writes:eD > > :Is it possible to view the Password History on OpenVMS V7.1-2 ? > >t > >   Nope.  > > G > > :Is there a tool to decrypt the file or somewhere that the user whon is@ > > :requesting this can view the passwords that they have used? > >i > >   Nope.a > >hD > >   If you wanted to throw enough processor cycles at the problem,	 you coulde@ > >   certainly search for obvious passwords using a "dictionary attack", ortC > >   you could search for cleartext passwords that hash out to thee
 same valueD > >   as the specified user's password(s).  But the password hashing	 algorithmgB > >   used on OpenVMS is quite deliberately one-way in nature, and intended to0@ > >   provide an appropriately "random" distribution of the hash results -- theE > >   algorithm is a hash, and not encryption.  The hashing algorithmn is based? > >   on a Purdy polynomial, and this choice was intended to bec difficult to< > >   reverse and difficult to duplicate the hash results... > > A > >   If the user wishes to re-use passwords, you can disable the- passwordF > >   history mechanism or lock in a particular password.  If the user wishesG > >   to re-use passwords more frequently than presently permitted, you1 can2E > >   shorten the window where re-use is prohibited.  (Details are in  thenC > >   security manual, as well as a description of why you probably  don'tw > >   want to do this.)  > > @ > >   By default, OpenVMS quite deliberately makes the re-use of	 passwords G > >   relatively difficult for users.  Details of choosing passwords --  or> > >   of choosing password phrases -- are also included in the
 manual.)  I'dtC > >   encourage spending more time with this user explaining system  securityE > >   and learning how to choose relatively secure passwords than I'dl spendsA > >   helping the user recover and/or remember their old passwordn choices. > >eC > >   Also please realize that I do not know if YOU are hacking thef systemD > >   password file -- resistance to password attacks was one of the reasonsoD > >   why the password system was designed with the hash...  OpenVMS triesr7 > >   not to keep (valid) cleartext passwords around...g > > E > To which I must add. An unsigned posting about security issues fromn an anonymousE > looking hotmail address doesn't inspire confidence in the integrityd of the	 > poster.U >  > ___  > Paul Sture
 > Switzerlanda >s >y     Sent via Deja.como http://www.deja.com/   ------------------------------  % Date: Wed, 31 Jan 2001 17:48:32 -0600i7 From: "David J. Dachtera" <djesys.nospam@earthlink.net>sB Subject: Re: Possible to view Password History on OpenVMS V7.1-2 ?- Message-ID: <3A78A450.D2382DBD@earthlink.net>n   crc_cole@hotmail.com wrote:l > , > Yes, I am hacking into an OpenVMS system.   ? What makes you think you can succeed where Kevin Mitnik failed?h   > I got nothing better to do.   C Try Monster.com. You can probably find a way to actually makes somed money with your skills.i  D > Man, I should have been more clever than to use a hotmail address.E > Then maybe someone would tell me how to hack into a OpenVMS system.s  	 Doubtful.a  F > I figured I could just go out, ask how to hack with my email address& > showing, and someone would tell me.   8 Do you *REALLY* think we're *THAT* f___ing stupid ???!!!   > Although I wanted to know thea? > HISTORY of a users (which is ME) password and not the currentS > password.c  / Storing unencrypted passwords is dumb - ALWAYS!e   > Thanks for the info though.    Hope we didn't help...   -- e David J. Dachtera  dba DJE Systems  http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/e  F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.    ------------------------------  % Date: Wed, 31 Jan 2001 21:06:27 -0500.0 From: "Sharkonwheels" <tonym@compusourceDOT.net>) Subject: R215F-BA Available in S. FloridaB, Message-ID: <YU3e6.186$Go6.111961@news2.mco>  K Anyone needs it, it's available for $100.00 Plus shipping. Kinda' heavy, soo beL ready. I hated to see it get dumped, so I bought it from the place, figuring+ someone out there HAD to have a use for it.w  G Has one RF31 (380MB) drive inside of it, includes the DSSI interconnecti cable.L This is an external enclosure, about 1" cube, but about 2.5-3' high. OverallG good condition, have no DSSI system to test with, so can't speak on the 	 conditiony
 of the drive.s   Tony tonymATcompusourceDOTnet   ------------------------------    Date: 31 Jan 2001 22:27:12 -05002 From: malmberg@eisner.decus.org (John E. Malmberg)- Subject: Re: R215F-BA Available in S. Florida + Message-ID: <SMA36wShWExg@eisner.decus.org>n  , In article <YU3e6.186$Go6.111961@news2.mco>,2 "Sharkonwheels" <tonym@compusourceDOT.net> writes:< > Anyone needs it, it's available for $100.00 Plus shipping.: > Kinda' heavy, so be ready. I hated to see it get dumped,) > so I bought it from the place, figuring?- > someone out there HAD to have a use for it.t >i< > Has one RF31 (380MB) drive inside of it, includes the DSSI > interconnect cable. 9 > This is an external enclosure, about 1" cube, but aboutl: > 2.5-3' high. Overall good condition, have no DSSI system= > to test with, so can't speak on the condition of the drive.l  ; How can it be in good condition when it has been compresseds7 to be in a 1 inch cube that is about 2.5 feet tall? :-)h   ------------------------------  % Date: Wed, 31 Jan 2001 19:44:03 -0800r! From: Tom Linden <tom@kednos.com> - Subject: RE: R215F-BA Available in S. Floridat9 Message-ID: <EJEDILJIPPJOIEFOLDHLGEHJCFAA.tom@kednos.com>e   You need to add water.   > -----Original Message-----; > From: John E. Malmberg [mailto:malmberg@eisner.decus.org] + > Sent: Wednesday, January 31, 2001 7:27 PMi > To: Info-VAX@Mvb.Saic.Come/ > Subject: Re: R215F-BA Available in S. Floridae >  > . > In article <YU3e6.186$Go6.111961@news2.mco>,4 > "Sharkonwheels" <tonym@compusourceDOT.net> writes:> > > Anyone needs it, it's available for $100.00 Plus shipping.< > > Kinda' heavy, so be ready. I hated to see it get dumped,+ > > so I bought it from the place, figuring'/ > > someone out there HAD to have a use for it.i > >s> > > Has one RF31 (380MB) drive inside of it, includes the DSSI > > interconnect cable.o; > > This is an external enclosure, about 1" cube, but aboutV< > > 2.5-3' high. Overall good condition, have no DSSI system? > > to test with, so can't speak on the condition of the drive.r > = > How can it be in good condition when it has been compressedo9 > to be in a 1 inch cube that is about 2.5 feet tall? :-)a >    ------------------------------  % Date: Wed, 31 Jan 2001 17:24:19 -0300t) From: fabio_compaq@ep-bc.petrobras.com.bri Subject: RWASTL Message-ID: <OF30C4772B.0EDF59B1-ON032569E5.006FF4CB@ep-bc.petrobras.com.br>  5 Do anyone know a way to stop a process in RWAST . . .a6 If there is a freeware-program to do that ....etc ....   Regardsl   FC   ------------------------------    Date: 31 Jan 2001 17:05:48 -0500, From: koehler@eisner.decus.org (Bob Koehler) Subject: Re: RWAST+ Message-ID: <EL6xc+E6Umkc@eisner.decus.org>j  x In article <OF30C4772B.0EDF59B1-ON032569E5.006FF4CB@ep-bc.petrobras.com.br>, fabio_compaq@ep-bc.petrobras.com.br writes: > 7 > Do anyone know a way to stop a process in RWAST . . . 8 > If there is a freeware-program to do that ....etc .... >   # If it's in RWAST it is "stopped".     0 See FAQ MGMT30 for what you really want to know.  F ----------------------------------------------------------------------? Bob Koehler                     | Computer Sciences Corporatione= NASA GSFC Flight Software       | Federal Sector, Civil GroupeE                                 | please remove ".aspm" when replyingd   ------------------------------  # Date: Wed, 31 Jan 2001 22:02:51 GMT 1 From: "Mark D. Jilson" <jilly@clarityconnect.com>u Subject: Re: RWAST2 Message-ID: <3A788C77.DC54BCCE@clarityconnect.com>  G Since there are a dozen or so different things that can cause a processVC to get into RWAST you really need to look at the process w/ SDA andEB discover the underlying reason for the RWAST.  DSNlink has a great/ article on this if you have a support contract.   * fabio_compaq@ep-bc.petrobras.com.br wrote: > 7 > Do anyone know a way to stop a process in RWAST . . .-8 > If there is a freeware-program to do that ....etc .... > 	 > Regardsj >  > FC   -- BD Jilly	- Working from Home in the Chemung River Valley - Lockwood, NY0 	- jilly@clarityconnect.com			- Brett Bodine fan. 	- Mark.Jilson@Compaq.com			- since 1975 or so, 	- http://www.jilly.baka.com               -   ------------------------------  % Date: Wed, 31 Jan 2001 19:15:59 -0500m- From: JF Mezei <jfmezei.spamnot@videotron.ca>m Subject: Re: RWAST, Message-ID: <3A78AABE.3C2571B7@videotron.ca>  * fabio_compaq@ep-bc.petrobras.com.br wrote: > 7 > Do anyone know a way to stop a process in RWAST . . .i8 > If there is a freeware-program to do that ....etc ....  ; RWAST is VMS's equivalent to the Blue Screen of Death :-:-)e  N The big difference is that it is THEORETICALLY possible to recover from it. IfK you are dealing with TK50 or TK70, if you attempt to stop a process that istJ doing IO to the tape, then waiting long enough will usually free the state9 (the process must wait for the outstanding IO completes).n   ------------------------------  # Date: Wed, 31 Jan 2001 19:40:43 GMT # From: marinto <marinto@my-deja.com>,) Subject: Re: select() and getservbyname()i) Message-ID: <959pnq$tf7$1@nnrp1.deja.com>   + In article <USopXfKc4z6n@eisner.decus.org>,p5   malmberg@eisner.decus.org (John E. Malmberg) wrote:x. > >> In article <94s2lp$ofi$1@nnrp1.deja.com>,) > > marinto <marinto@my-deja.com> writes:iD > >> >I've written some tcpip routines for unix and am about to port them > > to: > >> >vms using system services. After a quick look in the table "OpenVMSA > >> >System Service and Equivalent C Socket Call" in the OpenVMS_? > >> >documentation I missed two functions that I need. The twor	 functions  > > arer# > >> >select() and getservbyname().l > >> > >s8 > Someone else has posted the getservbyname() reference. >S= > select() is also implemented, On OpenVMS it only works withf< > sockets.  