1 INFO-VAX	Tue, 02 Oct 2001	Volume 2001 : Issue 547       Contents: Re: Alternate vendor memory.
 Re: Backup
 Re: Backup0 RE: Best Way to do Interprocessing Communication2 BIG differences in Vax vs ALpha compile/link times6 Re: BIG differences in Vax vs ALpha compile/link times6 Re: BIG differences in Vax vs ALpha compile/link times6 Re: BIG differences in Vax vs ALpha compile/link times6 Re: BIG differences in Vax vs ALpha compile/link times6 Re: BIG differences in Vax vs ALpha compile/link times! Re: Combine (Team) network cards?  Re: DECTalk help needed - Re: DLT IV Tapes - How long should they last? 9 Re: Error on restore from tape - Directory Depth problem? 9 Re: Error on restore from tape - Directory Depth problem?  FAQ September 2001 edition Re: FAQ September 2001 editionI Re: Filtering IP Addresses? (was: Re: Point me in the right direction.... 0 Fw: Best Way to do Interprocessing Communication Re: Good bye VAX-6400's  Re: Good bye VAX-6400's  Re: Good bye VAX-6400's  Re: Good bye VAX-6400's L Re: I WILL QUIT USING VMS WHEN THE MILITARY DOES - IN 2015 OR MAYBE  LONGER!L Re: I WILL QUIT USING VMS WHEN THE MILITARY DOES - IN 2015 OR MAYBE  LONGER!K Re: I WILL QUIT USING VMS WHEN THE MILITARY DOES - IN 2015 OR MAYBE LONGER! P Re: I WILL QUIT USING VMS WHEN THE MILITARY DOES - IN 2015 OR MAYBE LONGER! LONGP Re: I WILL QUIT USING VMS WHEN THE MILITARY DOES - IN 2015 OR MAYBE LONGER! LONG  Re: Letter from Michael Cappelas  Re: Letter from Michael Cappelas  Re: Letter from Michael Cappelas  Re: Letter from Michael Cappelas  Re: Letter from Michael Cappelas, Re: Locate Filename containing logical block7 Re: Looking for Classic Oracle/VMS version matrix again 8 Re: ODS2 available -- reads ODS-2 disks on Windows, UNIX8 Re: ODS2 available -- reads ODS-2 disks on Windows, UNIX* Re: OpenVMS employment in southern France?G Re: OpenVMS Fundamenalists was (Re: Major VMS reseller says VMS legacy) G Re: OpenVMS Fundamenalists was (Re: Major VMS reseller says VMS legacy) " Re: OpenVMS, Apache and hackers..." RE: OpenVMS, Apache and hackers..." Re: OpenVMS, Apache and hackers..." Re: OpenVMS, Apache and hackers..." RE: OpenVMS, Apache and hackers..." Re: OpenVMS, Apache and hackers..." Re: OpenVMS, Apache and hackers...' Re: Point me in the right direction....  Re: Question on VMS Virus  Re: Question on VMS Virus  Ram Disk for VMS?  Re: Really Unhackable...????  Re: SCSI Cluster support on LVD?" Re: Sun goes after Alpha user baseA Re: To the tune of "It's still rock and roll to me" by Billy Joel  Re: VAX disaster tolerance, Re: VMS Itanium port speeded up. Tru64 dies.5 Re: VMScluster using fibrechannel and memory channel? 5 Re: VMScluster using fibrechannel and memory channel?  Re: VT emulation on Linux / Re: Warranty on ES45 reduced from 3 years to 1? > Re: [VMScluster] Cross Architecture Boot with Cluster_Config ?  F ----------------------------------------------------------------------  % Date: Mon, 01 Oct 2001 21:02:02 -0000 - From: wspencer@ap.nospam.org (Warren Spencer) % Subject: Re: Alternate vendor memory. 7 Message-ID: <912DA54B6warrenspencer1977@207.126.101.97>   9 graham@the-shades.demon.co.uk (Graham Harrison) wrote in  2 <ed1713eb.0109301405.15d7d546@posting.google.com>:   >Hi,   > % >Last question this evening - honest.  > ' >Has anyone got any experience of using & >Kingston compatable memory?  I put a ( >purchase request in for some memory for( >a DS20E and the bean counters tried to 
 >get me this.  > - >I wouldn't normally credit the bean counters 1 >cost saving with much worth - but the difference - >is staggering.  5548 versus 400 (yes 400)  > 7 >If Compaq don't support it, or it will cancel/increase 7 >my service contract then fair enough and Compaq memory 9 >it is - otherwise how do I justify it to the big cheese?  >  >Thanks. >Graham.  L For that kind of cost savings, you can afford to buy a backup memory kit or  two, and support it yourself.   G I've used non-DEC memory (including Kingston) in a variety of machines  E since 1983, and never had a memory problem that couldn't be fixed by   swapping in the backup kit.     G DEC field service (in North America) has always been cooperative.  The  K worst thing they've ever asked is to remove the 3rd part stuff temporarily  J to see if it is the source of the problem; obviously a reasonable request.     ws   --     Warren Spencer' Senior Software Engineer (not a writer)  The Associated Press  L ** My employer does not necessarily agree with my statements - neither do I  **   ------------------------------  % Date: Mon, 01 Oct 2001 23:37:07 +0100 1 From: Steve Reece <SYSTEM@ipl.demon.co.nospam.uk>  Subject: Re: Backup 6 Message-ID: <3BB8FE23.4C7CB2AF@ipl.demon.co.nospam.uk>  F IIRC, Glenn Everhart had an excellent workaround for doing just this. C It wasn't capable of dealing with multi-tape savesets but for other  stuff it worked fine.  It may be in the FAQ....   Steve.     Hoogenboom, Martin wrote:  >  > Hi,  >  > I have a simple question: D > Is it possible to backup a disk from one system, across a network,- > to a tape unit connecte to another system ? 0 > If yes, how what needs to run on the systems ? > = > Both systems are standalone Alpha's running OpenVMS 7.2-1H1  >  > TIA, > Martin Hoogenboom    --  G "A shadow fell over her face; clear, as if the composure were rent like E a veil.  And her lips parted, but only with a short intake of breath. A Then she said, 'Well, then you are right.  Indeed, we are even.'" % 		Louis, "Interview with the Vampire"    ------------------------------  # Date: Mon, 01 Oct 2001 23:03:29 GMT 2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman) Subject: Re: Backup 1 Message-ID: <5D6u7.901$YP.24202@news.cpqcorp.net>   j In article <3BB8FE23.4C7CB2AF@ipl.demon.co.nospam.uk>, Steve Reece <SYSTEM@ipl.demon.co.nospam.uk> writes:G :IIRC, Glenn Everhart had an excellent workaround for doing just this.  D :It wasn't capable of dealing with multi-tape savesets but for other :stuff it worked fine. :It may be in the FAQ....   E   It *is* in the OpenVMS FAQ, as I fixed a reported bug in the DCL of $   the section a couple of years ago.   :Hoogenboom, Martin wrote:E :> Is it possible to backup a disk from one system, across a network, . :> to a tape unit connecte to another system ?1 :> If yes, how what needs to run on the systems ?     N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  $ Date: Tue, 2 Oct 2001 15:36:13 +10006 From: "Arena, Steve" <Steve.Arena@pacificpower.com.au>9 Subject: RE: Best Way to do Interprocessing Communication Q Message-ID: <D750FFBD4936D111842000805F15EFA4043A87C4@meppb1.pacificpower.com.au>    IMHO  J Tibco Rendezvous Publish/Subscribe technology enables a very high level ofD process independance with very effective interprocess communication.B Processes can be on different nodes or OS, and comms can easily be
 monitored.  ( One process publishes others subscribe.    Regards  Steve    > -----Original Message-----2 > From:	Martin Hoogenboom [SMTP:martinh@xs4all.nl]% > Sent:	Tuesday, 2 October 2001 15:36  > To:	Info-VAX@Mvb.Saic.Com ; > Subject:	Fw: Best Way to do Interprocessing Communication  > ? > I think Event Flags were also lost. Oke, its simple and small C > communication, but it works and if your'e only wating for another  > process's signal ....  >  > ----- Original Message -----  . > From: <yyyc186@illegaltospam.mindspring.com> > To: <Info-VAX@Mvb.Saic.Com> ( > Sent: Monday, October 01, 2001 3:10 PM; > Subject: Re: Best Way to do Interprocessing Communication  >  > 6 > > In <3BB3AF77.A0D49FB@videotron.ca>, on 09/27/2001 A > >    at 07:00 PM, JF Mezei <jfmezei.spamnot@videotron.ca> said:  > > # > > You left off ACMS and MQ Series  > >  > > >John McLean wrote: H > > >> (a) Logical names - ideal for short messages,  same machine untilI > > >> (b) Distributed Lock manager (DSL) - up to 32 bytes (?) and can be J > > >> (c) Global sections - useful for sharing larger amounts of data andB > > >> (d) Files - last resort because the I/O is relatively slow. > >  > >  > > >You forgot: > > H > > >(e) Mailboxes. Works on same node. medium amounts of data. Good for% > > >many-to-one but not one-to-many. L > > >    (unless you create one mailbox for each process that talks to you).9 > > >Server doesn't know if a client disapears or apears.  > > L > > >(f) Decnet. Works accross nodes.  A single process can talk to multipleE > > >process (one-to-many, as well as many-to-one). Server knows when 	 > someone K > > >connects and disconnects. Server has identification of who connects to  > > >it. (username and node) > > J > > >(g) TCPIP. More or less the same as TCPIP, but with less asynchronous# > > >stuff and less identification.  > > J > > >(h) ICC  (Intra Cluster communications). Very efficient. One any node> > > >inside a cluster. But doesn't use standard QIO stuff thatH > > >mailbox,decnet,tcpip use. Requires system management to enable yourA > > >application to declare a service that others can connect to.  > > --  ? > > -----------------------------------------------------------  > > yyyc186@mindspring.com? > > -----------------------------------------------------------  > >  >    >    ------------------------------   Date: 1 Oct 2001 14:36:03 -0700 / From: rcyoung@aliconsultants.com (Robert Young) ; Subject: BIG differences in Vax vs ALpha compile/link times = Message-ID: <91437ce6.0110011336.3fcc1364@posting.google.com>   - Two systems: Vax 4000-600 & Alphaserver 2100a   C Compaq/DEC C compiler on both.  V6.4 on Alpha, V6.0 on Vax. We just F installed the Alpha this weekend so that is the reason for the version	 mismatch.   F Vax (VMS 7.1, ?MHz, 128Mb memory, RF-74 DSSI drives)compiles and linksC program in 14 minutes. On Alpha (VMS 7.3, 250MHz-2 processor, 512Mb F memory, RZ29D-VWs)it takes 90 minutes for same EXACT code, same linker
 file, etc.    E Code is written/compiled/linked first on Vax, then copied on Alpha to D be recompiled/linked there so we have versions for both platforms at
 this time.  C Any suggestions/recommendations would be helpful where we could get  the D work on Alpha done faster? It seems that the discrepancy is just too large ' between the systems. Thanks in advance.    Respond to email if possible   ------------------------------  # Date: Mon, 01 Oct 2001 22:45:09 GMT 2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)? Subject: Re: BIG differences in Vax vs ALpha compile/link times 1 Message-ID: <Vl6u7.898$YP.24250@news.cpqcorp.net>   o In article <91437ce6.0110011336.3fcc1364@posting.google.com>, rcyoung@aliconsultants.com (Robert Young) writes: . :Two systems: Vax 4000-600 & Alphaserver 2100a : > :Compaq/DEC C compiler on both.  V6.4 on Alpha, V6.0 on Vax...   :Vax ...compiles and linksH :program in 14 minutes. On Alpha ... it takes 90 minutes for same EXACT  :code, same linker file, etc.  ..F :Code is written/compiled/linked first on Vax, then copied on Alpha toE :be recompiled/linked there so we have versions for both platforms at  :this time.   E   Consider clustering, as you can operate off the same source pool...   D :Any suggestions/recommendations would be helpful where we could getI :the work on Alpha done faster? It seems that the discrepancy is just too . :large between the systems. Thanks in advance.  E   Profile your build and figure out where you are spending your time. E   Once you know if it is the CC, the LINK, the DCL command WAIT that  D   is invoked only during the OpenVMS Alpha build, or something else A   entirely different, we can then (try to) locate the difference. E   You will also want to identify what else (if anything) is active on E   the systems, and you will want to investigate system parameter and  '   process (quota) resource constraints.    :Respond to email if possible      Ask here, get an answer here.     N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  # Date: Mon, 01 Oct 2001 22:46:28 GMT ! From: Ian Parker <parker@gol.com> ? Subject: Re: BIG differences in Vax vs ALpha compile/link times & Message-ID: <T$ONWEAYKPu7EwVo@gol.com>  D In article <91437ce6.0110011336.3fcc1364@posting.google.com>, Robert) Young <rcyoung@aliconsultants.com> writes . >Two systems: Vax 4000-600 & Alphaserver 2100a > D >Compaq/DEC C compiler on both.  V6.4 on Alpha, V6.0 on Vax. We justG >installed the Alpha this weekend so that is the reason for the version 
