INFO-VAX Thu, 14 Jun 2007 Volume 2007 : Issue 322 Contents: Re: Another opportunity Re: Another opportunity Re: Another opportunity Re: Another opportunity Broken TSZ07 Re: BYPASS privilege !! Re: BYPASS privilege !! Re: BYPASS privilege !! Re: BYPASS privilege !! Encrypting a pdf file on vms 8.3 Re: Encrypting a pdf file on vms 8.3 Re: Encrypting a pdf file on vms 8.3 Re: Encrypting a pdf file on vms 8.3 RE: Encrypting a pdf file on vms 8.3 Re: Encrypting a pdf file on vms 8.3 RE: Encrypting a pdf file on vms 8.3 Re: Encrypting a pdf file on vms 8.3 Re: Encrypting a pdf file on vms 8.3 Re: Encrypting a pdf file on vms 8.3 Re: Encrypting a pdf file on vms 8.3 Re: ES45 Mod 2 Service Bulletin Re: help running OpenVMS on Itanium 2 2620 WinXP variant Re: help running OpenVMS on Itanium 2 2620 WinXP variant Re: help running OpenVMS on Itanium 2 2620 WinXP variant Re: OT: Carla Re: Question for the Group Re: Question for the Group Re: Question for the Group Re: Question for the Group Re: Question for the Group Re: Question for the Group Re: Question for the Group Re: Question for the Group Re: Question for the Group Re: Question for the Group Re: Question for the Group RPC and clnttcp_create SSL FTP Re: SSL FTP TCPIP Services SFTP vms 8.3 to Red Hat Linux AS4 Re: vms 8.3 to Red Hat Linux AS4 Re: vms 8.3 to Red Hat Linux AS4 Re: vms 8.3 to Red Hat Linux AS4 Re: VMS analogue of FBSD and linux hier(7) man pages Re: VMS analogue of FBSD and linux hier(7) man pages Re: VMS analogue of FBSD and linux hier(7) man pages Re: VMS analogue of FBSD and linux hier(7) man pages Re: VMS analogue of FBSD and linux hier(7) man pages Re: VMS analogue of FBSD and linux hier(7) man pages Re: VMS analogue of FBSD and linux hier(7) man pages Re: VMS analogue of FBSD and linux hier(7) man pages Re: VMS analogue of FBSD and linux hier(7) man pages Re: What's the implication of this development if any... Re: ZFS and OS X 10.5 (Leopard) ---------------------------------------------------------------------- Date: Thu, 14 Jun 2007 08:11:36 +0200 From: "P. Sture" Subject: Re: Another opportunity Message-ID: In article , Michael Kraemer wrote: > david20@alpha2.mdx.ac.uk schrieb: > > > > It's much too long since I setup VMS on a VAX system so I can't really > > comment on how well integrated TCPIP is into it's installation procedures. > > On Alpha the installation/upgrade procedure asks whether to install/upgrade > > DEC TCPIP services in the same way it asks whether you want to > > install/upgrade > > DECNET Phase IV or DECNET Phase V. > > That's similar to the VAX, but a VAX needs additional steps before TCPIP > works. > > > > >>On my VS4000s I had to edit sys$system:modparams.dat > >>to add a MIN_INTSTKPAGES=12, otherwise it wouldn't run. > >>(this was nowhere described, it took me several hours to figure it out). > >>Plus run some autogen stuff. > >>Plus the three UCX PAKs (no other OS I know of needs such dongles > >>just to be able to communicate with the rest of the world). > > > > DECNET also requires a license PAK. > > TCP/IP on most other systems does not need PAKs. > > >>Of course you can't just ftp the license.com since ftp is not yet > >>working. Brilliant. > > > > > > Apart from systems which were designed to install via FTP (and which > > therefore > > start up a restricted FTP process) is there any system which allows ftp > > access > > during installation ? Even systems which install via ftp don't, as far as I > > am > > aware, allow you to just ftp up a file in the middle of the installation. > > I tried to point out that it's a major nuisance in general > to type those silly authorization codes and checksums > (sth which is not necessary on most other systems). > One can partially avoid this by just running the license file as a DCL > script. This would require to ftp this file before, but hey, TCPIP needs > a PAK. No PAK, no TCPIP; no TCPIP, can't ftp the PAKs. Vicious circle so > to say. Or did I miss sth here ? Stick the license file on a) a VMS formatted floppy if you have access to another system. b) burn a CD (and might as well put some other stuff such as Freeware on there while you are at it. c) hook up another system via the serial port and a terminal emulator (increasingly hard now a PC might not have a serial