It is documented both in the Compaq TCPIP manual > and in the Compaq C Manual.5 >51 > The TCPIP documentation can be found online at: $ > http://www.openvms.compaq.com:8000 > . > The Compaq C documenation is also online at: >M+ > http://www.openvms.compaq.com/commercial/   E /me already knows that select() is implemented on vms...but I thought F that I should use system services when writing my library. Then I need< to know how to do what select() does with system services :)   >1 > -JohnL > wb8tyw@qsl.network > Personal Opinion Only. >L     Sent via Deja.com_ http://www.deja.com/   ------------------------------  # Date: Wed, 31 Jan 2001 19:54:29 GMT;# From: marinto <marinto@my-deja.com> ) Subject: Re: select() and getservbyname();) Message-ID: <959qhh$u78$1@nnrp1.deja.com>   - In article <27JAN200116253790@gerg.tamu.edu>, *   carl@gerg.tamu.edu (Carl Perkins) wrote:) > marinto <marinto@my-deja.com> writes...e- > }  carl@gerg.tamu.edu (Carl Perkins) wrote:r, > }> marinto <marinto@my-deja.com> writes...D > }> }I've written some tcpip routines for unix and am about to port them > }to>: > }> }vms using system services. After a quick look in the table "OpenVMSA > }> }System Service and Equivalent C Socket Call" in the OpenVMSe? > }> }documentation I missed two functions that I need. The twot	 functions. > }are# > }> }select() and getservbyname().o > }> } > }> }Thanks in advance. > }> } > }> }/Marinto > }>G > }> The getservbyname() function is present in the C libraries, so you  > }don'tB > }> need to replace it with anything else. From the on-line help: > }> > }> CCn > }> > }>   Socket_Routines > }> > }>     getservbyname > }> > [...]t > }bH > }Ok, I guess that solves my problem with getservbyname(). But I'm justE > }curios...if one should use a language where one haven't got accesse to a? > }similar function, how could it be done with system services?i >pG > There is no such language (or, at least, there isn't supposed to be).;F > You can call the C library routines from any language (although someF > will be easier than others - it depends on how much work you have toC > do to use the returned data structure and pass it null terminatedL- > strings by reference). Such is life on VMS.  >m< > You might be able to do it in other ways. If this call has equivalents,D > the exact syntax probably depends on which IP stack you are using. ConsultND > the documentation and/or online help (HELP MULTINET PROGRAMMING in the variouscF > parts doesn't indicate any equivalent). It would presumably be via a QIO.D > There should be some examples for your IP stack, but not necessary onesH > that do this specific thing, on your system somewhere. Multinet has anA > examples subdirectory off the multinet root directory. Compaq's. product 2 > probably has a subdirectory off of SYS$EXAMPLES. >fG > }> On the other hand, select() isn't (at least not in my somewhat oldw
 > }versionC > }> of C). It isn't the "VMS way" of doing things. Others may haves > }suggestionsF > }> as to how to work around this, but it probably depends on exactly > }what you  > }> are trying to do. > },D > }In the unix version of my wait function I can wait until there is dataF > }to read from a client, data to read from a pipe (mailbox), a client isE > }ready to receive data, a client has disconnected or another client" > }tries to connect. > } B > }My wait function can already wait for data on several mailboxesH > }(IO$M_WRTATTN with an ast that sets an event flag and also sets whichD > }mailbox it was set for, then I just wait for the event flag to be set).t > }/Marintoo >kG > It turns out that there is a select() function available. I missed iti9 > the first time (probaby had a typo in my HELP command):e >f > CC >  >   Socket_Routinest >* >     select >iC >            Allows you to poll or check a group of sockets for I/OwH >            activity. It can check what sockets are ready to be read or? >            written, or what sockets have a pending exception.k >h >            Syntax: > # >                 #include <time.h>c >bC >                 int select(int nfds, int *readfds, int *writefds, > >                               int *execptfds, struct timeval
 *timeout); > D > It doesn't say it here, but this routine is only valid for networkD > connections (that would be the "n" in the "nfds", I expect) and is > not useable for other things.l >(E > Doing it all with ASTs and/or event flags would be the typical "VMS F > way" - mostly ASTs for anything even moderatly complex (a thread perD > connection is another way, these days). You set everything up such thatB > it fires off an AST when it completes. This allows you to handle multipleF > data streams in one process and one thread. Instead of using a bunch ofF > event flags, one common way to do it is that when the AST fires off, itE > puts the relevant data in the correct queue (for example, you mights wantD > a queue to put new connection info into, one to put completed read dataC > into, and one to put completed write data into via functions likehA > LIB$INSQTI and LIB$REMQHI) and then wakes up (via SYS$WAKE) thetD > main process (which was hibernating via SYS$HIBER) then exits, theA > main process then processes what is in the queue(s) and resumes G > hibernating when it finds them all empty. There's an example of this, 1 > communicating via DECnet rather than TCP/IP, ini SYS$EXAMPLES:DB_SERVER.C.cE > It is possible to translate this fairly directly into an equivalentcE > that uses TCP/IP. In order to do this, you need to use QIO calls to D > communicate with your IP stack rather than the socket library (youE > should be able to still use some socket library calls - such as thenE > one above - mixed in if you are carefull and don't do processing att1 > AST level that could change the relevant data).l >mD > For example, using Multinet you'd have a call that looks something9 > like this in there somewhere late in the initial setup:t >eC > /* Wait for a connection, fire