 >mismatch. > G >Vax (VMS 7.1, ?MHz, 128Mb memory, RF-74 DSSI drives)compiles and links D >program in 14 minutes. On Alpha (VMS 7.3, 250MHz-2 processor, 512MbG >memory, RZ29D-VWs)it takes 90 minutes for same EXACT code, same linker  >file, etc.  >  > F >Code is written/compiled/linked first on Vax, then copied on Alpha toE >be recompiled/linked there so we have versions for both platforms at  >this time.  > D >Any suggestions/recommendations would be helpful where we could get >theE >work on Alpha done faster? It seems that the discrepancy is just too  >large( >between the systems. Thanks in advance. >  >Respond to email if possible   G I'm sure others will respond with clearer and more correct suggestions, G but I do seem to remember something about working set quotas and linker @ memory requirements being much larger on Alpha.  You may need to% increase your ws quotas on the Alpha.    --  
 Ian Parker   ------------------------------   Date: 1 Oct 2001 20:27 CDT' From: carl@gerg.tamu.edu (Carl Perkins) ? Subject: Re: BIG differences in Vax vs ALpha compile/link times , Message-ID: <1OCT200120274010@gerg.tamu.edu>  % Ian Parker <parker@gol.com> writes... E }In article <91437ce6.0110011336.3fcc1364@posting.google.com>, Robert * }Young <rcyoung@aliconsultants.com> writesH }>Vax (VMS 7.1, ?MHz, 128Mb memory, RF-74 DSSI drives)compiles and linksE }>program in 14 minutes. On Alpha (VMS 7.3, 250MHz-2 processor, 512Mb H }>memory, RZ29D-VWs)it takes 90 minutes for same EXACT code, same linker }>file, etc.  E }>Any suggestions/recommendations would be helpful where we could get  }>the F }>work on Alpha done faster? It seems that the discrepancy is just too }>large ) }>between the systems. Thanks in advance.  }> }>Respond to email if possible } H }I'm sure others will respond with clearer and more correct suggestions,H }but I do seem to remember something about working set quotas and linkerA }memory requirements being much larger on Alpha.  You may need to & }increase your ws quotas on the Alpha. }  }--  }Ian Parker   I In general, memory related process quotas should be at *least* doubled on H an Alpha compared to a VAX. That would be, to begin with, the WSDEFAULT,J WSQUOTA, WSEXTENT, and PGFLQUOTA parameters for each account in the sysuaf9 file (as mainipulated by AUTHORIZE), to be more specific.    --- Carl   ------------------------------   Date: 1 Oct 2001 20:08:12 -0700 / From: rcyoung@aliconsultants.com (Robert Young) ? Subject: Re: BIG differences in Vax vs ALpha compile/link times < Message-ID: <91437ce6.0110011908.3461869@posting.google.com>  < Nothing else running on the system (outside of normal system processes).   F Clustering is planned,(we have 2 system set up now in fact), but thereE are a few standalones at present. When we get the time.....(as we all  have said many times before)...   @ I'll get back with the profile info soon (1-2 days max) I think.  < I only mentioned email for a response as I would "notice" it
 sooner....   ------------------------------  % Date: Tue, 02 Oct 2001 00:13:52 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca> ? Subject: Re: BIG differences in Vax vs ALpha compile/link times , Message-ID: <3BB93EF2.3CA0C383@videotron.ca>  M I have found that SHOW PROC/CONT/ID=xxxx helps debug some  performance issues L at times. You'll get a feel for how much the working set initially grows andN then how much paging occurs, and how much time the process spends in COM state versus LEF etc.    ------------------------------  # Date: Mon, 01 Oct 2001 19:06:54 GMTn2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)* Subject: Re: Combine (Team) network cards?1 Message-ID: <i93u7.883$YP.23941@news.cpqcorp.net>e  u In article <ed1713eb.0110010430.31fe2937@posting.google.com>, graham@the-shades.demon.co.uk (Graham Harrison) writes: \ :Hamlyn Mootoo <univms@bigfoot.com> wrote in message news:<3BB7AB31.C551AFCE@bigfoot.com>... : C :> Are you already running Full Duplex on your (I assume) Ethernet?  :r7 :Yes, running 100Mb full duplex but its not enough.... e  0   Try gigabit Ethernet?  Try splitting the load?  D   NetRAIN is not available in TCP/IP Services V5.0A, V5.1, nor V5.2.:   This support is planned, but is not presently available.    N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  # Date: Mon, 01 Oct 2001 18:22:18 GMT 2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)  Subject: Re: DECTalk help needed1 Message-ID: <uv2u7.879$YP.24093@news.cpqcorp.net>t  W In article <3BB4D8F6.6D285E2C@osfn.org>, "Michael L. Umbricht" <mikeu@osfn.org> writes:s :Hoff Hoffman wrote: :> as :> In article <00A02A58.A8D0DF32@SendSpamHere.ORG>, system@SendSpamHere.ORG (Brian Schenkenberger, VAXman-) writes:Dc :> :In article <9otdib$3oc$1@info.cs.uofs.edu>, bill@triangle.cs.uofs.edu (Bill Gunshannon) writes:s :> .. @ :> :>Is there anywhere I can find ROM images for the EPROMs in a :> :>DECTalk DTC01?? :> eI :>   I'm checking with the resident DTC01 engineering pack-rat, to see ifl< :>   he has any EPROMs or copies of the EPROM images around.  H   The DECtalk materials were included when the group was sold off to the<   Smart Modular, including the associated tools and widgets.  $     http://www.smartmodulartech.com/  I :Where can I find the error codes for the LED's on the back of the DTC01?e  C   What I have had access to was previously included on the OpenVMS  6   Freeware.  This included examples and documentation.   N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  % Date: Mon, 01 Oct 2001 20:48:07 -0500o1 From: "David J. Dachtera" <djesys.nospam@fsi.net>-6 Subject: Re: DLT IV Tapes - How long should they last?' Message-ID: <3BB91CD7.4C8ABE0D@fsi.net>0   Todd Nelson wrote: > M > Thanks for the hint about lifetime warranties...  I will have to check into-J > it.  The only thing is this:  If the tape is bad, I cannot erase it, andN > therefor cannot send it anywhere but the garbage... to protect the data from3 > being used by people who are not authorized, etc.o  @ Bulk erase them. VHS degaussers show work very nicely, if you're* persistent. Watch trhe duty cycle, though.   -- s David J. DachteraD dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/w   ------------------------------  % Date: Sun, 30 Sep 2001 23:04:38 -0700e* From: "Mitchell Troy" <mtroy@jpl.nasa.gov>B Subject: Re: Error on restore from tape - Directory Depth problem?- Message-ID: <9pajed$itp$1@nntp1.jpl.nasa.gov>u  I Several people have asked for the exact commands sorry, I did not include  them initially:a   Backup command:u% $  MOUNT /FOR /NOUNLOAD MKB500:USRSAVoE $  BACKUP /REWIND/verify /RECORD /IGNORE=INTERLOCK /IGNORE=NOBACKUP -oK    /LIST=SYS$COMMON:[SYSMGR.BACKUPS.LIST]FULL_data_disk_'MONTH''DAY'.BCK  -e/       DISK$data:[000000...]*.*;*  MKB500:USRSAVl $  DISMOUNT MKB500   Restore command: $mount mkb0: /forT8 $ backup mkb0:*.*/save_set dkb300:[000000...]/log/verify  G I have not yet tried [*...] as the output specifier for the restore buts will.a  0 Thanks again for any other help you can provide,
 Mitchell troy  mtroy@jpl.nasa.gov    < "Chris Scheers" <chris@applied-synergy.com> wrote in message6 news:754a27c1.0109301907.85e7c8b@posting.google.com...7 > "Mitchell Troy" <mtroy@Jpl.nasa.gov> wrote in message ) news:<9p7kfm$801$1@nntp1.jpl.nasa.gov>....J > > We are running OpenVMS 6.2. Recently our user disk crashed. We are now( > > trying to restore from tape and get: > > ' > > %BACKUP-S-CREDIR, created directory  > >wK DKB300:[000000.COHARA.FORTRAN.TEMPLATES.SRC.EXTRACT.CAT.JOSH_65.BIG_PIX_X2]rD > > %BACKUP-I-INVDEEPDIR, output directory [000000...] not valid for' > > deepdirectories, try rooted logicallJ > > %BACKUP-E-OPENOUT, error opening DKB300:[000000...]CAT.EXE;1 as output' > > -RMS-F-DIR, error in directory namehF > > %BACKUP-E-OPENOUT, error opening DKB300:[000000...]*.*;* as output) > > -RMS-F-RLF, invalid related NAM blockp > > J > > The last message just repeats and it appears no more data is extracted fromI > > tape. The problem as far as I can tell is we can't create more then 8u levels1 > > of directory's. Is there a solution for this?l > >n > > Thanks,t > > Mitchell Troyv > > mtroy@jpl.nasa.gov > F > It depends on your backup.  What was the command used to create your; > backup?  What command are you using to restore the files?f >oB > What is the output specifier for your restore?  If you are using > [000000...], try using [*...]E >4 > Good luck!   ------------------------------  # Date: Mon, 01 Oct 2001 21:29:43 GMTI2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)B Subject: Re: Error on restore from tape - Directory Depth problem?1 Message-ID: <bf5u7.895$YP.24167@news.cpqcorp.net>a  Z In article <9pajed$itp$1@nntp1.jpl.nasa.gov>, "Mitchell Troy" <mtroy@jpl.nasa.gov> writes:J :Several people have asked for the exact commands sorry, I did not include :them initially: :  :Backup command: ..F :$  BACKUP /REWIND/verify /RECORD /IGNORE=INTERLOCK /IGNORE=NOBACKUP -L :   /LIST=SYS$COMMON:[SYSMGR.BACKUPS.LIST]FULL_data_disk_'MONTH''DAY'.BCK  -0 :      DISK$data:[000000...]*.*;*  MKB500:USRSAV .. :Restore command:n ..9 :$ backup mkb0:*.*/save_set dkb300:[000000...]/log/verify-  G   You are unfortunately going to get to learn about the structure of an-K   OpenVMS system disk and about the (old/ODS-2) RMS directory depth limits.m  J   Because of the way this particular BACKUP was created, you will have to G   reset all directory backlinks (all directory aliases) on this system aI   disk manually, due to the files-oriented BACKUP originally used.  This 8J   will include SYSCOMMON/VMS$COMMON entries, but there may well be others.  C   If I had to rebuild this system disk and could not use this as aneH   opportunity to reload cleanly (either V6.2 or more current software), D   I'd probably dump this whole mess onto another disk, and then pickB   through the directory trees in the carcass to rebuild the "real"   system disk.  A   I'd transfer the non-SYS, non VMS$COMMON directories over.  I'daB   then transfer over one [SYSn...] root, and I'd SET FILE/ENTER to?   add VMS$COMMON.DIR based on [SYSn]SYSCOMMON.DIR.  If you havevC   other SYSx... roots, you'll have to haul those over individually,,B   remove the contents of that root's [SYSx.SYSCOMMON] and then addC   the alias to the VMS$COMMON.DIR.  Once you are done, DFU or otherl@   tools can be used to make the VMS$COMMON.DIR entry the primary    entry among all alias entries.  B   There may well be other (site-specific) directory alias entries.  C   Deep directories will require you to use concealed rooted logical0D   names to "hide" levels -- you'll end up having to set up a logicalA   name for the various (deep) roots.  On this version of OpenVMS,(C   you can have up to two sets of eight levels of directories, with  D   the upper eight "hidden" behind the concealed rooted logical name.  D   You may well have to rebuild your queue database and possibly someE   other files that were open when the BACKUP was made, since the use .F   of /IGNORE=INTERLOCK (better known as "allow data inconsistencies") *   implies that this was an on-line BACKUP.  I   I would strongly encourage you to use BACKUP/IMAGE/[NO]ALIAS next time.oG   BACKUP/IMAGE handles all this stuff transparently, which is why that dB   approach is the documented approach for system disks.  I'd also F   encourage you to try the restoration, just to make sure it works -- F   that step would have shown how much "fun" you are now going to have 	   here.  r    N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  % Date: Tue, 02 Oct 2001 00:52:46 +0100 1 From: Steve Reece <SYSTEM@ipl.demon.co.nospam.uk>c# Subject: FAQ September 2001 edition-6 Message-ID: <3BB90FDE.1454CE9F@ipl.demon.co.nospam.uk>  F Last time I looked the VMS web site had an edition of the VMS FAQ fromG about the 10th Sep 2001 on it but I've been unable to find a plain text- version of this release.  E Looking at the VMS website a moment or two ago, the September releasejH seems to have been replaced by the April release again so I'm confused!!  9 Did I miss it in the newsgroup or did events overtake it?:   Thanks in advance. Steve. --  G "A shadow fell over her face; clear, as if the composure were rent likenE a veil.  And her lips parted, but only with a short intake of breath.iA Then she said, 'Well, then you are right.  Indeed, we are even.'"p% 		Louis, "Interview with the Vampire"O   ------------------------------  $ Date: Mon, 1 Oct 2001 21:08:52 -0400. From: "warren sander" <sander@ma.ultranet.com>' Subject: Re: FAQ September 2001 editione+ Message-ID: <9pb46h$79v$1@bob.news.rcn.net>r  4 The last faq I posted on the web site was april 2001  2 I'm waiting for a new version from the maintainers --B ------------------------------------------------------------------6 Warren Sander                        OpenVMS MarketingK Compaq Computer Corporation          Work:  warren.sander@remove.compaq.com-L 200 Forest Street MR01-3/J1          Personal: sander@remove.ma.ultranet.com3 Marlboro, MA 01752                   (508) 467-4875B5    My opinions are my own and I only speak for myselfs,          Read http://www.openvms.compaq.com/B ------------------------------------------------------------------   ------------------------------  # Date: Mon, 01 Oct 2001 23:12:20 GMT62 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)R Subject: Re: Filtering IP Addresses? (was: Re: Point me in the right direction....1 Message-ID: <oL6u7.902$YP.24242@news.cpqcorp.net>   ~ In article <um6u7.8867$cp1.1134823@newsread2.prod.itd.earthlink.net>, "Andy Bustamante" <a_c_bustamante@earthlink.net> writes: :Assuming Compaq's TCPIP 5.?M :If you use /accept anything not listed will be rejected.  