1 INFO-VAX	Sun, 03 Sep 2006	Volume 2006 : Issue 482       Contents: Re: Getting into VMSP Re: OT - running IE in Windows (was:Re: Thoughts on the book: DEC is  dead, long4 Re: Thoughts on the book: DEC is dead, long live DEC4 Re: Thoughts on the book: DEC is dead, long live DEC4 RE: Thoughts on the book: DEC is dead, long live DEC4 Re: Thoughts on the book: DEC is dead, long live DEC. Re: We're not dead yet, we're feeling better !. Re: We're not dead yet, we're feeling better !. Re: We're not dead yet, we're feeling better !. Re: We're not dead yet, we're feeling better !. Re: We're not dead yet, we're feeling better !  F ----------------------------------------------------------------------  % Date: Sat, 02 Sep 2006 17:39:02 -0400 2 From: Bob Willard <BobwBSGS@TrashThis.comcast.net> Subject: Re: Getting into VMS : Message-ID: <nN6dndwCIIz5ZGTZnZ2dnUVZ_tSdnZ2d@comcast.com>   Richard B. Gilbert wrote:  > faceman28208@yahoo.com wrote:  > I >> While going Wikipedia looking for some work-related material I clicked J >> off a few links and ended up on the VMS world. I have not touched a VMS. >> system in over a decade (since I left DEC). >>J >> I noticed that many of the things Digital used to keep under locked andI >> key (e.g. SDL) or made difficult for customers to use (e.g. BLISS) are J >> freely available. I see there is a hobbyist program as well. I see that) >> Alpha is dead and VMS has been ported.  >>G >> Let's suppose that I wanted to buy/build a VMS, can one be had for a D >> reasonable price? (Or is it like the old days when DEC gouged its, >> customers in its quest for high margins.) >>4 >> How would one get a home VMS system these days??? >>E >> Do you have to get some special HP System or has HP had the common 2 >> sense to open up VMS to off-the-shelf hardware? >>I >> I used to have a Vaxstation 3100 at home back in the days when I was a I >> Digit. I remember how difficult DEC made it for employees to buy a VMS 
 >> system. >>G >> After all these years of working on M$-DOG and Eunuchs it is amazing J >> how far we have regressed since the height of VMS. One can only imagineF >> where the industry could have been now if Digital had had some real >> management. >> >>> From the olde days:  >>I >> Q: What's the difference between Digital Equipment and the Boy Scouts? , >> A: The Boy Scouts have adult supervision. >> >> g,  >> > G > You can pick up used DEC "kit" on e-Bay.  I have an old Alphastation  E > 200/166 that I would part with for not too much money.  There is a  ( > vigorous market in used DEC equipment! > # > q.  How do you identify a DEC CE? K > a.  He is the guy who is changing all his tires to see which one is flat!  > & > q.  How do you tell if he's your CE? > a.  His spare is flat!  
 And this -  = Q. What's the difference between a PRO350 and a bowling ball? & A. The bowling ball has more software. --   Cheers, Bob    ------------------------------  % Date: Sat, 02 Sep 2006 20:00:09 +0200 / From: Paul Sture <paul.sture.nospam@hispeed.ch> Y Subject: Re: OT - running IE in Windows (was:Re: Thoughts on the book: DEC is  dead, long J Message-ID: <paul.sture.nospam-E919A5.20000902092006@mac.sture.homeip.net>   In article  H <FA60F2C4B72A584DBFC6091F6A2B8684019BCAC9@tayexc19.americas.cpqcorp.net> , )  "Main, Kerry" <Kerry.Main@hp.com> wrote:     E > [Just make sure you have the original OS CD's before you try this.]  >   9 Nah. Just make sure you have a decent backup solution :-)     D BTW Kerry, can you please learn to chop irrelevant quotes from your C posts? It gets quite tiresome ploughing through too excess quoting.    TIA    --  
 Paul Sture   ------------------------------  $ Date: Sat, 2 Sep 2006 19:01:44 -0400) From: "Neil Rieck" <n.rieck@sympatico.ca> = Subject: Re: Thoughts on the book: DEC is dead, long live DEC < Message-ID: <44fa0bf1$0$24200$9a6e19ea@news.newshosting.com>  * <faceman28208@yahoo.com> wrote in message < news:1157147437.150915.111850@e3g2000cwe.googlegroups.com... >  [...snip...] > D > 4. Dipping the pen in the company ink.-- I've never seen a companyI > where so many managers were boinking their employees (literally). Using G > corporate resources to carry on extramarital affairs was accepted and 7 > tolerated. I've never seen that at any other company.  >   ! Talk to someone working for Apple    [...snip...] > G > 6. Problem #5 was compounded by the fact that in the customer contact C > areas, due to the fact the "metrics" covered margin, there was no C > investment in equipment. You were expected to help customers with I > machines you had never seen and have 10 developers doing development on  > one Microvax II. >   K I remember getting customer training at 129 Parker St, in Maynard. We were  K in the labs working on a bunch of uVAX4300 machines (one machine for every  M two students). We went for afternoon coffee and upon our return we were told  F that one of the machines had been shipped out to deal with a customer J emergency. I don't remember hearing an apology or anyone getting a credit L voucher. The 2 students with the missing machine were just forced to double L up. (I remember thinking that the hand-writing-might be on the wall; we all 9 knew that Olson had been given the push before this time)    > F > 8. Digital had an image problem. The company refused to advertise onB > television. In spite of the fact it was the 2nd largest computerG > company, many people had no idea who we were. The Massachusetts Mafia I > never understood this problem because every back road in that state had  > a digital office.  >   H They advertised for a long time (years) on the PBS program "the Nightly  Business Report".    >  [...snip...] > E > Keeping with the customer value....I remember all the headaches DEC F > caused with the moronic license PAK scheme it came up with. "To helpE > the customer manage their licenses" we told them as we screwed them  > with that useless headache.  >   L Almost every company makes bone-headed decisions but companies that produce G both large amounts of both hardware and software seem to make the most  I interesting ones.  IBM almost lost everything when they tried to have it  K both ways (anybody remember the PC s/w fiascos from PC-DOS though to OS2?)  L but they've dumped all that stuff onto Lenovo and gone back to their larger M system roots. Although they have made large sums of money on their operating  D systems, this software is primarily used to leverage the selling of L hardware. When selling large amounts of hardware, I've heard that the price F of IBM software can be negotiated very low as long as the customer is M willing to sign an NDA about the whole deal. (Hey, once the stuff is written  L the hard work has been done; anything else it can do for your business is a 7 plus). To people who are really cheap they offer LINUX.   F To your point, I remember hearing DEC people talking about cloned VAX J hardware showing up in Russia as one of the reasons for licence PAKs. I'm D not sure that this was a valid reason for pissing off your domestic J customers. You would have thought that once they developed this scheme it K would have been easier to issue time-limited licenses to Universities with  I zero paranoia. (Almost everyone will agree that raising the next crop of  M plants on your fertilizer is the best way to ensure your company will have a   future in IT).  
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.! http://www3.sympatico.ca/n.rieck/    ------------------------------  $ Date: Sat, 2 Sep 2006 19:22:59 -0400) From: "Neil Rieck" <n.rieck@sympatico.ca> = Subject: Re: Thoughts on the book: DEC is dead, long live DEC < Message-ID: <44fa10ec$0$24196$9a6e19ea@news.newshosting.com>  2 "Main, Kerry" <Kerry.Main@hp.com> wrote in messageN news:FA60F2C4B72A584DBFC6091F6A2B8684019BCA93@tayexc19.americas.cpqcorp.net...  J With all the talk about the book "DEC is dead, long live DEC" some people 6 here seem to be convinced that DEC was full of losers.  L This week I'm upgrading some Alpha Servers from OpenVMS-7.3-2 to 8.2 and as M anyone who has done something like this already knows, you are provided with  I a bunch of firmware updates to make most of your peripherals, as well as  I your console, compatible with the new OS. (before we update our DS20e we  M copy the system onto a junk AS2100 and upgrade that system in order to fully  K test our in-house applications; I just brought up 8.2 on a machine that is   11 years old; Amazing)  J I am convinced that if the OS was one of the LINUX distros or MS Windows, H that I'd be experiencing hardware-software compatibility problems which H might require my employer to run out and purchase more-up-to-date newer  peripherals.  L DEC-Compaq-HP might have had their share of questionable upper managers but < the technology produced by their engineers was/is top-notch.  