accept AST routine when connectionh attempte >  * is detected. */ > = >     status= SYS$QIO(0, cp_tcp_chan, IO$_ACCEPT_WAIT, &iosb,r< >                     &cp_tcp_accept, cm, 0, 0, 0, 0, 0, 0); >tC > This causes the "cp_tcp_accept" function to be run as an AST when E > someone tries to connect. You never have to have anything elsewhereaC > in the program hang around specifically waiting for a connect, ordC > polling to see if there has been a connection. When there is one,-E > the specified function is executed as an AST. That routine then (inmD > my data queue paradigm) assigns a channel, accepts the connection,C > inserts an entry into a queue with data about the new connection,rB > does a QIO like the one above to prepare for the next connectionC > attempt, calls SYS$WAKE to wake up the main process, and returns. B > The main process then wakes up, checks the queues, finds the new@ > connection data, and issues the first QIO to read from the newC > connection (and perhaps does a write QIO with any initial "hello"e > information).  >1  C Since the programs that uses my library should work the same way onaD both unix and vms, I can't use asts outside my library. I'm thinking@ bout implementing threads though but thats for later (the serverF program Im writing now only has to handle 8 clients at a time...). ButF for future programs that are to be written with my library, threads is* a good idea. (Who needs fork() anyways ;).   > Fun with asynchronous IO.' > 
 > --- Carl >      Sent via Deja.com  http://www.deja.com/   ------------------------------    Date: 31 Jan 2001 17:03:59 -0500, From: koehler@eisner.decus.org (Bob Koehler)) Subject: Re: select() and getservbyname()w+ Message-ID: <7vE5pl4d2w3t@eisner.decus.org>z  O In article <959pnq$tf7$1@nnrp1.deja.com>, marinto <marinto@my-deja.com> writes:a > G > /me already knows that select() is implemented on vms...but I thoughteH > that I should use system services when writing my library. Then I need> > to know how to do what select() does with system services :) >   H One can do a $QIO with function IO$_READVBLK and TCPIP$C_MSG_PEEK as P4.H There's a corresponding socket call parameter MSG_PEEK which can by usedF as a flag to recv (which also does a eventually does a $QIO).  Both ofB these are discussed in "DIGITAL TCP/IP Services for OpenVMS System# Services and C Socket Programming".   D Generally if actually doing socket I/O via system services one wouldE just $QIO a IO$_READVBLK and let it complete asynchonously instead of2 emulating UNIX I/O concepts.  A Using select() one typically does whatever processing is possibletB without I/O, then when I/O is required one sits on select().  When@ select() completes you are guaranteed the corresponding I/O willF complete without waiting.  Any time sensitive processing should followD the select().  Then you loop back to the begining of this paragraph.  H When using VMS concepts one generally $QIO's some asynchronous I/O, thenL does whatever processing is possible without I/O.  When I/O is required one G sets on a $WFLOR until one has completed.  You can stick time sensitiveeG processing in an AST.  Then you loop back to the top of this paragraph.I  F Since there are multiple was to synch up your processing with your I/OE (event flags, ASTs, locks, ...) you can vary the VMS approach to suite your needs.   F ----------------------------------------------------------------------? Bob Koehler                     | Computer Sciences Corporation = NASA GSFC Flight Software       | Federal Sector, Civil GroupSE                                 | please remove ".aspm" when replyinge   ------------------------------   Date: 31 Jan 2001 17:59:47 GMT3 From: gartmann@immunbio.mpg.de (Christoph Gartmann)t# Subject: System version X6N6-B6D...g0 Message-ID: <959jqj$ebl$1@n.ruf.uni-freiburg.de>   Hello,  O today I did a stupid thing: I installed the patch VMS62TO71U2 and didn't notice M that I already had installed it half a year ago :-( . Now the system comes upeL with a version of "X6N6-B6D". Ok, I thought I put VMS712_UPDATE-V0300 on topG of it as a did when I installed VMS62TO71U2 for the first time. But thetL installation crashes nearly at the end. Instead of running 7.1-2 as before IG have this X6N6-B6D. Is there a way to get rid of this without restoringhD the system disk from tape? I tried a PRODUCT REMOVE VMS62TO71U2, butK PolyCenter claims it is not there although it is shown in the history list.    Regards,    Christoph Gartmanny  H -- --------------------------------------------------------------------+H | Max-Planck-Institut fuer      Phone   : +49-761-5108-464   Fax: -452 |H | Immunbiologie                                                        |H | Postfach 1169                 Internet: gartmann@immunbio.mpg.de     |H | D-79011  Freiburg, FRG                                               |H +--------- http://www.immunbio.mpg.de/home/english/menue.html ---------+   ------------------------------  % Date: Wed, 31 Jan 2001 15:31:52 -0500u0 From: David Beatty <David.Beatty.NOSPAM@sas.com>' Subject: Re: System version X6N6-B6D... 2 Message-ID: <3XV4OpcIXlM3VPpS8IVukIZp1=Xe@4ax.com>  F I accidentally did a similar thing to a workstation by installing someB patches on V7.2-1 that were already there.  I figured the cleanestB thing to do would be to restore; I already had a base system disk.  
