1 INFO-VAX	Mon, 30 Oct 2000	Volume 2000 : Issue 607       Contents: Re: Backup tape question Re: Backup tape question Re: Disk mounting problem  Re: Disk mounting problem  Re: Disk mounting problem  RE: Disk mounting problem  Re: Disk mounting problem  Re: eBay (guess what) again. Re: eBay (guess what) again. Re: eBay (guess what) again. Re: eBay (guess what) again. Re: Ethernet Problems? File signatures under VMS  Re: File signatures under VMS  Re: File signatures under VMS $ Re: FTP from VMS to the demon's seed/ Re: including form setup mods in PDF conversion 4 Re: Narrow compatible SCSI controller for OpenVMS... NAT for VMS ! Re: Nice job - porting VNC to VMS ! Re: Nice job - porting VNC to VMS  Re: open an Imac? ( Re: OpenVMS Disk Services for Windows NT( Re: OpenVMS Disk Services for Windows NT= Re: Oracle/VMS performance problems with Multithreaded Server  Re: PHP for VMS?" Problems zipping big Oracle export& Re: Problems zipping big Oracle export& Re: Problems zipping big Oracle export& Re: Problems zipping big Oracle export Re: Rdb on VMS reference?  Re: Rdb on VMS reference?  Re: Rdb on VMS reference?  Re: Rdb on VMS reference?  RE: Rdb on VMS reference?  SHAREWARE: txt2pdf 4.1 Re: Sun "uptime" belly laugher Re: Sun "uptime" belly laugher RE: Sun "uptime" belly laugher Re: Sun "uptime" belly laugher Re: Sun "uptime" belly laugher Re: Sun "uptime" belly laugher RE: Sun "uptime" belly laugher Re: Sun "uptime" belly laugher Sun Moon Re: Sun Moon Re: TEK 4111 emulator?F Tieing Oracle 8 usage to a single username (was: Urgent Help Needed..) Re: Urgent Help Needed..... . VIOC max transfer blocks before cache bypassed2 Re: VIOC max transfer blocks before cache bypassed( Re: Virtual addresses with Multinet-4.2? Re: VMSINSTALL Question  Re: VMSINSTALL Question  Re: What does DCPS really do ?' Re: What VMS format are these files in? ' Re: What VMS format are these files in?   F ----------------------------------------------------------------------  % Date: Mon, 30 Oct 2000 09:05:07 +0000 2 From: Chris Sharman <Chris.Sharman@CCAgroup.co.uk>! Subject: Re: Backup tape question . Message-ID: <39FD39C3.16482CF6@CCAgroup.co.uk>   Steven Xie wrote:  >  > Hello there, > J > I have a question to ask regarding to the backup tape. I have a dlt tape > drive C > attached to my alpha server (OS version is V7.2-1). When I do the  > backup, I H > do  backup/norewind... so every time when the back up get executed, it > willD > continue with last time's backup. My question is, how do I know my
 > backup tape I > is getting close to the end and is there a way to determine that I need  > change > the tape?   = The tip I got here (can't remember who from, sorry) was to do G f$getd(...,"opcnt") to count tape operations before & after. Operations D * blocksize ( / 2 if verifying) is about the best guide available toA data written. If your DLT tape is doing compression, though, it's F anyone's guess what your capacity will be. Our TZ87 can vary from 17Gb? to 19Gb on consecutive nights, on broadly the same data. You're E presumably doing a 'set mag/skip=end' to position before the backup ?   F If you're doing backup in batch, I don't know of any way of preventingC it requesting a second tape: you have to do reply/abort or whatever $ (interactively or programmatically).   Chris    ------------------------------  % Date: Mon, 30 Oct 2000 07:19:37 -0500 2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>! Subject: Re: Backup tape question 7 Message-ID: <200010300719_MC2-B8DF-7AE7@compuserve.com>   E         BACKUP will skip to the end of the tape on its own unless you  specify /REWIND.  % Message text written by Chris Sharman 
 <Big snip> >You're H presumably doing a 'set mag/skip=3Dend' to position before the backup ?<
 <big snip>   ------------------------------    Date: 30 Oct 2000 08:36:53 -0500, From: koehler@eisner.decus.org (Bob Koehler)" Subject: Re: Disk mounting problem+ Message-ID: <jI3N8ESTG9SM@eisner.decus.org>   E In article <39FAAA68.30378.589DF6E@localhost>, stan@stanq.com writes:  > G > No, but you can access the files using DECnet.  Maybe you can create   > a logical name:  > 1 > 	$ define/sys/exec DISK$XYZ XYZ::SYS$SYSDEVICE:  >  > Haven't tried this, tho...  F Yes, this works (for all functions supported by DECnet).  You can evenD embed access information in the logical name if you don't care about@ others seeing it (it's fairly hard to see someone else's process
 logicals).  3    $define remote node"""user pass"""::device:[dir]     $set default remote:   E And it works for any system that supports DECnet FAL, not just VMS (I 5 used to edit all my TOPS-20 files with EDT this way).   F ----------------------------------------------------------------------? Bob Koehler                     | Computer Sciences Corporation = NASA GSFC Flight Software       | Federal Sector, Civil Group E                                 | please remove ".aspm" when replying    ------------------------------  % Date: Mon, 30 Oct 2000 14:18:42 +0100 . From: "David Otten" <david.otten(at)gtech.com>" Subject: Re: Disk mounting problem) Message-ID: <8tjsfm$l941@news1.gtech.com>   I SLS's offering in this regard was bought from Touch Technologies: TTI_RTF  and TTI_RDEV   http://www.ttisms.com/rdf.html  J Allows you to use remote tape drives as though they were local. I THINK it9 would allow shared access of disks, but I never tried it.      Hope this helps    David.Otten@nospam.gtech.com   ------------------------------  # Date: Mon, 30 Oct 2000 16:29:12 GMT / From: Mike Price <mike.price@littlewoods.co.uk> " Subject: Re: Disk mounting problem) Message-ID: <8tk7kc$nls$1@nnrp1.deja.com>   D We had a contractor in recently that mentioned a third party productG that has been produced as a rival to shadowing - it doesn't require the   remote nodes to be in a cluster.A Can't remember who produced it as we have proper shadowing but it ' sounded like it would do what you want.   & Does anyone know what the product is??* if not try a search for VMS and shadowing?   Mike --B All opinions are my own and do not necessarily reflect those of my employer    & Sent via Deja.com http://www.deja.com/ Before you buy.    ------------------------------  % Date: Mon, 30 Oct 2000 12:03:04 -0500 ; From: "Everhart, Glenn (FUSA)" <GlennEverhart@FirstUSA.com> " Subject: RE: Disk mounting problemN Message-ID: <4B279CF3578CD211B945009027178017046D00B9@swilnts809.wil.fusa.com>  * There are free remote tape and remote disk+ packages on sigtapes. tcp/ip remote tape is & also there. You get sources; I use em.   -----Original Message-----1 From: Chuck Chopp [mailto:ChuckChopp@rtfmcsi.com] ) Sent: Saturday, October 28, 2000 10:09 AM  To: Info-VAX@Mvb.Saic.Com " Subject: Re: Disk mounting problem     Bruce Pagram wrote:    > Good morning > B > Is it possible to mount a volume on one cluster from a different* > cluster?  NFS is out due to lack of UCX. >  > Thanks in advance  >  > Bruce   D You might try going to http://www.advsyscon.com and looking at theirB DeviceShare product.  It allows you to share disk and tape devicesE between non-clustered systems over the network via TCP/IP and DECnet. 4 It might allow you to do what you are wanting to do.     -- Chuck Chopp   8 ChuckChopp@rtfmcsi.com            http://www.rtfmcsi.com0                                   ICQ # 22321532@ RTFM Consulting Services Inc.     864 801 2795 voice & voicemail2 103 Autumn Hill Road              864 801 2774 fax4 Greer, SC  29651                  800 400 4935 pagerC                                   8004004935@alphapage.airtouch.com    ------------------------------  # Date: Mon, 30 Oct 2000 16:42:12 GMT % From: A.Greig@virgin.net (Alan Greig) " Subject: Re: Disk mounting problem2 Message-ID: <39fda1e6.1820017586@news.newsguy.com>  E On 30 Oct 2000 08:36:53 -0500, koehler@eisner.decus.org (Bob Koehler)  wrote:    F >And it works for any system that supports DECnet FAL, not just VMS (I6 >used to edit all my TOPS-20 files with EDT this way).  C A slightly easier way would have been to use the TOPS-20 version of > EDT would it not? Although I'm not sure when this first becameE available and it didn't ship with TOPS-20 but came on the TOPS-20/VMS 1 integration tape which could be ordered for free.   D But back then my editor of choice on TOPS-20 was EMACS. That was the2 version of EMACS that could execute TECO commands.   --
 Alan Greig   ------------------------------  % Date: Mon, 30 Oct 2000 10:35:18 +0100 = From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com> % Subject: Re: eBay (guess what) again. ) Message-ID: <39FD40D6.200E97AA@gtech.com>    "Webb, William W" wrote:N > http://dailynews.yahoo.com/htx/cn/20001023/tc/ebay_blames_upgrade_for_system > _troubles_1.html >   > Monday October 23 10:00 PM EDT) > eBay blames upgrade for system troubles  > " > By Troy Wolverton, CNET News.com > C > eBay apologized to customers Monday for a system upgrade that has E > led to intermittent outages and frustrating service problems during  > the past two weeks.  > B >  In a note to members on its annoucements board, eBay apologizedE > for the problems and blamed them on an upgrade of its database that  > it began in August.   C It sounds as if eBay are simply not capably of achieving acceptable  uptime.    Arne   ------------------------------  % Date: Mon, 30 Oct 2000 17:48:48 +0000 0 From: andrew harrison <andrew.nospam@uk.sun.com>% Subject: Re: eBay (guess what) again. * Message-ID: <39FDB480.8738A244@uk.sun.com>   Jordan Henderson wrote:  > , > In article <39F961BA.9CB2B675@uk.sun.com>,4 > andrew harrison  <andrew.nospam@uk.sun.com> wrote: > >Warren Spencer wrote: > >>5 > >> wwebb1@email.usps.gov (Webb, William W) wrote in C > >> <D46FE9B132FB9B44AEC242A96E4AB7502CB284@rlghncst625.usps.gov>:  > >> > >> -- snip --  > >>F > >> >As part of the upgrade, eBay has begun to divide its system intoF > >> >different databases and onto different systems. The goal is thatG > >> >if one feature on the system goes down, the rest of the site will + > >> >still be available, the company said.  > >> > > >> > >> -- snip --  > >>N > >> It appears they are partitioning the application to increase availabilityM > >> and scalability.  If so, IMHO, it's a clumsy and course hack that likely J > >> would be unneccessary it the application was hosted on OpenVMS, whereD > >> galaxy and clusters provide native partitioning and incremental > >> scalability.  > >>C > >So if you had an Oracle Database running on say a 2 node OpenVMS C > >cluster for performance reasons. How would you tune the database B > >if for example you wanted to add another 2 nodes and run OPS on? > >4 rather than 2 nodes. How would you ensure that the queries , > >are evenly balanced across all the nodes. > > @ > >Of course you don't have to partion the database but rememberB > >you want the best performance ideally you are aiming for linear@ > >scalability into your single DBMS by adding additional nodes. > > ? > >Get back to me when you have though it out and please please B > >don't come back with OpenVMS clusters just do it transparently, > >they don't. > >  > < > This is how Andrew spins here, knocking down one issue and; > ignoring the real issue at hand.  Here he focusses on the < > scaleability aspect that wasn't actually mentioned by eBay; > in the quoted paragraph, ignored the availability aspect, < > which is what was at issue and was given as the reason for
 > the change.  > 5 Jordan it doesn't make any difference. OPS gives you  < better availability and if you don't care about scalability 8 then you don't need to partition your users or you data.  9 If however you want availability and scalability then for 7 scalability reasons you need to partition your data and 5 your users which is a much trickier proposition than    simply running OPS on two nodes.  8 Strangely this isn't spin as any Oracle OPS trained DBA  will tell you.  8 > What we can gather from eBay's statements is that they< > continue to have availability problems with their back-end= > database servers (which are all Sun/Oracle, I believe), and @ > they are trying yet something else to improve this sorry stateA > of affairs.  You'd think that Sun would make sure that they are C > a shining example of availability since the embarrassing failures B > of the past, but apparently they've been unable to do this, even@ > after much talk about how Sun engineers were working with eBay > to improve the situation.  > 6 Does the post say this, more FUD based on no evidence 
 on your part.   : You have consigned yourself to the Rob Young backing choir  which is a very sad place to be.  < > Of course, we can't _prove_ that Sun HW/SW has anything to; > do with this (not surprising with Sun's history of having > > customers sign NDAs in order to get service), but there sure7 > are more and more indications that Sun reliability is A > questionable.  And, it's not just people in comp.os.vms who say > > this.  I redirect everyone's attention to the recent Gartner > statements reported here:  >     R > http://x63.deja.com/getdoc.xp?AN=686121920&CONTEXT=972659765.1119944705&hitnum=0 > @ > Back to the issue of partitioning for scaleability, since this? > is all Andrew wants to talk about.  Sure, it's a great thing. ? > I wonder if we could get Andrew to admit that Galaxy provides B > flexibility in this area that Sun cannot?  Naaah, he's just here" > to tear down Compaq and OpenVMS. >   8 Right Jordan so you are happy to try to FUD Java on the : basis of what Sun might do but hasn't since Java appeared,4 you continue to try to FUD Sun over eBay and at the 5 same time you accuse me of trying to tear Compaq and  