port). > > > > Version 8.2 and later of OPENVMS don't require the Open3D license for 3D > > support and don't support installation of the Open3D layered product which > > is > > probably why it isn't on the hobbyist CD. Unfortunately OpenVMS 8.2 also > > dropped support for some graphics cards. > > see > > http://h71000.www7.hp.com/doc/82FINAL/6674/6674pro_retired.html#retirementch > > > > and > > > > http://h71000.www7.hp.com/doc/82FINAL/6674/6674pro_hardware2.html#graphicsbo > > ards > > Well, 7.3-2 (alpha) was all I had as hobbyist when I tried it last time. > Ask nicely offline and something can be organised. :-) -- Paul Sture ------------------------------ Date: Thu, 14 Jun 2007 10:29:48 +0000 (UTC) From: m.kraemer@gsi.de (Michael Kraemer) Subject: Re: Another opportunity Message-ID: In article <07061322585169_202003EE@antinode.org>, sms@antinode.org (Steven M. Schweda) writes: > From: Michael Kraemer > > > [...] but I think VMS always requires 512-byte SCSI. > > Many people seem to believe that VMS requires a CD-ROM drive which > has a 512-byte-block jumper, but I do not. I no longer (exclusively) > use the oldest hardware on the planet, but I've used some Toshiba CD-ROM > drives (XM5401B, XM-5701B, XM-6201B), and, I believe, some others (NEC?) > which had no block-size jumper, at least with Alphas, and perhaps with > VAXes, too. The drive does, I believe, need to respond properly to a > SCSI command which sets the block size, and the computer needs to be > smart enough to send the command, but it seems that many drives and > computers qualify. > interesting. > > I've noticed this on non-VAXes, too. And not only with CD-ROM > drives. as the number of systems I've installed grows, I'm inclined to say it's sort of black magic. Or just trial and error. > > Wasn't someone compiling a list of which CD-ROM drives work on what? you may find sth here: http://sites.inka.de/pcde/help/dec-cdrom-list.txt > I claim that a list of which drives fail on what would be shorter. > ------------------------------ Date: Thu, 14 Jun 2007 11:22:11 GMT From: rdeininger@mindspringdot.com (Robert Deininger) Subject: Re: Another opportunity Message-ID: In article , david20@alpha2.mdx.ac.uk wrote: >In article , Michael Kraemer writes: >>david20@alpha2.mdx.ac.uk schrieb: >> >>> Well it's installed as part of the standard OS installation procedure along >>> with in the same way as DECNET, X-Windows etc >> >>The best you can say is that they (TCPIP,Gfx) are on the same >>installation CD, but they're not well integrated at all. >>On VAX the installation is a PITA. The VAX installation process was frozen a long time ago, alas. This is an argument that VMS engineering lost to another part of Digital. >It's much too long since I setup VMS on a VAX system so I can't really >comment on how well integrated TCPIP is into it's installation procedures. >On Alpha the installation/upgrade procedure asks whether to install/upgrade >DEC TCPIP services in the same way it asks whether you want to install/upgrade >DECNET Phase IV or DECNET Phase V. >>On my VS4000s I had to edit sys$system:modparams.dat >>to add a MIN_INTSTKPAGES=12, otherwise it wouldn't run. >>(this was nowhere described, it took me several hours to figure it out). Stuff like that was documented in release notes; you probably missed it. This sort of manual labor has been mostly weeded out these days. If there are layered product release notes telling you to edit MODPARAMS.DAT today on Alpha or Itanium, it's a bug that should be fixed. >>Plus run some autogen stuff. Running Autogen has been a documented part of an installation or upgrade since almost forever. >>Plus the three UCX PAKs (no other OS I know of needs such dongles >>just to be able to communicate with the rest of the world). Yes, you've beat the dead horse here with your opinion of VMS PAKs. The PAKs are much simplified nowadays on Itanium. > >DECNET also requires a license PAK. > >>Of course you can't just ftp the license.com since ftp is not yet >>working. Brilliant. >Version 8.2 and later of OPENVMS don't require the Open3D license for 3D >support and don't support installation of the Open3D layered product which is >probably why it isn't on the hobbyist CD. Unfortunately OpenVMS 8.2 also >dropped support for some