 Good luck!   Davidb  @ On 31 Jan 2001 17:59:47 GMT, gartmann@immunbio.mpg.de (Christoph Gartmann) wrote:   >Hello,e >3P >today I did a stupid thing: I installed the patch VMS62TO71U2 and didn't noticeN >that I already had installed it half a year ago :-( . Now the system comes upM >with a version of "X6N6-B6D". Ok, I thought I put VMS712_UPDATE-V0300 on topnH >of it as a did when I installed VMS62TO71U2 for the first time. But theM >installation crashes nearly at the end. Instead of running 7.1-2 as before IiH >have this X6N6-B6D. Is there a way to get rid of this without restoringE >the system disk from tape? I tried a PRODUCT REMOVE VMS62TO71U2, buteL >PolyCenter claims it is not there although it is shown in the history list. >I	 >Regards,e >   Christoph Gartmann >bI >-- --------------------------------------------------------------------+ I >| Max-Planck-Institut fuer      Phone   : +49-761-5108-464   Fax: -452 |SI >| Immunbiologie                                                        |uI >| Postfach 1169                 Internet: gartmann@immunbio.mpg.de     |mI >| D-79011  Freiburg, FRG                                               |oI >+--------- http://www.immunbio.mpg.de/home/english/menue.html ---------+i   ------------------------------   Date: 31 Jan 2001 20:22:20 GMT3 From: gartmann@immunbio.mpg.de (Christoph Gartmann)t' Subject: Re: System version X6N6-B6D...b0 Message-ID: <959s5s$hfc$1@n.ruf.uni-freiburg.de>  N In article <959jqj$ebl$1@n.ruf.uni-freiburg.de>, I (Christoph Gartmann) wrote:P >today I did a stupid thing: I installed the patch VMS62TO71U2 and didn't noticeN >that I already had installed it half a year ago :-( . Now the system comes upM >with a version of "X6N6-B6D". Ok, I thought I put VMS712_UPDATE-V0300 on top-H >of it as a did when I installed VMS62TO71U2 for the first time. But theM >installation crashes nearly at the end. Instead of running 7.1-2 as before IwH >have this X6N6-B6D. Is there a way to get rid of this without restoringE >the system disk from tape? I tried a PRODUCT REMOVE VMS62TO71U2, butRL >PolyCenter claims it is not there although it is shown in the history list.  K I must correct myself: I did not install the same patch but a newer versionoM of the same patch. After a second reboot I got rid of the strange OS version.fL But now DECwindows doesn't start. It gets stuck. And the VMS712_UPDATE-V03007 still cannot be installed (PCSI just keeps crashing)...-   Regards,    Christoph Gartmann0  H -- --------------------------------------------------------------------+H | Max-Planck-Institut fuer      Phone   : +49-761-5108-464   Fax: -452 |H | Immunbiologie                                                        |H | Postfach 1169                 Internet: gartmann@immunbio.mpg.de     |H | D-79011  Freiburg, FRG                                               |H +--------- http://www.immunbio.mpg.de/home/english/menue.html ---------+   ------------------------------    Date: 31 Jan 2001 20:58:53 -05002 From: malmberg@eisner.decus.org (John E. Malmberg)O Subject: Re: TCPIP 5.0A FTP Server Rejecting Connections - was RE: Where to getv+ Message-ID: <NdyNeZhSZM9+@eisner.decus.org>n  / In article <3A789BCA.76ED3429@am1.ericsson.se>,.1 Gloria Griffith <qusgagh@am1.ericsson.se> writes:c > M > I have a problem on a OVMS system that will not allow any ftp access to the5L > machine. The machine can be accessed by telnet, set host etc., but the ftpP > command fails. I have compared TCPIP configuration on the problem machine withL > the working machines and I have been unable detect any error. I have TCPIP > version 5.A installed. >tO > This is the error I get when I try to ftp to this DS20E Alpha Server, running- > Open VMS 7.1-2 >r > $ Ftp gc04bs2 > %TCPIP-E-FTP_NETERR, I/O error on network device6 > -SYSTEM-F-REJECT, connect to network object rejected
 > $ UptimeN > OpenVMS V7.2-1  on node GC04BS  31-JAN-2001 12:50:48.00  Uptime  32 01:44:33 > $t$ [Other ftp client dialog edited out]  C That the FTP command line works on a system and the FTP server doesa@ not has no connection to each other.  They are separate programs" and have a separate configuration.  B I would recommend checking to see on server gc04bs to see if there2 is a file SYS$DEVICE:[TCPIP$FTP]TCPIP$FTP_RUN.LOG.  + This may indicate the cause of the problem.s  D I would also recommend checking that your SYS$MANAGER:SYLOGIN.COM onG gc04bs has world execute access and the same for any command files thatm it runs.  C Also make sure that all of the commands that only work on terminalsVC are not executed for non-interactive processes like the ftp server._F The file SYS$MANAGER:SYLOGIN.TEMPLATE supplied with your system should. have the instructions in it on how to do this.  A Of course you could compare the file and it's protection with thee system that works.  H And finally there is the possibility that the FTP Server was not set up.  F You can use the @sys$manager:tcpip$config.com to check it's status and