If you use /REJECTp$ :any other host will be accepted.... ..H :TCPIP> disable service telnet            !Don't telnet in and try this!? :TCPIP> set service telnet /accept=network=(10.0.0.0:255.0.0.0)  :TCPIP> enable service telneth ..T :"Robin" <rlb@austin.rr.com> wrote in message news:3BB5E33E.3ACC12A@austin.rr.com...E :> Is it possible to restrict Telnet and FTP access via IP address on  :> an OpenVMS 7.3 Alphaserver?  =   Or via (brute-force) DCL-based IP filtering in SYLOGIN.COM.c  F   (This approach clearly isn't the best way for all cases, but as the G   question is rather terse, I'll happily provide an equally terse (and tF   possibly quite wrong for this application) answer.  :-)  Put anotherE   way, we need some background on the question -- what problem(s) areeD   you looking to solve?  A firewall (filtering router) and/or betterE   control over the login source for the user and/or encryption might  C   be better (and alternative) approaches, particularly if you have l5   untrusted users around and/or an untrusted network.   N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  $ Date: Tue, 2 Oct 2001 06:35:56 +0200- From: "Martin Hoogenboom" <martinh@xs4all.nl> 9 Subject: Fw: Best Way to do Interprocessing Communicationa3 Message-ID: <002801c14afb$c19cccd0$8149130a@WWS006>   = I think Event Flags were also lost. Oke, its simple and smallHA communication, but it works and if your'e only wating for another  process's signal ....M   ----- Original Message -----  , From: <yyyc186@illegaltospam.mindspring.com> To: <Info-VAX@Mvb.Saic.Com>f& Sent: Monday, October 01, 2001 3:10 PM9 Subject: Re: Best Way to do Interprocessing Communicationo    4 > In <3BB3AF77.A0D49FB@videotron.ca>, on 09/27/2001 ? >    at 07:00 PM, JF Mezei <jfmezei.spamnot@videotron.ca> said:  > ! > You left off ACMS and MQ Series  >  > >John McLean wrote:eF > >> (a) Logical names - ideal for short messages,  same machine untilG > >> (b) Distributed Lock manager (DSL) - up to 32 bytes (?) and can belH > >> (c) Global sections - useful for sharing larger amounts of data and@ > >> (d) Files - last resort because the I/O is relatively slow. >  >  > >You forgot: > F > >(e) Mailboxes. Works on same node. medium amounts of data. Good for# > >many-to-one but not one-to-many.-J > >    (unless you create one mailbox for each process that talks to you).7 > >Server doesn't know if a client disapears or apears.e > J > >(f) Decnet. Works accross nodes.  A single process can talk to multipleK > >process (one-to-many, as well as many-to-one). Server knows when someoneeI > >connects and disconnects. Server has identification of who connects to  > >it. (username and node) > H > >(g) TCPIP. More or less the same as TCPIP, but with less asynchronous! > >stuff and less identification.Z > H > >(h) ICC  (Intra Cluster communications). Very efficient. One any node< > >inside a cluster. But doesn't use standard QIO stuff thatF > >mailbox,decnet,tcpip use. Requires system management to enable your? > >application to declare a service that others can connect to.n > -- >= > -----------------------------------------------------------p > yyyc186@mindspring.com= > -----------------------------------------------------------t >   g   ------------------------------  # Date: Mon, 01 Oct 2001 19:10:58 GMT + From: Jeff Campbell <jcampbell@ins-msi.com>   Subject: Re: Good bye VAX-6400's+ Message-ID: <3BB8B7FF.EF05AABA@ins-msi.com>   ) Got any extra you'd consider selling? 8-)m  = I've got a rack mount TU58-CA with both drive pucks gone. 8-(U  
 Jeff Campbelld n8wxs@arrl.net     Brian Tillman wrote: > H > >No, real pain with the TU58's is getting replacements when the rubber > >drive wheel goes bad. > D > We've gotten to a point where we make our own rubber drive wheels. > --C > Brian Tillman                   Internet: tillman_brian at si.comiC > Smiths Aerospace                          tillman at swdev.si.coml? > 3290 Patterson Ave. SE, MS      Addresses modified to preventy> > Grand Rapids, MI 49512-1991     SPAM.  Replace "at" with "@": >        This opinion doesn't represent that of my company   ------------------------------  % Date: Mon, 01 Oct 2001 21:15:00 +0100 + From: "antonio.carlini" <arcarlini@iee.org>-  Subject: Re: Good bye VAX-6400's' Message-ID: <3BB8CEC4.9B3F6F5F@iee.org>o   Bart Zorn wrote: > J > There was once upon a time a VAXstation 3540, which contained up to 4 MV% > II's (or similar, like the MV 3200).  + The VAXstation 3520 had a KA60 module whichp( used two CVAX processors - these are the( ones used in the MicroVAX 3400/3600/3900" systems (and the VS3200 & VS3500)./ Put in two KA60s and you get a VAxstation 3540.g   No UV2, no VAXBI :-)   Antonios   -- i   ---------------n- Antonio Carlini             arcarlini@iee.orgp   ------------------------------  % Date: Mon, 01 Oct 2001 21:16:55 +0100-+ From: "antonio.carlini" <arcarlini@iee.org>r  Subject: Re: Good bye VAX-6400's' Message-ID: <3BB8CF37.6542749E@iee.org>m   Fred Kleinsorge wrote: > F > The Lynx was the wrong machine to put those graphics on, and you are! > corrent, almost none were sold.h  1 Out of interest, what would the *correct* machine / have been? None of the (other) machines with a a& VAXBI would have fit the bill (or even. under the desk!). Perhaps the CVAX stuff might have been a better fit?t   Antonios   -- n   --------------- - Antonio Carlini             arcarlini@iee.orgn   ------------------------------   Date: 2 Oct 2001 05:06:38 GMTP- From: Joe Heimann <heimann@nog.ecs.umass.edu>t  Subject: Re: Good bye VAX-6400's, Message-ID: <9pbi0u$4os$2@odo.ecs.umass.edu>  + Bart Zorn <B.Zorn@truebit.nospam.nl> wrote:-J > There was once upon a time a VAXstation 3540, which contained up to 4 MV% > II's (or similar, like the MV 3200)   C 4 MV 3 processors, 2 in a Vaxstation 3520.  2.8? VUP per processor.    JoeM   > Bart Zorn   < > "Jonathan Boswell" <jsb@ost.cdrh.fda.gov> wrote in message, > news:3BB8838E.46652D7C@ost.cdrh.fda.gov... >> "antonio.carlini" wrote:H6 >> > I only ever saw one VS8000 "in the field". It was8 >> > a KA825 (1.2VUPs, same CPU as the VAX 8250). It was7 >> > indeed VAXBI based and came with an RD53 (or RD54)@ >>J >> Hmmm.  I thought the VS8000 had 3 Microvax II's inside.  Was there ever > such a8 >> VAXstation, or has my memory invented this ex nihilo?   ------------------------------  # Date: Mon, 01 Oct 2001 21:18:35 GMTi4 From: "Terry C. Shannon" <terryshannon@mediaone.net>U Subject: Re: I WILL QUIT USING VMS WHEN THE MILITARY DOES - IN 2015 OR MAYBE  LONGER! = Message-ID: <L45u7.63375$vq.11577764@typhoon.ne.mediaone.net>k  3 "John McLean" <mcleanj@dplanet.ch> wrote in messageh$ news:3BB8DB94.1C60B87F@dplanet.ch... >a >i > "Terry C. Shannon" wrote:o   > > >l > > L > > It gets worse. Seems that the next US aircraft carrier (due 2007/2008 orH > > so), CVN such-and-such, will sport a Windoze-based battle management system.nK > > I am sure that rivals and potential enemies of the USA are delighted bye this > > Stupid Strategy Trick. > 8 > Now that's what I call a Microsoft fRight sTimulator ! >-   No doubt. I can see it now....  @ DO YOU WANT TO ACTIVATE THE PHALYNX CLOSE-IN AIR DEFENSE SYSTEM?  I DO YOU *REALLY* WANT TO ACTIVATE THE PHALYNX CLOSE-IN AIR DEFENSE SYSTEM?n   ETC...  I And the last thing the guys and gals in the Combat Information Center saw  was...   CLIPPIT THE DANCING PAPERCLIP.   ------------------------------  # Date: Mon, 01 Oct 2001 22:17:51 GMTv+ From: jcring@switch.com (John C. Ring, Jr.)jU Subject: Re: I WILL QUIT USING VMS WHEN THE MILITARY DOES - IN 2015 OR MAYBE  LONGER!n, Message-ID: <9papgj$p08$1@usenet.switch.com>  s In article <L45u7.63375$vq.11577764@typhoon.ne.mediaone.net>, "Terry C. Shannon" <terryshannon@mediaone.net> wrote:  >n >No doubt. I can see it now....o >eA >DO YOU WANT TO ACTIVATE THE PHALYNX CLOSE-IN AIR DEFENSE SYSTEM?  >dJ >DO YOU *REALLY* WANT TO ACTIVATE THE PHALYNX CLOSE-IN AIR DEFENSE SYSTEM? >w >ETC...y  L Probably, but SELF-DESTRUCT will work by hitting the "Microsoft"+"D" key at O the same time, and THAT will work without confirmation.  Just like hitting the nK Microsoft+??? key undocks my laptop without confirmation.  And then nicely - tells me that I'm undocked.1  7 I've since pried the "Microsoft" key off my keyboard...>   ------------------------------   Date: 1 Oct 2001 14:34:51 -0700 ( From: bob@instantwhip.com (Bob Ceculski)T Subject: Re: I WILL QUIT USING VMS WHEN THE MILITARY DOES - IN 2015 OR MAYBE LONGER!= Message-ID: <d7791aa1.0110011334.495c665c@posting.google.com>5  r tessier-ashpool@usa.net (Chris Bardell) wrote in message news:<9f261edc.0110010534.79163b0e@posting.google.com>...o > bob@instantwhip.com (Bob Ceculski) wrote in message news:<d7791aa1.0109291946.256c80b1@posting.google.com>...  > I HAVE BEEN ON VMS NOW FOR2 > > OVER 15 YEARS AND HAVE YET TO HAVE AN OS CRASH > E > I love OpenVMS as much as you do, but surely the above is bullsh!t?h > F > OpenVMS is a million times more stable than just about all other OSs< > (witness the majority of telecoms billing, lotteries, chip# > fabrications etc. using OpenVMS).e > 7 > But to suggest/imply OpenVMS never crashes is untrue.M  = its true ... i'll say it again ... in my 15 years now on bothm7 vax and alpha vms, i have "never" had a vms os crash!!!S; the only crash i ever had was on a vaxstation 4000-90a whentA the clock chip went out ... that was a hardware induced crash ...i5 no software vms crashes ever!!!  and that is no b.s.!.  ? i took a one month stint years ago to try an as400 for the heckp< of it after some training since at the time the local market< money wise for as400 progs was sky high and crashed on os400= twice in one month ... called ibm and they didn't have a cluec@ as to why ... just took 3 days windows 2000 adv. server training; for a nt replacement on a client with a special app and the 9 class server crashed 3 times in the 3 days of training!!!oA i never will use as my server anything other than vms again!  ande that is no b.s.e   ------------------------------  % Date: Mon, 01 Oct 2001 23:09:40 +0200l& From: John McLean <mcleanj@dplanet.ch>Y Subject: Re: I WILL QUIT USING VMS WHEN THE MILITARY DOES - IN 2015 OR MAYBE LONGER! LONGe* Message-ID: <3BB8DB94.1C60B87F@dplanet.ch>   "Terry C. Shannon" wrote:o > 4 > "Keith Brown" <kbrown780@isd.net> wrote in message+ > news:trfdsar2l50pe1@corp.supernews.com...  > > Philip Lewis wrote:- > >-G > > > Did I not recently hear that one of those wonder ships running NT@	 > crasheddN > > > every system, one after another and ended up lying dead in the water for > 3r) > > > days, and needed a tow into port ??s > > >p > > >  > >a > > You heard correctly. > >n > J > It gets worse. Seems that the next US aircraft carrier (due 2007/2008 orN > so), CVN such-and-such, will sport a Windoze-based battle management system.N > I am sure that rivals and potential enemies of the USA are delighted by this > Stupid Strategy Trick.  6 Now that's what I call a Microsoft fRight sTimulator !     John McLeana   ------------------------------  % Date: Mon, 01 Oct 2001 21:03:01 -050001 From: "David J. Dachtera" <djesys.nospam@fsi.net>oY Subject: Re: I WILL QUIT USING VMS WHEN THE MILITARY DOES - IN 2015 OR MAYBE LONGER! LONG ' Message-ID: <3BB92055.B522D533@fsi.net>e   Jan-Erik Sderholm wrote:t >  > Not even a spare ship ?? :-)  E They didn't upgrade to the latest-and-greatest before putting to sea,aG nor did they download the service pack for that upgrade after they wereU two or three days out.  D Of course, I did hear that there was a rather nasty virus associated3 with that screen saver of Saddam buggering Osama...c   -- t David J. Dachtera  dba DJE Systemsl http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/r   ------------------------------  # Date: Mon, 01 Oct 2001 18:27:07 GMTu4 From: "Terry C. Shannon" <terryshannon@mediaone.net>) Subject: Re: Letter from Michael Cappelas'= Message-ID: <%z2u7.15206$xG6.6032143@typhoon.ne.mediaone.net>N  @ "Christoph Gartmann" <gartmann@immunbio.mpg.de> wrote in message* news:9pa3q1$seb$1@n.ruf.uni-freiburg.de...C > In article <2Y_t7.854$YP.23634@news.cpqcorp.net>, "Sue Skonetski"[$ <susan.skonetski@compaq.com> writes: > J > >I have just received this and thought you might want to see it.  Please be7 > >aware that not everyone (Compaq) may have seen this.  > [...] ? > >  a.. Our commitment is to continue implementing our OpenVMSl plan-of-record, D > >which includes ongoing support on the upcoming EV7 and EV79 Alpha systems,E > >as well as a continuation of the port and release on Itanium-basede systems.L > >We will, as previously stated, migrate our OpenVMS application portfolio,K > >ensuring the ongoing operating environment for our OpenVMS customers andc > >partners. >cK > What is meant with "migrating"? This sounds slightly different to what iso" > said about Himalaya or Unix... . >h  J I suspect that it's a matter of semantics. VMS, Tru64, and NSK all face an. architectural migration. Which is nothing new:   PDP--> VAX (VMS) VAX-- Alpha (VMS)t MIPS--> Alpha (Unix) CISC-->MIPS (NSK)[ Alpha-->IA64 (Tru64 Bravo port)A   ------------------------------  % Date: Mon, 01 Oct 2001 15:22:49 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca>a) Subject: Re: Letter from Michael Cappelasn, Message-ID: <3BB8C282.30354210@videotron.ca>   Sue Skonetski wrote:N > architectures.  As our valued customer, I want to emphasize that during thisM > transition period, and following the completion of the merger, your successe6 > and investment protection are foremost in our minds.  I So Compaq is down to just a single customer ????? ;-) ;-) :-) :-) :-) ;-)   M > As our track record over multiple decades clearly proves, we understand the-C > importance of your commitment to our mission critical platforms.    J But your track record over the years proves that Compaq doesn't understandN that Customers also need Compaq to keep to its own commitments. The Way CompaqJ handled the Alpha murder has removed much credibility from any Compaq road maps, speeches, commitments.  K Furthermore, a letter from Capellas means nothing. I want to see what Carlyb) says about VMS before I believe anything.e  7 > Below are several clear statements we can make today:n   And let Carly change tomorrow.  J > most available, scalable systems at the lowest cost.  Specifically, thisD > includes our announced plans for EV7 and EV79 version of the Alpha  N Is this a typo ? I was under the impression that it was EV69 to be followed byI EV7 at which point the engineers were to be let go (moved to Intel). Will,C engineers remain at Compaq past EV7 to produce an EV79 afterwards ?a  J > Therefore, it is our intent to converge Tru64 UNIX and HP/UX to a singleK > UNIX offering on Itanium, and to insure a smooth transition and migration.: > from Tru64 UNIX Alpha to this converged UNIX on Itanium.  I While it is good for Compaq to admit that Tru64 is dead (otherwise Compaq-K would have lost even more credibility), I think that Compaq should hurry to . provide details on what this will really mean.  K And I think that Compaq would have more credibility if it outright admittedSN that HP-UX would be improved to include some of the features of Tru64 to allowL Tru64 customers to migrate to HP-UX. The "converge" terminology just adds toM the confusion and shows that Compaq doesn't really want to admit the reality.-  K If I were Compaq, I would hurry to provide details on exactly what of Tru64- will be ported to HP-UX B (API, clustering, ?) And Compaq should also identify ISVs who haveO applications that run only on Tru64 to get them to start porting to HP-UX ASAP.n   ------------------------------  % Date: Mon, 01 Oct 2001 23:21:57 +0200u& From: John McLean <mcleanj@dplanet.ch>) Subject: Re: Letter from Michael Cappelasa* Message-ID: <3BB8DE75.529B050E@dplanet.ch>   Sue Skonetski wrote: >  > Dear Newsgroup,l >    .... (snip)t  M > As our track record over multiple decades clearly proves, we understand thelC > importance of your commitment to our mission critical platforms. e  H Another fragment of credibility gone.  Compaq has been around since 1982D (ie. about 2 decades) and only ever dealt with PC's until about 1996C when Tandem was purchased.  Hands up all those people who associatea mission-critical with PC's ??    .... (snip)     
 > Regards, >  > Michael Capellas >  > Chairman and CEO >  > Compaq Computer Corporationa >          John McLean    ------------------------------  % Date: Mon, 01 Oct 2001 21:23:33 -0500o1 From: "David J. Dachtera" <djesys.nospam@fsi.net>g) Subject: Re: Letter from Michael Cappelaso' Message-ID: <3BB92525.91191F72@fsi.net>h   Sue Skonetski wrote: >  > Dear Newsgroup,n > L > I have just received this and thought you might want to see it.  Please be6 > aware that not everyone (Compaq) may have seen this.  H My response is that coming from Compaq, or any representative of Compaq,< the word "commitment" is entirely meaningless at this point.  D It is wisely said that actions speak louder than words. It is thenceH that I shall draw my information as to where Compaq's commitments exist.  H ...of course, if anyone from the Compaq (or HP) management team wants toH pass along any more words, I will certainly review them for content as a' means of gathering further information.a  C For the present, however, I would recommend that both Compaq and HPxF refrain from publicly stating any more "commitments", until experienceH in the market place has at least embyronically re-established a positive track record for them.   -- l David J. Dachteran dba DJE Systemso http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------  % Date: Mon, 01 Oct 2001 22:42:02 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca>:) Subject: Re: Letter from Michael Cappelasc, Message-ID: <3BB92974.8313A2F4@videotron.ca>   "David J. Dachtera" wrote:E > For the present, however, I would recommend that both Compaq and HPoH > refrain from publicly stating any more "commitments", until experienceJ > in the market place has at least embyronically re-established a positive > track record for them.  K I disagree to a certain extent. While Commitments from Compaq may be not benL work the bits they are written on, the lack of such commitments would mean aK lot more. So Compaq/HP must make those commitments, but be fully aware thata5 customers will take all fo them with a grain of salt.-  N But as you said, actions speak louder than words. Instead of promising to portN VMS products to that IA64 thing, perhaps Compaq should publish the list of allG the software products that they plan on porting, and be honest and alson/ publish those products that will NOT be ported.w  L Once that is done, then they can start working on getting the strategic ISVs! to commit to ports very publicly.)  N Now, that is only to achieve a level of trust in Compaq not killing VMS in theJ next 5 years. It will prevent increased rate of customers leaving VMS.  IfK Compaq really wants to impress customers, its shoudl not only commit to notdK killing VMS, but also commit to marketing it and making it a core "industryoM standard" product. Unless that happens, customers will continue to think that J Compaq doesn't care about VMS and won't do anything more than necessary to/ keep the profitable VMS customers from leaving.-   ------------------------------  % Date: Mon, 01 Oct 2001 23:14:50 +0200-& From: John McLean <mcleanj@dplanet.ch>5 Subject: Re: Locate Filename containing logical block:* Message-ID: <3BB8DCCA.84FE0289@dplanet.ch>  ! norm.raphael@jamesbury.com wrote:  > - > I have a shadowset on RZ28-M's that logs anc5 > unrecoverable media error on logical block 00334339f4 > whenever I do a full backup from another system in3 > the cluster to which its members are mscp-served.c > 8 > Is there a way to tell which file contains that block? > * > (Analyze/READ does not show any errors.)    G You might get some idea if you look at the contents of the disk at that  point.  ? Try DUMP/BLOCK=(START:334339,COUNT:1)  <disk_id>  (eg. DKA200:).  H If you use the physical device instead of a filename, you should get the logical block.  B Change the value of COUNT to see a few more blocks if you want to.     John McLeant   ------------------------------   Date: 1 Oct 2001 20:21:57 -0700 / From: rcyoung@aliconsultants.com (Robert Young)c@ Subject: Re: Looking for Classic Oracle/VMS version matrix again= Message-ID: <91437ce6.0110011921.354f256d@posting.google.com>   E Check out the Lillian Hobbs book on Rdb. It gives you a lot if usefulr info.r  iF You can also go to metalink.oracle.com, register, and get into the Rdb@ info. Sometimes it takes a bit of looking on the pull down menus however.  A You may also want to join the JCC listserver on Rdb. It is a veryVF popular place to ask questions and post problems. Many of Oracle's RdbD "gurus" frequent it. Put LIST and HELP as separate lines in the body@ of an email to oraclerdb-request@jcc.com if I remember correctly   ------------------------------  # Date: Tue, 02 Oct 2001 01:59:57 GMT0- From: goathunter@goatley.com (Hunter Goatley) A Subject: Re: ODS2 available -- reads ODS-2 disks on Windows, UNIX 1 Message-ID: <3bb91f23.103660786@news.process.com>o  2 On Mon, 1 Oct 2001 11:55:11 -0400, "Brian Tillman", <tillman_brian@notnoone.notnohow.com> wrote:  J >Any change you could provide the file WNASPI32.DLL in the zip file, since >you provide ODS2_WIN32.EXE?  G Well, that seems to be the big question about ASPI in general.  An ASPIaJ interface is needed, and you can find some out there, but legally, I don'tF think they're supposed to be redistributed (at least, I'm not going toG try it).  The CDex (my favorite (and free) CD ripper/MP3 encoder) has a.$ different pointer to an ASPI driver.   Hunter ------9 Hunter Goatley, Process Software, http://www.process.com/o9 goathunter@goatley.com     http://www.goatley.com/hunter/    ------------------------------  # Date: Tue, 02 Oct 2001 02:06:53 GMT - From: goathunter@goatley.com (Hunter Goatley)sA Subject: Re: ODS2 available -- reads ODS-2 disks on Windows, UNIXc1 Message-ID: <3bb92103.104140736@news.process.com>I  P On Tue, 02 Oct 2001 01:59:57 GMT, goathunter@goatley.com (Hunter Goatley) wrote:  H >Well, that seems to be the big question about ASPI in general.  An ASPIK >interface is needed, and you can find some out there, but legally, I don'trG >think they're supposed to be redistributed (at least, I'm not going to?H >try it).  The CDex (my favorite (and free) CD ripper/MP3 encoder) has a% >different pointer to an ASPI driver.n >d? "The CDex [page]...."  and the URL is:  http://www.cdex.n3.net/D  H You can also find a link for it on my "W2K software for VMS users" page:  & http://www.goatley.com/hunter/w2k.html   Hunter ------9 Hunter Goatley, Process Software, http://www.process.com/ 9 goathunter@goatley.com     http://www.goatley.com/hunter/t   ------------------------------  $ Date: Mon, 1 Oct 2001 17:15:25 -04002 From: "Sue Skonetski" <susan.skonetski@compaq.com>3 Subject: Re: OpenVMS employment in southern France?n1 Message-ID: <X05u7.892$YP.24018@news.cpqcorp.net>h  + sent to a couple of folks I know in France,s   sue   9 "Christian Leue" <Christian.Leue@gmx.de> wrote in messagee+ news:B7DE6181.4291%Christian.Leue@gmx.de...H >N
 > Hi experts!D >ND > I've been trying to find the proper forum for VMS management-levelH > employment offers in southern France. The "traditional" online sources likeJ > monster.com aren't turning up much. Is it just because there are no more VMS-E > shops hiring? Or is it that employers are ashamed to admit they arei running. > VMS? >.  > Any tips would be appreciated! >T > ChristianE >H   ------------------------------  * Date: Mon, 1 Oct 2001 11:03:39 -0700 (PDT). From: Fabio Cardoso <fabiopenvms@yahoo.com.br>P Subject: Re: OpenVMS Fundamenalists was (Re: Major VMS reseller says VMS legacy)@ Message-ID: <20011001180339.43478.qmail@web20204.mail.yahoo.com>  3 What I know about Iranians is they are nice people,O/ despite their religious government. What is theA6 real different of a religious or democratic gov - both of them f... the people !!!t6 They have good, intelligent and dedicated students and0 scientists. And the "normal" population is kind.2 And I have sympathy for Kathami - he looks like my father ! :-)   Regardsy   FC=20     2 --- "Terry C. Shannon" <terryshannon@mediaone.net> wrote:3 > Funny thing about the Iranians... had a talk with  > some US Department of55 > State lawyers earlier this year; seems that Iran isj > trying to sue the US3 > Government for supplying "obsolete" DECsystem andE > PDP gear way back in the > 70's - 80's. >=20 > Go figure! >=20	 > cheers,p >=20	 > terry ss >=20 > ----- Original Message -----2 > From: "Fabio Cardoso" <fabiopenvms@yahoo.com.br>4 > To: "Terry C. Shannon" <terryshannon@mediaone.net>) > Sent: Monday, October 01, 2001 12:57 PMl4 > Subject: OpenVMS Fundamenalists was (Re: Major VMS > reseller says VMS legacy)  >=20 >=201 > > We are all OpenVMS Fundamentalists ... (????)e > >e4 > > I consider my self a "sunite" or those great (?)