 OpenVMS down.   : As I said in a previous posting get more mirror time look  long and hard.  8 The real damage isn't being done by me its being done by9 ill informed OpenVMS boosters posting stuff either about  : OpenVMS or about UNIX which is total B***S**T particularly1 when the masses fall for it hook line and sinker.   8 As for Galaxy, show me a customer reference for a large 7 database implimentation using Galaxies on Alphas and I  ) will sit up and take a bit more interest.   3 As Mr Heller once said, technology is like fish the 3 longer you leave it on the shelf the less appealing  it is.   Regards  Andrew Harrison  Enterprise IT Architect    ------------------------------  % Date: Mon, 30 Oct 2000 13:18:38 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca> % Subject: Re: eBay (guess what) again. , Message-ID: <39FDAD6D.45C99AB9@videotron.ca>   andrew harrison wrote:; > If however you want availability and scalability then for 9 > scalability reasons you need to partition your data and 6 > your users which is a much trickier proposition than" > simply running OPS on two nodes. > 9 > Strangely this isn't spin as any Oracle OPS trained DBA  > will tell you      OK, here is a serious question:   M Is this where 64 bit memory and VLM make a significant difference because it  ? allows you to scale a single database without partitioning it ?    ------------------------------    Date: 30 Oct 2000 13:52:32 -0500* From: young_r@eisner.decus.org (Rob Young)% Subject: Re: eBay (guess what) again. + Message-ID: <woLGt9RIrgch@eisner.decus.org>   ] In article <39FDB480.8738A244@uk.sun.com>, andrew harrison <andrew.nospam@uk.sun.com> writes:   8 > Does the post say this, more FUD based on no evidence  > on your part.  > < > You have consigned yourself to the Rob Young backing choir" > which is a very sad place to be. >   	 	Ha-ha-ha   > 	Go visit the other thread Andrew... you have the *beginnings* 	of a lot of work!   				Rob    ------------------------------  % Date: Mon, 30 Oct 2000 17:53:56 -0000 - From: wspencer@ap.nospam.org (Warren Spencer)  Subject: Re: Ethernet Problems? / Message-ID: <svrddk8u6kdl33@news.supernews.com>   C jweisen@my-deja.com (fooguy) wrote in <8tcg0h$mj$1@nnrp1.deja.com>:   A >All of the sudden today I'm seeing the following error on the OP 	 >Console:  > 9 >%%%%%%%%%%%  OPCOM  27-OCT-2000 11:28:36.42  %%%%%%%%%%% # >Message from user SYSTEM on SHRIKE @ >Event: Alignment Error from: Node LOCAL:.SHRIKE CSMA-CD Station
 >CSMACD-0,. >        at: 2000-10-27-11:28:36.420-04:00Iinf8 >        eventUid   4A1E5122-ABFC-11D4-963C-534852494B458 >        entityUid  A904F95B-A522-11D4-8003-AA00040005048 >        streamUid  AC3329C2-A522-11D4-8003-AA0004000504 >m >t >$9 >%%%%%%%%%%%  OPCOM  27-OCT-2000 11:31:11.84  %%%%%%%%%%%b# >Message from user SYSTEM on SHRIKE B >Event: Frame Check Error from: Node LOCAL:.SHRIKE CSMA-CD Station
 >CSMACD-0,. >        at: 2000-10-27-11:31:11.840-04:00Iinf8 >        eventUid   A6C185E8-ABFC-11D4-963D-534852494B458 >        entityUid  A904F95B-A522-11D4-8003-AA00040005048 >        streamUid  AC3329C2-A522-11D4-8003-AA0004000504 >FH >I assume it's an ethernet frame error. The system is connected directly9 >to a 100mbt switch, and the device stats look like this:p >i >$ SHOW DEV EWA0/FULLn > H >Device EWA0:, device type DE500, is online, network device, device is a >template only.  > ? >Error count                    0    Operations completed     0H? >Owner process                 ""    Owner UIC         [SYSTEM]i? >Owner process ID        00000000    Dev Prot S:RWPL,O:RWPL,G,W ? >Reference count                0    Default buffer size    512e >d >Any thoughts? >  >-- . >*********************************************) >"All I every wanted from life was to seet- >Larry Wall give Bill Gates a Perl Necklace."  >  >t' >Sent via Deja.com http://www.deja.com/y >Before you buy.  H Those look like physical errors to me. I'd bet on a faulty network card J somewhere on the local lan segment.  If you're using DECNet, issue a SHOW L EXEC COUNTERS command from the NCP prompt.  Are any of the "error" counters  greater than zero?  Which ones?e  C You can use a network sniffer to track down the MAC address of the aD offending network card - it could be any card attached to the local L segment.  Some sniffers have the ability to translate MAC addresses to node A names.  Otherwise, you have to pop open boxes to locate the card.    ws   -- n3 << What if there were no hypothetical questions? >>-   ------------------------------  # Date: Mon, 30 Oct 2000 13:58:26 GMT  From: linuxmtl@my-deja.com" Subject: File signatures under VMS) Message-ID: <8tjupu$fh3$1@nnrp1.deja.com>a  	 Good day,   D I am looking for a third party software (or any solution) that would> allow to monitor any changes to files to help us preserve data+ integrity in a highly secured environmment.   & Example: sys$manager:systartup_vms.com  5          To be able to trace any changes on this filew.          and, of course, who made the changes.   Regards,      & Sent via Deja.com http://www.deja.com/ Before you buy.I   ------------------------------    Date: 30 Oct 2000 09:34:00 -05009 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)h& Subject: Re: File signatures under VMS+ Message-ID: <hjr9UqjWJQd9@eisner.decus.org>c  F In article <8tjupu$fh3$1@nnrp1.deja.com>, linuxmtl@my-deja.com writes: > Good day,o > F > I am looking for a third party software (or any solution) that would@ > allow to monitor any changes to files to help us preserve data- > integrity in a highly secured environmment.u > ( > Example: sys$manager:systartup_vms.com > 7 >          To be able to trace any changes on this filee0 >          and, of course, who made the changes.  E Tracking who does the changes is a normal capability of VMS Auditing.   B For an "inspect and compare" capability, you could use a CD-R copy with the DIFFERENCES command.-  M For something fancier, LJK/Security from my company provides such a function.<A Mail me your postal address and I will ensure you get literature.-   ------------------------------  # Date: Mon, 30 Oct 2000 14:56:53 GMTk8 From: DCantor@remove.three.words.shore.net (Dave Cantor)& Subject: Re: File signatures under VMS< Message-ID: <V_fL5.44063$P82.5556980@news1.rdc1.ct.home.com>  E In article <8tjupu$fh3$1@nnrp1.deja.com>, linuxmtl@my-deja.com wrote:-F > I am looking for a third party software (or any solution) that would@ > allow to monitor any changes to files to help us preserve data- > integrity in a highly secured environmment.n  t( > Example: sys$manager:systartup_vms.com  @7 >          To be able to trace any changes on this files0 >          and, of course, who made the changes.  N Make a list of the files you want to track (in a file, of course), and record J the highest block number, first free byte, and checksum of the file.  You L might want to record the FID, too, and owner UIC and SOGW protection codes,  and perhaps the ACL, if any.  O In order to check integrity, read the list of files, and for each entry, check uI that the file in question exists and that the version number recorded in yH your list of files is the current highest version number.  If it passes M those checks then check to see if it has the right highest block number (use -M F$FILE_ATTRIBUTES).   If it doesn't, the file has changed.  If it does, then  N check the first free byte number.  If that is different, the file has changed.L If not, then compare the checksum (using the undocumented CHECKSUM command).B (For image files, use CHECKSUM /IMAGE.)   See the VMS FAQ list at O http://Eisner.decus.org/vms/faq.htm, item MGMT46 for a (very) brief exposition m of the CHECKSUM command.  M Another attribute you might want to check is the revision number, so you can oK check for changes to the file attributes, even though the data in the file t itself is not changed.  M Of course, you need to protect the file which contains the list of files and  O attributes very carefully.  A skilled subverter will modify your list of files -L along with whatever other files they change so that you will not detect the O change.  You might keep the file attribute list offline (like on your PC), and DN upload it when (and only when) you perform the check, then delete it from the # VMS system after the check is done.e  M For tracing who made the changes, you'll need to turn on security auditing.  i> See the printed and/or on-line documentation and help files.      N For complete security, disconnect all network and modem connections from your 2 VMS system, and allow access only from OPA0:.  :-)     Dave C.i1 20 years managing VMS systems and still chugging.i   ------------------------------  # Date: Mon, 30 Oct 2000 13:55:05 GMTo- From: "Richard D. Piccard" <piccard@ohio.edu>-- Subject: Re: FTP from VMS to the demon's seed ( Message-ID: <39FD7DB7.CB35173C@ohio.edu>  G Actually, we find the visibility through the FTP software of ALL extant9- versions of the file to be a GOOD THING (TM).w  E It permit all users to download an earlier version of a Web page wheni@ they realize they have hosed it up on their PC -- the Web server functions as a backup.  G In six years, we have never been asked by any of our users to restore adE Web page file from backup tapes.  I'm sure it will happen eventually,:A but clearly it is much less common than would be the case if they G couldn't see and access the prior versions.  We tell them not to remove G the old versions until a few days after uploading the changes, for this  very reason.  D Visibility of all versions of each file also permits relatively easyG manual housekeeping for those who don't want to confront a VMS $ prompt- to give the purge command.  #                                 RDPt     JF Mezei wrote:    > Jerry Leslie wrote: * > > : DEFINE/SYSTEM TCPIP$FTP_NO_VERSION 1 > > UCX$FTP_NO_VERSION >c/ > Is there a reason why this is not a default ?2   --B ==================================================================B Dick Piccard                           Academic Technology ManagerB piccard@ohio.edu                                 Computer ServicesB http://oak.cats.ohiou.edu/~piccard/                Ohio University   ------------------------------   Date: 30 Oct 2000 16:45:39 GMT2 From: mathog@seqaxp.bio.caltech.edu (David Mathog)8 Subject: Re: including form setup mods in PDF conversion, Message-ID: <8tk8jj$5jk@gap.cco.caltech.edu>  F In article <8tetga$qh1$1@nnrp1.deja.com>, hnbowman@my-deja.com writes:1 >i am looking for a report distribution solution. 1 >what i want to do is to convert my OpenVMS printe2 >files into a PDF or PS format.  i have found many1 >tools that will do this, but i cannot find a wayt1 >to retain the formatting that is provided by the . >OpenVMS form setup modules.  the PDF/PS files1 >must be able to be formatted using the CPI, LPI,k/ >and LINES modules from a control library.  anyh >ideas?e  K Maybe you could redirect your existing queue (or a duplicate of it) so that-J the processed output file goes into a second queue on your system which isJ stopped.  The processed files could then be pulled out via a DCL procedure9 and run through Ghostscript (or whatever else you have.) m   Regards,   David Mathog mathog@seqaxp.bio.caltech.edum? Manager, sequence analysis facility, biology division, Caltech .   ------------------------------  % Date: Mon, 30 Oct 2000 06:52:00 -0500m) From: yyyc186.illegaltospam_@flashcom.nete= Subject: Re: Narrow compatible SCSI controller for OpenVMS...u9 Message-ID: <39fd611b$1$lllp186$mr2ice@news.flashcom.com>t  9 In <v04210124b620aa0dbcb7@[140.186.88.130]>, on 10/28/00 I6    at 12:08 PM, Dick Munroe <munroe@csworks.com> said:   Watch Ebay.h  F I bought the same dual (50pin + 68pin) SCSI controller on their for my' alpha PC164 that cost me $200+ for $50.s   Roland  J >Anybody have an recommendations?  I still have a bunch of narrow  devicesH >that I don't want to throw away.  Are there any PCI OpenVMS  compatibleI >controllers out there for OpenVMS?  Preferably not costing  an arm and aD >leg?B   >TIA   >Dick Munroe >--o, >Dick Munroe			(E) mailto:munroe@csworks.com. >Cottage Software Works, Inc.	(O) 978 815 1622 >PMB 361				(F) 617 489 0328- >464 Common St.			(W) http://www.acornsw.com/@ >Belmont, Ma. 02478    --  ; -----------------------------------------------------------gD yyyc186@flashcom.net              To Respond delete ".illegaltospam"6                             MR/2 Internet Cruiser 2.2a8                             For a Microsoft free univers; -----------------------------------------------------------    ------------------------------  # Date: Mon, 30 Oct 2000 13:24:58 GMT> From: trevor_deja@my-deja.com> Subject: NAT for VMS) Message-ID: <8tjsr5$dpb$1@nnrp1.deja.com>s   Hi  H I little while ago I mentioned here that I had written a NAT package for- VAX VMS, and some expressed interest in this.   5 I've called this package NATUCX and it's available at   ! www.galifrey.pwp.blueyonder.co.uka  E I don't claim it to be perfect but it does work.  I've had it runningb> for about a year ok and I've tested it on a second machine ok.  H It may be of interest to those who connect their VAX to the internet viaA an ISP, and would also like to give other computers access to theo Internet via the VAX.      Regards, Trevorl   trevor_deja@my-deja.como        & Sent via Deja.com http://www.deja.com/ Before you buy.u   ------------------------------  % Date: Mon, 30 Oct 2000 11:08:38 +0100e= From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com>o* Subject: Re: Nice job - porting VNC to VMS( Message-ID: <39FD48A6.1139514@gtech.com>   Larry Kilgallen wrote:Q > In article <39F8CAA5.F12DCBEF@isd.net>, Keith Brown <kbrown780@isd.net> writes:iK > > Having just tried the latest release of the VNC viewer on VMS I have toe4 > > say GREAT JOB to Hunter Goatley and David North. > F > Certainly that is unlikely if proponents such as yourself decline toH > mention in the newsgroup what it does.  I for one am unlikely to chase0 > URLs for something whose purpose is not clear.  3 VNC is a cross-platform "PC Anywhere"-like product.f  C You can view screen of another computer and send keyboard and mouseu input.   Arne   PS: I like VNC !   ------------------------------    Date: 30 Oct 2000 07:49:27 -05009 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)u* Subject: Re: Nice job - porting VNC to VMS+ Message-ID: <4McK9ov1C5ve@eisner.decus.org>s  h In article <39FD48A6.1139514@gtech.com>, Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com> writes: > Larry Kilgallen wrote:R >> In article <39F8CAA5.F12DCBEF@isd.net>, Keith Brown <kbrown780@isd.net> writes:L >> > Having just tried the latest release of the VNC viewer on VMS I have to5 >> > say GREAT JOB to Hunter Goatley and David North.s >> hG >> Certainly that is unlikely if proponents such as yourself decline tofI >> mention in the newsgroup what it does.  I for one am unlikely to chasen1 >> URLs for something whose purpose is not clear.r > 5 > VNC is a cross-platform "PC Anywhere"-like product.p  G It is important when describing things in comp.os.vms to avoid assumingnG that people have your own background in other operating systems.  There = is no "PC Anywhere" product that comes with VMS or Macintosh.h  E > You can view screen of another computer and send keyboard and mouseu > input.   That's better :-)e   ------------------------------  % Date: Fri, 27 Oct 2000 17:41:28 -0400i$ From: "Tim Kidder" <kkmfg@iserv.net> Subject: Re: open an Imac?> Message-ID: <iqgL5.1816$g66.119044@news-east.usenetserver.com>  L Sure, acquire a radial arm saw w/ approximately a 6" cutting disk. Place theG iMac on the table of the saw and carefully remove the side of the unit.s  H This really is not the correct newsgroup to ask the question in (and youL undoubtably know it). If I didn't hate the MacOS so much (and to some extentA macs in general) then I might not have given you a hard time. ;-)n    7 Bob Koehler <koehler@eisner.decus.org> wrote in message.