 to enable it.   J If the above, including using the tcpip$config.com to disable and reenableE the FTP server do not fix the problem, then the following informatione could be of help:m   $TCPIP SHOW SERVICE FTP /FULLt  % $DIR/SECURITY SYS$MANAGER:SYLOGIN.COMh   -Johnh wb8tyw@qsl.network Personal Opinion Only.   ------------------------------  % Date: Wed, 31 Jan 2001 15:52:13 -0500s& From: "Rod Prince" <prince@wserve.com> Subject: TCPIP configuration- Message-ID: <980974369.69142@night.wserv.com>j  K Using OpenVMS 7.2 (AXP), UCX 5.0, can I configure a single NIC for multiple;K IP interfaces?  If so, where do I look for information on how to set it up?m  
 Rod Prince prince@wserv.com   ------------------------------  $ Date: Thu, 1 Feb 2001 11:54:54 +1100/ From: "Phil Howell" <phowell@snowyhydro.com.au>m  Subject: Re: TPU$WORK work files. Message-ID: <or2e6.124$%3.4259@ozemail.com.au>  8 JF Mezei <jfmezei.spamnot@videotron.ca> wrote in message% news:3A73299E.6EA8865@videotron.ca...aK > If your sys$login has a version limit of 2, whenever you try to edit moreo thanL > 2 files at the same time, you get a nasty error message to the effect that theM+ > TPU$WORK.TPU$WORK file cannot be created.s >aI > There are, of course many ways to deal with this, so I am curious abouto what4 > methods are most used , drawbacks advantages etc ? >># > Some of the way I could think of:s >eJ > create a directory with no version limit and define SYS$SCRATCH to go to it.  >eI > create a TPU$WORK logical to point to a file that already exists with a> > version limit of 0K > (hoping that file never gets deleted otherwise next time you use TPU, they& > directory's version limit kicks in). >oH > always use /NOWORK when starting TPU (unless you know you're editing a very > large file).4 I use $define tpu$work sys$scratch:tpu$work.'pid5_8'+ where pid5_8 is derived from the proc_index)% so is unique while the process existse Phil   ------------------------------  % Date: Wed, 31 Jan 2001 15:08:20 -0500h/ From: "Webb, William W" <wwebb1@email.usps.gov>d' Subject: RE: Try this on Linux or NT/MEeI Message-ID: <D46FE9B132FB9B44AEC242A96E4AB7502CB3B0@rlghncst625.usps.gov>    Steve-    * We've heard it in brief many times before.  3 I, for one, would be *extremely interested* in what 4 we in the States refer to as "all the gory details."   WWWebb     -----Original Message-----0 From: Info-VAX-Request@Mvb.Saic.Com at INTERNET ) Sent: Wednesday, January 31, 2001 1:58 PMr6 To: Webb, William W; Info-VAX@Mvb.Saic.Com at INTERNET' Subject: RE: Try this on Linux or NT/MEn    / In article ?002569E5.004C98D3.00@quegw01.btyp?,l'   Steve.Spires@yellowpages.co.uk wrote:eH ? Contact:   Tel: 3063  -  IS - Infrastructure, 1st Floor, Bridge Street Plazaf ? 2 ? No. This is NOT an urban legend. 'Tis very true. ?yE ? I'm having an email discussion with their system manager right now.  ?n ? Steve Spires ?s   Steve,  D Could you get him to post a message to this newsgroup confirming the) story for all the unbelievers out there ?.  @ Someone asked me for further details and evidence for this storyB a few months ago (They wanted to use it to bolster support for VMSD in their company). Unfortunately I only had it second-hand from thisH newsgroup and wasn't able to point him to anyone who knew the definitive6 story or the exact name of the train company involved.  
 David Webb VMS and Unix team leader CCSS Middlesex University          ? ? Christof Brass ?brass@infopuls.com? on 31/01/2001 04:20:50 AM  ?/" ? To:        Info-VAX@Mvb.Saic.Com- ? cc:         (bcc: Steve Spires/YellowPages)iG ? From:      Christof Brass ?brass@infopuls.com?, 31 January 2001, 4:20  a.m. ?o  ? Re: Try this on Linux or NT/ME ?  ? Michael Moroney wrote: ? ? ' ? ? Don Sykes ?don@alphase.com? writes:  ? ? A ? ? ?The application tracks all 911 calls (thousands per day) and  controlsG ? ? ?all police dispatching for the city (pop ~300,000). I just checkeda and E ? ? ?the application and OS have been running continuously, 24/7, fort over 13,= ? ? ?months. It was last rebooted ?14-DEC-1999 06:37:31.00? ! 4 ? ? ?Can any other OS point to successes like this ? ? ?tD ? ? The apparent uptime record is a VMS system running a railroad in Ireland,@ ? ? it had an uptime of 18 YEARS.  Unfortunately it was recently shutdown for ? ? an upgrade.a ? ?t	 ? ? -Mikeo ?v> ? I regard this as a modern urban legend in its shortest form.@ ? No power outage in Ireland for that long?? I don't believe it.B ? Any useful work done on that machine without mistakenly shuttingB ? down, switching off or otherwise screwing up?? I don't think so. ? What version of VMS? ? Why upgrading now?" ? Questions, questions, questions. ?g ?o     Sent via Deja.comi http://www.deja.com/   ------------------------------  % Date: Wed, 31 Jan 2001 16:32:29 -0500o% From: "John Vottero" <John@mvpsi.com>n' Subject: Re: Try this on Linux or NT/MEe/ Message-ID: <t7h14jasbjts85@news.supernews.com>d  L I would also like to hear from sites with VMSclusters which have been up forH years.  