graphics cards. >see >http://h71000.www7.hp.com/doc/82FINAL/6674/6674pro_retired.html#retirementch > >and > >http://h71000.www7.hp.com/doc/82FINAL/6674/6674pro_hardware2.html#graphicsboards > >This appears to be another HP development since under COMPAQ and DEC VMS was >well known for it's continued support for decades old hardware with the >latest versions of the OS. You can call it an "HP development" if it makes you feel better, but that (retiring some old graphics cards) was one decision made entirely within VMS engineering. >>All-in-all, installation procedures, support of vendor's >>own hardware as well as integration of standard components >>just plain sucks, when compared with the competition. Keep in mind that you are asking for FREE, hobbyist support of NON-standard graphics cards which all PAYING customers left behind many years ago. In addition to the cost of building and testing the old Open3D stuff (which is painful since it was never part of the VMS build) there may still be royalty costs which HP would have to incur to provide this to hobbyists. ------------------------------ Date: Thu, 14 Jun 2007 05:27:01 -0700 From: Andrew Subject: Re: Another opportunity Message-ID: <1181824021.196953.41000@g37g2000prf.googlegroups.com> On 13 Jun, 15:18, koeh...@eisner.nospam.encompasserve.org (Bob Koehler) wrote: > In article <5d7lceF342pf...@mid.individual.net>, b...@cs.uofs.edu (Bill Gunshannon) writes: > > In article <1181640747.768522.7...@r19g2000prf.googlegroups.com>, > > IanMiller writes: > >> paragraph 3.6.1 is talking about Trusted Operating Systems with > >> mandatory access control. This has not been seen in VMS land since > >> SEVMS. > > > I thought regular VMS had "mandatory access control"? I guess I > > was wrong. :-( But it still fits into the general model proposed > > by this paper. > > The "mandatory access control" available in SEVMS and specified in > vertain government applications is not the kind of file control > seen in VMS, UNIX, Windows, or other general purpose OS. It includes > concepts like write-up (to higher security level) and read-down (to > lower security level). That is not correct, MAC support is available for a number of general purpose platforms including Linux and Solaris. http://www.sun.com/software/solaris/ds/trusted_extensions.jsp For example provides Mandatory Access Control on Solaris 10. Regards Andrew Harrison ------------------------------ Date: 14 Jun 2007 12:44:12 GMT From: bill@cs.uofs.edu (Bill Gunshannon) Subject: Broken TSZ07 Message-ID: <5dcrgrF34gimdU1@mid.individual.net> Some may remember my mentioning having a TSZ07 with problems. I have thoroughly test it and have determined it is either the Tach Sensor on the Take-up motor or the the logic board it connects to. Is there any chance that anyone has a tape drive that these parts could be canabalized from that wold be willing to do it? I assume there were other Tape Units that used the same mechanicals and logic except without the SCSI interface. Obviously, parts from one of these would probably work as well. I will likely ask in other DEC groups as well, but I seem to find the most expertise here. bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves bill@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ Date: Thu, 14 Jun 2007 07:31:35 +0200 From: "P. Sture" Subject: Re: BYPASS privilege !! Message-ID: In article , Tad Winters wrote: > JF Mezei wrote in news:12634$467036b6 > $cef8887a$30475@TEKSAVVY.COM: > > > Bob Koehler wrote: > >> I can't answer this one, but I can tell you that I ran a VMS > >> system with SYSTEM disuser'ed for yeaqrs with no ill affects. > > > > > > I ran with mine not disusered, but with nointeractive, no dialup, no > > batch etc. > > > > However, when I restructured my startup procedures to have spawned > > subprocesses and batch jobs, it was very inconvenient to have those > > restrictions and I re-enabled SYSTEM. > > > > I ran with no restrictions on SYSTEM on several systems for years. I had > set the password to some random long string to which I paid no attention > when I set it. I would generally submit (re)start procedures for MultiNet, > Samba, and even Watcher on occasion, under SYSTEM. If auditors asked, I > always told them it was necessary for the proper