% news:LpRy$k2tWqx7@eisner.decus.org...o >yI >   Before I dice into a flood of Mac newgroup messages, does anyone herea3 >   happend to know how to open the case on a iMac?  > H > ----------------------------------------------------------------------A > Bob Koehler                     | Computer Sciences Corporationa? > NASA GSFC Flight Software       | Federal Sector, Civil Group G >                                 | please remove ".aspm" when replyingr   ------------------------------  % Date: Mon, 30 Oct 2000 12:58:02 -0200i) From: fabio_compaq@ep-bc.petrobras.com.brs1 Subject: Re: OpenVMS Disk Services for Windows NTiL Message-ID: <OFF247D7C3.5EE72DFE-ON83256988.00521FA2@ep-bc.petrobras.com.br>  K I was just suggesting - If it is a Storage Product, should work under a SANu	 too, ok ?n   Regards,   Fabio C.                  6 Uwe Zessin <zessin@my-deja.com> em 28/10/2000 11:50:29L                                                                             L                                                                             L                                                                                 >                                                               >                                                               >                                                               >  Para:    Info-VAX@Mvb.Saic.Com                               >                                                               >  cc:      (bcc: Fabio dos Santos Cardoso/E-P-BC/Contratada)   >                                                               >                                                               >                                                               >  Assunto: Re: OpenVMS Disk Services for Windows NT            >                                                                             7 In article <OF09CA6482.FCB92949-ON83256985.0063F0EA@ep-  bc.petrobras.com.br>,h,   fabio_compaq@ep-bc.petrobras.com.br wrote:D > This product  looks like commited to the SAN estrategy  of Compaq,= > but I didnt find it related in the SAN homepage ! Did you ?  >i; > http://www.openvms.compaq.com/openvms/brochures/OPENVMSD/   > EH? NTDS goes via LAN, not SAN. Both offer block-oriented disk> access, but that's all. NTDS is even oder that SAN support for OpenVMS.   --
 Uwe Zessin    & Sent via Deja.com http://www.deja.com/ Before you buy.p   ------------------------------  # Date: Mon, 30 Oct 2000 16:21:04 GMTd% From: Uwe Zessin <zessin@my-deja.com>I1 Subject: Re: OpenVMS Disk Services for Windows NT ) Message-ID: <8tk758$n71$1@nnrp1.deja.com>8  7 In article <OFF247D7C3.5EE72DFE-ON83256988.00521FA2@ep-  bc.petrobras.com.br>, ,   fabio_compaq@ep-bc.petrobras.com.br wrote:G > I was just suggesting - If it is a Storage Product, should work under  > a SAN too, ok ?4  G It works by placing a container-file on an ODS-2 volume - I assume that4E it works on ODS-5 volumes, too. That volumes _might_ be fibre channel G connected disks, but I don't see this as a 'commitment' (your words) to  Compaq's SAN strategy.  F It just happens to work on such a disk like a compiler happens to workB there, too. Does that make the compiler a storage product? I don't& think you will try to tell me that ;-)   --
 Uwe Zessin    & Sent via Deja.com http://www.deja.com/ Before you buy.d   ------------------------------  % Date: Mon, 30 Oct 2000 13:20:21 -0500e& From: "Syltrem" <syltrem@videotron.ca>F Subject: Re: Oracle/VMS performance problems with Multithreaded Server8 Message-ID: <nSiL5.3548$Zq6.121484@wagner.videotron.net>   Oh! Oh!t  L When I read this, it sounded an alarm in my head! How could 7336 be the lastI "real" version of Oracle on OpenVMS, given the current agreements betweenrJ Oracle and Compaq? And how about the fact that the customer said they willG move to Unix? Hopefully not only because they think Oracle will be more  supported on that platform...-  K In short, OpenVMS is FULLY supported by Oracle, in details, see the commentiG on that from Alan Belancik, OpenVMS Relationship Manager with Oracle atS Compaq.e    H  As the OpenVMS Relationship Manager with Oracle, I am in a position toK tell you that our two groups are working very closely together. Our goal iseL to offer customers high performance solutions that are based on AlphaServersL running the Oracle database. These have been put to use in many of the world? 's most mission critical  applications requiring extremely hight
 availability. C Your question dealt with the nature of the Oracle database and its I applicability to the OpenVMS world. I can assure you that there is a very I astute team of engineers and managers in place who work on the porting of<J the Oracle database to OpenVMS. These people are very well acquainted withK the OpenVMS operating system and work to ensure that the products available I from Oracle take advantage of OpenVMS's unique attributes. Our pioneeringmL work in clustering has been reflected in Oracle's Parallel Server offerings.J Engineers at Oracle have been working on the interplay between new OpenVMSB technologies, such as the Galaxy Software architecture for dynamicL partitioning, and the new "WildFire" series of AlphaServers, the GS80, GS160
 and GS320.K True, the porting of the Oracle database begins with a set of common code,/L however once it is released to the various "platform groups" with in Oracle,I they begin to customize the final product to reflect the features of thatl* operating system and hardware combination.J Unfortunately, there is always some confusion in the field. We at DigitalL Equipment, and now Compaq, have unfortunately contributed to that confusion.H One item of confusion would be the perceived differences between VMS andK OpenVMS. Some people have erroneously concluded that VMS is for VAX systems J and OpenVMS is for AlphaServer systems. Actually, OpenVMS was the new nameJ applied to ALL VMS in order to better reflect its adherence to many of theL world's standards for openness. I mention this because you reference versionK 7.3.3 6 as being the last real VMS version. It is true that Oracle databasedD products were not ported to support VAX systems once Oracle made theL transition from their version 7.x products to their 8.x version products. ItL is very close to being correct that version 7.3.3.6 was the last version forF VAX systems. However, I do not see how this can be said of AlphaServer systems running OpenVMS.H Nonetheless, I will confer with my Oracle colleagues to see if they canI share any insight into this contention. Naturally, it is in any company'soI best interest to try to standardize the core of their product in order to G reduce operating expenses. Still, they need to be able to customize thehE final product to best suit the platform it is intended for as well astI optimizing its performance on that platform. We have been working closelyaH with the engineers in the porting group and in the performance groups to ensure that this is the case.sJ We will continue to work with Oracle as our premier database partner. OurK user surveys show that approximately 88% of our OpenVMS customers who use atJ commercial database use one from Oracle. We will continue to invest in ourL relationship. One benefit that we expect from this relationship is a productC from Oracle that is best suited for our AlphaServer systems running 	 OpenVMS..rI Please let me know if more information comes to light or how I can be ofpI further assistance. Again, thank you for bringing this information to ourmG attention. We share your belief that OpenVMS continues to be one of thef+ world's best operating environments - EVER!u Thanks and merci milles fois,e Alan
 Alan Belancikf Oracle Relationship Manageru OpenVMS System Software GroupS 603-884-0363 fax: 603-884-2006s    9 "I.A. Saez @tue.nl>" <"i.a.saez.<secret> wrote in messagee  news:39F91C4B.ECB2555E@tue.nl...
 > Malcolm, >eJ > We have the same problem with Oracle 7.3.4.4 on OpenVMS 7.1. We have now > a mix G > of MTS and dedicated connections. Dedicated for long running apps andbH > MTS for short connections. Oracle did investigate the problem and toldF > me it still exists in 8.1.6 for OpenVMS. Oracle 7.3.3.6 was the last7 > real VMS version (probably made by VMS especialists).d >  >r >  See (some) of ther tar: >a >e >  27-SEP-00 12:35:39g >o >a >o >  27-SEP-00 13:24:41  >n: >  To reproduce the slow MTS performance (for development). >  - The following patches are installed here:G >  686524-(Base BUG#544858) ORA-602 RUNNING PRO*C WITH TO_DATE AND BINDe
 >  VARIABLES. D >  1055554-(Base BUG#553138) ORA-12612 ON W95 CLIENTS AFTER APPLYING >  520734(430972.rG >  1218450-(Base BUG#845454) ORA-1403 REFERRING TO CERTAIN PL/SQL TABLEc >  ELEMENTS.1 >  1326425- ORA-1013 when sqlnet.expire_time > 0.M' >  1309610-(Base BUG#1021962) MTS SLOW.:6 >  1326725- Listener/Dispatchers consumes Buffered IO.D >  - Used the following MTS parameters in V7.3.4.4 INIT.ORA (OTS14): >0 >1 >  27-SEP-00 14:19:34c >d >  mts_listener_address =e4 > "(ADDRESS=(PROTOCOL=tcp)(HOST=nlvms2)(PORT=1521))": >  MTS_DISPATCHERS="(ADDRESS=(PARTIAL=TRUE)(PROTOCOL=TCP)\, >  (HOST=nlvms2)(PORT=5000))(DISPATCHERS=1)": >  MTS_DISPATCHERS="(ADDRESS=(PARTIAL=TRUE)(PROTOCOL=TCP)\, >  (HOST=nlvms2)(PORT=5001))(DISPATCHERS=1)" >  mts_servers = 2 >  mts_service = ots14B >  - Restart the instance --> will register the dispatchers at the
 > listener) >  - Now within the database using SCOTT:fE >  - Create a BIGEMP table from EMP with a large number of rows in it ? >  (I used BIGEMP with 458752 rows) ---> BIGEMP.DMP export file J >  - Create two connect aliases, one for MTS (OTS14) and one for DEDICATED > server
 >  (OTS14D) :  >  ots14.world = >  (DESCRIPTION =e >  (ADDRESS_LIST =
 >  (ADDRESS =  >  (PROTOCOL = TCP)i >  (Host = nlvms2) >  (Port = 1521) >  ) >  ) >  (CONNECT_DATA = >  (SID = ots14) >  (GLOBAL_NAME = ots14.world) >  ) >  ) >  ots14d.world =u >  (DESCRIPTION =  >  (ADDRESS_LIST =
 >  (ADDRESS =  >  (PROTOCOL = TCP)  >  (Host = nlvms2) >  (Port = 1521) >  ) >  ) >  (CONNECT_DATA = >  (SID = ots14) >  (GLOBAL_NAME = ots14.world) >  (SERVER = DEDICATED)  >  ) >  ). >  - Create the following test script TEST.SQL3 >  select to_char(sysdate, 'hh24:mi:ss') from dual;A >  set termout off;n+ >  select * from bigemp where ename='KING';  >  set termout on;3 >  select to_char(sysdate, 'hh24:mi:ss') from dual;  >  spool a.a >  select count(*) from bigemp;-
 >  spool off; 3 >  select to_char(sysdate, 'hh24:mi:ss') from dual;t >  exitD* >  - and reproduce the problem as follows:4 >  $ sqlplus scott/tiger@ots14d --> dedicated server
 >  SQL> @testd >  14:08:16 --> start time. >  14:08:55 --> end of query with where clause >  count(*) --> 458752 >  14:09:11 --> end timeI >  Total elapsed time : 39 seconds for the first part, 16 seconds for ther > second >  Now with MTS:0 >  $ sqlplus scott/tiger@ots14 --> shared server
 >  SQL> @testn >  14:12:15 --> start time. >  14:18:12 --> end of query with where clause >  count(*) --> 458752 >  14:18:29 --> end timeG >  Total elapsed time : 357 (!!) seconds for the first part, 17 secondse	 > for theo< >  second part, a huge difference with DEDICATED connection. >a >  >  27-SEP-00 15:47:41l >o >  .C >  Tested with V8.1.6.0 on OpenVMS (same configuration, same data):e< >  The following patches are installed on this installation:C >  patch#1335503 - ORA-1013 on queries when sqlnet.expire_time > 0.h6 >  $ sqlplus scott/tiger@ots16d ----> dedicated server
 >  SQL> @test, >  15:39:41 --> start time. >  15:40:16 --> end of query with where clause >  count(*) --> 458752 >  15:40:29 --> end timeI >  Total elapsed time : 35 seconds for the first part, 13 seconds for theO >  second part >  Now with MTS:2 >  $ sqlplus scott/tiger@ots16 ----> shared server
 >  SQL> @test1 >  15:45:26 --> start time. >  15:47:34 --> end of query with where clause >  count(*) --> 458752 >  15:47:47 --> end timeJ >  Total elapsed time : 128 seconds for the first part, 13 seconds for the >  second partE >  Conclusion: MTS with V816 is better performing than with V7344 butf	 > elapsedv8 >  time differs a lot from dedicated server connections.0 >  Above needs to be transferred to development. >V >  >  27-SEP-00 17:06:13  > E >  Advised customer to not use MTS where elapsed time is impacted andi
 > databaseJ >  links are accessed. Customer will setup prespawned server (although notH >  error free) this evening. AGreed on assisting in tuning such prespawn >  environment when needed...- >- >- >  27-SEP-00 17:33:19- >- >- >- >  04-OCT-00 12:47:50t > D >  Awaiting the results of the case as distributed to development... >  >2 >  12-OCT-00 13:02:28  > B >  Problem is reproduced by development and under investigation... >- >- >  13-OCT-00 15:06:49- >-" >  Discussed status with customer:J >  - Problem is under control as a mix is made with both MTS and dedicated1 >  (for the long running jobs) server connections$H >  - Explained that I do not expect a solution on short term based on myF >  contacts with development. As the problem also reproduces with V816H >  I expect the problem to be fixed in a future release (and if possible >  wil be backported)aJ >  - Customer stated that a move will be made to Unix within the next half > year >  === Agreed ===eH >  We decided to close the problem for now as a 'working' environment is > setupeE >  now. Will update this TAR with the 'internal' developments on thiso > matter and? >  therefore leave status at 14 'Partner (=development) Action'd >@ >d >. >o > Kind regards,a >j > Ivan Saezw > i.a.saez.scheihing@tue.nlu >  > Malcolm Dunnett wrote: >- >  >e > >,> > >   I'm running Oracle 8.1.6 on a VMS 7.2 system. I've tried@ > > setting up the MultiThreaded server on this system. It works> > > but the performance with my programs is terrible ( queriesH > > that take 1 second with a dedicated server process might take 5 or 6> > > with the Multithreaded server, for example ) [ ie all I doB > > to get the performance difference is change the client connect@ > > to include a "SERVER=DEDICATED" parameter]. The programs are> > > written using Oracle7 OCI calls, they fetch single rows of> > > data at a time from the database - I realize they could beB > > more efficient using array fetch, but it's never been an issue$ > > with dedicated server processes. > >oD > >   I realize this is a pretty broad question without more detail,C > > but does anyone have similar experiences ( or alternately, good:D > > experiences with the MTS on a VMS system ). Any ideas if there's+ > > anything I could do to make MTS useful?e > >- > >4L ============================================================================ =?> > > Malcolm Dunnett      Malaspina University-College   Email: dunnett@mala.bc.caL > > Information Systems  Nanaimo, B.C. CANADA V9R 5S5     Tel: (250)755-8738   ------------------------------  # Date: Mon, 30 Oct 2000 11:22:21 GMTI% From: A.Greig@virgin.net (Alan Greig)G Subject: Re: PHP for VMS?>2 Message-ID: <39fd591e.1801385214@news.newsguy.com>  @ On Fri, 27 Oct 2000 23:52:07 GMT, winston@SSRL.SLAC.STANFORD.EDU. ("Alan Winston - SSRL Admin Cmptg Mgr") wrote:    H >I could try hiring an experienced consultant to build DBD::Oracle on myH >system, but that would immediately become a money problem, and I'd look; >pretty stupid if it didn't talk properly to Rdb after all.   E We had an experienced Oracle Developer recently write a GUI (X based)wC application on VMS using Oracle Develper 2000 which used RDB as thenB back end rather that native Oracle. It all works perfectly but theD version of the RDB/Oracle interface shipping with RDB is not currentC or even complete. After talks with Oracle RDB support we discoveredgE that they can supply a much more complete version but only if you askr@ for it. I believe this is the version that will ship with RDB 8.   Worth getting hold of.   --