There must be more than one site that migrated from VAX to AlphaF without shutting down the cluster.  I would love to hear the story(s).    G <d.webb@mdx.ac.uk> wrote in message news:959lsh$pnt$1@nnrp1.deja.com...h1 > In article ?002569E5.004C98D3.00@quegw01.btyp?,,) >   Steve.Spires@yellowpages.co.uk wrote:oJ > ? Contact:   Tel: 3063  -  IS - Infrastructure, 1st Floor, Bridge Street > Plazao > ?t4 > ? No. This is NOT an urban legend. 'Tis very true. > ?hG > ? I'm having an email discussion with their system manager right now.t > ?s > ? Steve Spires > ?a >  > Steve, >nF > Could you get him to post a message to this newsgroup confirming the+ > story for all the unbelievers out there ?d >)B > Someone asked me for further details and evidence for this storyD > a few months ago (They wanted to use it to bolster support for VMSF > in their company). Unfortunately I only had it second-hand from thisJ > newsgroup and wasn't able to point him to anyone who knew the definitive8 > story or the exact name of the train company involved. >t > David Webb > VMS and Unix team leader > CCSS > Middlesex University >e >n >  >A >wA > ? Christof Brass ?brass@infopuls.com? on 31/01/2001 04:20:50 AMl > ?r$ > ? To:        Info-VAX@Mvb.Saic.Com/ > ? cc:         (bcc: Steve Spires/YellowPages)sI > ? From:      Christof Brass ?brass@infopuls.com?, 31 January 2001, 4:20c > a.m. > ?t" > ? Re: Try this on Linux or NT/ME > ?y > ? Michael Moroney wrote: > ? ?n) > ? ? Don Sykes ?don@alphase.com? writes:e > ? ?tC > ? ? ?The application tracks all 911 calls (thousands per day) and 
 > controlsI > ? ? ?all police dispatching for the city (pop ~300,000). I just checkedm > andcG > ? ? ?the application and OS have been running continuously, 24/7, forf	 > over 13w? > ? ? ?months. It was last rebooted ?14-DEC-1999 06:37:31.00? !w6 > ? ? ?Can any other OS point to successes like this ? > ? ?hF > ? ? The apparent uptime record is a VMS system running a railroad in
 > Ireland,B > ? ? it had an uptime of 18 YEARS.  Unfortunately it was recently > shutdown for > ? ? an upgrade.n > ? ?d > ? ? -Miket > ?f@ > ? I regard this as a modern urban legend in its shortest form.B > ? No power outage in Ireland for that long?? I don't believe it.D > ? Any useful work done on that machine without mistakenly shuttingD > ? down, switching off or otherwise screwing up?? I don't think so. > ? What version of VMS? > ? Why upgrading now?$ > ? Questions, questions, questions. > ?: > ?5 >  >m > Sent via Deja.coms > http://www.deja.com/   ------------------------------  % Date: Thu, 01 Feb 2001 00:19:16 +0100c) From: Brass Christof <brass@infopuls.com>d' Subject: Re: Try this on Linux or NT/MEr7 Message-ID: <20010201001916.H21381@mozart.infopuls.com>t  O On Wed, Jan 31, 2001 at 01:56:33PM +0000, Steve.Spires@yellowpages.co.uk wrote:rN > Contact:   Tel: 3063  -  IS - Infrastructure, 1st Floor, Bridge Street Plaza >  > 2 > No. This is NOT an urban legend. 'Tis very true. > E > I'm having an email discussion with their system manager right now.K >  > Steve Spires >  >  >  > > > I regard this as a modern urban legend in its shortest form.@ > No power outage in Ireland for that long?? I don't believe it.B > Any useful work done on that machine without mistakenly shuttingB > down, switching off or otherwise screwing up?? I don't think so. > What version of VMS? > Why upgrading now?" > Questions, questions, questions.  % Could you comment about my questions?n   ------------------------------  # Date: Wed, 31 Jan 2001 23:52:25 GMT(/ From: "Tom Simpson" <simpsont@xxx.mediaone.net>h' Subject: Re: Try this on Linux or NT/MEtD Message-ID: <Zy1e6.1505$En4.30640@typhoon.jacksonville.mediaone.net>  K I'd be happy if I could go a whole 60 days without crashing one node or thet othereF in my 2 node ES40 cluster!  The Alpha systems aren't even close to the stability of the VAX6000 series (IMHO).   -Tom  . "Don Sykes" <don@alphase.com> wrote in message% news:3A7712ED.978DAC4F@alphase.com...nG > The Oakland police dept went into production with their 911 emergencyrH > dispatch system in 1991. It's now a DEC 7000 Model 710 cluster running > VMS Ver.V6.2.oG > The application tracks all 911 calls (thousands per day) and controls-H > all police dispatching for the city (pop ~300,000). I just checked andJ > the application and OS have been running continuously, 24/7, for over 13: > months. It was last rebooted "14-DEC-1999 06:37:31.00" !1 > Can any other OS point to successes like this ?-J > I know I have to reboot my NT 4.0 W/S every few days. Linux might last a > few weeks. >h >t >    ------------------------------  % Date: Wed, 31 Jan 2001 17:12:10 -0600@/ From: Gloria Griffith <qusgagh@am1.ericsson.se> $ Subject: Re: Where to get TeX/LaTeX?