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.8 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html   ------------------------------  $ Date: Sat, 2 Sep 2006 21:34:34 -0400' From: "Main, Kerry" <Kerry.Main@hp.com> = Subject: RE: Thoughts on the book: DEC is dead, long live DEC T Message-ID: <FA60F2C4B72A584DBFC6091F6A2B8684019BCAEA@tayexc19.americas.cpqcorp.net>   > -----Original Message-----3 > From: Neil Rieck [mailto:n.rieck@sympatico.ca]=20 ! > Sent: September 2, 2006 7:23 PM  > To: Info-VAX@Mvb.Saic.Com ? > Subject: Re: Thoughts on the book: DEC is dead, long live DEC  >=204 > "Main, Kerry" <Kerry.Main@hp.com> wrote in message@ > news:FA60F2C4B72A584DBFC6091F6A2B8684019BCA93@tayexc19.america > s.cpqcorp.net... >=20B > With all the talk about the book "DEC is dead, long live DEC"=20 > some people=208 > here seem to be convinced that DEC was full of losers. >=20	 [snip...]    Well, I certainly hope not.    :-)   A While Digital's directions could certainly have gone in different ? directions had there been steadier hands at the wheel, the core E strengths of Digital like Engineering and Customer Service focus were  very much leading edge.=20  F For me, while Digital certainly has had its up's and down's (yes, likeH most companies, everyone has their own little horror story), its cultureE was one which I do not think anyone would hesitate to go back to. One B example is the Digital and other notes files which still continuesH today.  You can go back and read these Digital notesfile entries all theJ way back to July 1989 when that particular notesfile was first created.=20   Regards    ------------------------------  % Date: Sat, 02 Sep 2006 21:40:06 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> = Subject: Re: Thoughts on the book: DEC is dead, long live DEC , Message-ID: <44FA3275.558525D6@teksavvy.com>   Michael Kraemer wrote:. > For that PAK crap alone DEC deserved to die.: > Even in its mild hobbyist incarnation it's a major PITA.    H Arguments can be made both ways. PAKs were a great way to normalise yourH inventory of software and make sure DEC had the same list as you did. InH large corporations, it wasn't so obvious what licences belonged to which department.   N Had there not been a PAK system, things would have gotten far worse over time.  D And PAKs are what allowed condist to exist. While consist may now beG moot due to internet distribution capabilities, it was of great use for E a long time (and still is because internet distribution isn't quite a  reality for most customers).    F Now, had DEC made it easier to buy/manage/consolidate the licences viaE the web or other interface instead of having to fight for 3 months to H find one person able to place an order and issue a pak, or just update a- PAK, then things would have been much better.    ------------------------------  % Date: Sat, 02 Sep 2006 14:07:36 -0700 ' From: David Mathog <mathog@caltech.edu> 7 Subject: Re: We're not dead yet, we're feeling better ! + Message-ID: <edcrqo$6i0$1@naig.caltech.edu>   - David Turner, Island Computers US Corp wrote:  <SNIP>  > when we decided to 8 > go with a non-alpha box with dual p/s and RAID5 + ADG.  A Ah yes, RAID, that panacea of uptime - except when the controller > glitches and a couple of hundred Gb of data suddenly disappear< for good.  Another site's RAID failure bit me recently while: setting up a Mandriva machine.  That linux flavor supportsA URPMI and my favorite mirror for it is at ftp://mirrors.usc.edu/, @ but curl (the download agent for urpmi) wouldn't connect.  So in a browser I entered:     http://mirrors.usc.edu/   5 and found that currently it says (this will certainly  change eventually!):  G NOTE: Mirrors is currently re-syncing its mirrors due to a failed RAID  G array. Please be patient as we pull down large quantities of data over  I the next week or so. We're returning a 503 on directories not yet synced   for your convenience.   A The thing is, everybody I know who runs RAID has had this type of @ catastrophic event, some of them several times, and these events; take *much* longer to recover from than other types of disk > failures because it's not just one disk that must be restored,% but the entire contents of the array.   B So RAID5 if you must, but a fast tape drive and recent backup tape< will still be required to save your bacon in some instances.   Regards,   David Mathog mathog@caltech.edu   ------------------------------  % Date: Sat, 02 Sep 2006 20:54:10 -0500 % From: Dan Foster <usenet@evilphb.org> 7 Subject: Re: We're not dead yet, we're feeling better ! 5 Message-ID: <slrnefkde2.hok.usenet@zappy.catbert.org>   T In article <edcrqo$6i0$1@naig.caltech.edu>, David Mathog <mathog@caltech.edu> wrote: > 3 > Another site's RAID failure bit me recently while < > setting up a Mandriva machine.  