1 INFO-VAX	Fri, 11 Jun 2004	Volume 2004 : Issue 322       Contents:* Re: =?iso-8859-1?Q?NEC=B4s?= Itanium Blade= Re: DCL enhancement - REQUEST/REPLY - how to capture a value.  DecServer password reset Re: DecServer password reset% Re: Experience with solid state disks % Re: Experience with solid state disks % Re: Experience with solid state disks % Re: Experience with solid state disks % Re: Experience with solid state disks % Re: Experience with solid state disks % Re: Experience with solid state disks  Re: few questions  Re: few questions 	 Re: Merci , Re: Opportunity for OVMS in mainstream press( Opportunity for OVMS in mainstream press, Re: Opportunity for OVMS in mainstream press0 Re: Quadrics adds lower-end supercomputer switch# Re: Start Ingres with VAX time back # Re: Start Ingres with VAX time back # Re: Start Ingres with VAX time back # Re: Start Ingres with VAX time back 3 Re: Sun To Open Source Solaris ... sort of ...maybe  Re: T4 and friends Re: T4 and friends Re: T4 and friends Re: VMS Media and DOCs for Free  Re: VMS Media and DOCs for Free  Re: VMS Media and DOCs for Free  Re: VMS Media and DOCs for Free  Re: VMS Media and DOCs for Free  What might have been RE: What might have been Re: What might have been Re: What might have been2 Re: Why does idle telnet session get disconnected?2 Re: Why does idle telnet session get disconnected?2 Re: Why does idle telnet session get disconnected?2 Re: Why does idle telnet session get disconnected?2 Re: Why does idle telnet session get disconnected?2 Re: Why does idle telnet session get disconnected?2 Re: Why does idle telnet session get disconnected?2 Re: Why does idle telnet session get disconnected?2 Re: Why does idle telnet session get disconnected?: [OpenVMS Alpha V7.3-2] How to keep OPA0 operator enabled ?# Re: [OpenVMS V7.3-2] Mount question J [TCPIP] How is the routing table updated when ROUTED or GATED aren't used?N Re: [TCPIP] How is the routing table updated when ROUTED or GATED aren't used?P Re: [TCPIP] How is the routing table updated when ROUTED or GATED aren't used? a  F ----------------------------------------------------------------------  % Date: Thu, 10 Jun 2004 20:45:33 -0500 2 From: David J Dachtera <djesys.nospam@comcast.net>3 Subject: Re: =?iso-8859-1?Q?NEC=B4s?= Itanium Blade + Message-ID: <40C90EBD.C7C15E44@comcast.net>    Fabio Cardoso wrote: > C > I think its time to liberate OpenVMS from the HP hardware only !   C Many of us out here agree. However, hp/VMS is still in total denial F about IA-32 (multi-hundred-billion dollar a year market, VMS's share =F $0), and everyday goes deeper into denial about x86-64, the soon-to-be+ defacto (read: "industry") 64-bit standard.    D.J.D.   P.S.: H I never belived I would hear myself say it, but perhaps total, obsessiveE greed at hp/VMS would be a good thing if it would get them over their  x86 psychosis.   ------------------------------  % Date: Thu, 10 Jun 2004 20:27:09 -0500 2 From: David J Dachtera <djesys.nospam@comcast.net>F Subject: Re: DCL enhancement - REQUEST/REPLY - how to capture a value.* Message-ID: <40C90A6D.4AD4669@comcast.net>   Dick Piccard wrote:  > G > If you go to google and search only in comp.os.vms for "answer.c" you @ > will quickly find Ehud Gavron's 1991 posting that provides theF > software you need (less than 250 lines of well-commented "C" source,F > plus .CLD, and .HLP).  It puts the reply value into a symbol of your	 > choice.  > > > We have been using it unchanged since 1991 to provide a moreF > interactive batch backup environment for our operators to use.  Yes,D > we did have to re-compile and re-link when we migrated from VAX to2 > Alpha, but no source code changes were required.  H If you would care to post an archive (saveset, .ZIP, etc.) with the textG files and VAX and ALpha .OBJs, that would be useful for folks who don't  have a C compiler.  A What we *REALLY* need a freeware C->Macro32 translator. Every VMS + distro. ever shipped (I think) has MACRO32.    D.J.D.   ------------------------------  % Date: Thu, 10 Jun 2004 23:22:52 -0400 0 From: "Glen Thompson" <glen.thompson@cox.netnet>! Subject: DecServer password reset . Message-ID: <jA9yc.13774$ye1.10519@lakeread05>  H I have a DecServer 700-16 that was purchased used and kept in storage asD a backup.  I tried to put it into service but the previous owner hadH changed the default password.  Checking on-line I found the instructionsB to reset it by holding in the reset button while powering on.  OneD manual said to hold the reset until E was shown on the indicator.  IC think I've tried every possible combination I could think of.  Even 7 tested the switch to verify that it was making contact.   B Anyone have any other ideas to try?  Unless I can get the password cleared, the device is useless.    thanks,  glen   ------------------------------  # Date: Fri, 11 Jun 2004 04:13:08 GMT % From: "Mike Naime" <mnaime@kc.rr.com> % Subject: Re: DecServer password reset 7 Message-ID: <ojayc.11165$Fd.5218@twister.rdc-kc.rr.com>   E Generally, pressing and holding the reset button for a 20 count while G powering on has worked for me on countless decserver 700's in the past. K This shoud wipe out everything and reset to factory defaults.  If the otehr G properties are not being reset, then maybe the reset button/function is  broken?    Mike  9 Glen Thompson <glen.thompson@cox.netnet> wrote in message ( news:jA9yc.13774$ye1.10519@lakeread05...J > I have a DecServer 700-16 that was purchased used and kept in storage asF > a backup.  I tried to put it into service but the previous owner hadJ > changed the default password.  Checking on-line I found the instructionsD > to reset it by holding in the reset button while powering on.  OneF > manual said to hold the reset until E was shown on the indicator.  IE > think I've tried every possible combination I could think of.  Even 9 > tested the switch to verify that it was making contact.  > D > Anyone have any other ideas to try?  Unless I can get the password! > cleared, the device is useless.  > 	 > thanks,  > glen >  >    ------------------------------    Date: 10 Jun 2004 17:24:28 -0700/ From: prosullivan@aol.com (prosullivan@aol.com) . Subject: Re: Experience with solid state disks= Message-ID: <a14b767a.0406101624.765f1106@posting.google.com>   N > > Is anyone using solid state disks in a VMS/Oracle environment?  If so, who >  is & > > the vender and how do you like it? > > # > > Any input would be appreciated.  > A ESE7x drives - very quick seek time, but battery-backed (10 years F life). Used to be a good idea, but will only fit into a SBB enclosure,@ (no EVA, but early HSG80 SBB will be OK), and, yuk, only -VW (40F megabit/sec SCSI ultra-wide transfer rate, compared with 2048 Mb todayB on EVA). Pick one up - they wobble internally when you carry them.D We tried one out for a pathworks share that was causing a problem :)   ------------------------------    Date: 10 Jun 2004 17:27:31 -0700/ From: prosullivan@aol.com (prosullivan@aol.com) . Subject: Re: Experience with solid state disks= Message-ID: <a14b767a.0406101627.5a9284bf@posting.google.com>   F But first, you need to quantify that your redo logs are the bottleneckB before throwing hardware at it. And unfortunately, the EVA doesn't have% any performance monitoring features.    F - Not quite: watch out for the SAN monitor which creates a process perE device for EVA's on VMS. For me, I use PerfCap PAWZ VMS collector and F capacity planner which does get me FC card usage, actual disk response> time, and logical/physical correlation on a per process basis.   ------------------------------  % Date: Thu, 10 Jun 2004 20:39:59 -0500 2 From: David J Dachtera <djesys.nospam@comcast.net>. Subject: Re: Experience with solid state disks+ Message-ID: <40C90D6F.14B32302@comcast.net>    David Harrold wrote: > N > We are looking at adding solid state disks to our environment.  Our DBAs areK > looking at using solid state disk for the redo logs to improve the Oracle  > performance. > M > Our current configuration is 3 GS1280s (16 CPUs and 64GB of memory), Oracle P > 8.1.7.3, 2 EVA5000s and VMS V7.3-1.  We will be upgrading to V7.3-2 in another
 > 2 weeks.  E E-mail me privately. We have a config. very similar to yours and ours E screams on ESA arrays. How to de-mung the reply-to should be obvious.    D.J.D. NMH    ------------------------------    Date: 10 Jun 2004 21:08:37 -0600+ From: young_r@encompasserve.org (Rob Young) . Subject: Re: Experience with solid state disks3 Message-ID: <XtoqBDiqAsGe@eisner.encompasserve.org>   c In article <5e9ec09li7fb67118h4gphj44bg9sf6318@4ax.com>, David Harrold <DHarrold@wi.rr.com> writes:  > N > We are looking at adding solid state disks to our environment.  Our DBAs areK > looking at using solid state disk for the redo logs to improve the Oracle  > performance. > M > Our current configuration is 3 GS1280s (16 CPUs and 64GB of memory), Oracle P > 8.1.7.3, 2 EVA5000s and VMS V7.3-1.  We will be upgrading to V7.3-2 in another