operation of the system. > Besides, I had managers who were not part of IT who seemed to always be > allowed to retain SETPRV by going to the CIO. You'd have thought that > wouldn't fly between SOX and HIPAA audits, but it's still like that, and > I'm no longer an employee. BTDT with a project leader who went ballistic when we took ALL privileges away from him. He appeared to view as many privileges as possible as a status symbol. :-( The immediate problem there was that the application testers would receive kits that wouldn't work without BYPASS - aarrgghh! Going back to PDP days, the auditors we had to deal with then had only come across batch systems before. That was an interesting learning curve on both sides. -- Paul Sture ------------------------------ Date: Thu, 14 Jun 2007 07:40:25 +0200 From: "P. Sture" Subject: Re: BYPASS privilege !! Message-ID: In article , Tad Winters wrote: > AEF wrote in news:1181695669.097631.262900 > @x35g2000prf.googlegroups.com: > > > > > [The following is based in part on other posts in this thread.] > > > > > > It sounds like nothing short of pulling the plug would satisfy these > > SOX auditors: > > > > "Can you run the system without running the system?" > > > > "Is there any system function that _absolutley_ requires the > > system to be running?" > > > > > > Still Yet Again: > > > > "Who's guarding the guards?" This will probably always be a > > problem. > > > > > > AEF > > > > I used to tell my supervisor that I _could_ provide the information the > auditors wanted, but since the non-disclosure of the information was part > of the security of the system, I'd have to immediately kill those auditors. > ;-) Hehe. Many years ago, we had a rule with visiting DEC or IBM support guys that we'd alternate between which company would buy lunch. An auditor announced that if he paid, his company would add 10% and charge us as part of their final bill. We paid... -- Paul Sture ------------------------------ Date: Thu, 14 Jun 2007 00:08:28 -0700 From: "Bart.Zorn@gmail.com" Subject: Re: BYPASS privilege !! Message-ID: <1181804908.838700.34200@z28g2000prd.googlegroups.com> On Jun 14, 7:47 am, DeanW wrote: > On 6/11/07, Richard B. Gilbert wrote: > > > JF Mezei wrote: > > > In the end, isn't it still true that for a functional system, you still > > > need to trust at least one system manager who could still wreak havok on > > > your system if he truly wanted to ? > > > > Or can a system truly be locked down to a point where the system manager > > > cannot do his job without supervision from the security folks ? > > > Yes, it can! It may take me days to remember exactly what it's called > > but there is a secondary password that can be required to log in to an > > account; IOW two passwords, only one of which is known to the system > > manager. I've never known a site that actually used this feature but > > it's there! > > $ HELP SET PASS /SYSTEM or HELP SET PASS /SECONDARY: > [snip happened] > The /SECONDARY and /SYSTEM qualifiers are incompatible. > > So you can set secondary passwords... but possibly not for SYSTEM. The /SYSTEM qualifier has nothing to do with the SYSTEM account. $ help set password /system SET PASSWORD /SYSTEM Requires the SECURITY privilege. Changes the system password rather than a user password. [ S n i p . . . ] Bart Zorn ------------------------------ Date: Thu, 14 Jun 2007 10:44:26 -0700 From: DeanW Subject: Re: BYPASS privilege !! Message-ID: <3f119ada0706141044g19b51077k4a6e3e02328be7c7@mail.gmail.com> On 6/14/07, Bart.Zorn@gmail.com wrote: > On Jun 14, 7:47 am, DeanW wrote: > The /SYSTEM qualifier has nothing to do with the SYSTEM account. That should teach me to post after 6pm... ------------------------------ Date: Thu, 14 Jun 2007 09:15:44 -0400 From: Chuck Aaron Subject: Encrypting a pdf file on vms 8.3 Message-ID: Group, Is there VMS software that will encrypt pdf files, and your recommendations please? Thanks in Advance, Chuck ------------------------------ Date: Thu, 14 Jun 2007 08:32:21 -0500 (CDT) From: sms@antinode.org (Steven M. Schweda) Subject: Re: Encrypting a pdf file on vms 8.3 Message-ID: <07061408322106_202003EE@antinode.org> From: Chuck Aaron > Is there VMS software that will encrypt pdf files, and your > recommendations please? There's