 Alan Greig   ------------------------------  # Date: Mon, 30 Oct 2000 15:38:29 GMTt/ From: Mike Price <mike.price@littlewoods.co.uk>i+ Subject: Problems zipping big Oracle exporto) Message-ID: <8tk4lk$ksb$1@nnrp1.deja.com>   B We have been having problems zipping some big Oracle export files.B After some helpful advise from this group (thanks to all ) we have/ downloaded the most recent version of Zip (2.2)o   We are still getting problems    We are on VMS 7.1-2o   The errors are as follows   ( EXODCS> zip x.zip enquiry_data_piece.dmp    adding: ENQUIRY_DATA_PIECE.DMP& zip warning: no such file or directoryG         zip warning: could not open for reading: ENQUIRY_DATA_PIECE.DMPt#         zip warning: zip file empty   B Just so you know, here is a full dir of the file in question. I amD running from an account with full priv (inc bypass sysprv et al) and( with fairly big wsdef and pagefile quota    & EXODCS> dir/ful enquiry_data_piece.dmp  ) Directory RAID_DATA2:[ORACLE7.DBA.EXPORT]0  2 ENQUIRY_DATA_PIECE.DMP;1                  File ID:	 (752,3,0)s1 Size:      8213320/8213328    Owner:    [ORACLE7]o" Created:    4-OCT-2000 02:00:09.53& Revised:   30-OCT-2000 15:17:26.22 (5) Expires:   <None specified>  Backup:    <No backup recorded>  Effective: <None specified>  Recording: <None specified>  File organization:  Sequential Shelved state:      OnlineC File attributes:    Allocation: 8213328, Extend: 256, Global buffer  count: 0$                     No version limit2 Record format:      Fixed length 4096 byte records Record attributes:  None RMS attributes:     None Journaling enabled: None> File protection:    System:RWD, Owner:RWD, Group:R, World:RWED Access Cntrl List:  None  ( Total of 1 file, 8213320/8213328 blocks.    ANy suggestions would be welcome     TIA    Mike --B All opinions are my own and do not necessarily reflect those of my employer    & Sent via Deja.com http://www.deja.com/ Before you buy.    ------------------------------  # Date: Mon, 30 Oct 2000 17:21:51 GMT + From: Craig A. Berry <calepine@my-deja.com>i/ Subject: Re: Problems zipping big Oracle export ) Message-ID: <8tkan5$qh1$1@nnrp1.deja.com>   ) In article <8tk4lk$ksb$1@nnrp1.deja.com>,92   Mike Price <mike.price@littlewoods.co.uk> wrote:D > We have been having problems zipping some big Oracle export files.D > After some helpful advise from this group (thanks to all ) we have1 > downloaded the most recent version of Zip (2.2)v  * It looks here as though the latest is 2.3:  7 <http://www2.wku.edu/scripts/fileserv/fileserv.com?ZIP>i    3 > Size:      8213320/8213328    Owner:    [ORACLE7]p  F You're pushing the 4GB limit of an individual file to be zipped.  See:  8 <ftp://ftp.freesoftware.com/pub/infozip/FAQ.html#limits>    & Sent via Deja.com http://www.deja.com/ Before you buy.s   ------------------------------  % Date: Mon, 30 Oct 2000 18:28:17 +0000g From: SysAdmin <djesys@fsi.net>m/ Subject: Re: Problems zipping big Oracle export0' Message-ID: <39FDBDC1.737FB9F2@fsi.net>a   Mike Price wrote:o > D > We have been having problems zipping some big Oracle export files.D > After some helpful advise from this group (thanks to all ) we have1 > downloaded the most recent version of Zip (2.2)r >  > We are still getting problemss >  > We are on VMS 7.1-2  >  > The errors are as followst > * > EXODCS> zip x.zip enquiry_data_piece.dmp  ( Unless you have set up a symbol like so:   $ ZIP :== ddcu:<dir>ZIP/VMSp  D ...the file as stored in the archive may not be usable on an OpenVMS system if you try to UNZIP it.  " >   adding: ENQUIRY_DATA_PIECE.DMP( > zip warning: no such file or directoryI >         zip warning: could not open for reading: ENQUIRY_DATA_PIECE.DMP0% >         zip warning: zip file emptyr  B Is the file open for write or exclusive access by another process?  tD > Just so you know, here is a full dir of the file in question. I amF > running from an account with full priv (inc bypass sysprv et al) and* > with fairly big wsdef and pagefile quota > ( > EXODCS> dir/ful enquiry_data_piece.dmp > + > Directory RAID_DATA2:[ORACLE7.DBA.EXPORT]c > 4 > ENQUIRY_DATA_PIECE.DMP;1                  File ID: > (752,3,0)c3 > Size:      8213320/8213328    Owner:    [ORACLE7]a [snip]  E As others have pointed out, the size of this file exceeds what can be  expressed in a signed longword:e   $ WRITE SYS$OUTPUT %X7FFFFFFFg
 2147483647   $ A = 8213320*512 
 $ SH SYM A6   A = -33427456   HEX = FE01F000   Octal = 37600370000  1 You need V2.3 of ZIP from Hunter's FILESERV site.l  ; Personally, I still can't believe that people are trying tocE compress/archive files of that size. I would have thought the compute.G power requirements would be prohibitive: ZIP typically eats up (almost)lG an entire CPU, VAX or Alpha, and a file of that size could take tens ofd- hours (VAX) or hours (Alpha) to compress, no?    David J. Dachterao   ------------------------------  % Date: Mon, 30 Oct 2000 18:35:33 +0000e From: SysAdmin <djesys@fsi.net>o/ Subject: Re: Problems zipping big Oracle exportO' Message-ID: <39FDBF75.F401803A@fsi.net>n   SysAdmin wrote:M [snip] > $ ZIP :== ddcu:<dir>ZIP/VMSh   That should be:i   $ ZIP :== $ddcu:<dir>ZIP/VMS   ..., of course!q   David J. Dachterao   ------------------------------  % Date: Mon, 30 Oct 2000 15:33:36 +0000 - From: Tim Llewellyn <tim.llewellyn@bbc.co.uk>n" Subject: Re: Rdb on VMS reference?) Message-ID: <39FD94D0.3FF543A4@bbc.co.uk>r   Alan Greig wrote:n   >  >tG > Oracle have dropped plans to release a commercial RDB for NT. However D > there is an unsupported RDB for NT available for download from the= > Oracle web site.  Free to any licensed users of RDB on VMS." >t  H Hmmm, just why, if you had a pucker Rdb for VMS, would you want to run a free,t unsupported one on NT?      --y6 Tim Llewellyn, OpenVMS Infrastructure, Remarcs Project0 MedAS at the BBC, Whiteladies Road, Bristol, UK.A Email tim.llewellyn@bbc.co.uk. Home tim.llewellyn@cableinet.co.uk   A I speak for myself only and my views in no way represent those ofb MedAS or the BBC.n   ------------------------------  % Date: Mon, 30 Oct 2000 17:25:47 +0100n= From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com> " Subject: Re: Rdb on VMS reference?) Message-ID: <39FDA10B.7D59FF06@gtech.com>f   Tim Llewellyn wrote:J > Hmmm, just why, if you had a pucker Rdb for VMS, would you want to run a > free,  > unsupported one on NT?   Training/playing-around.   Arne   ------------------------------  # Date: Mon, 30 Oct 2000 16:28:12 GMTo% From: A.Greig@virgin.net (Alan Greig)." Subject: Re: Rdb on VMS reference?2 Message-ID: <39fd9e62.1819117652@news.newsguy.com>  1 On Mon, 30 Oct 2000 15:33:36 +0000, Tim Llewellynt  <tim.llewellyn@bbc.co.uk> wrote:   >  >n >Alan Greig wrote: >e >> >>H >> Oracle have dropped plans to release a commercial RDB for NT. HoweverE >> there is an unsupported RDB for NT available for download from the > >> Oracle web site.  Free to any licensed users of RDB on VMS. >> >eI >Hmmm, just why, if you had a pucker Rdb for VMS, would you want to run ac >free, >unsupported one on NT?  >e  E I think that the NT code is RDB8 which isn't out for VMS yet (in lateeD Beta) so if you grab the NT version you can have a play with all theD new features and management front end tools that come with RDB 8 andD see just exactly how much of a help/annoyance they are. There's alsoE no check that you have a VMS license before you can download it so itt$ may just be there for legal reasons.  D Several people should really be shot in the RDB management team. DEC, spent a fortune porting RDB to OSF-1/DigitalF Unix/Tru-64/whatever-its-called-tomorrow only to sell it to Oracle whoE then dropped it. Oracle then spent a fortune porting it to NT only to D drop it (although blaming Compaq for forcing them to do so). I'm not2 sure if the NT port was initiated before the sale.  E Mind you with the message at that time being that VMS was legacy theye$ probably thought they had no choice.  E Still I don't find it amusing that VMS support contract fees followedmF by Oracle support fees were thrown away in this manner. Or that Oracle@ RDB 8 was scheduled to appear first on NT then on VMS. No wonder Oracle charges so much!e  
 What a waste.a   > --7 >Tim Llewellyn, OpenVMS Infrastructure, Remarcs Projecti1 >MedAS at the BBC, Whiteladies Road, Bristol, UK.oB >Email tim.llewellyn@bbc.co.uk. Home tim.llewellyn@cableinet.co.uk >tB >I speak for myself only and my views in no way represent those of >MedAS or the BBC. >h >a   --
 Alan Greig   ------------------------------  % Date: Mon, 30 Oct 2000 17:22:22 +0000a- From: Tim Llewellyn <tim.llewellyn@bbc.co.uk> " Subject: Re: Rdb on VMS reference?) Message-ID: <39FDAE4E.FD0BD86A@bbc.co.uk>o   Arne Vajhj wrote:   > Tim Llewellyn wrote:L > > Hmmm, just why, if you had a pucker Rdb for VMS, would you want to run a	 > > free,A > > unsupported one on NT? >3 > Training/playing-around. >F  J I guess  so as as long as the NT version is feature (and bug!) compatible.! I don't know if this is the case.E   --6 Tim Llewellyn, OpenVMS Infrastructure, Remarcs Project0 MedAS at the BBC, Whiteladies Road, Bristol, UK.A Email tim.llewellyn@bbc.co.uk. Home tim.llewellyn@cableinet.co.ukn  A I speak for myself only and my views in no way represent those of1 MedAS or the BBC.s   ------------------------------  % Date: Mon, 30 Oct 2000 12:55:01 -0500 ) From: "Ebinger . Eric" <EEbinger@drc.com>(" Subject: RE: Rdb on VMS reference?B Message-ID: <7162F87E9EF4D311BA9900805FC1D3AE7A622C@and02.drc.com>   > -----Original Message-----6 > From: A.Greig@virgin.net [mailto:A.Greig@virgin.net]G > then dropped it. Oracle then spent a fortune porting it to NT only totF > drop it (although blaming Compaq for forcing them to do so). I'm not4 > sure if the NT port was initiated before the sale.  5 Oracle's statement at the time of the cancellation ofB. Rdb on NT was that Rdb on NT relied heavily on6 the Bliss compiler for NT.  Compaq dropped support for6 the Bliss compiler on ALL versions of NT when Alpha NT was dropped by Compaq.    4 This explanation has the great virtue of fitting the3 facts.  It has the great disadvantage of not having( any black helicopters.  :^)      Eric Ebinger   ------------------------------  # Date: Mon, 30 Oct 2000 13:40:00 GMT(, From: SANFACE Software <sanface@sanface.com> Subject: SHAREWARE: txt2pdf 4.1 ) Message-ID: <8tjtnf$eji$1@nnrp1.deja.com>   5 txt2pdf is a very flexible and powerful PERL5 program @ (5 penguins at LinuxBerg, 5 cows at TuCows Mac, BeOS and PSION).5 It's a converter from text files to PDF format files.r Why do you need txt2pdf?% Most of your documents are text filestF Usually, your reports from legacy applications, DBs, ERP applications, datawarehouse are textualfG txt2pdf is a PERL5 script, so you can use it in every operating systemso supported by PERL5, : if you prefer we distribute also executables for these OS:E Windows,Linux,Solaris,AIX,HP-UX,FreeBSD. Inside the Windows version ar VB GUI (Visual txt2pdf)n= txt2pdf is a native converter, you don't need to pass through1 PostScript formatT2 txt2pdf is specific for text to PDF conversion, soC you can mark coloured (using PERL regular expression) frases in theg produced PDF filesF you can mark bold, italic, bolditalic frases in the produced PDF files% you can add page number in every page > you can add text at the beginning and at the end of every file" you can add a border to every page; every word like http://... ftp://... mailto:... https://...h. file:... ldap:... news:...  will become an URL> you can create a link to a specific page within a PDF document http://...pdfdocument#pdfmarkfC every word like mime:... will become a link that launch the correct3 application and opens the file, you can use background and foreground layersG every predefined encodings, supported by the PDF format, is supported +e the Unix default! txt2pdf supports STDIN and STDOUTc9 EPD 1.0 support. You can use EPD inside background layer.s  the fee for every licence is $85C SANFACE Software is going to give you a free licence for every goodl idea or for every good modifya   txt2pdf is shareware5 The txt2pdf source code is our company core business.q