That linux flavor supportsC > URPMI and my favorite mirror for it is at ftp://mirrors.usc.edu/, B > but curl (the download agent for urpmi) wouldn't connect.  So in > a browser I entered:  F I hate to say it, but many PC RAID controllers are, well, um, how do ID say it nicely and in a non-prejudicial way? I can't think of one, soF I'll just say that they aren't robust because many employs a number ofA dubious tricks (to save money) that don't hold up well under real  failure situations.   B I can dig up citations and details from my notes if you ever want.  H If I have to use RAID on x86, I only stick with *known* _solid_ HBAs andG vendors. Period. Even if I have to buy one, even if the machine already  has a bundled RAID controller.  9 Even if it's for home gear, and out of my pocket. Always.   I > NOTE: Mirrors is currently re-syncing its mirrors due to a failed RAID  I > array. Please be patient as we pull down large quantities of data over  K > the next week or so. We're returning a 503 on directories not yet synced   > for your convenience.   F That would actually make me wonder about their operations. No offense;0 they're probably good, but still have to wonder.  H Consider this: if you had a RAID-5 array (or LUN) and a hot swap disk...G you would need *3* drives to fail before the array is a goner, assuming ) it's a disk-related failure (usually is).   C How do you lose 3+ disks in a very short period of time, short of a 7 power surge, in ordinary circumstances and not know it?   E At a former employer... they lost one too many drives in their RAID-5 H set because of a misconfiguration in their monitoring/alert notification _and_ their backup system.  C The next 18 months were a living h@ll, with tremendous impact. They A begged then-DEC to help them salvage data. They were told, no, it  couldn't be done.   > Eventually, about 16(?) months later, someone got the ear of aH sufficiently senior person at DEC. DEC then got engineers working on theF issue. Surprise: they were able to restore about 80% of the data usingG special tools to walk on-disk data structures and rebuild various maps.   D But the pain was extreme, and cost a serious amount of money. PeopleE were fired or quietly decided to retire early. They also subsequently D beefed up their notification system and did drills, including backup tape restoration drills. :-)  C Around here, as soon as a drive starts emitting warning messages or G errors, our on-call gets paged by our monitoring system and entire team H is emailed. We then move *very* quickly on drive replacement. So far, we= haven't had an entire array fail due to too many dead drives.   @ Even then, we have multiple arrays, backup software, tests, etc.  C > The thing is, everybody I know who runs RAID has had this type of B > catastrophic event, some of them several times, and these events= > take *much* longer to recover from than other types of disk @ > failures because it's not just one disk that must be restored,' > but the entire contents of the array.   $ Yeah, entire array failure is nasty.  D > So RAID5 if you must, but a fast tape drive and recent backup tape> > will still be required to save your bacon in some instances.  E Funny thing is, latest tape drive technology can be actually somewhat = faster than hard drives (for sequential I/O only, obviously).   G We've had interesting situations in architecting a good backup solution C with such fast tape drives because previously we funnelled all data D through disk-based temporary storage pools. Some adjustments helped.  H I would, though, never do RAID-5 or any other technology as a sole meansE of saving my gluteus maximus (aka 'butt'). Always in conjunction with 1 other tools, and preferrably, in multiple layers.    -Dan   ------------------------------  % Date: Sat, 02 Sep 2006 21:33:53 -0400 3 From: "Richard B. Gilbert" <rgilbert88@comcast.net> 7 Subject: Re: We're not dead yet, we're feeling better ! : Message-ID: <9M2dnQapBcuZrGfZnZ2dnUVZ_rydnZ2d@comcast.com>   David Mathog wrote: / > David Turner, Island Computers US Corp wrote:  > <SNIP> >  > when we decided to  > 9 >> go with a non-alpha box with dual p/s and RAID5 + ADG.  >  > C > Ah yes, RAID, that panacea of uptime - except when the controller @ > glitches and a couple of hundred Gb of data suddenly disappear> > for good.  Another site's RAID failure bit me recently while< > setting up a Mandriva machine.  That linux flavor supportsC > URPMI and my favorite mirror for it is at ftp://mirrors.usc.edu/, B > but curl (the download agent for urpmi) wouldn't connect.  So in > a browser I entered: >  >  http://mirrors.usc.edu/ > 7 > and found that currently it says (this will certainly  > change eventually!): > I > NOTE: Mirrors is currently re-syncing its mirrors due to a failed RAID  I > array. Please be patient as we pull down large quantities of data over  K > the next week or so. We're returning a 503 on directories not yet synced   > for your convenience.  > C > The thing is, everybody I know who runs RAID has had this type of B > catastrophic event, some of them several times, and these events= > take *much* longer to recover from than other types of disk @ > failures because it's not just one disk that must be restored,' > but the entire contents of the array.  > D > So RAID5 if you must, but a fast tape drive and recent backup tape> > will still be required to save your bacon in some instances. > 