 > > number4 > > of OpenVMS SysAdmins which believe OVMS/Alpha is > the 4 > > best option of OS. Until the arriving of Itanium > (the > > real Islam expansion). > >V6 > > And there are here, the Iranians which believe VAX > are.5 > > the best thing in the world - of course it WAS inv > > the last century :-) > >S4 > > And the Talebans (PDP owners) which want to live- > > in the past and dont want the progress...a6 > > These are the dangerous people - they dont want to. > > buy new systems (OVMS) , so they dont help  > > the OVMS market  grow up.... > >t > >c > > Hmmm > >  > >t > > Regards  > >e > > FC > >a > > But there are some > >06 > > --- "Terry C. Shannon" <terryshannon@mediaone.net>
 > > wrote: > > > 3 > > > "Bill Gunshannon" <bill@triangle.cs.uofs.edu>  > wroten > > > in message* > > > news:9p9orq$ki$1@info.cs.uofs.edu... > > > > In article > > >h4 > <kpJt7.14273$xG6.5507195@typhoon.ne.mediaone.net>, > > > >  "Terry C. Shannon"  > <terryshannon@mediaone.net>m
 > > > writes: 
 > > > > |>/ > > > > |> Yep, the tide is receding for "oldero	 > Digital0  > > > systems" a.k.a. VAXen. The	 > > > VAXj5 > > > > |> installed base has declined to ~150K unitsH > > > worldwide.
 > > > > |> > > > >E6 > > > > And I'll bet that doesn't even include my 10!!	 > > > :-)h > > > >o > > > > bill > > > >n > > >w3 > > > That very well may be the case. I suspect the@  > > > installed base numbers are2 > > > estimates, and/or that they are derived from$ > > > maintenance contract renewals,5 > > > etc. Suffice it to say that system censustaking  > is > > > an inexact science (forf6 > > > proof, look at the disparity between the numbers > > > provided by IDC and  > > > Gartner/Dataquest).  > > >t5 > > > In any event, look for CPQ to issue revised VMS0 > > > installed base numbers4 > > > Pretty Soon Now. The 450K figure is way out of > date;  > > > the real numbern) > > > undoubtedly is significantly lower.d > > >s > > >p > >r > >  > > =3D=3D=3D=3D=3DwM > > =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=- =3D=3D > > F=E1bio dos Santos Cardoso > > OpenVMS System Manager > > Rio de Janeiro - Brazilm > > fabiopenvms@yahoo.com.brM > > =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=g =3D=3D > >56 > > __________________________________________________ > > Do You Yahoo!?0 > > Listen to your Yahoo! Mail messages from any > phone. > > http://phone.yahoo.com >=20     =3D=3D=3D=3D=3DvL =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3Dd F=E1bio dos Santos Cardoso OpenVMS System Manager Rio de Janeiro - Brazilr fabiopenvms@yahoo.com.brL =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3Di  2 __________________________________________________ Do You Yahoo!?3 Listen to your Yahoo! Mail messages from any phone.  http://phone.yahoo.com   ------------------------------  % Date: Mon, 01 Oct 2001 21:28:05 -0500M1 From: "David J. Dachtera" <djesys.nospam@fsi.net>1P Subject: Re: OpenVMS Fundamenalists was (Re: Major VMS reseller says VMS legacy)& Message-ID: <3BB92635.B9BEE6C@fsi.net>   Fabio Cardoso wrote: > 5 > What I know about Iranians is they are nice people,s1 > despite their religious government. What is theI8 > real different of a religious or democratic gov - both > of them f... the people !!! 8 > They have good, intelligent and dedicated students and2 > scientists. And the "normal" population is kind.4 > And I have sympathy for Kathami - he looks like my > father ! :-)   Umm, Fabio?   A You're not making friends with posts like that, at least not IMO,l anyway.t   FWIW...e   -- m David J. Dachterap dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/n   ------------------------------  $ Date: Mon, 1 Oct 2001 15:10:41 -0400+ From: "Rick Barry" <barry@star.zko.dec.com>a+ Subject: Re: OpenVMS, Apache and hackers...c1 Message-ID: <_b3u7.884$YP.24054@news.cpqcorp.net>N  5 "Bob Ceculski" <bob@instantwhip.com> wrote in message 7 news:d7791aa1.0109300827.7393e183@posting.google.com...h: > Michael Austin <miaustin@bellsouth.net> wrote in message) news:<3BB60C15.94A3B64D@bellsouth.net>...eE > > There were some recent threads on the security of OpenVMS.  WhileeF > > building my web-hosting service, after finally being able to use aI > > dynamic DNS service  for my ADSL/PPPoE connection, I discovered in myoL > > Apache access_log.; file over 21,000 attempts in the past 3 weeks to use7 > > some WINNT hacks to gain access to the system as hee5 > > laughed..."AH-HAHAHAAA.... no access for you!!!".. > >oJ > > I was at a major chemical company in August when the last really nastyG > > virus and it's offspring were hitting...  I informed them that they L > > could solve ALL of the issues by running OpenVMS/Apache as their primary2 > > web servers.  But, alas, they never listen.... > >o > > Michael Austin4 > > First DBA Source, Inc. -- www.firstdbasource.comI > > www.spacelots.com -- a Web Hosting Service...coming soon to a browsert > > near you...  >dJ > a better webserver to use is purveyor for vms ... it is more secure thanF > apache because only mailbox priviledges are given to the workers ...E > that is not true with apache workers ... also purveyor is a processnC > based webserver and unlike apache is easier to tune for increased F > workloads ... also we did performance tests with purveyor and apacheB > under tcpware and purveyor won hands down!  tcpware beats ucx orJ > multinet as an ip stack because it is based on the vms kernel, not unix!I > purveyor also has remote browser management interface that is terrific!qG > it also has a wonderful built in proxy server that alta vista used tonF > use several years ago ... dec used purveyor when it ran its vms siteD > also ... i have found an xml product that i can call out of my cgiI > scripts to serve xml pages ... it is a solid product ... i have watchedo> > my logs getting hit the last 2 weeks by nimda and just yawn!  @ Compaq Secure Web Server (CSWS) is based on Apache and is also aF process-based web server, requiring only TMPMBX and NETMBX privileges.K Tuning is fairly straight-forward, though easier if you're already familiart with tuning Apache.   L CSWS runs on Compaq's TCP/IP stack as well as TCP/IP stacks from TCPware andC Multinet. We've worked closely with Process Software on a couple of E occasions to correct problems to ensure CSWS will run on those TCP/IPo stacks.d  I CSWS is fully supported under your existing OpenVMS support contract. The > product is free of charge, but completely supported by Compaq.  G We'll continue to add features and stay in synch with the latest Apache-  Software Foundaiton base-levels.  G The OpenVMS web site (http://www.openvms.compaq.com) is served by CSWS.u  # Coming in the next release of CSWS:a   - Apache 1.3.20D - PHP 4.0 and mod_php>
 - mod_rewritek - mod_proxy  - Perl 5.6.1 and mod_perl 1.25  
 Rick Barry CSWS Project Teamv Compaq Computer Corporationr
 Nashua, NH   ------------------------------  $ Date: Mon, 1 Oct 2001 16:37:44 -0400* From: WILLIAM WEBB <WWEBB1@email.usps.gov>+ Subject: RE: OpenVMS, Apache and hackers....- Message-ID: <0033000036962586000002L062*@MHS>   3 =0AHey Rick-- Is there a hobbyist license for CSWS?a   WWWebb   > -----Original Message-----1 > From: Info-VAX-Request@Mvb.Saic.Com at INTERNET,( > Sent: Monday, October 01, 2001 3:24 PMD > To: Webb, William W Raleigh, NC; Info-VAX@Mvb.Saic.Com at INTERNET- > Subject: RE: OpenVMS, Apache and hackers...0 >c >07 > "Bob Ceculski" <bob@instantwhip.com> wrote in messagep9 > news:d7791aa1.0109300827.7393e183@posting.google.com...e< > > Michael Austin <miaustin@bellsouth.net> wrote in message+ > news:<3BB60C15.94A3B64D@bellsouth.net>...eH > > > There were some recent threads on the security of OpenVMS.  While=  H > > > building my web-hosting service, after finally being able to use = ao: > > > dynamic DNS service  for my ADSL/PPPoE connection, I > discovered in my? > > > Apache access_log.; file over 21,000 attempts in the paste > 3 weeks to use9 > > > some WINNT hacks to gain access to the system as hea7 > > > laughed..."AH-HAHAHAAA.... no access for you!!!".e > > > ? > > > I was at a major chemical company in August when the lastl > really nasty? > > > virus and it's offspring were hitting...  I informed thema > that theyr= > > > could solve ALL of the issues by running OpenVMS/Apacher > as their primary4 > > > web servers.  But, alas, they never listen.... > > >  > > > Michael Austin6 > > > First DBA Source, Inc. -- www.firstdbasource.com> > > > www.spacelots.com -- a Web Hosting Service...coming soon > to a browser > > > near you...  > >M; > > a better webserver to use is purveyor for vms ... it is- > more secure thanH > > apache because only mailbox priviledges are given to the workers ..= .eH > > that is not true with apache workers ... also purveyor is a process=  E > > based webserver and unlike apache is easier to tune for increasednH > > workloads ... also we did performance tests with purveyor and apach= elD > > under tcpware and purveyor won hands down!  tcpware beats ucx or: > > multinet as an ip stack because it is based on the vms > kernel, not unix!.> > > purveyor also has remote browser management interface that > is terrific!; > > it also has a wonderful built in proxy server that altae > vista used to H > > use several years ago ... dec used purveyor when it ran its vms sit= e F > > also ... i have found an xml product that i can call out of my cgi> > > scripts to serve xml pages ... it is a solid product ... i > have watched@ > > my logs getting hit the last 2 weeks by nimda and just yawn! > B > Compaq Secure Web Server (CSWS) is based on Apache and is also aH > process-based web server, requiring only TMPMBX and NETMBX privileges= . < > Tuning is fairly straight-forward, though easier if you're > already familiar > with tuning Apache.o > = > CSWS runs on Compaq's TCP/IP stack as well as TCP/IP stackse > from TCPware andE > Multinet. We've worked closely with Process Software on a couple oftH > occasions to correct problems to ensure CSWS will run on those TCP/IP=  	 > stacks.t > = > CSWS is fully supported under your existing OpenVMS support. > contract. Thee@ > product is free of charge, but completely supported by Compaq. > ; > We'll continue to add features and stay in synch with theo > latest Apachem" > Software Foundaiton base-levels. >-9 > The OpenVMS web site (http://www.openvms.compaq.com) ism > served by CSWS.a >d% > Coming in the next release of CSWS:e >n > - Apache 1.3.20c > - PHP 4.0 and mod_php. > - mod_rewritec
 > - mod_proxya  > - Perl 5.6.1 and mod_perl 1.25 >e > Rick Barry > CSWS Project Teama > Compaq Computer Corporation  > Nashua, NH >=   ------------------------------   Date: 1 Oct 2001 15:03:39 -0700 ( From: bob@instantwhip.com (Bob Ceculski)+ Subject: Re: OpenVMS, Apache and hackers... = Message-ID: <d7791aa1.0110011403.75316ad4@posting.google.com>o  \ "John Vottero" <John@mvpsi.com> wrote in message news:<trh02ueklun806@news.supernews.com>...7 > "Bob Ceculski" <bob@instantwhip.com> wrote in message09 > news:d7791aa1.0109300827.7393e183@posting.google.com...e< > > Michael Austin <miaustin@bellsouth.net> wrote in message, >  news:<3BB60C15.94A3B64D@bellsouth.net>...G > > > There were some recent threads on the security of OpenVMS.  While:H > > > building my web-hosting service, after finally being able to use aK > > > dynamic DNS service  for my ADSL/PPPoE connection, I discovered in my.N > > > Apache access_log.; file over 21,000 attempts in the past 3 weeks to use9 > > > some WINNT hacks to gain access to the system as heo7 > > > laughed..."AH-HAHAHAAA.... no access for you!!!".b > > > L > > > I was at a major chemical company in August when the last really nastyI > > > virus and it's offspring were hitting...  I informed them that they N > > > could solve ALL of the issues by running OpenVMS/Apache as their primary4 > > > web servers.  But, alas, they never listen.... > > >o > > > Michael Austin6 > > > First DBA Source, Inc. -- www.firstdbasource.comK > > > www.spacelots.com -- a Web Hosting Service...coming soon to a browser( > > > near you...u > >:L > > a better webserver to use is purveyor for vms ... it is more secure thanH > > apache because only mailbox priviledges are given to the workers ...G > > that is not true with apache workers ... also purveyor is a process  > 9 > My Apache worker processes only have NETMBX and TMPMBX.  > E > > based webserver and unlike apache is easier to tune for increaseduH > > workloads ... also we did performance tests with purveyor and apacheD > > under tcpware and purveyor won hands down!  tcpware beats ucx orL > > multinet as an ip stack because it is based on the vms kernel, not unix!K > > purveyor also has remote browser management interface that is terrific! I > > it also has a wonderful built in proxy server that alta vista used totH > > use several years ago ... dec used purveyor when it ran its vms siteF > > also ... i have found an xml product that i can call out of my cgiK > > scripts to serve xml pages ... it is a solid product ... i have watchedo@ > > my logs getting hit the last 2 weeks by nimda and just yawn! > F > Didn't Process retire Purveyor?  Have they released the source code? > L > If Apache on OpenVMS doesn't provide the performance you need then I wouldL > suggest looking as the OSU server or WASD.  