 We trust you. $ You can test text2pdf and modify it.E You can't use a modify version of txt2pdf for production purpose. Youa+ can't resell txt2pdf or a modify version ofn* it without SANFACE Software authorization.C You can't copy part of it to include in your source without SANFACEd Software authorization.A   What's new in this version  A EPD ( http://EPD.sourceforge.net ) support in the background filee   Test txt2pdf 4.1!e) You can find it at http://www.sanface.com    -- SANFACE Software= Your technology glasses. We help you see your full potential.e http://www.sanface.com mailto:sanface@sanface.com% WAP       http://www.sanface.com/wap/-! i-mode  http://www.sanface.com/i/-    & Sent via Deja.com http://www.deja.com/ Before you buy.    ------------------------------  % Date: Mon, 30 Oct 2000 10:08:04 +0000 0 From: andrew harrison <andrew.nospam@uk.sun.com>' Subject: Re: Sun "uptime" belly laugher-* Message-ID: <39FD4884.7FEC2ABC@uk.sun.com>   "D.Webb" wrote:l > _ > In article <39F95D82.606B75A8@uk.sun.com>, andrew harrison <andrew.nospam@uk.sun.com> writes:e > >Jordan Henderson wrote: > >>2 > >> In article <972438729.683982@sj-nntpcache-5>,6 > >>   shawbrow@delilah.cisco.com (Shawn Brown) wrote:0 > >> > In article <8t22fk$f71$1@nnrp1.deja.com>,4 > >> > Jordan Henderson  <jordan@my-deja.com> wrote: > >> > >I > >> > >But, we've been told that you can patch Solaris with no downtime,a0 > >> > >just like you can with OpenVMS clusters. > >> >J > >> > Sure you can.  You can patch a suncluster or a vcs cluster by doing# > >> > rolling upgrades.  La de da.p > >> >, > >> > >What is it that you are doing wrong? > >> >K > >> > Don't be a jackass.  Apparently that's fairly difficult for you, butn > >> > lets try.  OK?i > >>D > >> I was just asking a question.  I don't know much about patchingJ > >> Solaris, but you seemed to be providing a data point that it's not as, > >> seemless as some would have us believe. > >>C > >Solaris 8 allows you to patch it while it is running. One postereA > >on this group pointed out that this is used to apply temporaryeF > >patches to the running OS. But you can of course patch the image onB > >disk as well so that your "temporary" patch becomes "permanent" > >when the OS is rebooted.h > >g > E > As I asked for last time you posted this. Please supply evidence of E > this ability to patch the image on disk (whilst the system is stilleL > running). VMS can in general do this (Though I am not sure VMS engineeringP > would go so far as to claim it can always be done) because running images willG > still refer to the previous version. However solaris does not support, > multiple versions of files.e > K > If the images you are referring to patching are read into memory when the P > system boots and the files are never therafter accessed then yes you should beG > able to do what you claim. However I would suspect that the domain ofg> > applicability for such patching would be relatively limited. > : All the Sun patching systems allow you to patch the image : on disk. You are supposed to reboot a node after applying ; the patches but this would only be necessary if you didn't >4 have the patched image already running which you do.  7 I can also apply a patch to a Jumpstart image which areO, generally used to upgrade OS images en mass.  = Now most people would probably still use the patch the image i3 and then reboot method even with Solaris 8 if they t5 had a cluster just as you would with OpenVMS but the d0 option does exist to patch the running image and avaoid the reboot.  A > >Of course Shawn may not be running Solaris 8 in which case youSC > >upgrade a node in the cluster and re-boot it just as you do withh > >OpenVMS.c > >t > 	 > Andrew,e > E > I don't know very much about Solaris Clusters but do multiple nodescK > share any of the main unix partitions (/, /usr, /var etc) in the same way H > that multiple VMS systems can share a system disk or does each Solaris= > cluster node have to have all its own OS  executables etc ?  > = Each cluster node has its own executables though most people -; also have a Jumpstart server which is used as a source for P; all the OS images. This of course makes it easy to upgrade n> an indevidual node but does mean that you need to ensure that 9 the nodes are configured the same. Hence why most people p use Jumpstart.   Regardsf Andrew Harrisoni Enterprise IT Architectw   ------------------------------  % Date: Mon, 30 Oct 2000 10:18:31 +0000t0 From: andrew harrison <andrew.nospam@uk.sun.com>' Subject: Re: Sun "uptime" belly laughern* Message-ID: <39FD4AF7.96F507ED@uk.sun.com>   Paul Sture wrote:n > =    > ___In articleoJ > <910612C07BCAD1119AF40000F86AF0D805284930@kaoexc3.kao.cpqcorp.net>, Ker= ry Main  > wrote: > =p  - > From: "Main, Kerry" <Kerry.Main@compaq.com>d > > Newsgroups: comp.os.vms + > > Subject: RE: Sun "uptime" belly laughern) > > Date: Sat, 28 Oct 2000 10:53:34 -0500  > >  > > Andrew,q > > # > > Combining two replies in one ..e > > A > > >>> Solaris 8 allows you to patch it while it is running. <<<D > > J > > So does OpenVMS. OpenVMS has static parameters and dynamic parameters= =2EAJ > > Changing the static ones requires a reboot. Changing dynamic ones doe= s not. > >eJ > > So does Solaris 8. Please correct me if I am mistaken, but ALL OS's o= n ALLdJ > > platforms have some key kernel mode parameters or cache type sizes th= at if < > > it is desired to change them, then a reboot is required. > >bJ > > >>> Kerry this is just marketing. You cannot do upgrades to a running=  system J > > with no impact to the users because users have to be disconnected fro= m thea. > > running node that needs to be upgraded.>>> > >wJ > > Not with OpenVMS clusters. We have Customers that today do this all t= he	 > > time.b > >.J > > One great feature of OpenVMS clusters is that this can be done with Z= EROwJ > > application availability impact. ie. no fail-over of processes, resou= rces,/E > > no restarting of applications, no disconnecting active users etc.a > >oJ > > As you have heard me explain many times before - (you must be tired o= f me > > explaining this by now)r > >eJ > > Using the shared disk architecture (which allows users to log into an= ynJ > > system for full read-write access to any files on any disks at the sa= meJ > > time) and the TCPIP V5 load brokers load balancing (standard DNS only=  doesvJ > > round robin), you simply disable logins on the system that is planned=  to beJ > > shutdown. This allows current connections / users to continue working= , butsJ > > all NEW connections are directed at other systems in the cluster. A n= umberOJ > > of hours later (next day?), when all db/user connections are all on o= therJ > > systems in the cluster, then that system which needs maint (tuning st= aticJ > > parameters, OS upgrades using 2 system disk rolling upgrade, OS patch= es, HWJ > > replace ect) can be shut down with ZERO application availability impa= ct.P > >mJ > Kerry, can we please nail this one shut. (I'm getting tired of the same=  old# > argument between you and Andrew).  > =e  J > Andrew is totally correct when he says that _I_ lose my sessions when a=  givenJ > node is closed. As a system manager I do not log out at the end of a da= y (WitheJ > an average of 80 sessions, would you?), although obviously my workstati= on iso > well and truly locked. > =   B Quite so Kerrys point about people being concerned about scheduled
 downtime =  J is not solved by OpenVMS by itself because upgrading a node will require =  G the people connected to that node either being disconnected or loging =E   out themselves.:  F Of course you can acheive a seamless failover from one node to anotherD without any downtime but to do this you need apps/middleware etc and> those apps/middleware etc also exist on non OpenVMS platforms.    J > But then, all those sessions are sitting at the command line prompt, wh= ich is$ > not the normal end user interface. > =s  J > When it comes to clients us=EEng OPS, or an assortment of middleware ta= lking toJ > something else, it is really not a problem. One node bombs due to a har= dwareRJ > problem (yes, we _do_ admit to the occasional one Andrew), the other no= de(s)19 > simply take over. The end user sees no loss of service.t > =t    D Quite so but unfortunately this is not unique to OpenVMS. Everyone =  A agrees that OpenVMS as a platform makes this easier but you can =1  3 acheive the same effect on other platforms as well.l  ? This has always been my point, Kerrys assertion that nodes in =l  = OpenVMS clusters can be rebooted with no impact ignores the =   4 sessions connected to the rebooted node. Of course =  > you can log them out (scheduled downtime for those sessions) =  8 but you can also do this with almost any other Cluster =   technology known to man.   Regardso Andrew Harrisons Enterprise IT Architect    ------------------------------  % Date: Mon, 30 Oct 2000 07:03:07 -0600c+ From: "Main, Kerry" <Kerry.Main@compaq.com>s' Subject: RE: Sun "uptime" belly laugher N Message-ID: <910612C07BCAD1119AF40000F86AF0D805284949@kaoexc3.kao.cpqcorp.net>   Andrew,   F >>> Quite so Kerrys point about people being concerned about schedule= dVF downtime is not solved by OpenVMS by itself because upgrading a node = willF require the people connected to that node either being disconnected o= rl logging out themselves. >>>R  ( Hey, I think its starting to sink in :-)  F >>> This has always been my point, Kerrys assertion that nodes in Ope= nVMS6 clusters can be rebooted with no impact ignores the=20+ sessions connected to the rebooted node.>>>w  F Oops, maybe not. Let me explain a little slower and in more detail. := -)  F When normal users go home at night or finish a shift or disconnect fr= om@ their modem, they log out. This means their current sessions are9 disconnected from whatever systems they are connected to.l   Do we agree on this?  F While there are exceptions where some users staying logged into a sys= tem for F days, most shops would say this is a security risk. Especially since = mostF peoples screensaver password is less than 3-4 characters and almost n= everF (ok, never) changes. So, this is why inactive connections after x hou= rs areF either killed by the system (Hitman stuff) or the application itself = timest' out by design and drops the connection.   C So, lets be clear - this is normal practice, not something special.?  F Assuming availability is a concern and OpenVMS Clusters are in place = - once again:  F - system planned to be shutdown in the next 8-16 hours (whatever - co= uld be= longer if long batch job running) has its logins disabled.=20oF - Current user sessions are not impacted. When done, they logout and = go homek' (or whatever) assuming normal users.=20eF - Since the system to be shutdown has had its logins disabled, the TC= PIPwF load broker updates the 8.1.2 BIND compliant DNS - (even a Sun box) t= o noF longer include that system in the list of available systems participa= ting insF the DNS TCPIP cluster alias. (I think this also works with TCPware an= du Multinet, but need to confirm)6 - all new sessions go to other systems in the cluster.F - when all sessions have moved to the other systems and there are no = longerF any active sessions on the node to be shutdown, then that system can = be, shutdown. No user notices, no fuss, no pain.F - when that system is rebooted, logins are re-enabled. The load broke= r seesF the logins re-enabled, updates the DNS with a "this system is again p= art ofF the TCPIP Cluster alias" record update message and that system again = starts5 to participate in the dynamic DNS load sharing again.   F Some Customers today tend to do things like disabling logins on the n= ode toF be shutdown before they go home at night so in the morning, they can = shut that system down.    No user impact.=20  2 No active users sessions are disconnected. Period.   No middleware. =20  2 Works with host based or 2 or 3 tier applications.  - Users do not even know a system was shutdown.c  F Works for any OpenVMS cluster application / db that uses a standard T= CPIP+ DNS which is at least BIND 8.1.2 compliant.   F Key to this concept working is to be able to log into any system in t= he> cluster, see all devices, files as if they were one system.=20  B With 2 system disks in place, even OS rolling upgrades can be done" transparently with no user impact.  F Also key note is that this concept will not work (see next point) on = any F platform that has devices and files allocated to one system ie. one t= hat-F transparently "serves" those files/devices to other cluster members, = becauseeF if the system that owns those devices/ files needs to be shutdown, th= enF those files and devices ownership needs to be failed-over to another = system# which will disconnect active users.o  F Hey, I am not saying this will solve world hunger, but this is, imho,=  anrF advantage of a shared everything cluster architecture ie. planned sys= temtF downtimes with ZERO impact on application availability and no users b= eing
 disconnected.I  F And, yes, for three tier apps, one might be able to design and add so= meF middleware logic to handle this for architectures that are shared not= hingF like UNIX and NT. This assumes you think about what the clients are d= oing,CF and how to stall them while the disks are failed oveer and applicatio= ns areF restarted on another node in the background, but what about the curre= ntF connections to that system being shut down? It gets a little hairier.=  What-F about 2 tier applications with no middleware logic? What about host b= ased0 applications (yes, there are still many around).  B Again, all of this can be transparent with current OpenVMS clusterF applications that are based on TCPIP DNS access to the cluster. Works=  with. DECnet as well btw.t   Cool eh? =20h Back to you Andrew.d   :-)    Regards,  
 Kerry Main Senior Consultants Compaq Canada Inc. Professional Servicesf Voice: 613-592-4660r Fax  :  819-772-7036 Email: Kerry.Main@Compaq.com       -----Original Message-----9 =46rom: andrew harrison [mailto:andrew.nospam@uk.sun.com]d Sent: October 30, 2000 5:19 AM To: Info-VAX@Mvb.Saic.Com ' Subject: Re: Sun "uptime" belly laugher-     Paul Sture wrote:  >=20 > ___In article F > <910612C07BCAD1119AF40000F86AF0D805284930@kaoexc3.kao.cpqcorp.net>,=  Kerry Main > wrote: >=20- > From: "Main, Kerry" <Kerry.Main@compaq.com>b > > Newsgroups: comp.os.vms + > > Subject: RE: Sun "uptime" belly laugherM) > > Date: Sat, 28 Oct 2000 10:53:34 -0500  > >f > > Andrew,o > >f# > > Combining two replies in one ..t > >oA > > >>> Solaris 8 allows you to patch it while it is running. <<<l > >oF > > So does OpenVMS. OpenVMS has static parameters and dynamic parame= ters.pF > > Changing the static ones requires a reboot. Changing dynamic ones=  doesp not. > >sF > > So does Solaris 8. Please correct me if I am mistaken, but ALL OS= 's on. ALLdF > > platforms have some key kernel mode parameters or cache type size= s that if< > > it is desired to change them, then a reboot is required. > >iF > > >>> Kerry this is just marketing. You cannot do upgrades to a run= ning systemF > > with no impact to the users because users have to be disconnected=  fromr thei. > > running node that needs to be upgraded.>>> > >eF > > Not with OpenVMS clusters. We have Customers that today do this a= ll the	 > > time.o > >tF > > One great feature of OpenVMS clusters is that this can be done wi= th ZERO.C > > application availability impact. ie. no fail-over of processes,s