 > 2 weeks. > O > Is anyone using solid state disks in a VMS/Oracle environment?  If so, who is $ > the vender and how do you like it? > ! > Any input would be appreciated.  > 	 > Thanks,  >   ; 	Too bad you weren't talking about 2 nodes.  You could very B 	cheaply have a RamDisk on each node and shadow them.  I certainly? 	wouldn't recommend 3 RamDisks :-).  With 2 RamDisks, two nodes  	fly, one node doesn't.   ? 	Probably your best way to go (just as cheap or expensive - ha) B 	as solid state is quite expensive is to use PermaCache.  However,$ 	PermaCache isn't supported for VMS.  ) 	So you are looking at solid state, okay.     	With 64 GBytes of ram, how much> 	is that dedicated to Oracle SGA?  Enough?  Of course.  So youB 	are trying to speed up truly random reads as the EMC storage will> 	pre-fetch for you on sequential reads and catch those writes.  = 	So maybe you are bottlenecked on reads and seeing high queue A 	lengths?  If so, with 7.3-2 go to triple-shadowing those red-hot > 	disks across metas that are sharing different physical disks.  > 	I'm in rambling mood , but I'm curious as to what problem you@ 	are solving with a solid state disk?  Here is a thread from an + 	earlier post, re: solid state performance:   Z http://groups.google.com/groups?selm=YrbWQa1roZXz%40eisner.encompasserve.org&output=gplain   				Rob    ------------------------------    Date: 10 Jun 2004 22:08:10 -0600+ From: young_r@encompasserve.org (Rob Young) . Subject: Re: Experience with solid state disks3 Message-ID: <PnyrtZ6oLRWE@eisner.encompasserve.org>   a In article <XtoqBDiqAsGe@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes:     9 	Excuse my oversight... re-reading the post , I see where @ 	you say "redo logs."  You're talking about high write activity,< 	just make sure they are spread across enough physical disks9 	(i.e. metas not hypers).  As this link below points out:   [  http://groups.google.com/groups?selm=YrbWQa1roZXz%40eisner.encompasserve.org&output=gplain     =  Not versus a RamDisk, as Bill points out you have to trundle E  the filesystem then go out to a controller talking to an SSD.  Very  B  relevant, I am seeing writes take 1.2 milliseconds to controller 	  cache.     A 	I highly doubt you will be able to get better than 1 millisecond # 	on writes with a solid state disk.    				Rob     e > In article <5e9ec09li7fb67118h4gphj44bg9sf6318@4ax.com>, David Harrold <DHarrold@wi.rr.com> writes:  >>  O >> We are looking at adding solid state disks to our environment.  Our DBAs are L >> looking at using solid state disk for the redo logs to improve the Oracle >> performance.  >>  N >> Our current configuration is 3 GS1280s (16 CPUs and 64GB of memory), OracleQ >> 8.1.7.3, 2 EVA5000s and VMS V7.3-1.  We will be upgrading to V7.3-2 in another  >> 2 weeks.  >>  P >> Is anyone using solid state disks in a VMS/Oracle environment?  If so, who is% >> the vender and how do you like it?  >>  " >> Any input would be appreciated. >>  
 >> Thanks, >>   > = > 	Too bad you weren't talking about 2 nodes.  You could very D > 	cheaply have a RamDisk on each node and shadow them.  I certainlyA > 	wouldn't recommend 3 RamDisks :-).  With 2 RamDisks, two nodes  > 	fly, one node doesn't.  > A > 	Probably your best way to go (just as cheap or expensive - ha) D > 	as solid state is quite expensive is to use PermaCache.  However,& > 	PermaCache isn't supported for VMS. > + > 	So you are looking at solid state, okay.  > " > 	With 64 GBytes of ram, how much@ > 	is that dedicated to Oracle SGA?  Enough?  Of course.  So youD > 	are trying to speed up truly random reads as the EMC storage will@ > 	pre-fetch for you on sequential reads and catch those writes. > ? > 	So maybe you are bottlenecked on reads and seeing high queue C > 	lengths?  If so, with 7.3-2 go to triple-shadowing those red-hot @ > 	disks across metas that are sharing different physical disks. > @ > 	I'm in rambling mood , but I'm curious as to what problem youB > 	are solving with a solid state disk?  Here is a thread from an - > 	earlier post, re: solid state performance:  > \ > http://groups.google.com/groups?selm=YrbWQa1roZXz%40eisner.encompasserve.org&output=gplain > 	 > 				Rob  >    ------------------------------  % Date: Fri, 11 Jun 2004 00:18:31 -0400 * From: "Bill Todd" <billtodd@metrocast.net>. Subject: Re: Experience with solid state disks2 Message-ID: <YsedncUt0ff1r1TdRVn-sA@metrocast.net>  8 "Rob Young" <young_r@encompasserve.org> wrote in message- news:PnyrtZ6oLRWE@eisner.encompasserve.org... 5 > In article <XtoqBDiqAsGe@eisner.encompasserve.org>, - young_r@encompasserve.org (Rob Young) writes:  >  > : > Excuse my oversight... re-reading the post , I see whereA > you say "redo logs."  You're talking about high write activity,   F Not just 'high write activity', but high append-only (i.e., sequential6 write) activity, perhaps typically in smallish chunks.  = > just make sure they are spread across enough physical disks  > (i.e. metas not hypers).  I It's not clear that small sequential writes benefit noticeably from being H spread out over many disks, unless the striping is necessary because theG streaming bandwidth of a single disk is insufficient (i.e., log data is I generated at more than about 30 MB/sec, which is a fairly impressive clip K for simple transaction logs but which redo logs could exceed assuming heavy I insertion activity where all the new data winds up in the log - if that's " how Oracle handles its redo logs).  I Of course, if there are multiple such logs, you'd likely want a dedicated K mirrored disk pair for each of them (though if the aggregate bandwidth were H 20 MB/sec or less you still might be able to get away with a single pairG fronted by a moderately intelligent stable mirrored write-back cache)..   L By the way, you've made several EMC-flavored references, but the poster said he's running an EVA5000.   ...   B > I highly doubt you will be able to get better than 1 millisecond$ > on writes with a solid state disk.  K That could still offer several times the performance of a physical write to G a magnetic disk.  And I suspect that there are configurations (possibly K using fewer intermediaries than the one you're familiar with does) offering I somewhat lower latency than 1 ms.  ISTR reported one-way FC communication K latencies as good as around 10 us., for example - and there's really little K reason for a *great* deal more overhead than that in the processing at both  ends.    - bill   ------------------------------    Date: 10 Jun 2004 23:40:42 -0600+ From: young_r@encompasserve.org (Rob Young) . Subject: Re: Experience with solid state disks3 Message-ID: <i2fyF5+FQqAR@eisner.encompasserve.org>   _ In article <YsedncUt0ff1r1TdRVn-sA@metrocast.net>, "Bill Todd" <billtodd@metrocast.net> writes:  > : > "Rob Young" <young_r@encompasserve.org> wrote in message/ > news:PnyrtZ6oLRWE@eisner.encompasserve.org... 6 >> In article <XtoqBDiqAsGe@eisner.encompasserve.org>,/ > young_r@encompasserve.org (Rob Young) writes:  >> >>; >> Excuse my oversight... re-reading the post , I see where B >> you say "redo logs."  You're talking about high write activity, > H > Not just 'high write activity', but high append-only (i.e., sequential8 > write) activity, perhaps typically in smallish chunks. > > >> just make sure they are spread across enough physical disks >> (i.e. metas not hypers).  > K > It's not clear that small sequential writes benefit noticeably from being J > spread out over many disks, unless the striping is necessary because theI > streaming bandwidth of a single disk is insufficient (i.e., log data is K > generated at more than about 30 MB/sec, which is a fairly impressive clip M > for simple transaction logs but which redo logs could exceed assuming heavy K > insertion activity where all the new data winds up in the log - if that's $ > how Oracle handles its redo logs). > K > Of course, if there are multiple such logs, you'd likely want a dedicated M > mirrored disk pair for each of them (though if the aggregate bandwidth were J > 20 MB/sec or less you still might be able to get away with a single pairI > fronted by a moderately intelligent stable mirrored write-back cache)..  > N > By the way, you've made several EMC-flavored references, but the poster said > he's running an EVA5000. >   = 	Oh my.  Not only is my reading very poor on this one, it was  	at an all time low:  K looking at using solid state disk for the redo logs to improve the Oracle   
 performance.     P Our current configuration is 3 GS1280s (16 CPUs and 64GB of memory), Oracle     $ 8.1.7.3, 2 EVA5000s and VMS V7.3-1.   5 	I had it in my brain that EMC was involved.  Sheesh.    > ...  > C >> I highly doubt you will be able to get better than 1 millisecond % >> on writes with a solid state disk.  > M > That could still offer several times the performance of a physical write to I > a magnetic disk.  And I suspect that there are configurations (possibly M > using fewer intermediaries than the one you're familiar with does) offering K > somewhat lower latency than 1 ms.  ISTR reported one-way FC communication M > latencies as good as around 10 us., for example - and there's really little M > reason for a *great* deal more overhead than that in the processing at both  > ends.  >   A 	Yes - but.  With the EVA having a LUN across many disks, coupled ; 	with the fact the write has to go out and back and acked,  A 	realistically is there any OS doing writes under 1 ms to a solid  	state disk?  > 	I don't see how there can be an advantage or where the win is= 	at.  Random read IOs, okay.  But writes to write-back cached < 	controllers?  Maybe if as you allude to above, it is direct; 	attached - okay.  But 3 nodes are involved the solid state > 	disk has to hang off a switch so the write path would be HBA,0 	switch, SSD and back again.  1 ms or so, right?   				Rob    ------------------------------  % Date: Thu, 10 Jun 2004 11:09:45 -0700 + From: "Barry Treahy, Jr." <Treahy@MMaz.com>  Subject: Re: few questions' Message-ID: <40C8A3E9.7040001@MMaz.com>    Nom de Plume wrote:    >Bob Koehler wrote:  >    > 1 >>soccer13player@yahoo.com (Nom de Plume) writes:  >>     >>E >>>Again from a user/programmer perspective (the point that was being  >>>made), I feel that:	 >>>        >>> - >>   I think we'll have to agree to disagree.  >>     >> > E >Cool.  Even fellow VMS bigots, errr proponents, don't always have to  >agree.  ;)  > A >I kind of like to think of myself as a missionary converting the ? >unenlighted UNIX wanks, errr wonks, to VMS.  "Unix is an older C >operating system by about a decade", "Unix isn't as stable", "Unix F >doesn't have real clustering", "just about any tool available on UNIXC >is available on OpenVMS", and *"its easy to get around on OpenVMS, D >HELP is better than man and the written documentation is AWESOME."* >  >    > F and VMS isn't as prevalent!  Even with all of VMS's strengths, it has F lost traction in the market place.  As said countless times by myself D and others, the best product doesn't always out perform the better, C good, or terrible products in the market place...  Case and point,  