 > Regards, >  > David Mathog > mathog@caltech.edu  B It has always been true that RAID (1 or 5) protects you from disk I failure, not from controller failure or software errors.   It's not, and  E never has been, a substitute for backup.  A file deleted in error is  I still deleted from a RAID array.  A file corrupted by software errors is    still corrupted on a RAID array.  F You may not need to restore from backup quite as frequently on a RAID G array but if you need to, your need is just as great as if the failure   had been on a JBOD.    ------------------------------  % Date: Sat, 02 Sep 2006 22:53:11 -0400 ( From: Bill Todd <billtodd@metrocast.net>7 Subject: Re: We're not dead yet, we're feeling better ! G Message-ID: <d8ydnYaXvqkF3mfZnZ2dnUVZ_tqdnZ2d@metrocastcablevision.com>    Richard B. Gilbert wrote:    ...   D > It has always been true that RAID (1 or 5) protects you from disk K > failure, not from controller failure or software errors.   It's not, and  * > never has been, a substitute for backup.  = Not in and of itself, anyway.  But in combination with other   technologies, it can be.      A file deleted in error is K > still deleted from a RAID array.  A file corrupted by software errors is  " > still corrupted on a RAID array.  H But in both those cases a recent snapshot is considerably more valuable A than an older backup which must be laboriously retrieved from an   off-line medium.  E Beyond snapshots lies 'continuous data protection' (CDP), in which a  H full audit trail of updates (from a consistent backup point or from Day F Zero) is maintained separately from the running system, such that the I system state at any earlier point in time (not just at explicit snapshot  D points) can be reconstructed, on line.  This offers protection from F *any* failure (at worst, you just have to back things up to the point G just before it occurred), while the RAID gives the running system high  E availability in the absence of failures which it can't itself handle.   F Backup is on the way out:  it's becoming infeasible at today's system I sizes even to protect against catastrophic failure (many businesses just  F can't afford to wait days for a full restore to occur), and snapshots B have already supplanted it for recovery from user and application G oopses.  Its replacements aren't mature yet, but they're closing in on  E what the eventual post-backup solution will be (a solution which may   cover archiving needs as well).    - bill   ------------------------------  % Date: Sat, 02 Sep 2006 22:33:42 -0400 ( From: Bill Todd <billtodd@metrocast.net>7 Subject: Re: We're not dead yet, we're feeling better ! G Message-ID: <7MednaUebvSbomfZnZ2dnUVZ_qWdnZ2d@metrocastcablevision.com>    Dan Foster wrote:    ...   J > Consider this: if you had a RAID-5 array (or LUN) and a hot swap disk...I > you would need *3* drives to fail before the array is a goner, assuming + > it's a disk-related failure (usually is).  > E > How do you lose 3+ disks in a very short period of time, short of a 9 > power surge, in ordinary circumstances and not know it?   G All the hot-swap disk lets you do is replace the failed disk while the  < array is still running:  perhaps you meant a hot spare disk?  I A hot spare also allows normal processing to continue while it's rebuilt  G but only reduces the time it takes to bring the array back up to snuff  E from the time it would take if you replaced the disk manually:  if a  G second disk fails before that happens, your data is still toast.  With  H disks in the hundreds-of-GB size range even at full tilt (i.e., leaving H no bandwidth to continue normal operations) it takes hours to rebuild a H failed disk, so a hot spare is primarily useful only in situation where I manually replacing the failed disk within a few hours (using hot-swap if  I the array must be kept running) is not an option (i.e., if the time that  B the array must operate in degraded mode significantly exceeds the G rebuild time after the failed disk has been replaced, since the latter  9 will place at least as much unusual stress on the disks).   F If you want robustness in a parity RAID even if two disks should fail ) within the rebuild time, you need RAID-6.    - bill   ------------------------------   End of INFO-VAX 2006.482 ************************