 resources,E > > no restarting of applications, no disconnecting active users etc.  > >.F > > As you have heard me explain many times before - (you must be tir= ed of  me > > explaining this by now)o > >aF > > Using the shared disk architecture (which allows users to log int= o any F > > system for full read-write access to any files on any disks at th= e sameF > > time) and the TCPIP V5 load brokers load balancing (standard DNS = only doesF > > round robin), you simply disable logins on the system that is pla= nned tof beF > > shutdown. This allows current connections / users to continue wor= king,  butuF > > all NEW connections are directed at other systems in the cluster.=  A numberF > > of hours later (next day?), when all db/user connections are all = on othertF > > systems in the cluster, then that system which needs maint (tunin= gu staticF > > parameters, OS upgrades using 2 system disk rolling upgrade, OS p= atches,n HWF > > replace ect) can be shut down with ZERO application availability = impact.f > >eF > Kerry, can we please nail this one shut. (I'm getting tired of the = same oldg# > argument between you and Andrew).p >=20F > Andrew is totally correct when he says that _I_ lose my sessions wh= en a giveneF > node is closed. As a system manager I do not log out at the end of = a dayw (WithoF > an average of 80 sessions, would you?), although obviously my works= tation is > well and truly locked. >=20B Quite so Kerrys point about people being concerned about scheduled downtime=20-F is not solved by OpenVMS by itself because upgrading a node will requ= ire=20F the people connected to that node either being disconnected or loging= =20  out themselves.o  F Of course you can acheive a seamless failover from one node to anothe= raD without any downtime but to do this you need apps/middleware etc and> those apps/middleware etc also exist on non OpenVMS platforms.    F > But then, all those sessions are sitting at the command line prompt= , whichh is$ > not the normal end user interface. >=20F > When it comes to clients us=EEng OPS, or an assortment of middlewar=	 e talking  toE > something else, it is really not a problem. One node bombs due to at hardwareF > problem (yes, we _do_ admit to the occasional one Andrew), the othe= r' node(s)d9 > simply take over. The end user sees no loss of service.I >=20  C Quite so but unfortunately this is not unique to OpenVMS. Everyone=  =20oB agrees that OpenVMS as a platform makes this easier but you can=203 acheive the same effect on other platforms as well.a  @ This has always been my point, Kerrys assertion that nodes in=20> OpenVMS clusters can be rebooted with no impact ignores the=205 sessions connected to the rebooted node. Of course=20m? you can log them out (scheduled downtime for those sessions)=20e9 but you can also do this with almost any other Cluster=20k technology known to man.   Regardsa Andrew Harrisone Enterprise IT Architecta   ------------------------------  % Date: Mon, 30 Oct 2000 15:06:14 +0000g0 From: andrew harrison <andrew.nospam@uk.sun.com>' Subject: Re: Sun "uptime" belly laughera* Message-ID: <39FD8E66.F5E0D694@uk.sun.com>   "Main, Kerry" wrote: > =   	 > Andrew,v > =l  H > >>> Quite so Kerrys point about people being concerned about scheduledJ > downtime is not solved by OpenVMS by itself because upgrading a node wi= llH > require the people connected to that node either being disconnected or > logging out themselves. >>>u > =s  * > Hey, I think its starting to sink in :-) > =p  5 Kerry do you have a clue what you are talking about ?c  J > >>> This has always been my point, Kerrys assertion that nodes in OpenV= MS5 > clusters can be rebooted with no impact ignores thee- > sessions connected to the rebooted node.>>>a > =r  J > Oops, maybe not. Let me explain a little slower and in more detail. :-)=   > =O  9 Try not to be patronising when it is clear that you don'to! know what you are talking about ?t  J > When normal users go home at night or finish a shift or disconnect from=  B > their modem, they log out. This means their current sessions are; > disconnected from whatever systems they are connected to.e > =n   > Do we agree on this? > =n   Yes assuming that:B A	They are connected users in the terms that you use, many systemsA 	don't have connected users they have applications that connect =t  * 	and remain connected to a backend system.> B	You arn't running a system thats hosting users from multiple> 	time zones. Many of the systems I have worked on have users =  : 	in Tokyo, London and New York and for these systems there0 	is no quiet time, there is no night and no day.  J > While there are exceptions where some users staying logged into a syste= m for J > days, most shops would say this is a security risk. Especially since mo= stJ > peoples screensaver password is less than 3-4 characters and almost nev= erJ > (ok, never) changes. So, this is why inactive connections after x hours=  areJ > either killed by the system (Hitman stuff) or the application itself ti= mes ) > out by design and drops the connection.i > =a    E You seem to think that users connect directly to backend systems, odd- thats =4  B what used to happen but the majority of big backend systems that =  @ I work on don't have any users (except the DBA's and the admins)@ the users connect to mid tier servers. Perhaps this is where you are getting confused.     = I won't bother responding to indevidual points in the rest ofg; you posting except to say that it is all scheduled downtimen= and the you can also, particularly if you are allowed to kick>? users/connected apps off a node do this with any cluster that =    I can think of.M  @ The fact that you don't seem to realise that this is scheduled =  A downtime and the fact that you seem to think that the capabilites=E that you talk about are only available for OpenVMS illustrates that =e  1 you really should not be posting on this subject.e  ; Let me ask you a question, why do you think that people aret< asking for 99.9XXXX% uptime ? Could it not have something to= do with the fact that many systems have active users 24 hours < a day. For this kind of system where is your window when youC can do maintenance. There isn't one hence the 99.9XXX% requirement.d  E > So, lets be clear - this is normal practice, not something special.e > =a  J > Assuming availability is a concern and OpenVMS Clusters are in place - = once > again: > =s  J > - system planned to be shutdown in the next 8-16 hours (whatever - coul= d be< > longer if long batch job running) has its logins disabled.J > - Current user sessions are not impacted. When done, they logout and go=  homei& > (or whatever) assuming normal users.J > - Since the system to be shutdown has had its logins disabled, the TCPI= PhJ > load broker updates the 8.1.2 BIND compliant DNS - (even a Sun box) to = noJ > longer include that system in the list of available systems participati= ng inbH > the DNS TCPIP cluster alias. (I think this also works with TCPware and  > Multinet, but need to confirm)8 > - all new sessions go to other systems in the cluster.J > - when all sessions have moved to the other systems and there are no lo= ngerJ > any active sessions on the node to be shutdown, then that system can be=  . > shutdown. No user notices, no fuss, no pain.J > - when that system is rebooted, logins are re-enabled. The load broker = seesJ > the logins re-enabled, updates the DNS with a "this system is again par= t ofJ > the TCPIP Cluster alias" record update message and that system again st= arts7 > to participate in the dynamic DNS load sharing again.t > =i  J > Some Customers today tend to do things like disabling logins on the nod= e toJ > be shutdown before they go home at night so in the morning, they can sh= ut > that system down.a > =    > No user impact.h > =   4 > No active users sessions are disconnected. Period. > =    > No middleware. > =k  4 > Works with host based or 2 or 3 tier applications. > =r  / > Users do not even know a system was shutdown.l > =o  J > Works for any OpenVMS cluster application / db that uses a standard TCP= IP- > DNS which is at least BIND 8.1.2 compliant.r > =g  J > Key to this concept working is to be able to log into any system in the=  = > cluster, see all devices, files as if they were one system.b > =   D > With 2 system disks in place, even OS rolling upgrades can be done$ > transparently with no user impact. > =   J > Also key note is that this concept will not work (see next point) on an= yoJ > platform that has devices and files allocated to one system ie. one tha= tSJ > transparently "serves" those files/devices to other cluster members, be= causeaJ > if the system that owns those devices/ files needs to be shutdown, then=  J > those files and devices ownership needs to be failed-over to another sy= stem% > which will disconnect active users.  > =t  J > Hey, I am not saying this will solve world hunger, but this is, imho, a= npJ > advantage of a shared everything cluster architecture ie. planned syste= m J > downtimes with ZERO impact on application availability and no users bei= ng > disconnected.  > =n  J > And, yes, for three tier apps, one might be able to design and add some=  J > middleware logic to handle this for architectures that are shared nothi= ngJ > like UNIX and NT. This assumes you think about what the clients are doi= ng, J > and how to stall them while the disks are failed oveer and applications=  areJ > restarted on another node in the background, but what about the current=  J > connections to that system being shut down? It gets a little hairier. W= hatgJ > about 2 tier applications with no middleware logic? What about host bas= ed2 > applications (yes, there are still many around). > =e  D > Again, all of this can be transparent with current OpenVMS clusterJ > applications that are based on TCPIP DNS access to the cluster. Works w= ithn > DECnet as well btw.n > =   
 > Cool eh? > =    > Back to you Andrew.I > =k   > :-)t > =o  
 > Regards, > =    > Kerry Main > Senior Consultant  > Compaq Canada Inc. > Professional Servicesn > Voice: 613-592-4660p > Fax  :  819-772-7036 > Email: Kerry.Main@Compaq.com > =>   > -----Original Message-----9 > From: andrew harrison [mailto:andrew.nospam@uk.sun.com]r  > Sent: October 30, 2000 5:19 AM > To: Info-VAX@Mvb.Saic.Com ) > Subject: Re: Sun "uptime" belly laugherh > =u   > Paul Sture wrote:c > >t > > ___In articlesJ > > <910612C07BCAD1119AF40000F86AF0D805284930@kaoexc3.kao.cpqcorp.net>, K= erry > Main
 > > wrote: > > / > > From: "Main, Kerry" <Kerry.Main@compaq.com>u > > > Newsgroups: comp.os.vms - > > > Subject: RE: Sun "uptime" belly laughera+ > > > Date: Sat, 28 Oct 2000 10:53:34 -0500  > > >r
 > > > Andrew,h > > >S% > > > Combining two replies in one ..e > > >=C > > > >>> Solaris 8 allows you to patch it while it is running. <<<o > > >fJ > > > So does OpenVMS. OpenVMS has static parameters and dynamic paramete= rs.iJ > > > Changing the static ones requires a reboot. Changing dynamic ones d= oesa > not. > > >tJ > > > So does Solaris 8. Please correct me if I am mistaken, but ALL OS's=  on  > ALLuJ > > > platforms have some key kernel mode parameters or cache type sizes = that > if> > > > it is desired to change them, then a reboot is required. > > >dJ > > > >>> Kerry this is just marketing. You cannot do upgrades to a runni= ng > systemJ > > > with no impact to the users because users have to be disconnected f= romm > theD0 > > > running node that needs to be upgraded.>>> > > >mJ > > > Not with OpenVMS clusters. We have Customers that today do this all=  the > > > time.h > > > J > > > One great feature of OpenVMS clusters is that this can be done with=  ZEROoE > > > application availability impact. ie. no fail-over of processes,t > resources,G > > > no restarting of applications, no disconnecting active users etc.  > > >aJ > > > As you have heard me explain many times before - (you must be tired=  oft > me > > > explaining this by now). > > >aJ > > > Using the shared disk architecture (which allows users to log into = anygJ > > > system for full read-write access to any files on any disks at the = sameJ > > > time) and the TCPIP V5 load brokers load balancing (standard DNS on= ly > doesJ > > > round robin), you simply disable logins on the system that is plann= ed tot > beJ > > > shutdown. This allows current connections / users to continue worki= ng,  > but=J > > > all NEW connections are directed at other systems in the cluster. A=   > numberJ > > > of hours later (next day?), when all db/user connections are all on=   > othertH > > > systems in the cluster, then that system which needs maint (tuning > staticJ > > > parameters, OS upgrades using 2 system disk rolling upgrade, OS pat= ches,t > HWJ > > > replace ect) can be shut down with ZERO application availability im= pact.r > > >yJ > > Kerry, can we please nail this one shut. (I'm getting tired of the sa= me > oldr% > > argument between you and Andrew).n > >sJ > > Andrew is totally correct when he says that _I_ lose my sessions when=  a > givennJ > > node is closed. As a system manager I do not log out at the end of a = dayl > (WithsJ > > an average of 80 sessions, would you?), although obviously my worksta= tion > is > > well and truly locked. > > D > Quite so Kerrys point about people being concerned about scheduled
 > downtimeJ > is not solved by OpenVMS by itself because upgrading a node will requir= eeG > the people connected to that node either being disconnected or logingc > out themselves.e > ==  H > Of course you can acheive a seamless failover from one node to anotherF > without any downtime but to do this you need apps/middleware etc and@ > those apps/middleware etc also exist on non OpenVMS platforms. > =h  J > > But then, all those sessions are sitting at the command line prompt, = whichw > is& > > not the normal end user interface. > >eJ > > When it comes to clients us=EEng OPS, or an assortment of middleware = talking  > toG > > something else, it is really not a problem. One node bombs due to ai
 > hardwareH > > problem (yes, we _do_ admit to the occasional one Andrew), the other	 > node(s)g; > > simply take over. The end user sees no loss of service.i > >n > =u  D > Quite so but unfortunately this is not unique to OpenVMS. EveryoneA > agrees that OpenVMS as a platform makes this easier but you can 5 > acheive the same effect on other platforms as well.  > =t  ? > This has always been my point, Kerrys assertion that nodes ins= > OpenVMS clusters can be rebooted with no impact ignores the64 > sessions connected to the rebooted node. Of course> > you can log them out (scheduled downtime for those sessions)8 > but you can also do this with almost any other Cluster > technology known to man. > =   	 > RegardsS > Andrew Harrisona > Enterprise IT Architecte   -- =   Andrew Harrison  Enterprise IT Architect0   ------------------------------  % Date: Mon, 30 Oct 2000 11:41:55 -0500 ' From: "Bill Todd" <billtodd@foo.mv.com>c' Subject: Re: Sun "uptime" belly laugheru( Message-ID: <8tk8fa$n92$1@pyrite.mv.net>  4 Main, Kerry <Kerry.Main@compaq.com> wrote in messageH news:910612C07BCAD1119AF40000F86AF0D805284949@kaoexc3.kao.cpqcorp.net...   ...y  H >Key to this concept working is to be able to log into any system in the< >cluster, see all devices, files as if they were one system.  G Something far from unique to VMS (at least as far as the file system is  concerned).d  C >With 2 system disks in place, even OS rolling upgrades can be done # >transparently with no user impact.n  I >Also key note is that this concept will not work (see next point) on anygI >platform that has devices and files allocated to one system ie. one thatiE >transparently "serves" those files/devices to other cluster members,t becauserH >if the system that owns those devices/ files needs to be shutdown, thenL >those files and devices ownership needs to be failed-over to another system$ >which will disconnect active users.  L This is something I've explained I don't know how many times to you, but youJ still persist in spewing the same FUD.  So try to get it through your head