 Windows...     Barry    --    > Barry Treahy, Jr                       E-mail: Treahy@MMaz.com> Midwest Microwave                          Phone: 480/314-1320> Vice President & CIO                         FAX: 480/661-7028                            ------------------------------   Date: 10 Jun 2004 18:47:38 GMT, From: bill@gw5.cs.uofs.edu (Bill Gunshannon) Subject: Re: few questions* Message-ID: <2iromaFq79goU1@uni-berlin.de>  = In article <f401eb7f.0406100938.4006d899@posting.google.com>, 0 	soccer13player@yahoo.com (Nom de Plume) writes: > Bob Koehler wrote:1 >>soccer13player@yahoo.com (Nom de Plume) writes:  >> >  G >> > Again from a user/programmer perspective (the point that was being  >> > made), I feel that: >>  . >>    I think we'll have to agree to disagree. > F > Cool.  Even fellow VMS bigots, errr proponents, don't always have to > agree.  ;) > B > I kind of like to think of myself as a missionary converting the@ > unenlighted UNIX wanks, errr wonks, to VMS.  "Unix is an older' > operating system by about a decade",    C "Unix has about 10 more years of serious research and development."   ? >                                      "Unix isn't as stable",    @ Statements made here and personal experience do not support this> conjecture.  Under the right controlled circumstances both are? of equal stability.  I used to have Unix servers that stayed up ? for a year at a time and only went down then because we used to @ have an annual, scheduled, campus wide, 8-12 hour power shutdown> and my UPSes were not designed to keep things going that long.C By counter-point, the VMS machines here are often rebooted, usually 8 on a weekend, but sometime in the middle of the workday.  D >                                                              "Unix! > doesn't have real clustering",    B Don't have any real experience in this venue, but then it probably8 depends on the definition of the term "real clustering".  G >                                "just about any tool available on UNIX  > is available on OpenVMS",   @ Office Suite?  WYSIWYG Desktop Publishing? I could just start at@ the beginning of the FreeBSD Ports tree and run through them all? but why bother.  even the people here are always lamenting that A they need to port some application to VMS from Unix because there  is no VMS equivalent.   D >                           and *"its easy to get around on OpenVMS, > HELP is better than man    Matter of opinion.F >                          and the written documentation is AWESOME."*  D Also a matter of opinion.  I have always found the VMS documentationC to be cumbersome and difficult to find answers in.  Of course, with E Unix I can resort to any of a hundred third party books in just about A any style I might find acceptable.  Can the same be said for VMS?   C It's really time VMS bigots^H^H^H^H^H^Hproponents admited that both D VMS and Unix have their strengths and their weaknesses.  There is noB silver bullet for computing needs.  You need to pick the tool thatD does the job best in a particular environment. Unix isn't the answerB for everything and neither is VMS.  And sometimes, neither of themD is the right answer.  The universal is that Linux is never the right answer.  :-)   bill   --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------  % Date: Fri, 11 Jun 2004 04:23:45 +0800 , From: Paul Repacholi <prep@prep.synonet.com> Subject: Re: Merci- Message-ID: <87smd3novi.fsf@prep.synonet.com>   4 "Dave Weatherall" <djw-nothere@nospam.nohow> writes:  C >> And Poles, Czechs, Greeks, Belgians, Nowegians, Luxembourg. Heck  >> who have I left out?    > The Dutch.  2 They certainly had a very hard time of it in WWII.  0 The one also forgoten are the German Resistance.   --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076* comp.os.vms,- The Older, Grumpier Slashdot. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------  # Date: Thu, 10 Jun 2004 21:01:44 GMT - From: JF Mezei <jfmezei.spamnot@teksavvy.com> 5 Subject: Re: Opportunity for OVMS in mainstream press @ Message-ID: <bb412d240ac8adc5c1dedad2c316f895@news.teranews.com>   (email privately seperately)   Kenneth Farmer wrote: 8 > 2. What applications does OpenVMS tend to be used for?K accounting, web serving, SMTP server,  some office automation, development, ( network monitoring, FTP and many others.  F > 3. Why has it hung on so well despite so many reports of its demise?  L BECAUSE IT WORKS. And because it doesn't require much maintenance and is notT prone to viri or system directory corruption. Robust file system that you can trust.  8 > 4. Why is OpenVMS so well suited to disaster recovery?  H Because its disaster recovery features and clustering are built-into theR operating system instead of being products added as applications on top of the OS.  1 > 5. Anything else you wish to add about OpenVMS?   G People who understand VMS understand its technical superiority and also L understand that its lack of popularity is enturely the fault of its owner(s)K who refuse to pit this superior product against competitors' and their own.   G When you need a job done reliably and you have the software for it, the ( popularity of the OS isn't so important.   ------------------------------  # Date: Thu, 10 Jun 2004 20:01:38 GMT 6 From: "Kenneth Farmer" <kfarmer@NOSPAM.spyderbyte.com>1 Subject: Opportunity for OVMS in mainstream press > Message-ID: <C63yc.19792$wH4.1168205@twister.southeast.rr.com>  L OK folks...here's an opportunity OVMS in the press.  I've been approached byK a mainstream publisher/writer.  I could answer these questions but I'd like K to make this a community effort, strength in numbers and knowledge.  I need K OpenVMS customers to answer these questions.  From there I will compile the & answers with the help of a few others.   A few bullet points...  C - Please send your responses to me via email,  kfarmer/openvms.org. H - Be careful about what you say as far as sensitive info, company names,> etc.  But company names and their usage would be very helpful.H - I want to try and have something back to this guy by Monday/Tuesday of
 next week.    = 1. What stats do you have on OpenVMS usage in the enterprise?   6 2. What applications does OpenVMS tend to be used for?  D 3. Why has it hung on so well despite so many reports of its demise?  6 4. Why is OpenVMS so well suited to disaster recovery?  / 5. Anything else you wish to add about OpenVMS?      Ken    -- Kenneth R. Farmer  OpenVMS.org    ------------------------------  % Date: Thu, 10 Jun 2004 21:10:04 -0500 2 From: David J Dachtera <djesys.nospam@comcast.net>5 Subject: Re: Opportunity for OVMS in mainstream press + Message-ID: <40C9147C.7E2614A9@comcast.net>    Kenneth Farmer wrote:  > N > OK folks...here's an opportunity OVMS in the press.  I've been approached byM > a mainstream publisher/writer.  I could answer these questions but I'd like M > to make this a community effort, strength in numbers and knowledge.  I need M > OpenVMS customers to answer these questions.  From there I will compile the ( > answers with the help of a few others. >  > A few bullet points... > E > - Please send your responses to me via email,  kfarmer/openvms.org. J > - Be careful about what you say as far as sensitive info, company names,@ > etc.  But company names and their usage would be very helpful.J > - I want to try and have something back to this guy by Monday/Tuesday of > next week. > ? > 1. What stats do you have on OpenVMS usage in the enterprise?   E For reasons known only to hp/VMS, these stat's are kept very close to B the vest. In addition, hp/VMS refuses to promote its product. Here> again, the reason defies all logic and divination. Perhaps VMS= management has fallen victim to some form of group psychosis.   8 > 2. What applications does OpenVMS tend to be used for?  F General computing of yesteryear, though much of this has fallen victimF to the rumors of VMS's demise combined with Compaq's total destructionF of its own and VMS's credibility, which hp has not even endeavoured to
 recapture.  E Large-scale database servers such as back-end servers for large-scale A desktop applications in finance, healthcare and other industries.   ? Some places still use VMS workstations for graphic rendition of A extremely complex data, such as engineering and atmospheric data.   F > 3. Why has it hung on so well despite so many reports of its demise?  D It's secure - search the web for references. You'll find, not at theG least, mention of DEFCON where VMS was found to be so unhackable it was  not invited back.   H It's stable - uptime is measured with a calendar instead of a stopwatch.  E It's scalable - the same o.s. binaries are installed on workstations, E middle-tier servers and enterprise class systems. Only the tuning and  "layered software" differs.   A VMS invented clustering almost twenty years ago. Newer clustering 9 efforts are mere shadows of what VMS clustering provides. @ "Long-distance" clustering or "wide-area" clustering can provide@ redunancy and 24/7/365 availability over hundreds of kilometres.  8 > 4. Why is OpenVMS so well suited to disaster recovery?  A A. Because of the ease of restoration. Any "image" backups can be G restored on any other same-architecture system with suitably configured B storage. No "building the system before you can build the system".  % B. Because of "wide-area" clustering.   5 C. Because of host-based RAID-1 and other strategies.   1 > 5. Anything else you wish to add about OpenVMS?   D OpenVMS could be a very well positioned replacement for less secure,H less stable, less scalable, less "available" products. All it would takeG is an owner/vendor who will let it "be all that it can be", rather than F "hiding it's light under a bushel" ("BRRRRR!! It's lost something kind of crucial!" -Godspell).   -- David J Dachtera dba DJE Systems  http://www.djesys.com/  F Ken: I will e-mail this to you when it shows up in Google. There is no6 info. here which relates directly to my full-time gig.   ------------------------------    Date: 10 Jun 2004 16:56:06 -07001 From: keithparris_NOSPAM@yahoo.com (Keith Parris) 9 Subject: Re: Quadrics adds lower-end supercomputer switch = Message-ID: <cf15391e.0406101556.55bc3f19@posting.google.com>   r fabiopenvms@yahoo.com.br (Fabio Cardoso) wrote in message news:<f30679fb.0406030342.979440e@posting.google.com>...A > Anyone using this technology with OpenVMS systems /clustering ? ; > The news mention Itanic systems in a specific lab in USA.   E Quadrics, like Myrinet, is a proprietary interconnect popular in High C Performance Technical Computing (HPTC) environments, where a single A (typically numerical-calculation-intensive) problem is divided up A across hundreds or thousands of systems in a UNIX- or Linux-based  "scalability cluster".  ? While it would theoretically be possible to run OpenVMS Cluster @ traffic over such an interconnect, there are no plans to supportC another proprietary cluster interconnect, although investigation is ? underway about the possibility of using a new industry-standard  interconnect in the future.    ------------------------------    Date: 10 Jun 2004 17:31:12 -0700/ From: prosullivan@aol.com (prosullivan@aol.com) , Subject: Re: Start Ingres with VAX time back= Message-ID: <a14b767a.0406101631.69fe447b@posting.google.com>    > > J > > How can we check this type of problem( the validity period) so that ItH > > should not be possible to start Ingres by setting the VAX time back. > > 	 > > regds  > > Sukanta  > 6 Hey, don't worry because  Ingres is now Open-Source...L http://news.com.com/CA+open-sources+Ingres+database/2100-7344_3-5219373.html   ------------------------------  # Date: Fri, 11 Jun 2004 01:37:33 GMT " From: "Tom M" <kryios@comcast.net>, Subject: Re: Start Ingres with VAX time back, Message-ID: <x18yc.78282$Ly.30994@attbi_s01>  & <prosullivan@aol.com> wrote in message7 news:a14b767a.0406101631.69fe447b@posting.google.com...  > > > L > > > How can we check this type of problem( the validity period) so that ItJ > > > should not be possible to start Ingres by setting the VAX time back. > > >  > > > regds 
 > > > Sukanta  > > 8 > Hey, don't worry because  Ingres is now Open-Source... > L http://news.com.com/CA+open-sources+Ingres+database/2100-7344_3-5219373.html  K I don't believe the version of Ingres being released as opensource will run L on a VAX.  Newer versions of Ingres transitioned to IEEE format floats and ID believe that limits new VMS releases to Alpha and Itanium platforms.   ------------------------------   Date: 11 Jun 2004 02:11:54 GMT From: healyzh@aracnet.com , Subject: Re: Start Ingres with VAX time back, Message-ID: <cab4da01976@enews2.newsguy.com>  ! Tom M <kryios@comcast.net> wrote: M > I don't believe the version of Ingres being released as opensource will run N > on a VAX.  Newer versions of Ingres transitioned to IEEE format floats and IF > believe that limits new VMS releases to Alpha and Itanium platforms.  J Which brings up the question, as a Hobbyist, does it being Opensource meanK that I can now get the source and build it on OpenVMS?  Or is this only the * source for something like UNIX or Windows?   		Zane   ------------------------------  # Date: Fri, 11 Jun 2004 03:57:22 GMT " From: "Tom M" <kryios@comcast.net>, Subject: Re: Start Ingres with VAX time back+ Message-ID: <C4ayc.7625$2i5.7193@attbi_s52>   & <healyzh@aracnet.com> wrote in message& news:cab4da01976@enews2.newsguy.com...# > Tom M <kryios@comcast.net> wrote: K > > I don't believe the version of Ingres being released as opensource will  run J > > on a VAX.  Newer versions of Ingres transitioned to IEEE format floats and I H > > believe that limits new VMS releases to Alpha and Itanium platforms. > L > Which brings up the question, as a Hobbyist, does it being Opensource meanI > that I can now get the source and build it on OpenVMS?  Or is this only  the , > source for something like UNIX or Windows? >  > Zane  L I don't know exactly what is going to be released in source form.  The pressL release said version 3 (which is beyond the currently released version), butK it didn't say which pieces are included (i.e. precompilers,  report writer,  etc).   J I think most of the source is pretty common between platforms so I imagineL the OpenVMS source will also be available.  It will probably be a few months before the source is available.    ------------------------------  % Date: Thu, 10 Jun 2004 22:11:18 -0400 * From: "Bill Todd" <billtodd@metrocast.net>< Subject: Re: Sun To Open Source Solaris ... sort of ...maybe2 Message-ID: <kZadnZ2aXpsKiVTdRVn-sA@metrocast.net>  K "Andrew Harrison SUNUK Consultancy" <Andrew_No.Harrison_No@nospamn.sun.com> ; wrote in message news:ca7h18$13m$1@new-usenet.uk.sun.com...  > Bill Todd wrote:J > > "Andrew Harrison" <andrew_._remove_harrison@su_n.com> wrote in message. > > news:ca1f2f$gth$1@new-usenet.uk.sun.com... > >  > >>Bill Todd wrote: > >> > >>3 > >>>"John Smith" <a@nonymous.com> wrote in message + > >>>news:pemdnQneyfJ_bV3dRVn-jg@igs.net...  > >>>  > >>>  > >  > > L http://story.news.yahoo.com/news?tmpl=story&ncid=1209&e=5&u=/nf/24366&sid=95 > >  > >>>>573734 > >>>  > >>> L > >>>That's nice, I guess.  But I'm more interested in the new 'dynamic fileK > >>>system' that they just announced, which at least claims to support the  > > 	 > > kinds  > > A > >>>of features I've been suggesting for years:  anyone seen any  substantive ; > >>>information on it yet (I didn't find any more than the A > >>>press-release/high-level-manager comments stuff at sun.com)?  > >>>  > >>C > >>Access to Solaris Dynamic Filesystem AKA ZFS can be had through A > >>the Software Express Program which gives people with existing ? > >>Solaris licenses access to the latest builds of Solaris 10.  > >  > > J > > What I'm looking for is access to *information* about the Dynamic FileJ > > System.  I did hear mentioned somewhere that it was a rebadged/updatedI > > version of a previous Sun file system, but don't know if that's true.  > >  > E > I will see if I can get you some technical documentation that isn't  > covered by an NDA.  K That would be great:  the little that's been said suggests some significant * features that I'd like to know more about.   - bill   ------------------------------  % Date: Thu, 10 Jun 2004 19:57:23 -0500 2 From: David J Dachtera <djesys.nospam@comcast.net> Subject: Re: T4 and friends + Message-ID: <40C90372.CB4B7066@comcast.net>    Michael Austin wrote:  >  > David J Dachtera wrote:  >  > <snip>E > > Well, out here in the real world, hand-outs at meetings are still  > > more-or-less mandatory.  > I > This depends on your pointy-haired managers... :)  *I* have not been in I > a performance meeting in years where a printed graph was mandatory.  We I > were usually in a room with multiple wall-projectors and if a graph was I > needed, it was point-click and on the screen -- next to the screen with A > a mon proc/top{whatever} running.  I guess some of us have been  > spoiled... :)  > 4 > > And for for past reference, managers like to seeC > > hard copy with the original dates on them (saves arguments over H > > authenticity), not the date the graph was most recently committed to > > hard-copy. > " > For starters... go to my websiteI > << http://www.firstdbasource.com/t4/t4chart.php >> and select the first I > file T4_ALPHA1_01APR2004_0000_2359.ZIP;1 and click the radio button for  > [MON.SYST] and hit submit. > F > You will note in the bottom right-hand corner of the each graph, the0 > date. Is it the printed date or the data date?  F Aren't you forgetting something? Print out a web page, and examine the; bordering margins. That question should then answer itself.   I > Having worked in a shop with more than 200 systems, I used the T4 tools I > to make my life easier without having to sort through 200+ graphs a day H > (for each parameter).  When working in an extremely large environment,J > that becomes unmanagable.  That shop initially used the CA {PCSA|VPM} orF > whatever it was called and WAS printing those histograms every day..F > Once we reached 75 or 80 systems, that became too many and the printD > portion was removed and a web-page was constructed to view them asB > necessary.  I also constructed DCL scripts to extract only dailyF > averages and could graph them using some point-and-click magic in anD > Excel spreadsheet.  My pointy-haired managers didn't care what theF > "print" date was, they were interested and were given the data date.; > Dave, you have got to train your managers better :) :) :)  > H > Not only did I have access to this site, but everyone with performanceF > responsibility did -- including the managers and application supportH > personnel.  If they wanted to see how a certain system was performing,I > point-click -bada-bing-bada-boom he had the data.  Sometimes, given the H > "new" bread of pointy-haired managers, this is much easier than havingN > to figure out who's desk or on what floor is that darn graph filing cabinet.  ; Store them on disk until you need to produce the hard-copy.   J > I don't believe I said that your methods were off-base.  There are timesG > when automating those print-outs can be useful.  I was merely showing H > you the possibilities. Can the current tools be adapted to do what youG > want.. Hey, with a bit of programming and $$$$, anything is possible. I > But as I also stated that takes time and there are those of us that are / > working on it.  It takes time -- and money...   D Unless you are acquiring equipment/software you don't curently have,A where's the monetary expense? (I know - it should be obvious, but  examine it anyway).   I > While I am taking the tools provided by HP and beginning to use them in J > ways thought not possible and writing the wrappers that make them usefulF > and databases to store the data such that data mining can occur moreE > easily, I currently don't have anyone funding my efforts.  Building + > tools is great, but then, so is eating...   C Say: Marketing. Promotion and distribution (two concepts currentl;y 2 beyond hp/VMS) are how you recoup your investment.  