There are both open source andA > support is very good (especially when you factor in the price).t  E my apache workers blew up in certain cgi apps and compaq support madeoI us give them "oper" privs as a result ... i don't remember the details....  F we tried osu web server and had numerous error and tuning problems ...A for example hanging threads that we could not get to time out ...yE there is no support for osu web server and the proxy server was lousye? as mr jones even states in docs that it is not intended to be a.E production proxy server, and as we added pseudo sites performance andOE errors increased ... was not professional, had no http error failovero; page support and building and using ssl was a nightmare ...sC you are correct saying that process no longer supports purveyor ... @ the product however has never errored on us in 12 months now ...A and we found support when we were writing a dll from a company in.E new zealand ... it is still a popular web server there and in englandoE from what i hear ... process will sell the code for aprice by specialhI request ... i tried to get compaq to buy the thing when they were looking H for a web server to port to vms but they picked the cheap route and wentE with a freebie in apache ... again our performance testing shows thatiG purveyor, being a processed based not thread based web server, blisters-C both osu web server and apache on tcpware in head to head tests ...dH we spent a year testing these three platforms and for speed, reliabilityJ and ease of use, purveyor won hands down ... we have a terrific web serverE in purveyor ... why would i want to step down to inferior products???nF purveyor costs only $995 and believe me it is well worth the money ...G i wish compaq would have listened to me ... thread based web servers in!K my experience do not run as well on vms as process based one like purveyor.hO did you know that purveyor was the first commercial web server not only for vmsyM but for nt as well and that its dll logic was picked up and jointly developedm2 by microsoft into IIS, minus the bugs of course :)   ------------------------------  $ Date: Mon, 1 Oct 2001 18:30:36 -0400% From: "John Vottero" <John@mvpsi.com>g+ Subject: Re: OpenVMS, Apache and hackers...b/ Message-ID: <trhrkdnppuha25@news.supernews.com>e  D The license for CSWS is included in OpenVMS.  If you have an OpenVMS* Hobbyist License, you have a CSWS license.  7 "WILLIAM WEBB" <WWEBB1@email.usps.gov> wrote in messagey' news:0033000036962586000002L062*@MHS...   0 Hey Rick-- Is there a hobbyist license for CSWS?   WWWebb   > -----Original Message-----1 > From: Info-VAX-Request@Mvb.Saic.Com at INTERNETe( > Sent: Monday, October 01, 2001 3:24 PMD > To: Webb, William W Raleigh, NC; Info-VAX@Mvb.Saic.Com at INTERNET- > Subject: RE: OpenVMS, Apache and hackers...M >l > 7 > "Bob Ceculski" <bob@instantwhip.com> wrote in messageI9 > news:d7791aa1.0109300827.7393e183@posting.google.com...h< > > Michael Austin <miaustin@bellsouth.net> wrote in message+ > news:<3BB60C15.94A3B64D@bellsouth.net>...vG > > > There were some recent threads on the security of OpenVMS.  While H > > > building my web-hosting service, after finally being able to use a: > > > dynamic DNS service  for my ADSL/PPPoE connection, I > discovered in my? > > > Apache access_log.; file over 21,000 attempts in the pastm > 3 weeks to use9 > > > some WINNT hacks to gain access to the system as he 7 > > > laughed..."AH-HAHAHAAA.... no access for you!!!".  > > >:? > > > I was at a major chemical company in August when the last  > really nasty? > > > virus and it's offspring were hitting...  I informed thema > that theyh= > > > could solve ALL of the issues by running OpenVMS/Apache  > as their primary4 > > > web servers.  But, alas, they never listen.... > > >  > > > Michael Austin6 > > > First DBA Source, Inc. -- www.firstdbasource.com> > > > www.spacelots.com -- a Web Hosting Service...coming soon > to a browser > > > near you...d > > ; > > a better webserver to use is purveyor for vms ... it isc > more secure thanH > > apache because only mailbox priviledges are given to the workers ...G > > that is not true with apache workers ... also purveyor is a processaE > > based webserver and unlike apache is easier to tune for increasedwH > > workloads ... also we did performance tests with purveyor and apacheD > > under tcpware and purveyor won hands down!  tcpware beats ucx or: > > multinet as an ip stack because it is based on the vms > kernel, not unix! > > > purveyor also has remote browser management interface that > is terrific!; > > it also has a wonderful built in proxy server that alta  > vista used tonH > > use several years ago ... dec used purveyor when it ran its vms siteF > > also ... i have found an xml product that i can call out of my cgi> > > scripts to serve xml pages ... it is a solid product ... i > have watched@ > > my logs getting hit the last 2 weeks by nimda and just yawn! >@B > Compaq Secure Web Server (CSWS) is based on Apache and is also aH > process-based web server, requiring only TMPMBX and NETMBX privileges.< > Tuning is fairly straight-forward, though easier if you're > already familiar > with tuning Apache.  >p= > CSWS runs on Compaq's TCP/IP stack as well as TCP/IP stacks= > from TCPware andE > Multinet. We've worked closely with Process Software on a couple ofoG > occasions to correct problems to ensure CSWS will run on those TCP/IPo	 > stacks.D >== > CSWS is fully supported under your existing OpenVMS support3 > contract. The_@ > product is free of charge, but completely supported by Compaq. >t; > We'll continue to add features and stay in synch with theh > latest Apache-" > Software Foundaiton base-levels. > 9 > The OpenVMS web site (http://www.openvms.compaq.com) is. > served by CSWS.u >c% > Coming in the next release of CSWS:M >r > - Apache 1.3.20V > - PHP 4.0 and mod_php3 > - mod_rewrites
 > - mod_proxy   > - Perl 5.6.1 and mod_perl 1.25 >u > Rick Barry > CSWS Project Teams > Compaq Computer Corporatione > Nashua, NH >=   ------------------------------  # Date: Mon, 01 Oct 2001 22:49:17 GMT.2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)+ Subject: RE: OpenVMS, Apache and hackers...h1 Message-ID: <Np6u7.899$YP.24141@news.cpqcorp.net>h  Z In article <0033000036962586000002L062*@MHS>, WILLIAM WEBB <WWEBB1@email.usps.gov> writes:  1 :Hey Rick-- Is there a hobbyist license for CSWS?   ,   There is specifically no license for CSWS.     Rick Barry wrote:y  > :> CSWS is fully supported under your existing OpenVMS support; :> contract. The product is free of charge, but completely   :> supported by Compaq...4  ?   A hobbyist can do what most any other licensed OpenVMS users V)   can do: download CSWS and configure it.3  N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  # Date: Mon, 01 Oct 2001 23:01:59 GMTu2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)+ Subject: Re: OpenVMS, Apache and hackers...s1 Message-ID: <HB6u7.900$YP.24163@news.cpqcorp.net>k  h In article <d7791aa1.0110011403.75316ad4@posting.google.com>, bob@instantwhip.com (Bob Ceculski) writes:F :my apache workers blew up in certain cgi apps and compaq support madeJ :us give them "oper" privs as a result ... i don't remember the details...  F   Any CGI script needs sufficient privileges for whatever it is doing.C   This has nothing to do with the CGI environment, and would apply iB   equally to the Purveyor or any other web server CGI environment.  B :... i tried to get compaq to buy the thing when they were lookingI :for a web server to port to vms but they picked the cheap route and wents :with a freebie in apache   G   The commercial web server market cratered several years ago -- as oneoJ   example, we had shipped Netscape FastTrack for several OpenVMS releases,H   and Netscape got out of that business (and several other web servers).  I   When Apache was chosen, we picked the most popular web server availablecK   to port and to package (and to support).  For all its benefits, Purveyor tK   wasn't it.  (I am aware of OpenVMS web servers that are still on or were yG   once running Purveyor here within OpenVMS Engineering, and there are aI   various folks familiar with it.)  Purveyor works nicely.  The on-going e=   Apache web server development, however, is far more active.   N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  $ Date: Mon, 1 Oct 2001 21:20:42 -0400. From: "warren sander" <sander@ma.ultranet.com>+ Subject: Re: OpenVMS, Apache and hackers...a+ Message-ID: <9pb4sn$aih$1@bob.news.rcn.net>t  K The OpenVMS external web server ran on Purveyor for 5 years, on Sept 10th Ic switched over to CSWSr  I I had no problems with purveyor and actually liked it and would have beeno= content to stay on it forever execpt for a couple of issues..c  L 1) server side includes required special extensions and the new compaq brand templates wereL much easier to impliment using includes. I don't think all you would like to start using .htp as themF extension for all the web pages instead of .html or in some cases .htm   2) Perl was acting up.  H 3) No support, (or very little). I got some and really can't complain as* the folks at process always helped me out.3 but it was getting harder to get problems addressed-  K 4) new features were needed to support the future we are envisoning for the1 OpenVMS web server  L I needed to go to a fully supported web server environment, on a new version" of OpenVMS (7.3 upgraded from 6.2)  G OpenVMS Engineering/Marketing needed to be able to show the OpenVMS web  server running on what weS are now supporting.    --B ------------------------------------------------------------------6 Warren Sander                        OpenVMS MarketingK Compaq Computer Corporation          Work:  warren.sander@remove.compaq.comwL 200 Forest Street MR01-3/J1          Personal: sander@remove.ma.ultranet.com3 Marlboro, MA 01752                   (508) 467-4875m5    My opinions are my own and I only speak for myself ,          Read http://www.openvms.compaq.com/B ------------------------------------------------------------------   ------------------------------  # Date: Mon, 01 Oct 2001 22:45:46 GMTD6 From: "Andy Bustamante" <a_c_bustamante@earthlink.net>0 Subject: Re: Point me in the right direction....E Message-ID: <um6u7.8867$cp1.1134823@newsread2.prod.itd.earthlink.net>e   Assuming Compaq's TCPIP 5.?AL If you use /accept anything not listed will be rejected.  If you use /REJECTG any other host will be accepted.  We had an Alpha outside of a firewallpE (customer network, long ugly story) which I configured not to talk to 
 strangers.  9 See the help on SET SERVICE /ACCEPT & SET SERVICE /REJECTe    G TCPIP> disable service telnet            !Don't telnet in and try this! > TCPIP> set service telnet /accept=network=(10.0.0.0:255.0.0.0) TCPIP> enable service telnet     -- Andy Bustamante  Remove the ASCII 95s to reply       , "Robin" <rlb@austin.rr.com> wrote in message& news:3BB5E33E.3ACC12A@austin.rr.com...D > Is it possible to restrict Telnet and FTP access via IP address on > an > OpenVMS 7.3 Alphaserver?   ------------------------------  $ Date: Mon, 1 Oct 2001 16:35:58 -05003 From: "David Silvers" <david.silvers@aspentech.com>k" Subject: Re: Question on VMS Virus. Message-ID: <9pandj$cs2$1@selma.aspentech.com>  K Hmmm -- actually, given privilege, any competent VMS coder could create andtE install a virus on VMS - but there's the rub 'given privilege', as toaK propagation years ago it would have been much easier to send something like 1 this around an improperly secured DECnet network.   ? It really depends upon the vigilance of the systems managers...n  + <david20@alpha2.mdx.ac.uk> wrote in messageo% news:9pa9ht$r5b$1@aquila.mdx.ac.uk...oE > In article <9pa7j8$9v8$1@joe.rice.edu>, leslie@clio.rice.edu (Jerryi Leslie) writes:  > >sfm (sfm1115@bjc.org) wrote:2J > >: I have never heard of a system running OpenVMS to obtain a virus, butK > >: recently we found that a NT Server which provides a software link to 1 + > >: of our servers had many virus's on it.y > >:D > >: Is there software I can obtain to check for virus's on my alpha, > >: servers or should I not worry about it. > >kL > >See the section titled "MGMT2.  Are there any known viruses for OpenVMS?" > >in the OpenVMS FAQ: > > ; > >   http://www.openvms.compaq.com/wizard/openvms_faq.htmll > > C > >IMO, it's a non-issue unless your VMS systems are serving shareso5 > >to Windows PCs by running SAMBA or Pathworks, etc.c > >f > J > There are a few other ways in which a VMS system might spread viruses to > vulnerable PC systems. >t > eg > J > If user's have mail stored on the VMS system and read it from a PC usingJ > POP or IMAP then you need to have scanned the incoming mail for viruses. > L > If a user can upload files via FTP to a directory which is then served outG > by a VMS based web-server - you need to scan those files for viruses.k >rJ > In short you might want to use Sophos if you have any situation where PC files-L > are being loaded onto a VMS system and then can be accessed by PC systems.J > (note. The Sophos product can be run against any files on the VMS system notDH > just on Pathworks shares - skim reading of the documentation sometimes leavesL > people with the erronious impression that you need to be running pathworks to > use Sophos). >. > David Webb > VMS and Unix team leader > CCSS > Middlesex University   ------------------------------   Date: 1 Oct 2001 18:40:29 -0500e- From: Kilgallen@SpamCop.net (Larry Kilgallen) " Subject: Re: Question on VMS Virus3 Message-ID: <3DvcPIKXcmsL@eisner.encompasserve.org>m  d In article <9pandj$cs2$1@selma.aspentech.com>, "David Silvers" <david.silvers@aspentech.com> writes:M > Hmmm -- actually, given privilege, any competent VMS coder could create andmG > install a virus on VMS - but there's the rub 'given privilege', as totM > propagation years ago it would have been much easier to send something like 3 > this around an improperly secured DECnet network.   