software to encrypt almost anything, depending on exactly what you want. General stuff: http://h71000.www7.hp.com/doc/83FINAL/6679/6679pro_002.html#enceas http://antinode.org/dec/sw/gnupg.html If you mean the PDF-specific stuff, like in Acrobat, I know of none (but that doesn't prove much). ------------------------------------------------------------------------ Steven M. Schweda sms@antinode-org 382 South Warwick Street (+1) 651-699-9818 Saint Paul MN 55105-2547 ------------------------------ Date: Thu, 14 Jun 2007 13:48:17 -0000 From: IanMiller Subject: Re: Encrypting a pdf file on vms 8.3 Message-ID: <1181828897.196361.169190@i38g2000prf.googlegroups.com> You could use the the built in ENCRYPT and use AES ------------------------------ Date: Thu, 14 Jun 2007 10:00:03 -0400 From: "Richard B. Gilbert" Subject: Re: Encrypting a pdf file on vms 8.3 Message-ID: <467149E3.6080702@comcast.net> Chuck Aaron wrote: > Group, > > Is there VMS software that will encrypt pdf files, and your > recommendations please? > > Thanks in Advance, > Chuck The last time I looked, there was VMS software that would encrypt/decrypt any kind of file. ISTR that the license was about $5K! I looked into it at my last job because we were storing customer's credit card numbers. When management found out what the encryption software cost they decided that the risk was acceptable! ------------------------------ Date: Thu, 14 Jun 2007 10:57:31 -0400 From: "Koska, John C. \(LNG-ALB\)" Subject: RE: Encrypting a pdf file on vms 8.3 Message-ID: <0FE5E87C5F0AE84B8C667FDC5224F6DA0C1636@LNGDAYEXCP01VC.legal.regn.net> I thought Encryption for OpenVMS is built into OpenVMS 8.3 at no extra license purchase cost. Is that correct? :) jck=20 > -----Original Message----- > From: Richard B. Gilbert [mailto:rgilbert88@comcast.net]=20 > Sent: Thursday, June 14, 2007 10:00 AM > To: Info-VAX@Mvb.Saic.Com > Subject: Re: Encrypting a pdf file on vms 8.3 >=20 > Chuck Aaron wrote: > > Group, > >=20 > > Is there VMS software that will encrypt pdf files, and your=20 > > recommendations please? > >=20 > > Thanks in Advance, > > Chuck >=20 > The last time I looked, there was VMS software that would=20 > encrypt/decrypt any kind of file. ISTR that the license was=20 > about $5K! >=20 > I looked into it at my last job because we were storing=20 > customer's credit card numbers. When management found out=20 > what the encryption software cost they decided that the risk=20 > was acceptable! >=20 >=20 >=20 ------------------------------ Date: Thu, 14 Jun 2007 15:01:18 -0000 From: IanMiller Subject: Re: Encrypting a pdf file on vms 8.3 Message-ID: <1181833278.876634.198090@e9g2000prf.googlegroups.com> The ENCRYPT product is free since VMS V8.2 and included in VMS since V8.3 ------------------------------ Date: Thu, 14 Jun 2007 11:34:20 -0400 From: "Koska, John C. \(LNG-ALB\)" Subject: RE: Encrypting a pdf file on vms 8.3 Message-ID: <0FE5E87C5F0AE84B8C667FDC5224F6DA0C169A@LNGDAYEXCP01VC.legal.regn.net> Thank you for the confirmation. I was a tad worried there for awhile. I thought I had missed a dose of reality. :) jck=20 > -----Original Message----- > From: IanMiller [mailto:gxys@uk2.net]=20 > Sent: Thursday, June 14, 2007 11:01 AM > To: Info-VAX@Mvb.Saic.Com > Subject: Re: Encrypting a pdf file on vms 8.3 >=20 > The ENCRYPT product is free since VMS V8.2 and included in VMS since > V8.3 >=20 ------------------------------ Date: Thu, 14 Jun 2007 11:55:56 -0500 From: Ron Johnson Subject: Re: Encrypting a pdf file on vms 8.3 Message-ID: On 06/14/07 08:15, Chuck Aaron wrote: > Group, > > Is there VMS software that will encrypt pdf files, and your > recommendations please? http://www.accesspdf.com/pdftk/ The PDF Toolkit might compile on VMS. If so, it will "natively" encrypt PDF files. -- Ron Johnson, Jr. Jefferson LA USA Give a man a fish, and he eats for a day. Hit him with a fish, and he goes away for good! ------------------------------ Date: Thu, 14 Jun 2007 13:07:26 -0400 From: Chuck Aaron Subject: Re: Encrypting a pdf file on vms 8.3 Message-ID: It appears that the encryption built into vms 8.3 does not encrypt pdf files. Is there software that will work on a VMS and do this? Thanks. > ------------------------------ Date: Thu, 14 Jun 2007 12:23:36 -0500 From: Ron Johnson Subject: Re: Encrypting a pdf file on vms 8.3 Message-ID: On 06/14/07 12:07, Chuck Aaron wrote: > It appears that the encryption built into vms 8.3 does not > encrypt pdf files. Is there software that will work on a VMS > and do this? pdftk does not use external encryption libraries. -- Ron Johnson, Jr. Jefferson LA USA Give a man a fish, and he eats for a day. Hit him with a fish, and he goes away for good! ------------------------------ Date: Thu, 14 Jun 2007 19:26:08 +0200 From: "P. Sture" Subject: Re: Encrypting a pdf file on vms 8.3 Message-ID: In article , Chuck Aaron wrote: > Group, > > Is there VMS software that will encrypt pdf files, and your > recommendations please? > Are they also to be decrypted on VMS? -- Paul Sture ------------------------------ Date: Thu, 14 Jun 2007 12:07:47 GMT From: rdeininger@mindspringdot.com (Robert Deininger) Subject: Re: ES45 Mod 2 Service Bulletin Message-ID: In article , Hal Kuff wrote: >There was a critical ES45 bulletin, riser card or backplane replacement >in 2002 ... anyone remember or have that document? I remember it, but I don't have it. I expect HP services has it. If you didn't have the work done back then, you should. ------------------------------ Date: Thu, 14 Jun 2007 12:04:53 GMT From: rdeininger@mindspringdot.com (Robert Deininger) Subject: Re: help running OpenVMS on Itanium 2 2620 WinXP variant Message-ID: In article , "David Turner, Island Computers" wrote: >Can someone point us in the right direction to getting VMS up and running on >this RX2600? Sure, glad to help if you provide some information. Please identify the system. What does it call itself? You've written "2620 WinXP variant" in the header, and rx2600 in the message body. The first name doesn't mean anything. The first advice is to follow the instructions in the VMS Installation and Upgrade manual. Make sure you have the console device configured correctly. Stick the VMS V8.3 DVD in the drive and boot it. If it doesn't do what you want, post the error message(s). Please don't re-word or re-interpret the messages. :-) ------------------------------ Date: Thu, 14 Jun 2007 09:21:00 -0400 From: "David Turner, Island Computers" Subject: Re: help running OpenVMS on Itanium 2 2620 WinXP variant Message-ID: <1372g5hh2ef0od5@news.supernews.com> Well, before we do anything the orange light is flashing Apparently, from the downloaded docs it means "the system needs attention" We sat with it all night and tried to comfort it but the light is still on Any suggestions as to how we can give it the attention it needs to turn off that bloody flashing orange light? David "Robert Deininger" wrote in message news:rdeininger-1406070805040001@dialup-4.233.149.205.dial1.manchester1.level3.net... > In article , "David Turner, > Island Computers" wrote: > >>Can someone point us in the right direction to getting VMS up and running >>on >>this RX2600? > > Sure, glad to help if you provide some information. > > Please identify the system. What does it call itself? > > You've written "2620 WinXP variant" in the header, and rx2600 in the > message body. The first name doesn't mean anything. > > The first advice is to follow the instructions in the VMS Installation and > Upgrade manual. > > Make sure you have the console device configured correctly. > > Stick the VMS V8.3 DVD in the drive and boot it. If it doesn't do what > you want, post the error message(s). Please don't re-word or > re-interpret the messages. :-) ------------------------------ Date: Thu, 14 Jun 2007 15:53:17 -0000 From: Hein RMS van den Heuvel Subject: Re: help running OpenVMS on Itanium 2 2620 WinXP variant Message-ID: <1181836397.911102.186450@o11g2000prd.googlegroups.com> On Jun 14, 8:21 am, "David Turner, Island Computers" wrote: > Well, before we do anything the orange light is flashing > We sat with it all night and tried to comfort it but the light is still on :-) > Any suggestions as to how we can give it the attention it needs to turn off > that bloody flashing orange light? google: +rx2600 +attention +site:hp.com --> http://docs.hp.com/en/rx2600rx56xx_update/rx2600rx56xx_update.pdf page 7 >> Day Job? >> Stick with it... JF... (an other non-US readers)... just in case the cultural refence is lost... This refers to a common