D > Bottom line: all of this is possible. You had said this is a greatG > business opportunity... let me know when you are ready to sign on the G > dotted line - because I have a great deal of it almost ready - except B > for the impact printer request... they can never do a good graph > justice.  :) :)   ; Again, my point is: people don't have all day to sit around F pointing-and-clicking their way to graphs that the machine can produce@ in seconds. It's almost like we're having to re-invent the wholeG computer wheel all over again: the tools have become so onerous that we : need autoates to manipulate the tools in a time-efficient, cost-effective way.    > <snipped> G > > See, the WhineBloze way has become so engrained, that we can't even J > > argue against the urge to have people playing point-and-click all day. > >  > I > Well, the PC and WhineBloze is just a tool that happens to sit on every I > desk in the world (well, mostly).  Everyone wanted n-tier architecture, G > now they have it.  I merely exploit it to use applications that don't I > exist on VMS.  Yes, it's sad, but until that changes, I will exploit it # > to accomplish my tasks and goals.   H O.k. That's fine. Now, finish the job: implement RPCs in such a way thatH the whole PC-based process can be initiated from a VMS batch job or UN*XC cron job, from downloading thne MONITOR data all the way through to  producing the hard-copy graphs.   H ...or, if you're *REALLY* __R_E_A_L_L_Y__ dead-set against printing themD automagically, unattended, then at least cook up a dialogue/list boxG where the user can print a pre-produced graph at their own convenience. A Just select the graph from a drop down list of graphs plotted and & stored, then send them to the printer.  H (SHEEEEEEZZZ!!! I REALLY can't belive that BG has so corrupted the worldA that we have to learn EDP all over again, just to be productive!)   	 >  > Sad.  > G > True... but no one ever said the best technologies will always win...   8 They don't have to "win" - place or show will do nicely!   D.J.D.   ------------------------------  % Date: Thu, 10 Jun 2004 19:59:50 -0500 2 From: David J Dachtera <djesys.nospam@comcast.net> Subject: Re: T4 and friends + Message-ID: <40C90405.C52A2542@comcast.net>    JF Mezei wrote:  >  > David J Dachtera wrote: H > > One key item to remember: in order to be truly valuable, performanceL > > reporting solutions *MUST* be 100% automated. They *MUST* be able to run6 > > in batch, unattended, scheduled by whatever means. > > > The collection and perhaps some processing of data should beG > "background/automated. But it should be viewable on-line very easily.   5 Likewise, hard-copy should be one command/click away.   N > But accessing graphs and displays should be available interactively at will.M > It is rather pointless to automatically print reports that you may not use.   B Depends. Businesses do a lot of pointless computing everyday. But,& that's just the way of thw world now:    Q: Is it productive?0 A: Who cares?! ...as long as it's "easy to use".  P > Much better to have a list of available reports on-line, and view the ones youC > wish, and possibly print the one or two you really need on paper.   ) Exactly what I just suggested to Michael.    D.J.D.   ------------------------------  % Date: Thu, 10 Jun 2004 20:23:39 -0500 2 From: David J Dachtera <djesys.nospam@comcast.net> Subject: Re: T4 and friends + Message-ID: <40C9099B.3C5CCCA0@comcast.net>    Paul Sture wrote:  >  > JF Mezei wrote:  > > David J Dachtera wrote:  > > H > >>One key item to remember: in order to be truly valuable, performanceL > >>reporting solutions *MUST* be 100% automated. They *MUST* be able to run6 > >>in batch, unattended, scheduled by whatever means. > >  > > @ > > The collection and perhaps some processing of data should beI > > "background/automated. But it should be viewable on-line very easily.  > > P > > But accessing graphs and displays should be available interactively at will.O > > It is rather pointless to automatically print reports that you may not use. R > > Much better to have a list of available reports on-line, and view the ones youE > > wish, and possibly print the one or two you really need on paper.  > J > Moving the emphasis away from performance reporting to reporting for the > whole business here... > J > I came across the following product nearly 10 years ago. The idea behindJ > it then was that you archive your ASCII spoolfiles to optical disk (COLDG > - Computer Outper to Laser Disk), and then you can not only call them H > back for auditing purposes, but can reconstruct data long since purgedJ > from databases, or simply analyze it. Today's large disk capacities meanJ > you don't need the often expensive hardware/software solutions that were > on offer back then of course.  > K > http://www.datawatch.com/dataconversionsoftware/monarch-report-mining.asp   F Yeah. Saw that along time ago, too. Useful if you'cant access the data7 files any other way, but still a point-and-click thing.    Need automation...   D.J.D.   ------------------------------  # Date: Thu, 10 Jun 2004 18:35:17 GMT - From: JF Mezei <jfmezei.spamnot@teksavvy.com> ( Subject: Re: VMS Media and DOCs for Free@ Message-ID: <7c90547ce9f7caf7a1c696702dfb5bd4@news.teranews.com>   Alan Boyles wrote: > L > I have a whole bunch of distribution kits of VMS, VMS layered products andJ > documentation dating back to 1998 which I'd be willing to let go for the' > huge price of  0$.  You pay shipping.    Where are you located ?    ------------------------------  % Date: Thu, 10 Jun 2004 17:11:23 -0400 0 From: "Alan Boyles" <alan.boyles@mindspring.com>( Subject: Re: VMS Media and DOCs for Free0 Message-ID: <10chjerjb3gled5@corp.supernews.com>   I am in Atlanta, Ga. USA.   K I have received more responses than I have media so I won't be able to send * stuff to everyone.  Just out of curiosity,J would having this stuff available on the web be better than trying to send$ bits and pieces to different people?K BTW I'm keeping the new stuff for my alphastation hobbyist system, so  this . stuff will be older ( mid Q3 2003 and before).   Alan  : "JF Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message: news:7c90547ce9f7caf7a1c696702dfb5bd4@news.teranews.com... > Alan Boyles wrote: > > J > > I have a whole bunch of distribution kits of VMS, VMS layered products and L > > documentation dating back to 1998 which I'd be willing to let go for the) > > huge price of  0$.  You pay shipping.  >  > Where are you located ?    ------------------------------  # Date: Thu, 10 Jun 2004 21:24:48 GMT 1 From: Michael Austin <maustin@firstdbasource.com> ( Subject: Re: VMS Media and DOCs for Free< Message-ID: <Ak4yc.5416$Ck7.2120@newssvr23.news.prodigy.com>   Alan Boyles wrote:   > I am in Atlanta, Ga. USA.  > M > I have received more responses than I have media so I won't be able to send , > stuff to everyone.  Just out of curiosity,L > would having this stuff available on the web be better than trying to send& > bits and pieces to different people?  H HP really doesn't like/ won't allow them to be made "freely" available. C      They make a boat-load of $$$ by selling the ConDist.  I think  9 someone recently wanted to do that and was told no. IIRC.    Michael.   ------------------------------  % Date: Thu, 10 Jun 2004 17:41:14 -0400 0 From: "Alan Boyles" <alan.boyles@mindspring.com>( Subject: Re: VMS Media and DOCs for Free0 Message-ID: <10chl6ot2g34ed6@corp.supernews.com>  L You think they care about the old stuff?  I doubt you could get that pressed again anyway, even from HP.    Alan  > "Michael Austin" <maustin@firstdbasource.com> wrote in message6 news:Ak4yc.5416$Ck7.2120@newssvr23.news.prodigy.com... > Alan Boyles wrote: >  > > I am in Atlanta, Ga. USA.  > > J > > I have received more responses than I have media so I won't be able to send. > > stuff to everyone.  Just out of curiosity,I > > would having this stuff available on the web be better than trying to  send( > > bits and pieces to different people? > I > HP really doesn't like/ won't allow them to be made "freely" available. D >      They make a boat-load of $$$ by selling the ConDist.  I think; > someone recently wanted to do that and was told no. IIRC.  > 
 > Michael. >    ------------------------------   Date: 11 Jun 2004 00:11:49 GMT From: healyzh@aracnet.com ( Subject: Re: VMS Media and DOCs for Free+ Message-ID: <caatc50oj1@enews1.newsguy.com>   / Alan Boyles <alan.boyles@mindspring.com> wrote: N > You think they care about the old stuff?  I doubt you could get that pressed > again anyway, even from HP.    > Alan  H I know they care about the old stuff.  Also, the older ConDists have theL advantage of having certain layered products that are no longer available onF the new ones.  The fact that they've so far left the Hobbyist programs intact rather surprises me.    		Zane   ------------------------------  % Date: Thu, 10 Jun 2004 22:23:57 -0400 * From: "Bill Todd" <billtodd@metrocast.net> Subject: What might have been 2 Message-ID: <tr6dnV9DgqgTilTdRVn-jA@metrocast.net>  6 http://www.theregister.co.uk/2004/06/10/ibm_iseries_5/  G Sounds not all that different from what we urged Curly to do with VMS 4 L years ago.  Of course, it was less time-critical for the i-series, since IBMF hadn't so flagrantly neglected it.  But this still shows that 'legacy'J environments, if good enough, still have the potential not only to prosper but to *grow*.   - bill   ------------------------------  % Date: Thu, 10 Jun 2004 19:48:39 -0700 # From: "Tom Linden" <tom@kednos.com> ! Subject: RE: What might have been 9 Message-ID: <NDEMLKKEBOIFBMJLCECIKEHIDGAA.tom@kednos.com>   B You will note that the OS400 runs on an emulator on the underlyingF cpu, thus maintaining compatibility with the past.  You could probablyD still run System 3 code on this platform.  Not offering this kind ofE compatibility was the most serious mistake Digital made (note that it D wasn't Compaq or HP, not that they wouldn't also have screwed it up)         -----Original Message-----1   From: Bill Todd [mailto:billtodd@metrocast.net] '   Sent: Thursday, June 10, 2004 7:24 PM    To: Info-VAX@Mvb.Saic.Com    Subject: What might have been     8   http://www.theregister.co.uk/2004/06/10/ibm_iseries_5/  I   Sounds not all that different from what we urged Curly to do with VMS 4 :   years ago.  