I VMS Development created one some years ago to demonstrate SEVMS at DECUS. H The phrase I generally use is there have been no recorded sightings of a VMS virus "in the wild".  I The example they created, by the way, did not use privilege -- just writeh. access to a command procedure run by the user.   ------------------------------   Date: 1 Oct 2001 20:26:22 -0700./ From: rcyoung@aliconsultants.com (Robert Young)  Subject: Ram Disk for VMS?= Message-ID: <91437ce6.0110011926.2902d555@posting.google.com>e  D Isn't there some "free" Ram disk software for OpenVMS...maybe on theC freeware cd or in a utility directory?? I keep thinking I have seen D something somewhere....but I am getting old and feeble minded and do  not know where I saw it anymore!   ------------------------------   Date: 2 Oct 2001 00:23:10 GMTr* From: bleau@umtof.umd.edu (Lawrence Bleau)% Subject: Re: Really Unhackable...????!) Message-ID: <9pb1de$vo7$1@hecate.umd.edu>i  c In article <PyHTewtI4QoC@eisner.encompasserve.org>, koehler@encompasserve.org (Bob Koehler) writes:nW >In article <9p0egn$o2j$1@hecate.umd.edu>, bleau@umtof.umd.edu (Lawrence Bleau) writes:t >> tI >> Needless to say, these things just can't be done on any of today's VMS  >> systems.  >> . >iA >   Why not?  Certainly SYSTEM can be used to set any password tot, >   anything if the holder is stupid enough.  O I suspect the previous poster was referring to the fact that out-of-the-box VMSwO doesn'y allow simplistic passwords; the installation - and upgrade - procedures O force you to change the password on SYSTEM, SYSTEST, & FIELD to something else, M and not an obvious, simplistic choice.  Of course one cab use AUTHORIZE lateruI on to change the password to a simplistic one, but that's not the default H (out of box, unchanged) password policy, it's a very poor user-specified password policy.   Lawrence Bleau University of Maryland" Physics Dept., Space Physics Group 301-405-6223 bleau@umtof.umd.edus   ------------------------------  # Date: Mon, 01 Oct 2001 19:05:04 GMT 2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)) Subject: Re: SCSI Cluster support on LVD? 1 Message-ID: <A73u7.882$YP.24018@news.cpqcorp.net>n  u In article <ed1713eb.0109301353.34d9459d@posting.google.com>, graham@the-shades.demon.co.uk (Graham Harrison) writes:r  ( :Does anyone know if, and if not - when,* :Scsi clustering will be supported on LVD?  C   LVD (Low Voltage Differential) is not particularly relevent, the e<   particular controller and its support for TCQ is relevent.6   UltraSCSI with LVD is Ultra2 (or maybe Ultra3) SCSI.  + :I have a few scsi clusters to support, andr/ :we need to try and reduce the IO bottle necks.e  D   Pet peeve: there is no such thing as a "SCSI cluster", as you mustG   always have a cluster communications interconnect in addition to the tE   multi-host storage interconnect that is provided by SCSI.  In othercG   words, any valid cluster that is sharing a SCSI bus between hosts is s;   by definition a mixed-interconnect cluster configuration.b  . :I can upgrade the controllers on the external, :RAID to LVD quite reasonably (compaired to + :adding more disks)  But are any of the LVDo) :(Ultra2 or Ultra3) controlers supported?a  D   You need to be very careful when mixing SCSI hardware generations.  E   With V7.1-1H1 (and ECOs) or (far better) V7.1-2 and ECOs, multihostwD   UltraSCSI configurations are first supported.  One of the typical =   UltraSCSI controllers used (and supported) is the KZPBA-CB.   N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  # Date: Tue, 02 Oct 2001 02:44:34 GMTe* From: cjt & trefoil <cheljuba@prodigy.net>+ Subject: Re: Sun goes after Alpha user baseS+ Message-ID: <3BB92A3E.72F8AD36@prodigy.net>V   "Terry C. Shannon" wrote:e > 5 > "Dan Foster" <dsf@frontiernet.net> wrote in messagew1 > news:9p966a$2gau$1@node21.cwnet.roc.gblx.net...M. > > In article <3B713EBB.CDA83D5E@uk.sun.com>,6 > > andrew harrison  <andrew.nospam@uk.sun.com> wrote:J > > >> some time next year or early 2003, EV7 is likely to.  When will Sun > figure* > > >> out how to make a cache that works? > > >> > > >C: > > >We did some time ago. The ecache on E10Ks and smaller< > > >UltraII based machiens has been modified. This happened > > >late last year. > >9E > > Hmm. From published reports that I've read, Sun went as far as tolE > > 'request' that early sites with the ecache fixes do so under NDA?. > > < > > If true (in whatever form), that sure wasn't impressive. > K > I reported this matter a couple of years ago, back when disgruntled UE10KaH > users set up a "Disgruntled UE10K Users" Web site. Jaikumar Vijayan atI > ComputerWorld ran the story in midsummer 2000. Sun was not amused about 8 > being outed, but they ultimately resolved the problem. >  > > G > > My understanding is that it took Sun quite a while to get an handle E > > on the whole thing - admittedly I still don't know what the causelF > > is to this day, but it's sounding like a general chip design issue+ > > that just didn't get caught in testing.e > I > The cache in the 400MHz and 440MHz UltraSparc II CPUs didn't have error,K > correction capability. As noted about, the more recent CPUs do not suffero > from this deficiency.a  O That in and of itself might not have been serious, as I understand it, had one bO of their suppliers not produced cache that was especially susceptible to cosmicm4 rays, and in a year of exceptional sunspot activity.  
 Murphy's law.l   ------------------------------  # Date: Mon, 01 Oct 2001 21:48:12 GMTe  From: jlsue <jlsuexxxz@home.com>J Subject: Re: To the tune of "It's still rock and roll to me" by Billy Joel8 Message-ID: <a1phrtsgndqf62v06erk3ev6tsbc1g1ubd@4ax.com>  7 On Sat, 29 Sep 2001 12:02:45 -0500, "David J. Dachtera"! <djesys.nospam@fsi.net> wrote:   >You missed your calling...i >a >Hamlyn Mootoo wrote:l >>   >> ...  > That's because there's no calling standard outside of OpenVMS.   ------------------------------  % Date: Tue, 02 Oct 2001 00:34:39 +0100 1 From: Steve Reece <SYSTEM@ipl.demon.co.nospam.uk> # Subject: Re: VAX disaster tolerancee6 Message-ID: <3BB90B9F.736B48F9@ipl.demon.co.nospam.uk>  C Not sure if anyone has respondedto this or not yet so here goes....u  H It's important to bear in mind what the aim of Recoverall is.  It is NOTH a business continuity service or disaster recovery service.  There is noH provision within Recoverall to have a mobile machine room rolling out of< Compaq within 4 hours to get your site up and running again.> What Recoverall does provide is an insurance policy to get youA equivalent kit or cash for the kit that you've lost in a disasterm) scenario.  i.e. it's an insurance policy.e  = Payout on Recoverall can be up to 90 days after the disaster.p   Steve.     Keith Parris wrote:t > H > For those who are really serious about disaster recovery, Digital (andB > now Compaq) sell a service called Recover-All, which promises toG > replace equipment lost in a disaster.  This service isn't included insE > a regular hardware service agreement; with that, they only agree tov? > service the equipment if it breaks, not to replace it if it'si > destroyed by a disaster.E > ------------------------------------------------------------------- E > Keith Parris | parris at encompasserve dot org | VMS consulting on:mE > Clusters, Disaster Tolerance, Internals, Performance, Storage & I/O    -- tG "A shadow fell over her face; clear, as if the composure were rent like E a veil.  And her lips parted, but only with a short intake of breath.aA Then she said, 'Well, then you are right.  Indeed, we are even.'"7% 		Louis, "Interview with the Vampire"a   ------------------------------  % Date: Mon, 01 Oct 2001 22:43:37 +0100o1 From: Steve Reece <SYSTEM@ipl.demon.co.nospam.uk>s5 Subject: Re: VMS Itanium port speeded up. Tru64 dies.C6 Message-ID: <3BB8F199.7427CA8E@ipl.demon.co.nospam.uk>   John,yG Are you really sure you need clusters with all three architectures in?  : Do you really _need_ cluster-common disks across all three architectures?  E You could perhaps make use of DEC DFS for presenting cluster disks onr@ the VAX and/or Alpha systems up to the IPF systems.  It wouldn'tE necessarily get over the problem of your customers being confident inQ@ running mixed architecture clusters but it might ease the route.   Steve.     John Santos wrote: > * > On Mon, 17 Sep 2001, Hoff Hoffman wrote: >  > > In article <y4d74w505k.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>, Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de> writes:4 > > :Kilgallen@SpamCop.net (Larry Kilgallen) writes: > > :sL > > :> Mark Gorham's slides indicated the first early adopters release wouldM > > :> have DECnet Phase IV (he paused expecting applause) but no clustering.mM > > :> They are hedging about support (as distinguished from code) for triplet > > :> architecture clusters.. > > :tL > > :Quite to the contrary - having triple-architecture clusters is good for > > :exposing the bugs.h > >sJ > >   I covered this in my presentation as well -- the "fun" here involvesK > >   testing and support.  I do not presently know of any technical reasonpK > >   why an OpenVMS VAX, OpenVMS Alpha, and OpenVMS Itanium node could notbO > >   coexist in the same cluster.  I do know that we will be targeting OpenVMSmJ > >   Alpha and OpenVMS Itanium systems for our testing, and are presentlyH > >   evaluating customer requirements for triple-architecture clusters. > >gI > >   I'll see if I can get copies of the presentations onto the website.a > D > Not speaking for my company, but just personal opinion...  I'm notA > sure if we will port our products to IPF, but if we do, we willaC > certainly want a triple-architecture cluster (i.e. add an ItaniumeC > to our existing 1-VAX/2-Alpha cluster.)  There is little prospecteC > that ALL our VAX customers will have either gone away or migratedhD > to Alpha in the next 2 years, and we will probably need to supportD > all 3 architectures for a considerable time.  We will want to keep* > our single source library, not clone it. > C > I don't know if many production shops are in a similar situation,-@ > but I think lots of software providers are.  We could probablyA > live with limitations that a production shop might not toleratecA > (e.g. not being able to share a SYSUAF.DAT between a VAX and an2C > IPF, or having to run multiple queue managers, etc.) but we wouldRB > definitely need cluster disk serving across all 3 architectures,  > distributed lock manager, etc. >  > --
 > John Santose   --  G "A shadow fell over her face; clear, as if the composure were rent likehE a veil.  And her lips parted, but only with a short intake of breath.nA Then she said, 'Well, then you are right.  Indeed, we are even.'"u% 		Louis, "Interview with the Vampire"    ------------------------------   Date: 1 Oct 2001 13:57:22 -0500u9 From: kaplow_r@eisner.encompasserve.org.mars (Bob Kaplow)a> Subject: Re: VMScluster using fibrechannel and memory channel?3 Message-ID: <V0nMSc824$zk@eisner.encompasserve.org>p  T In article <3BB85BC8.1DAC893E@127.0.0.1>, Nic Clews <sendspamhere@127.0.0.1> writes:E >> > In working the optimum cluster size, someone wrote a program (or-L >> > something) in Ingres to calculate it. It allowed us to pick the "right"I >> > cluster size and set up partitions and volume structures to the datavL >> > requirements. Unfortunately I've no way of getting that into the publicK >> > domain, but if you're looking for a little freeware project, a clusteroJ >> > sizes versus file size analysis tool with 'what if' would be nice :-) >> -M >> Years ago, I learned to set cluster sizes based on the MEDIAN file size ongM >> the volume. IIRC it was the late Cliff Fischer (DEC and when I met him MTI N >> engineer) who taught me this technique. Most defraggers offer a free reportL >> utility (to convince you to buy the product) that will show # of files ofO >> each size. If most of your files are 8 blocks, don't set the cluster size to- >> 2, or to 100. > I > Perhaps I'm not looking closely enough at the report. Math(s) is not my J > strong point, so I'd welcome more information on doing this manually, or > of course a program!   $ FAU sys$sysdevice /filed  6                         Fragmentation Analysis Utility FRAGMENTATION_ANALYSIS V7.2-000uI Copyright (c) 1995 Executive Software International. All Rights Reserved.j? _DSA500:                                 1-OCT-2001 13:42:59.56hI Number of Usable Blocks:  8378028       Cluster Size           :        9c   N <<<<<<<<<<<<<<<<<<<<<<<<<<<<  Free Space Summary >>>>>>>>>>>>>>>>>>>>>>>>>>>>>   I Total Free Space Size  :  2787021       Smallest Free Space    :        9tI Number of Free Spaces  :    17544       Largest  Free Space    :     1179.I #Spaces =  80% of Total:     6707       Mean Size of Free Space:      158A  IN <<<<<<<<<<<<<<<<<<<<<<<<<<<<<  File Information >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>I Maximum Number of Files:   418901       # Reserved     Files   :        9eI Total Number of Files  :    15696       # Placed       Files   :        5-I Total Size of all Files:  5560875       # Multi-Hdr File Hdrs  :      715 I Smallest File Size     :        9       # Multi-Volume Files   :        0-I Largest  File Size     :   843534       # Directory    Files   :      570,I Mean Size of all Files :      354       # Zero Length Files    :     1046eI # Extent Headers       :      665       # Files with Frags >= 2:     1184 I Total File Fragments   :    71221       Lost Blks Only         :    30132 I Avg Fragments per File :     4.86       Total Split I/Os       : 24269602aI % of Disk Fragmented   :        0%      # of Unwanted Fragments:    56571@  iN <<<<<<<<<<<<<<<<<<<<<<<<<<<<  File Sizes Summary >>>>>>>>>>>>>>>>>>>>>>>>>>>>>  eI      Size Range           # Files            Size Range           # Files J --------------------     ---------      --------------------     ---------L        1 to        1  -          0           256 to      511  -        400  L        2 to        3  -          0           512 to     1023  -        328  L        4 to        7  -          0          1024 to     2047  -        190  L        8 to       15  -       7144          2048 to     4095  -        104  L       16 to       31  -       2499          4096 to     8191  -         65  L       32 to       63  -       2050          8192 to    16383  -         35  L       64 to      127  -        958         16384 to    32767  -          9  L      128 to      255  -        851         32768 to  8378028  -         17    K In my case this is kind of boring, since this disk has a cluster size of 9,yF there  isn't anything in the first three buckets. This distribution isK pretty typical, but IF I had a cluster size of 1, you'd probably see ratherPF than more files in each bucket as the size got smaller, a peak at someL point. The peak is the place it was recommended to set your cluster size to.  L For example if the three zeros above were  983, 3762, and 8219 then the peakJ value 8219 would correspond to the 4-7 block size. That would be the valueG I'd set my cluster size to. Sometimes you'll find a disk with a bimodaloG distribution, especially if you've got very different workloads storingeB files on the disk. I used to see that with SAS, where both command8 procedures and huge data sets existed onthe same device.   ------------------------------  $ Date: Mon, 1 Oct 2001 21:37:34 +0200- From: "Martin Hoogenboom" <martinh@xs4all.nl>g> Subject: Re: VMScluster using fibrechannel and memory channel?3 Message-ID: <044501c14ab4$d057fac0$8149130a@WWS006>r   ----- Original Message -----; From: "Bob Kaplow" <kaplow_r@eisner.encompasserve.org.mars>e To: <Info-VAX@Mvb.Saic.Com> & Sent: Monday, October 01, 2001 8:57 PM> Subject: Re: VMScluster using fibrechannel and memory channel?    5 > In article <3BB85BC8.1DAC893E@127.0.0.1>, Nic Clews   <sendspamhere@127.0.0.1> writes:G > >> > In working the optimum cluster size, someone wrote a program (oroF > >> > something) in Ingres to calculate it. It allowed us to pick the "right" K > >> > cluster size and set up partitions and volume structures to the dataMG > >> > requirements. Unfortunately I've no way of getting that into thes publicE > >> > domain, but if you're looking for a little freeware project, a- cluster-L > >> > sizes versus file size analysis tool with 'what if' would be nice :-) > >>L > >> Years ago, I learned to set cluster sizes based on the MEDIAN file size onK > >> the volume. IIRC it was the late Cliff Fischer (DEC and when I met himo MTInI > >> engineer) who taught me this technique. Most defraggers offer a free  reportK > >> utility (to convince you to buy the product) that will show # of fileso ofI > >> each size. If most of your files are 8 blocks, don't set the cluster  size to3 > >> 2, or to 100. > >oK > > Perhaps I'm not looking closely enough at the report. Math(s) is not myDL > > strong point, so I'd welcome more information on doing this manually, or > > of course a program! >v > $ FAU sys$sysdevice /file  > 8 >                         Fragmentation Analysis Utility! > FRAGMENTATION_ANALYSIS V7.2-000>K > Copyright (c) 1995 Executive Software International. All Rights Reserved.sA > _DSA500:                                 1-OCT-2001 13:42:59.56eK > Number of Usable Blocks:  8378028       Cluster Size           :        9  > 2 > <<<<<<<<<<<<<<<<<<<<<<<<<<<<  Free Space Summary >>>>>>>>>>>>>>>>>>>>>>>>>>>>>i >iK > Total Free Space Size  :  2787021       Smallest Free Space    :        9aK > Number of Free Spaces  :    17544       Largest  Free Space    :     1179oK > #Spaces =  80% of Total:     6707       Mean Size of Free Space:      158n >t1 > <<<<<<<<<<<<<<<<<<<<<<<<<<<<<  File Information  >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>K > Maximum Number of Files:   418901       # Reserved     Files   :        9nK > Total Number of Files  :    15696       # Placed       Files   :        5-K > Total Size of all Files:  5560875       # Multi-Hdr File Hdrs  :      715 K > Smallest File Size     :        9       # Multi-Volume Files   :        0uK > Largest  File Size     :   843534       # Directory    Files   :      570sK > Mean Size of all Files :      354       # Zero Length Files    :     1046>K > # Extent Headers       :      665       # Files with Frags >= 2:     1184oK > Total File Fragments   :    71221       Lost Blks Only         :    30132 K > Avg Fragments per File :     4.86       Total Split I/Os       : 24269602cK > % of Disk Fragmented   :        0%      # of Unwanted Fragments:    565712 >r2 > <<<<<<<<<<<<<<<<<<<<<<<<<<<<  File Sizes Summary >>>>>>>>>>>>>>>>>>>>>>>>>>>>>r >nK >      Size Range           # Files            Size Range           # FilesCL > --------------------     ---------      --------------------     ---------L >        1 to        1  -          0           256 to      511  -        400L >        2 to        3  -          0           512 to     1023  -        328L >        4 to        7  -          0          1024 to     2047  -        190L >        8 to       15  -       7144          2048 to     4095  -        104L >       16 to       31  -       2499          4096 to     8191  -         65L >       32 to       63  -       2050          8192 to    16383  -         35L >       64 to      127  -        958         16384 to    32767  -          9L >      128 to      255  -        851         32768 to  8378028  -         17 >aJ > In my case this is kind of boring, since this disk has a cluster size of 9,H > there  isn't anything in the first three buckets. This distribution isF > pretty typical, but IF I had a cluster size of 1, you'd probably see ratherH > than more files in each bucket as the size got smaller, a peak at someJ > point. The peak is the place it was recommended to set your cluster size to.  >-I > For example if the three zeros above were  983, 3762, and 8219 then thej peakL > value 8219 would correspond to the 4-7 block size. That would be the valueI > I'd set my cluster size to. Sometimes you'll find a disk with a bimodal I > distribution, especially if you've got very different workloads storinguD > files on the disk. I used to see that with SAS, where both command: > procedures and huge data sets existed onthe same device. >t  E Somewhere in the back of my mind i remember something about higher IOlJ rates due to smaller cluster sizes. It must be possible to find an optimumG but it's probably not just 'the smaller the better'. I would seek for abH somewhat larger size that can hold most of your files in one cluster but? doesn't have its drawbacks in performance for the larger files. L Probably the cluster size of 9 blocks isn't bad at all in the given example.  J There is also something about the actual use of the files. If you would beH using Oracle for example I would go for a cluster size equal to or at anA integer multiple of the database page size. My guess is somethingaE the same for other databases. For that matter it could be interestinge# what the 17 big files are used for.i@ You should also keep in mind the access frequency of files. Some< large database files on a disk may well be accessed far moreD frequently then the many small files. From a performance perspective6 you would then seek for a very much different optimum.  D I have worked in an environment where the average size of files on aB specific disk was 92 blocks (some 5000000 files, size varying fromB 60-150 blocks. We scored best (in performance and disk efficiency) with a cluster size of 32.G But it took a while to find it out. Mind you, a full restore to a newlyd> formatted disk with a new clustersize (/TRUNCATE) took us more than 48 hours.B This was of course after the upgrade to V7.2-1. Since that versionD you can use smaller cluster sizes for larger volumes. Before that we- had to use stripe sets to hold all that data.o   Regards, Martin Hoogenboom    ------------------------------   Date: 1 Oct 2001 19:38 CDT' From: carl@gerg.tamu.edu (Carl Perkins) " Subject: Re: VT emulation on Linux, Message-ID: <1OCT200119380836@gerg.tamu.edu>  - "remy.younes" <remy.younes@free.fr> writes...mK }There is a simple shell script called vmsterm which redefines the keyboardd8 }mapping and works just fine in a telnet window/console.L }Unfortunately i don't remember where i did get it from, nor where i put it. }Prehaps somebody else ?   Try:  4 http://www-personal.une.edu.au/~oahlefel/vmsterm.htm   --- Carl  F }"Jeffrey H. Coffield" <jeffrey@digitalsynergyinc.com> a crit dans le9 }message news: 3BB32E7A.AEBF2558@digitalsynergyinc.com...rH }> Has anyone found a good (and free) VT emulation for Linux? The best I }> have been able I }> to find is Eterm but it has some bugs in switching between numeric ando }> application }> keypad modes. }>   ------------------------------   Date: 01 Oct 2001 18:14:29 GMT) From: wkb@freebie.xs4all.nl (Wilko Bulte)r8 Subject: Re: Warranty on ES45 reduced from 3 years to 1?; Message-ID: <3bb8b284$0$19390$e4fe514c@newszilla.xs4all.nl>t  k In <69Kt7.14283$xG6.5530020@typhoon.ne.mediaone.net> "Terry C. Shannon" <terryshannon@mediaone.net> writes:t    ; >"Jack Patteeuw" <jjpatteeuw@peoplepc.com> wrote in messagel' >news:3BB76B0E.314569DB@peoplepc.com...l >> >> >> "Terry C. Shannon" wrote: >> >B >> > Good point. Anything to minimize the entry price is goodness. >>L >> Actually anyone buying "NEW" hardware is really missing the boat !!  Used >ES40's are G >> cheap, even with 833 Mhz processors. !!  I got a quote on an 833 Mhz- >processor (used)- >> for $8,500.  H >Yep, if you don't need a GHz of compute performance, or the increase inL >memory bandwidth, you can save Big Bucks by opting for pre-owned equipment. >>L >> Sun is dying right now because there is so much of "last years" equipment >on the usedK >> market (from all the door coms going out of business) that hardly anyone 
 >is buyingM >> their new Sun Blades.  (Ultra 1 w/Creator 3D, 256 Mb memory, 2 2Gb - $750)-  F >Finally came across some Sun Blades. They were in an Internet cafe in% >Amsterdam, down by Leidseplein (sp).:   Spelled just fine Terry ;-)r     --- |   / o / /_  _   		email: 	wilko@FreeBSD.org.1 |/|/ / / /(  (_)  Bulte		Arnhem, The Netherlands	c   ------------------------------  % Date: Mon, 01 Oct 2001 20:07:14 +0200m, From: Hans Bachner <Hans.Bachner@compaq.com>G Subject: Re: [VMScluster] Cross Architecture Boot with Cluster_Config ?h* Message-ID: <3BB8B0D2.9A0CF3A8@compaq.com>   Peter LANGSTOEGER wrote: >  <snip> > ... I only wanted to >  > a) have a clusterQ > b) use Alphas as bootserversJ > c) use an Alpha system disk for the bootservers and the Alpha satellites9 > d) use an served VAX system disk for the VAX satellitese >  > andk > P > e) use an Alpha bootserver to add another VAX satellite to the VAX system disk > ! > (a-d is in use here for years)     re: e)N Not quite, but close. I'm pretty sure (couldn't actually test it) that you canM invoke CLUSTER_CONFIG from a VAX satellite to add / remove other VAX nodes tohL the cluster, no need for a VAX bootserver. However, you must manually updateI the MOP or LANCP database on the Alpha bootserver to (no longer) boot thes added/removed node.Q  I I agree that it would be nice to have full cross architecture support foreN cluster management, I believe it is a fair assumption that as long as you haveM an Alpha bootserver booting VAX nodes that there's at least one VAX availabley for cluster management tasks.n   > and only maybe in addition > J > f) use a remote VAX or the Alpha Bootserver to do an OpenVMS VAX upgrade' >         on the served VAX system disk G > g) use the Alpha Bootserver to install OpenVMS VAX on a (served) disk   M Installation/upgrade does not work on MSCP served disks. The workaround still L is, depending on your disk types, to take the VAX system disk from the AlphaK bootserver's environment and plug it into the VAX, or image backup the booteN disk to a local VAX disk for the upgrade and afterwards restore the local disk to the bootdisk.   -- .. Hans Bachner, Compaq Austria / Global Services  --- speaking only for myself ---   ------------------------------   End of INFO-VAX 2001.547 ************************