/ Message-ID: <3A789BCA.76ED3429@am1.ericsson.se>i  K I have a problem on a OVMS system that will not allow any ftp access to theXJ machine. The machine can be accessed by telnet, set host etc., but the ftpN command fails. I have compared TCPIP configuration on the problem machine withJ the working machines and I have been unable detect any error. I have TCPIP version 5.A installed.  M This is the error I get when I try to ftp to this DS20E Alpha Server, runningM Open VMS 7.1-2   $ Ftp gc04bs0 %TCPIP-E-FTP_NETERR, I/O error on network device4 -SYSTEM-F-REJECT, connect to network object rejected $ UptimeL OpenVMS V7.2-1  on node GC04BS  31-JAN-2001 12:50:48.00  Uptime  32 01:44:33 $     4 However, I can ftp from the same server. (see below)     $ ftp gc01bm* 220 gc01bm FTP Server (Version 4.2) Ready.$ Connected to cowboy.exu.ericsson.se., Name (cowboy.exu.ericsson.se:system): system' 331 Username SYSTEM requires a Password 	 Password:  230 User logged in.- FTP> bye 221 Goodbye. $ uptimeL OpenVMS V7.2-1  on node GC04BS  31-JAN-2001 12:46:20.64  Uptime  32 01:40:06 $3   Gloria Griffith  VMS System Administrator office: 972-583-7052 cell:   214-725-0820 May the data be with you!        >- >    ------------------------------  % Date: Wed, 31 Jan 2001 14:33:18 -0600g1 From: "Dave Gudewicz" <david.gudewicz@abbott.com>.0 Subject: Re: Writes flushed to Disk on Dismount?8 Message-ID: <959svv$n1e$1@fizban.fizban.pprd.abbott.com>  I Ken has stated rather categorically, that HSG and HSZ type controllers do@L NOT flush their caches upon issuing a dismount command.   If the cache flushL timeout is set high, like Ken recommends, then its very likely that stuff isH sitting in the controller's cache after the dismount and not on the disk	 platters.f  I He also mentioned the disk's own cache (not the controller's) but that is 3 usually disabled when used in these type scenarios.t  I Ken also stated that HSJ controllers use the MSCP (not SCSI) protocol andrH they indeed DO flush cache upon issuing a dismount command by use of the AVAILABLE command.  K I have no way of easily testing this, but Ken is one who usually does speak - with some authority in matters such as these.y     Dave...R  9 "Malcolm Wade" <Malcolm.Wade@asx.com.au> wrote in message-@ news:A8854B7F33E5D3119FDA00508B6A8C6334B18A@ASX235.asx.com.au...D > > From: young_r@eisner.decus.org [mailto:young_r@eisner.decus.org], > > Sent: Saturday, January 27, 2001 3:06 AM > > To: Info-VAX@Mvb.Saic.Comc4 > > Subject: Re: Writes flushed to Disk on Dismount? > >  > > 8 > > In article <3A719A97.6090803@wi.rr.com>, Scott Vieth > > <svieth@wi.rr.com> writes:A > > > My best guess is that if you had an application doing a ton- > > of writes,! > > > the controllers could cache2= > > > a lot of blocks without having to start writing them to  > > disk after 45  > > > seconds.  By setting the; > > > timeout to 64k second, you are telling the controllere > > "Write these tor% > > > the disk when you get a chance.t) > > > There's no hurry.  Take your time."p > > >S > >17 > > But this isn't accurate and which raised my initiall2 > > thread.  What it really means is:  "hold these7 > > writes for 65K seconds before flushing"  of course,V; > > the reality is most of those get pushed out in a modestr: > > write environment.  But during low I/O times one could> > > imagine a write is in cache for a few hours before getting > > committed. >oK > Is this really the case?  In my HSG80 V8.5 CLI reference manual it statesDH > "Specifies how may seconds (1-65535) of idle time on a unit may elapseL > before the write-back cache flushes its entire contents to the idle units" >sJ > My reading of this is that if you're constantly writing to disk it stays inI > cache (until the cache is full obviously???) and only when the disk hast beenH > idle for CACHE_FLUSH_TIMER value does the cache get flushed.  This was > confirmed to me by CPQ local.o >GL > If you shut your system down or dismount the disk, after CACHE_FLUSH_TIMER" > time the cache will get flushed. > J > I actually changed my value down to 1 second.  By my reading, pushing it out G > to a large value doesn't make any sense; if you ain't using the disk;t flusha > the cache! >eK > Of course if you're really paranoid; before dismounting the shadow member J > issue a SET unit NOWRITEBACK_CACHE on the controller and turn it back on > again on the unit later. >-1 > Seems like we need a definitive answer from KB.e >e	 > Cheers,.	 > Malcolm. >l > Malcolm J. Wade:4 > Senior Systems Engineer, Australian Stock Exchange5 > Level 3, 20 Bridge St, Sydney, NSW, 2000, Australiao= > PO Box H224, Australia Square, Sydney, NSW, 2000, AustraliaaB > Tel: +61 2 9227 0263, Mobile: 0417 046 925, Fax: +61 2 9227 0023  > Email: Malcolm.Wade@asx.com.au > Web: http://www.asx.com/ >a   ------------------------------   End of INFO-VAX 2001.063 ************************