Of course, it was less time-critical for the   i-series, since IBM H   hadn't so flagrantly neglected it.  But this still shows that 'legacy'L   environments, if good enough, still have the potential not only to prosper   but to *grow*.     - bill         --- (   Incoming mail is certified Virus Free.<   Checked by AVG anti-virus system (http://www.grisoft.com).A   Version: 6.0.701 / Virus Database: 458 - Release Date: 6/7/2004    --- & Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).? Version: 6.0.701 / Virus Database: 458 - Release Date: 6/7/2004    ------------------------------  % Date: Thu, 10 Jun 2004 22:56:42 -0400 # From: "John Smith" <a@nonymous.com> ! Subject: Re: What might have been , Message-ID: <_qSdnZjsraNJvFTdRVn-gQ@igs.net>  5 "Bill Todd" <billtodd@metrocast.net> wrote in message , news:tr6dnV9DgqgTilTdRVn-jA@metrocast.net...8 > http://www.theregister.co.uk/2004/06/10/ibm_iseries_5/ > I > Sounds not all that different from what we urged Curly to do with VMS 4 J > years ago.  Of course, it was less time-critical for the i-series, since IBM H > hadn't so flagrantly neglected it.  But this still shows that 'legacy'L > environments, if good enough, still have the potential not only to prosper > but to *grow*.    H OS400 or whatever you want to call it still has *thousands* of supported4 applications from many hundreds (if not more) ISV's.  C Your final exam assignment for today is to contrast and compare the H foregoing with the state of the ISV market and applications for OpenVMS.  L You have 5 minutes to complete your answer. Quoting HP officials will result in a failing grade.    ------------------------------  % Date: Thu, 10 Jun 2004 23:57:26 -0400 * From: "Bill Todd" <billtodd@metrocast.net>! Subject: Re: What might have been 2 Message-ID: <EMudnYQuzJXqsFTd4p2dnA@metrocast.net>  . "Tom Linden" <tom@kednos.com> wrote in message3 news:NDEMLKKEBOIFBMJLCECIKEHIDGAA.tom@kednos.com... D > You will note that the OS400 runs on an emulator on the underlyingH > cpu, thus maintaining compatibility with the past.  You could probablyF > still run System 3 code on this platform.  Not offering this kind ofG > compatibility was the most serious mistake Digital made (note that it F > wasn't Compaq or HP, not that they wouldn't also have screwed it up)  L While that observation may have some validity (though I'd personally suggestL that Digital made significantly more serious mistakes after that point, suchJ as defining NT as the future and urging VMS customers to move there), it'sJ not relevant to the comment I made, which involved the ability to energizeD VMS sales in the Y2K time-frame (regardless of all that had happened previously).  L Of course, when DEC created VAX, it included the ability to run a great dealL of PDP-11 application code (for example, IIRC VAX used PDP-11 Datatrieve forG several years until DTR-32 was finished, and for that matter had to use J RMS-11 to support indexed files in VMS V1, since RMS-32 did not).  And theL move to Alpha included the VEST mechanism that allowed reasonable amounts ofK VAX code to execute in the new environment.  While these facilities may not K have been adequate for everyone, they seem to have helped sufficiently that L I'd question your characterization of the problem's severity (without in any/ way suggesting that no problem existed at all).    - bill   ------------------------------  # Date: Thu, 10 Jun 2004 18:38:57 GMT - From: JF Mezei <jfmezei.spamnot@teksavvy.com> ; Subject: Re: Why does idle telnet session get disconnected? @ Message-ID: <1d7c02044af798352c0be481af24e042@news.teranews.com>   Lawrence Bleau wrote: 8 > >    $ sysconfig -r inet tcp_keepidle 150  ! 75 secs.   F > This looks promising.  I can't change it just now, will have to waitD > until later to test it out, and will let the group know next week.  L sysconfig and if config are dynamic and affect the "live" system. No need toL shut down the IP stack. In fact, if the above ends up working, you will needT to add this command to your systems startup right after the TCPIP stack has started.   ------------------------------  + Date: Thu, 10 Jun 2004 20:19:04 +0000 (UTC) * From: bleau@UMTOF.UMD.EDU (Lawrence Bleau); Subject: Re: Why does idle telnet session get disconnected? 0 Message-ID: <caafno$1r8$1@grapevine.wam.umd.edu>  p In article <Fj8xc.17639$rz4.9118@news-server.bigpond.net.au>, "Matt Muggeridge" <Matt.Muggeridge@hp.com> writes:J >Quite possibly this is due to a change in the tcp_keepidle time which wasK >made to be RFC compliant.  I think, in UCX V4.2 the default value was 75s, G >whereas the RFC required a default of 2 hours.  IIRC, this change took G >effect as of V5.3.  In any case, your VPN is likely not seeing traffic . >within its idle timeout and so disconnecting. > J >You can try adjusting the VPN idle timeout to be greater than 2 hours, or >adjust tcp_keepidle with: > ) >    $ @sys$manager:tcpip$define_commands I >    $ sysconfig -r inet tcp_keepidle 150  ! 75 secs.  (I know some use a   >value of 600 = 5 minute polls).  9 This doesn't work, Matt.  I tried the following commands:   $ $ @sys$manager:tcpip$define_commands$ $ sysconfig -r inet tcp_keepidle 600 600: unknown attribute   Just a sysconfig gives: r usage: sysconfig [-v] [-i <index>] -c | -s | -m | -q | -Q | -d | -o | -r | -u [<subsys>] [<attributes> | <opcode>]  @ There's probably a simple syntax error in this, but I can't findF online documentation on sysconfig.  Can you - or someone else in c.o.v - help out here?  D Here's a theory we developed today, which maybe farfetched, but does? explain the observed data.  To recap, we observed the following  disconnection pattern:  4 System  VMS     TCPIP           Disconnects the boss# UMTOF   V7.3-2  V5.4 - ECO 1    YES # UMTOF2  V7.3-2  V5.4 - ECO 1    YES " UMACE   V7.1-2  V4.2 - ECO 1    NO" UMSMS   V7.1-2  V4.2            NO   Suppose the following:A 1) V4.2 of TCPIP (UCX) doesn't implement keepalive probes for its     TELNET server, . 2) TCPIP V5.4 does implement keepalive probes,
 3) either:E   a) something prevents the keepalive probes from reaching the boss's       computer, or @   b) something prevents responses to keepalive probes from being      received by UMTOF, or=   b) the boss's tcpip package doesn't implement responding to       keepalive probes.  C The TELNET server on, say, UMACE would never send a keepalive probe D packet, but would also never terminate the tcp connection because it= is idle.  The TELNET server on UMTOF, however, would send the E keepalive probe, nver receive a response to it, decide the connection  is dead, and disconenct it.   C This supposes a lot, but as I said, does explain the data.  Is this F theory one we should follow up?  Can anyone debunk it definitively, orF provide evidence pro or con?  For example, a definitive statement thatF UCX V4.2 telnet server does use keepalive probes would blow this away.  ? Btw, is there a way to determine what the default keepalive (or + keepidle) setting is, or even if it's used?    Lawrence Bleau University of Maryland" Physics Dept., Space Physics Group 301-405-6223 bleau@umtof.umd.edu    ------------------------------  # Date: Thu, 10 Jun 2004 20:56:10 GMT - From: JF Mezei <jfmezei.spamnot@teksavvy.com> ; Subject: Re: Why does idle telnet session get disconnected? @ Message-ID: <68e68f343444ac1468927ce14e2a45ad@news.teranews.com>   Lawrence Bleau wrote: & > $ sysconfig -r inet tcp_keepidle 600 > 600: unknown attribute   documentation available in:    $TCPIP HELP SYSCONFIG      Also, you can :   ) sysconfig -m    to display the subsystems D sysconfig -q inet to display all parameter for the "inet" subsystem.  A Summary of differences between TCPIP Services 5.1 and 5.3-2 (VAX) = (command used to generated compared output: SYSCONFIG -q inet     4 *** 5.1 ***                            *** 5.3-2 ***  >                                        icmp_rejectcodemask = 0B ipport_userreserved = 5000             ipport_userreserved = 65535F                                        ipport_userreserved_min = 491527                                        ipqmaxlen = 1024 /                                        ipqs = 1  route_round_robin = 0  ipv6router = 0 ipv6forwarding = 0 ipv6_tunnel_default_mtu = 1280 ipv4_tunnel_default_mtu = 12805                                        tcbhashnum = 1 8 tcbhashsize = 32                       tcbhashsize = 5129 tcbquicklisten = 0                     tcbquicklisten = 1 @ tcp_keepalive_default = 0              tcp_keepalive_default = 06 tcp_keepcnt = 8                        tcp_keepcnt = 8; tcp_keepidle = 150                     tcp_keepidle = 14400 9 tcp_keepinit = 150                     tcp_keepinit = 150 : tcp_keepintvl = 150                    tcp_keepintvl = 150< tcp_recvspace = 32768                  tcp_recvspace = 614409                                        tcp_rexmtmax = 128 9 tcprexmtthresh = 3                     tcprexmtthresh = 3 < tcp_sendspace = 32768                  tcp_sendspace = 61440< udp_recvspace = 41920                  udp_recvspace = 42080@                                        ovms_nobroadcastcheck = 0?                                        ovms_printf_to_opcom = 1   I Blank lines indicate parameter not present in this version but present in  other version.  E Parameters with same values were removed to simplify list, except for  keep* parameters   ------------------------------  # Date: Thu, 10 Jun 2004 21:18:18 GMT - From: JF Mezei <jfmezei.spamnot@teksavvy.com> ; Subject: Re: Why does idle telnet session get disconnected? @ Message-ID: <929b3d834e6b0979601d1e3da75cfeba@news.teranews.com>   Matt Muggeridge wrote:N > And just so my typo doesn't get propagated too far... there is a missing "=" > sign.  It should have been:  > 7 >      $ sysconfig -r inet tcp_keepidle=150  ! 75 secs.   