 this time.  K Taking Sun's cluster implementation as an example (since that's likely whateK you're thinking about in this discussion), your statement above is flat-out G false.  Individual members of Sun clusters 'serve' portions of the filedE system to the rest of the cluster using NFS.  This service includes afK failover mechanism by which a failed serving node's disks can be taken over K by a partner and the service restored *without* requiring that applications L accessing the data on those disks from other cluster members be disconnectedK (though the failover time was on the order of a minute, last I knew, so theiK applications do require some patience).  The failover is accomplished by IPtI address aliasing (IIRC the address is specific to the portion of the fileoK system served, rather than to the node serving it) and includes transparentnK reestablishment of any outstanding client byte-range locks (whether this issL an NFS Vx enhancement or a Sun-proprietary feature I'm not sure, but I thinkA I may have seen a reference to someone else - might have been thelH 'mission-critical Linux' guys - doing the same thing, so it may not be).  K AIX IIRC serves JFS around its RS/6000 clusters in a similar manner and hasaH the ability to fail over disks to partner server nodes, but I don't knowD whether it takes the final steps of continuous access via IP addressJ failover and lock retention.  Haven't check up on Veritas cluster softwareF recently, but it wouldn't surprise me if they support the same kind ofB thing:  it's just not leading-edge functionality any more, and VMSD supporters would be well-advised not to keep their heads in the sandJ thinking it's a major differentiator (regardless of what they may think of Andrew).   - bill   ------------------------------  % Date: Mon, 30 Oct 2000 12:23:29 -0500e2 From: rdeininger@mindspring.com (Robert Deininger)' Subject: Re: Sun "uptime" belly laugher L Message-ID: <rdeininger-3010001223300001@user-2ive7bp.dialup.mindspring.com>  { In article <910612C07BCAD1119AF40000F86AF0D805284949@kaoexc3.kao.cpqcorp.net>, "Main, Kerry" <Kerry.Main@compaq.com> wrote:t  I > Oops, maybe not. Let me explain a little slower and in more detail. :-)l > H > When normal users go home at night or finish a shift or disconnect fr= > omB > their modem, they log out. This means their current sessions are; > disconnected from whatever systems they are connected to.u >  > Do we agree on this?    I No!  No matter how ofter or how slowly you say it, it isn't the case that M "normal" users always log out.  All those windows on our screens have lots ofsL context information, and re-creating everything is a waste of time.  I closeO a window when the work it's supporting is done.  If something gets interrupted,d! it might last weeks on my screen.o  L If VMS really wants to brag about "no user impact", you need to dust off the virtual terminals, make them applicable to DECterm windows, SET HOST, telnet, etc, and then make them so they can be saved indefinitely when a system shutdownI occurs.  There would still be a small cost to the user, who would have to E log everything in again, but at least the old context would be there.4  J A lot of the DECwindows apps have been abandoned, and one feature they areO missing is the "New Desktop" ability to remember their context across sessions.o  K You seem to think in terms of transactions or web pages or whatever.  Thinke of an interactive session as a "transaction" that lasts several weeks or more, and you will see that even VMS clusters have an availability gap when systems( need to be rebooted for whatever reason.  H But VMS reboots are rare, and can usually be planned in advance.  UnlikeL that SSlithering SSun SSystem called SSolariSS, which is rapidly making ebayO into a joke.  It's been weeks since everything at ebay worked at the same time.tG I'll gladly take an occasional few minutes of window-shuffling instead.h   -- s Robert Deininger rdeininger@mindspring.comk   ------------------------------   Date: 30 Oct 2000 18:14:35 GMT1 From: bill@triangle.cs.uofs.edu (Bill Gunshannon)u' Subject: RE: Sun "uptime" belly laugher , Message-ID: <8tkdqb$1sdu$1@info.cs.uofs.edu>  N In article <910612C07BCAD1119AF40000F86AF0D805284949@kaoexc3.kao.cpqcorp.net>,.  "Main, Kerry" <Kerry.Main@compaq.com> writes: |>J |> Oops, maybe not. Let me explain a little slower and in more detail. :-) |> aJ |> When normal users go home at night or finish a shift or disconnect fromC |> their modem, they log out. This means their current sessions are < |> disconnected from whatever systems they are connected to. |> e |> Do we agree on this?q  B No, that is not always the case.  And wether or not it is would be@ irrelevant tot he discussion as it would apply equally to either OpenVMS, Unix or any other OS.   |>  O |> While there are exceptions where some users staying logged into a system forh7 |> days, most shops would say this is a security risk. t  K We have only your claim of this.  The fact is, you can only make this claimsI for your shop.  I consider people leaving sessions logged on no more of a M security risk than the user walking down to the coffe-machine without logging I off.  That kind of security is not overly important in mine and I am sureoH many other shops environments.  I have users who stay logged on from one system boot to the next.  M |>                                                      Especially since mostsL |> peoples screensaver password is less than 3-4 characters and almost never |> (ok, never) changes.   M Irrelevant.  screensaver passwords on any of my systems are the same as loginoH passwords.  I am sure the same is true for most anything beyond Windows.  O |>                       So, this is why inactive connections after x hours arehM |> either killed by the system (Hitman stuff) or the application itself times * |> out by design and drops the connection.  I I do not run an idle process killer, nor am I aware of any applications ItL support that would time the user out.  The real world may be a bit different than your perception of it.n   |>  F |> So, lets be clear - this is normal practice, not something special.  G N, this is normal practice where you are.  It is not necessarily normalm practice everywhere.   |> eN |> Assuming availability is a concern and OpenVMS Clusters are in place - once	 |> again:i |> eN |> - system planned to be shutdown in the next 8-16 hours (whatever - could be= |> longer if long batch job running) has its logins disabled.uN |>- Current user sessions are not impacted. When done, they logout and go home' |> (or whatever) assuming normal users.uK |> - Since the system to be shutdown has had its logins disabled, the TCPIP L |> load broker updates the 8.1.2 BIND compliant DNS - (even a Sun box) to noO |> longer include that system in the list of available systems participating inmI |> the DNS TCPIP cluster alias. (I think this also works with TCPware and ! |> Multinet, but need to confirm)t9 |> - all new sessions go to other systems in the cluster.sN |> - when all sessions have moved to the other systems and there are no longerJ |> any active sessions on the node to be shutdown, then that system can be/ |> shutdown. No user notices, no fuss, no pain. N |> - when that system is rebooted, logins are re-enabled. The load broker seesN |> the logins re-enabled, updates the DNS with a "this system is again part ofN |> the TCPIP Cluster alias" record update message and that system again starts8 |> to participate in the dynamic DNS load sharing again.  M So, tell me again, if we were to assume that users must log off at the end ofaM the day and idle processes are killed (this can be done on systems other than M VMS as well) how does this differ from say turning off logins on a Solaris or N BSD box, waiting for everyone to log off, making the upgrade and rebooting the3 machine, thus re-allowing users on that box again??s   |> eN |> Some Customers today tend to do things like disabling logins on the node toL |> be shutdown before they go home at night so in the morning, they can shut |> that system down.  R And how does this differ from doing the exact same thing on a non-OpenVMS system??   |> a |> No user impact. |> a5 |> No active users sessions are disconnected. Period.l |> l |> No middleware.   4 And not unique to OpenVMS.  Now I'm really confused.  K I mean, far be it from me to defend Solaris, but nothing you have describeddL is unique to OpenVMS.  If the same social rules are forced on the users of a+ UNIX cluster, then the same result applies.p  G But then, your original premise was wrong and therefore the result wills( probably not be what you claimed either.   bill   -- sJ 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: 30 Oct 2000 18:24:56 GMT' From: david20@alpha2.mdx.ac.uk (D.Webb)a' Subject: Re: Sun "uptime" belly laughere0 Message-ID: <8tkedo$855$1@aquila.news.mdx.ac.uk>  ] In article <39FD4884.7FEC2ABC@uk.sun.com>, andrew harrison <andrew.nospam@uk.sun.com> writes:e >"D.Webb" wrote: >> b` >> In article <39F95D82.606B75A8@uk.sun.com>, andrew harrison <andrew.nospam@uk.sun.com> writes: >> > >>  F >> As I asked for last time you posted this. Please supply evidence ofF >> this ability to patch the image on disk (whilst the system is stillM >> running). VMS can in general do this (Though I am not sure VMS engineering Q >> would go so far as to claim it can always be done) because running images willeH >> still refer to the previous version. However solaris does not support >> multiple versions of files. >> oL >> If the images you are referring to patching are read into memory when theQ >> system boots and the files are never therafter accessed then yes you should beaH >> able to do what you claim. However I would suspect that the domain of? >> applicability for such patching would be relatively limited.  >> d; >All the Sun patching systems allow you to patch the image d; >on disk. You are supposed to reboot a node after applying a< >the patches but this would only be necessary if you didn't 5 >have the patched image already running which you do.i >:8 >I can also apply a patch to a Jumpstart image which are- >generally used to upgrade OS images en mass.: >n> >Now most people would probably still use the patch the image 4 >and then reboot method even with Solaris 8 if they 6 >had a cluster just as you would with OpenVMS but the 1 >option does exist to patch the running image and@ >avaoid the reboot.e >r   Andrew,o  K I said to always be able to patch a running systems images. To me a running H system is not a system which has been put into single-user mode in orderI to allow you to do a patch. Generally speaking you cannot just patch the eL images on disk. Running Images maybe accessing the file(s) you are trying toK patch paging in the code you are altering or code immediately following thee alteration.   G You also mention using Jumpstart to upgrade a system and apply patches.r9 To use Jumpstart you boot the Jumpstart client machine in > Jumpstart mode (with a special boot command) which then duringJ the boot process request images and patches from the Jumpstart server. TheD patches are not "live" on the Jumpstart server unless they have beenH additionally installed on that server.  They sit on the Jumpstart server@ (usually in a separate Jumpstart partition) in patch kit format.K Once the upgrade of the OS or application of the patches have completed the M Jumpstart client shuts down again and reboots using the normal boot commands.rK The patches are applied when the system is in the equivalent of single-user I mode (during the later phases of the Jumpstart boot) ie when there is no eL chance of running images paging in the code which is being changed/replaced  as the patching is occurring.   B >> >Of course Shawn may not be running Solaris 8 in which case youD >> >upgrade a node in the cluster and re-boot it just as you do with >> >OpenVMS. >> > >> t
 >> Andrew, >> gF >> I don't know very much about Solaris Clusters but do multiple nodesL >> share any of the main unix partitions (/, /usr, /var etc) in the same wayI >> that multiple VMS systems can share a system disk or does each Solariss> >> cluster node have to have all its own OS  executables etc ? >> n> >Each cluster node has its own executables though most people < >also have a Jumpstart server which is used as a source for < >all the OS images. This of course makes it easy to upgrade ? >an indevidual node but does mean that you need to ensure that c: >the nodes are configured the same. Hence why most people  >use Jumpstart.s >S  g     >Regards >Andrew Harrison >Enterprise IT Architect      
 David Webb VMS and Unix team leader CCSS Middlesex University   ------------------------------  % Date: Mon, 30 Oct 2000 09:46:51 -0200') From: fabio_compaq@ep-bc.petrobras.com.br  Subject: Sun MoonuL Message-ID: <OFB595EE2C.8CD2D6FD-ON83256988.00408925@ep-bc.petrobras.com.br>   Check this site   D http://news.cnet.com/news//0-1003-200-3248189.html?tag=st.cn.sr.ne.1  5 About the launch of the SunCluster 3.0.  A rival ????s   Regards, Fabio C.l   ------------------------------  % Date: Mon, 30 Oct 2000 18:43:18 -0000 - From: wspencer@ap.nospam.org (Warren Spencer)e Subject: Re: Sun Moon / Message-ID: <svrga6jhcecd14@news.supernews.com>c   Hi,g  G I've reproduced some of the pro-OpenVMS/Anti-Solaris remarks from this bL CNETNEWS.COM article for us OpenVMS enthusiasts, and to spark another round 6 of heated bandwidth consumption from Kerry and Andrew:   ________________________    I Sun will announce "Full Moon," the code name for SunCluster 3.0 software tC for sharing jobs across several servers, said Andy Ingram, head of a0 marketing for Sun's Solaris operating system...   