E Is there good documentation somewhere on the impact of each and every , parameter in sysconfig as well as ifconfig ?  H And as far as your typo is concerned, we understand. Out on the beach inM winter, your fingers must have been shivering because the temperature dropped  below 30C :-) :-) :-) :-)   ------------------------------  # Date: Thu, 10 Jun 2004 21:05:04 GMT 0 From: "Matt Muggeridge" <Matt.Muggeridge@hp.com>; Subject: Re: Why does idle telnet session get disconnected? ; Message-ID: <424yc.1516$sj4.403@news-server.bigpond.net.au>   K > sysconfig and if config are dynamic and affect the "live" system. No need  to > shut down the IP stack.   I They are dynamic in the sense they affect the live system.  However after K making the change, only new connections are affected.  So in problem cases, : I recommend that the service should at least be restarted.  5 >In fact, if the above ends up working, you will need I > to add this command to your systems startup right after the TCPIP stack  has started.  1 Personally, I prefer to keep sysconfig changes in  TCPIP$ETC:SYSCONFIGTAB.DAT.     9 > > >    $ sysconfig -r inet tcp_keepidle 150  ! 75 secs.   L And just so my typo doesn't get propagated too far... there is a missing "=" sign.  It should have been:   5      $ sysconfig -r inet tcp_keepidle=150  ! 75 secs.    Matt.    --  = -------------------------------------------------------------  OpenVMS TCP/IP Engineering Enterprise Computing Group Hewlett-Packard Company  Gold Coast, AUSTRALIA = -------------------------------------------------------------     : "JF Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message: news:1d7c02044af798352c0be481af24e042@news.teranews.com... > Lawrence Bleau wrote: 9 > > >    $ sysconfig -r inet tcp_keepidle 150  ! 75 secs.  > H > > This looks promising.  I can't change it just now, will have to waitF > > until later to test it out, and will let the group know next week. > K > sysconfig and if config are dynamic and affect the "live" system. No need  toI > shut down the IP stack. In fact, if the above ends up working, you will  needI > to add this command to your systems startup right after the TCPIP stack  has started.   ------------------------------  # Date: Thu, 10 Jun 2004 21:19:39 GMT 0 From: "Matt Muggeridge" <Matt.Muggeridge@hp.com>; Subject: Re: Why does idle telnet session get disconnected? ; Message-ID: <Lf4yc.1523$sj4.293@news-server.bigpond.net.au>   ; > This doesn't work, Matt.  I tried the following commands:  > & > $ @sys$manager:tcpip$define_commands& > $ sysconfig -r inet tcp_keepidle 600 > 600: unknown attribute >   . Sorry, I was missing an "=" sign.  Try this...  (     $ sysconfig -r inet tcp_keepidle=600  4 You can get online help with:  TCPIP HELP SYSCONFIG.  L Though, I could not find a good explanataion of the "-r" flag that shows theL correct syntax.  I'll take that back to our doc writer so it can be fixed in a future release.    Matt.    --  = -------------------------------------------------------------  OpenVMS TCP/IP Engineering Enterprise Computing Group Hewlett-Packard Company  Gold Coast, AUSTRALIA = -------------------------------------------------------------     7 "Lawrence Bleau" <bleau@UMTOF.UMD.EDU> wrote in message * news:caafno$1r8$1@grapevine.wam.umd.edu...E > In article <Fj8xc.17639$rz4.9118@news-server.bigpond.net.au>, "Matt , Muggeridge" <Matt.Muggeridge@hp.com> writes:L > >Quite possibly this is due to a change in the tcp_keepidle time which wasH > >made to be RFC compliant.  I think, in UCX V4.2 the default value was 75s,I > >whereas the RFC required a default of 2 hours.  IIRC, this change took I > >effect as of V5.3.  In any case, your VPN is likely not seeing traffic 0 > >within its idle timeout and so disconnecting. > > L > >You can try adjusting the VPN idle timeout to be greater than 2 hours, or > >adjust tcp_keepidle with: > > + > >    $ @sys$manager:tcpip$define_commands K > >    $ sysconfig -r inet tcp_keepidle 150  ! 75 secs.  (I know some use a " > >value of 600 = 5 minute polls). > ; > This doesn't work, Matt.  I tried the following commands:  > & > $ @sys$manager:tcpip$define_commands& > $ sysconfig -r inet tcp_keepidle 600 > 600: unknown attribute >  > Just a sysconfig gives: J > usage: sysconfig [-v] [-i <index>] -c | -s | -m | -q | -Q | -d | -o | -r) | -u [<subsys>] [<attributes> | <opcode>]  > B > There's probably a simple syntax error in this, but I can't findH > online documentation on sysconfig.  Can you - or someone else in c.o.v > - help out here? > F > Here's a theory we developed today, which maybe farfetched, but doesA > explain the observed data.  To recap, we observed the following  > disconnection pattern: > 6 > System  VMS     TCPIP           Disconnects the boss% > UMTOF   V7.3-2  V5.4 - ECO 1    YES % > UMTOF2  V7.3-2  V5.4 - ECO 1    YES $ > UMACE   V7.1-2  V4.2 - ECO 1    NO$ > UMSMS   V7.1-2  V4.2            NO >  > Suppose the following:C > 1) V4.2 of TCPIP (UCX) doesn't implement keepalive probes for its  >    TELNET server, 0 > 2) TCPIP V5.4 does implement keepalive probes, > 3) either:G >   a) something prevents the keepalive probes from reaching the boss's  >      computer, or B >   b) something prevents responses to keepalive probes from being >      received by UMTOF, or? >   b) the boss's tcpip package doesn't implement responding to  >      keepalive probes. > E > The TELNET server on, say, UMACE would never send a keepalive probe F > packet, but would also never terminate the tcp connection because it? > is idle.  The TELNET server on UMTOF, however, would send the G > keepalive probe, nver receive a response to it, decide the connection  > is dead, and disconenct it.  > E > This supposes a lot, but as I said, does explain the data.  Is this H > theory one we should follow up?  Can anyone debunk it definitively, orH > provide evidence pro or con?  For example, a definitive statement thatH > UCX V4.2 telnet server does use keepalive probes would blow this away. > A > Btw, is there a way to determine what the default keepalive (or - > keepidle) setting is, or even if it's used?  >  > Lawrence Bleau > University of Maryland$ > Physics Dept., Space Physics Group > 301-405-6223 > bleau@umtof.umd.edu    ------------------------------  + Date: Thu, 10 Jun 2004 22:35:12 +0000 (UTC) * From: bleau@UMTOF.UMD.EDU (Lawrence Bleau); Subject: Re: Why does idle telnet session get disconnected? 0 Message-ID: <caann0$4sh$1@grapevine.wam.umd.edu>  n In article <Lf4yc.1523$sj4.293@news-server.bigpond.net.au>, "Matt Muggeridge" <Matt.Muggeridge@hp.com> writes:< >> This doesn't work, Matt.  I tried the following commands: >>' >> $ @sys$manager:tcpip$define_commands ' >> $ sysconfig -r inet tcp_keepidle 600  >> 600: unknown attribute  > / >Sorry, I was missing an "=" sign.  Try this...  > ) >    $ sysconfig -r inet tcp_keepidle=600   A Thanks, Matt, this worked.  I'll tell me boss to do the test now.   5 >You can get online help with:  TCPIP HELP SYSCONFIG.   R Not much, though, as you noticed.  Thanks for taking this back to the doc section.  K Speaking of docs, the online docs on managing tcpip didn't have anything on Y this.  The HELP topic TCPIP didn't even have a pointer to tell one to look in TCPIP HELP.   7 Btw, is there a similar command for my UCX V4.2 system?    Lawrence Bleau University of Maryland" Physics Dept., Space Physics Group 301-405-6223 bleau@umtof.umd.edu    ------------------------------  # Date: Fri, 11 Jun 2004 01:18:16 GMT 0 From: "Matt Muggeridge" <Matt.Muggeridge@hp.com>; Subject: Re: Why does idle telnet session get disconnected? < Message-ID: <sL7yc.1735$sj4.1374@news-server.bigpond.net.au>  7 "Lawrence Bleau" <bleau@UMTOF.UMD.EDU> wrote in message * news:caann0$4sh$1@grapevine.wam.umd.edu...  J > Speaking of docs, the online docs on managing tcpip didn't have anything onL > this.  The HELP topic TCPIP didn't even have a pointer to tell one to look in TCPIP HELP.  I I agree.  This is a VMS thing that I find increasingly frustrating.  With G the advent of search engines, I can pretend the world of information is 2 flat... i.e. I don't have to think hierarchically.  I In the case of VMS help, you need to first think of categorizing the item K you want help about.  At times, this can take several attempts to find what H you're after.  For instance, I would prefer to type "help ifconfig" thanI "help tcpip unix_commands", only to learn that I should have typed "tcpip G help ifconfig".  This is not unique to TCP/IP, as I remember back to my E early days of VMS and had to remember to type "help lexical f$search"  instead of "help f$search".   $ Maybe one day we can flatten help...   > 9 > Btw, is there a similar command for my UCX V4.2 system?   & No.  sysconfig was introduced in V5.x.   Matt   ------------------------------  # Date: Fri, 11 Jun 2004 01:21:32 GMT 0 From: "Matt Muggeridge" <Matt.Muggeridge@hp.com>; Subject: Re: Why does idle telnet session get disconnected? ; Message-ID: <wO7yc.1741$sj4.533@news-server.bigpond.net.au>   : "JF Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message: news:929b3d834e6b0979601d1e3da75cfeba@news.teranews.com...  G > Is there good documentation somewhere on the impact of each and every . > parameter in sysconfig as well as ifconfig ?  + It is coming...  I don't have dates though.    > J > And as far as your typo is concerned, we understand. Out on the beach inG > winter, your fingers must have been shivering because the temperature  dropped  > below 30C :-) :-) :-) :-)  I That's it.  The mercury almost fell out the bottom of the thermometer the L other night when it got down to 6C.  Any colder and it had better snow! :-)   Matt.    ------------------------------  + Date: Thu, 10 Jun 2004 21:41:25 +0000 (UTC) 6 From: peter@langstoeger.at (Peter 'EPLAN' LANGSTOEGER)C Subject: [OpenVMS Alpha V7.3-2] How to keep OPA0 operator enabled ? 0 Message-ID: <newscache$9y34zh$y2l$1@news.sil.at>  H As you might know, there seems to be still a bug in OPC$mumble handling.F Some days/weeks ago, I wrote of my OPERATOR.LOG on OpenVMS VAX V7.3...  B I have OPC$OPA0_ENABLE defined as TRUE in SYLOGICALS.COM and on myF Alpha this gets honored and OPA0 is an operator terminal. During boot.  G I'm using an operator window (you know, defining DECW$CONSOLE_SELECTION H as "WINDOW" in DECW$PRIVATE_APPS_SETUP.COM) and when it starts, which isD shortly after the STARTUP process frees OPA0, OPA0 has no longer theC operator enabled status. But a couple of seconds/minutes has again.  (Why the delay ?)   J Then I log into CDE and OPA0 (= operator window) keeps enabled until I logH out again. A newly created DECW$CONSOLE process is then owning OPA0: andJ the operator status of OPA0 is disabled again. Then, nothing more happens.  H Many many minutes later, the status is still disabled (and probably someK OPCOM messages are already lost). Is this really intended ? Don't think so. D eg. Why have an operator window on a not logged-on graphic display ?D What sense makes an operator window not being an operator terminal ?  C Do you have a workaround which would make sense ? eg. I could ran a B batch job permanently (or every couple of minutes) and enable OPA06 as an operator terminal, but this seems an overkill...   TIA    --   Peter "EPLAN" LANGSTOEGER % Network and OpenVMS system specialist  E-mail  peter@langstoeger.atF A-1030 VIENNA  AUSTRIA              I'm not a pessimist, I'm a realist   ------------------------------  + Date: Thu, 10 Jun 2004 17:17:12 +0000 (UTC) 6 From: peter@langstoeger.at (Peter 'EPLAN' LANGSTOEGER), Subject: Re: [OpenVMS V7.3-2] Mount question0 Message-ID: <newscache$xpr3zh$9hh$1@news.sil.at>  _ In article <+4jQ1eUu9ya5@cuebid.zko.dec.com>, brooks@cuebid.zko.dec.nospam (Rob Brooks) writes: 9 >peter@langstoeger.at (Peter 'EPLAN' LANGSTOEGER) writes: 4 >> brooks@cuebid.zko.dec.nospam (Rob Brooks) writes:& >>>(Peter 'EPLAN' LANGSTOEGER) writes:8 >>>> So, my current guess is a resources problem so far. >>>>  J >>>> PS: Can someone please add a DEVBUSYINSET to HELP/MESSAGE and add the/ >>>> ressources guess as a possible solutions ?  >>  E >>>With all due respect, it is highly unlikely that lack of resources  >>>is the cause of the problem.  >>  I >> That means, you are guessing that the shadow merge fixed the problem ?  > G >No.  I'm guessing that the locking protocol messed itself up such that C >the mounting system was unable to acquire (or convert) the lock to E >either EX or PW (I can't remember in which mode this lock is taken).  > O >I don't know how things got messed up, but it is the rebooting of the cluster, / >not the subsequent merge, that "fixed" things.   H But then again, I did form the cluster 4 times by booting the bootserverL and later did boot the satellite into and every 4 times I got the same errorH (ok, not completely, the first ones had the default /ASSIST and I had noK operator terminal). So, if you ask me, the cluster reboot alone did NOT fix I the problem (4 times), but an AUTOGEN of the satellite and a shadow merge E (and maybe something else) before the fifth boot did fix the problem.   F Maybe if the problem reappears further tests will make some more hints- but until then we're only guessing I think...    --   Peter "EPLAN" LANGSTOEGER % Network and OpenVMS system specialist  E-mail  peter@langstoeger.atF A-1030 VIENNA  AUSTRIA              I'm not a pessimist, I'm a realist   ------------------------------    Date: 10 Jun 2004 11:29:16 -0700$ From: n8wxs@arrl.net (Jeff Campbell)S Subject: [TCPIP] How is the routing table updated when ROUTED or GATED aren't used? = Message-ID: <1a40a0b2.0406101029.7492c379@posting.google.com>    Hi all,   ' OpenVMS 7.3 on a DS20, TCPIP 5.1 ECO 4.    Local net: 192.168.12.0/24.  Remote net: 192.168.10.0/24.  / T1 link between nets, with routers on each end:    Local router: 192.168.12.4 Remote router: 192.168.10.1   5 There is a WAN router on the local net: 192.168.12.1.   0 The VMS system default (and permanent) route is:     0.0.0.0  192.168.12.4   ; No other routes are defined. ROUTED and GATED are not used.   7 The local end T1 router has this default route defined:       0.0.0.0  0.0.0.0  192.168.12.1   Problem:  J Local access to machines and printers residing on the remote network failsK after random intervals of time. Of 50 plus addresses, one, then two, then 3 K or 4 will stop being accessible. Most of these addresses are PCs, which get K their IP addresses from a DHCP server on the remote network. But at least 1 G failing address is an HP printer with a manually configured (static) IP ; address. These printers are accessed using TCPIP$TELNETSYM.   L A TCPIP SHOW ROUTE command will show that all the failing addresses now haveI DH routes through the WAN router, 192.168,12,1, instead of the previously @ correct route through the permanent default route, 192.168.12.4.  H So, my question is: how does the TCPIP routing database get updated with; incorrect routes when neither ROUTED nor GATED are running?   C And, 8-) what can I do to keep this dynamic update from happening??    TIA,  
 Jeff Campbell  n8wxs@arrl.net   ------------------------------  # Date: Fri, 11 Jun 2004 01:03:34 GMT 0 From: "Matt Muggeridge" <Matt.Muggeridge@hp.com>W Subject: Re: [TCPIP] How is the routing table updated when ROUTED or GATED aren't used? < Message-ID: <Gx7yc.1720$sj4.1322@news-server.bigpond.net.au>  L This may be due to route redirects and is usually indicative of a network or router setup problem.   G A network diagram along with subnet addressing and host addresses would J help.  Output from "netstat -nr" on both the PC and server would shed moreF light on the situation, along with route information for the local and remote routers.   J If the Flags field in the output from "netstat -nr" contains "D", then theK route was dynamically created as the result of an ICMP redirect.  You don't I usually want to disable redirects, as they provide a very useful service.   F Without more info, it sounds like your VMS node attempts to send an IPJ datagram to a remote PC and first sends to the default (.4) router (as youE would expect it to).  When the default router determines the next hop L address, it forwards the datagram out the same interface that it received itJ on, which is the signal to send an ICMP redirect back to VMS telling it toH use 192.168.10.1 in the future.  VMS then attempts to send to the remoteL router directly, which of course fails, since it is not on the same network.  K I would be inclined to look very closely at the subnetting and router table  on the two routers.    Matt.    --  = -------------------------------------------------------------  OpenVMS TCP/IP Engineering Enterprise Computing Group Hewlett-Packard Company  Gold Coast, AUSTRALIA = -------------------------------------------------------------     1 "Jeff Campbell" <n8wxs@arrl.net> wrote in message 7 news:1a40a0b2.0406101029.7492c379@posting.google.com... 	 > Hi all,  > ) > OpenVMS 7.3 on a DS20, TCPIP 5.1 ECO 4.  >  > Local net: 192.168.12.0/24.  > Remote net: 192.168.10.0/24. > 1 > T1 link between nets, with routers on each end:  >  > Local router: 192.168.12.4 > Remote router: 192.168.10.1  > 7 > There is a WAN router on the local net: 192.168.12.1.  > 2 > The VMS system default (and permanent) route is: >  >   0.0.0.0  192.168.12.4  > = > No other routes are defined. ROUTED and GATED are not used.  > 9 > The local end T1 router has this default route defined:  > " >   0.0.0.0  0.0.0.0  192.168.12.1 > 
 > Problem: > L > Local access to machines and printers residing on the remote network failsK > after random intervals of time. Of 50 plus addresses, one, then two, then  3 I > or 4 will stop being accessible. Most of these addresses are PCs, which  get K > their IP addresses from a DHCP server on the remote network. But at least  1 I > failing address is an HP printer with a manually configured (static) IP = > address. These printers are accessed using TCPIP$TELNETSYM.  > I > A TCPIP SHOW ROUTE command will show that all the failing addresses now  haveK > DH routes through the WAN router, 192.168,12,1, instead of the previously B > correct route through the permanent default route, 192.168.12.4. > J > So, my question is: how does the TCPIP routing database get updated with= > incorrect routes when neither ROUTED nor GATED are running?  > E > And, 8-) what can I do to keep this dynamic update from happening??  >  > TIA, >  > Jeff Campbell  > n8wxs@arrl.net   ------------------------------  # Date: Thu, 10 Jun 2004 22:22:14 GMT 0 From: glen herrmannsfeldt <gah@ugcs.caltech.edu>Y Subject: Re: [TCPIP] How is the routing table updated when ROUTED or GATED aren't used? a + Message-ID: <qa5yc.6273$2i5.3309@attbi_s52>    Jeff Campbell wrote:
 (big snip)  J > So, my question is: how does the TCPIP routing database get updated with= > incorrect routes when neither ROUTED nor GATED are running?    Route redirect.   9 When a router discovers that the outgoing packet is going 6 out the same interface it came in on, it sends an ICMP5 route redirect message to the sending host.   It then 4 goes into the routing table as a host route, instead of a network route.   % It might be that you can turn it off.    -- glen    ------------------------------   End of INFO-VAX 2004.322 ************************                                                                                                                                                                                                                                                                                                                                                                                                                  has a problem in combination with MSCP ?M Btw removing /OVERIDE=NO_FORCED_ERROR on the satellite doesn't make a change.e   -- a Peter "EPLAN" LANGSTOEGER % Network and OpenVMS system specialisti E-mail  peter@langstoeger.atF A-1030 VIENNA  AUSTRIA              I'm not a pessimist, I'm a realist   ------------------------------  * Date: Tue, 8 Jun 2004 20:19:43 +0000 (UTC)6 From: peter@langstoeger.at (Peter 'EPLAN' LANGSTOEGER), Subject: Re: [OpenVMS V7.3-2] Mount question1 Message-ID: <newscache$hua0zh$