 -- snip --  H Sun, though the dominant seller of Unix servers, doesn't have much of a K track record in clustering compared with those of some competitors. Compaq iJ Computer, with its OpenVMS and Tru64 Unix operating systems and its ultra-H high-end Tandem computer, has long been a respected force in clustering  software...   
 -- snip --  = Ingram acknowledged the weakness of SunCluster 2.2. "We were rF investing...effort in 3.0 rather than continuing to flesh out all the   feature sets of 2.2," he said...  
 -- snip --  K Sun's clustering software is only half of the company's new "software-as-a- G service" philosophy... (where the heck did they get the "service" idea   [ms]?)   _________________________    ws   -- n= Is there significance to the choice of the name "Full Moon" ?(   ------------------------------  % Date: Mon, 30 Oct 2000 11:49:03 -0500o# From: Jim Agnew <agnew@hsc.vcu.edu>o Subject: Re: TEK 4111 emulator? + Message-ID: <39FDA67F.62C18393@hsc.vcu.edu>    I think msdos kermit does a tek emulation, and it's is possible to run it under win95... if you want a free solution, however, it's : not the best solution... but didn't see kermit mentioned..   Chris Scheers wrote: > E > I am looking for a TEK 4111 emulator for VAX/VMS V7.1 and eXcursionr > V3.0.s > J > The TEK display needs to appear on the PC screen.  Ideally, it should beH > addressable as a device from the VAX (although I have some flexibility	 > there).c > # > Does anyone have any suggestions?  > I > -----------------------------------------------------------------------l& > Chris Scheers, Applied Synergy, Inc. > D > Voice: 817-237-3360            Internet: chris@applied-synergy.com >   Fax: 817-237-3074    ------------------------------    Date: 30 Oct 2000 07:44:04 -05009 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)nO Subject: Tieing Oracle 8 usage to a single username (was: Urgent Help Needed..)t+ Message-ID: <JZ6sg69ZxcFH@eisner.decus.org>w  n In article <6DB7DDFF3075D41191F3001083FC5FB81803E0@LCSRVR1>, Suresh Kumar-T <sureshkumar@mascon.co.in> writes:  - Please use meaningful title in the newsgroup.r  L > Now the   problem is suppose we have 100 users. We want to create each and= > every user use  ops$USERS command. It will waste of  time .r  B If I understand, you are concerned about a one-time administrativeB effort for an identity that will be used for ongoing transactions.  I > So I want to create some common user ID for Eg. suresh. All other users M > should go this ID only. So I want to grant this(suresh)id to all users. Gots > it..  E That is _very_ bad for security, as you get no accountability for whonE did what (certainly Rdb has full access to VMS Auditing, Oracle 8 maywA be somewhat less, but you can hope for the day it might improve).i   > This is very urgent.  B It cannot be very urgent if you chose to put "urgent" in the title rather than a meaningful title.l   ------------------------------   Date: 30 Oct 2000 10:14 -0400y From: hein@eps.zko.dec.c*m$ Subject: Re: Urgent Help Needed.....& Message-ID: <30OCT200010142632@miasys>  p In article <6DB7DDFF3075D41191F3001083FC5FB81803E0@LCSRVR1>, Suresh Kumar-T <sureshkumar@mascon.co.in> writes... >Hi ,h  : sC > I want use operating system username and Password to Oracle also.t  :0 ><create user ops$suresh identified  externally>  : hK >Now the   problem is suppose we have 100 users. We want to create each and < >every user use  ops$USERS command. It will waste of  time .  < That would NOT be a waste a time. It is the proper solution.7 It gives full accountibility, security and flexibility.sE Assign those users RIGHTS at the VMS level to grant/deny file access. F Grant those users ROLES in ORACLE to grant/deny database transactions.  E If you ever get one 'bad' user in the bunch, then you will be able tosH trace / restrict that users activities with 'surgical precision' insteadM of disabling the whole bunch and ask all users for the new 'secret' password.i  E If it is so urgent, you could/should have typed in the commands (with-E the help of some script of course) before the first reply will be in.    Regards,
         Hein.R   ------------------------------  # Date: Mon, 30 Oct 2000 16:09:44 GMTr% From: A.Greig@virgin.net (Alan Greig).7 Subject: VIOC max transfer blocks before cache bypassed42 Message-ID: <39fd9b5e.1818345692@news.newsguy.com>  B I'm sure this must be documented somewhere but I can't immediatelyF find it. Could someone tell me what the limit is on I/O block transferF count before VIOC is bypassed on an Alpha (VMS 7.2-1).  Also if anyone3 knows if this changes in the extended cache in 7.3.a  E Also a pointer to where this is documented if possible. It's probablyo! obvious and I've just gone blind.t --
 Alan Greig   ------------------------------  # Date: Mon, 30 Oct 2000 17:50:49 GMT = From: system@SendSpamHere.ORG (Brian Schenkenberger, VAXman-)n; Subject: Re: VIOC max transfer blocks before cache bypassedc0 Message-ID: <009F25F8.7E1F82A6@SendSpamHere.ORG>  Z In article <39fd9b5e.1818345692@news.newsguy.com>, A.Greig@virgin.net (Alan Greig) writes: >oC >I'm sure this must be documented somewhere but I can't immediatelynG >find it. Could someone tell me what the limit is on I/O block transfertG >count before VIOC is bypassed on an Alpha (VMS 7.2-1).  Also if anyonec4 >knows if this changes in the extended cache in 7.3.  K I thought the value was 34 blocks but when I checked [.SYS.LIS]VCC_DATA.LIS $ I found that the value is 35 blocks:   CACHE$GL_MAXBLK::n2 	.LONG	35			; MAXIMUM BLOCK SIZE OF I/O INTO CACHE  H It could be that it was changed somewhere along the way from when it wasJ first introduced but I can find no reference in the module comments to in-
 dicate so.   >aF >Also a pointer to where this is documented if possible. It's probably" >obvious and I've just gone blind. >--  >Alan Greig    [.SYS.LIS]VCC_DATA.LIS  ;)   --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COM             iO city, n., 1. a place where trees are cut down and streets are named after them.s   ------------------------------  % Date: Mon, 30 Oct 2000 17:37:43 +0000  From: SysAdmin <djesys@fsi.net> 1 Subject: Re: Virtual addresses with Multinet-4.2?r' Message-ID: <39FDB1E7.45F8DDF6@fsi.net>-  " "Yehavi Bourvine (58-4279)" wrote: >  > Hello,O >   I am running Multinet-4.2 on a VMS system and would like to have virtual IP J > addresses in order to have more than one WEB servers running on the sameQ > machine on the same port. In the past I did this with faked SLIP interfaces. Is<' > there a more elegant way of doing so?p; >                                         Thanks! __Yehavi:i  E Well, FWIW, CSWS (a.k.a. Apache V1.3.12 for OpenVMS) supports virtuald hosting.   David J. Dachtera    ------------------------------  % Date: Mon, 30 Oct 2000 13:01:13 -0200w) From: fabio_compaq@ep-bc.petrobras.com.bre  Subject: Re: VMSINSTALL QuestionL Message-ID: <OF313D8058.6BD41287-ON83256988.0052697C@ep-bc.petrobras.com.br>  C Just to imagine  ... a few years ago I read that Install Shield wasy developing a version7 of it=B4s program to Solaris. Why not to OpenVMS too  ?e   Fabio C.                  0 richard_maher@my-deja.com em 26/10/2000 14:12:16H                                                                        =     =20 H                                                                        =     =20sH                                                                        =     =20     @                                                              =20@                                                              =20@                                                              =20@  Para:    Info-VAX@Mvb.Saic.Com                              =20@                                                              =20@  cc:      (bcc: Fabio dos Santos Cardoso/E-P-BC/Contratada)  =20@                                                              =20@                                                              =20@                                                              =20@  Assunto: Re: VMSINSTALL Question                            =20@                                                              =20           =l     Hi,a  F As others have pointed out, the Developer's Guide to VMSINSTALL is theH book you want and is in the Archived section of your VMS bookreader CDs= .>  D I for one stay well clear of PCSI, but the one thing it does do thatD VMSINSTALL doesn't, is give you a DEinstall procedure. But obviously6 you can just write your own and ship it will your kit.   Regards Richard Maher.  H PS. One thing I don't recomend is doing a PRODUCT REMOVE DECNET-OSI and=  " then trying to reboot your system.  7 In article <OF6AF87170.620CCB9B-ON8325697D.00572692@ep-s bc.petrobras.com.br>, ,   fabio_compaq@ep-bc.petrobras.com.br wrote:? > Is there a document / internet link describing how to develope
 a "script" ore > similar toF > install an application under OVMS using VMSINSTAL?. We have a lot of
 > in-houseF > developed applications and some of them the  installation process is
 > complex. >s
 > Regards, >s > FC >r >i    & Sent via Deja.com http://www.deja.com/ Before you buy.o       =t   ------------------------------  % Date: Mon, 30 Oct 2000 15:19:50 +0000o  From: steven.reece@quintiles.com  Subject: Re: VMSINSTALL QuestionH Message-ID: <OF0C4B2AAA.3B23EA0A-ON80256988.0053F356@qedi.quintiles.com>  H Could it be because most VMS system managers wouldn't trust it to do th= ings properly on VMS?H Added to which, I think tools like this need to retain the opportunity = toH use a serial console as well as a graphics one.  GS140s for example don= 't# have a graphics console by default.d   Steve.  B Fabio (fabio_compaq at ep-bc dot petrobras dot com dot br) wrote :F >>>Just to imagine  ... a few years ago I read that Install Shield was developing a version7 of it=B4s program to Solaris. Why not to OpenVMS too  ?a   Fabio C.<<<=   ------------------------------  % Date: Mon, 30 Oct 2000 10:26:22 +0100g= From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com> ' Subject: Re: What does DCPS really do ?h) Message-ID: <39FD3EBE.24AFD5B3@gtech.com>i   JF Mezei wrote:mZ > Ok, I know that DCPS has the ability to convert non-portscript print jobs to postscript. > N > And it seems that DCPS will query the printer to ensure it is there prior to > starting a print job.f > O > But is there a complete description of exactly what DCPS does to a postscript . > file as it is sent to a postscript printer ? > O > Does DCPS interpret the submitted postscript code to insert its own code herec
 > and there ?e > P > What sort of dialogue (and is there dialogue in between pages of the same job)2 > does DCPS print symbiont have with the printer ?  2 Basicly DCPS supplies two pieces of functionality:B   * converts non-PostScript to PostScript before sending it to the printerlG   * dialog with printer which among other things enables it to get pagen     count correctm   Arne   ------------------------------    Date: 30 Oct 2000 08:17:58 -0500, From: koehler@eisner.decus.org (Bob Koehler)0 Subject: Re: What VMS format are these files in?+ Message-ID: <i$i7obTG2BlF@eisner.decus.org>-  W In article <27OCT200011481546@gerg.tamu.edu>, carl@gerg.tamu.edu (Carl Perkins) writes:. > D > Even though the Fortran formatting may normally use a fixed numberD > of bytes (it doesn't have to, at least not in VAX or DEC or CompaqG > Fortran), the created file is still normally a variable record length>D > file for the basic formatted sequential access file (so it will be8 > for any output that is intended to be human readable)   E Nope.  I'm talking about how Fortran continaing CHARACTER data reads  ? files with fixed record length attribute (FAB$B_RFM=FAB$C_FIX).g  F ----------------------------------------------------------------------? Bob Koehler                     | Computer Sciences Corporation = NASA GSFC Flight Software       | Federal Sector, Civil GroupoE                                 | please remove ".aspm" when replying    ------------------------------    Date: 30 Oct 2000 09:20:08 -0500 From: briggs@eisner.decus.org 0 Subject: Re: What VMS format are these files in?+ Message-ID: <jR1LkbO36pDo@eisner.decus.org>   Z In article <cPizN1EqtBYd@eisner.decus.org>, koehler@eisner.decus.org (Bob Koehler) writes:I > If EDT can handle text lines stored as fixed lengtgh records, why can'tkI > FTP?  IMHO if I'm sending a fixed length record file in ASCII mode each F > record should be transmitted with the RFC 959 CRLF pair, only if I'mJ > sending it in binary mode should the file end up on my UNIX system as an# > undelimitted collection of bytes.O > H > Are EDT and I alone in the concept that a fixed length record of ASCII= > characters is a line of text?  IIRC Fortran thinks so, too.   B $ PRINT and $ TYPE agree with FTP.  A fixed length record of ASCIIB characters is not a line of text.  Unless, of course, the file has implied carriage control.b  J As presented on hardcopy output, a fixed length record of ASCII characters@ with null carriage control most certainly is not a line of text.F Anywhere else, the notion of what constitutes a "line of text" may not& always be as precise as we would wish.   	John Briggs   ------------------------------   End of INFO-VAX 2000.607 ************************