joke / tease here in the US where folks take a stab at a joke possily suggesting the are considering a career being a (night-time) comedian. But I'm sure you made David chuckle, despite his retort. Cheers, Hein. ------------------------------ Date: Thu, 14 Jun 2007 03:44:14 -0400 From: Bill Todd Subject: Re: OT: Carla Message-ID: Tom Linden wrote: > EE Times: Latest News > Did you know?: HP's Compaq deal not a bust As long as you take care not to look very closely. ... > How is HP doing? By the end of its current fiscal year Oct. 31, HP's > annual revenue is predicted to double, to about $100 billion, from $48.9 > billion seven years ago. Well, whoop-de-doo! Of course, a CAGR of about 10% (before any deduction for inflation...) doesn't sound quite as impressive, but still... Then again, HP *bought* most of that growth when it bought Compaq, rather than earned it in the marketplace, so a better metric would be to sum the HP and Compaq revenues before the announcement of the merger and use *that* as the base figure to determine what HP's actual growth has been (wasn't that total around $80 billion in 2001? if so, then in the intervening 6 years Carly's Visionary Accomplishment has grown more like 25%; I don't care enough to check the numbers myself, but anyone tempted to take this article at face value might be well advised to). ... > The revenue jump, especially in hardware, Hmmm - I wonder what hardware that would be, since IIRC HP's non-x86-based system revenues currently don't even match the levels that Compaq's comparable products all by themselves enjoyed prior to the Alphacide. x86, printers, and ink, perhaps? Save for the last, it's hardly surprising that the article prefers to focus on revenues rather than on profits. is a testimony to HP's > resilience as well as its R&D focus--annual R&D spending is about $3.6 > billion What a princely sum! Almost as much as HP spent all by itself before the Compaq acquisition (and not counting for inflation), IIRC - though, as with the earlier numbers, I'm not sufficiently motivated to check my memory - feel free to do so for me if it matters to you. - bill ------------------------------ Date: Thu, 14 Jun 2007 08:28:30 +0200 From: Michael Kraemer Subject: Re: Question for the Group Message-ID: Bob Koehler schrieb: >> >>This story is as old as 1988. >>BTW, a possible reason why VMS wasn't affected might by >>they didn't have TCP back then. Sometimes it seems to be >>an advantage to be behind. > > > We certainly did have IP back then. I think UCX came later, 1989 ? > >>A little googling would reveal that around the same >>time there certainly were a lot of successful VMS hacks going on, >>some of them are listed at >>http://www.geocities.com/SiliconValley/Lab/7378/hacker.htm >>Wasn't there this easy possiblity to break in with some >>field engineer password ? > > > Knowing a password will allow you to break in to any computer. If it's particularly easy to obtain such passwords, it is an inherent weakness of that platform. But I assume they have fixed that by now. ------------------------------ Date: Thu, 14 Jun 2007 09:20:28 +0000 (UTC) From: m.kraemer@gsi.de (Michael Kraemer) Subject: Re: Question for the Group Message-ID: In article <1181738553.834005.64810@a26g2000pre.googlegroups.com>, BaxterD@tessco.com writes: > > In my experience, the type of incident described above are simply the > result of poor cluster configuration. If you don't provide > redundency then these things will happen to anyone. The failures of > the technical staff who designed the configuration, be they local, HP/ > Compaq/Digital, cannot and should not be blamed on the OS. The guys in charge back then were quite some bigots, so I"m sure they did a lot to make their beloved system look better than the rest. So it might well be a weakness of the OS. > Since UNIX systems tend to share nothing, a "freeze" is sufficient for > them to survive, however in Real Clustering, nodes share everything > therefore a break in communication requires a much more complex > response. A proper configuration would ensure that communication > could not be interupted. I assess all that from the point of view of an end user. The experience was that the availability of VMS services wasn't any higher than that of Unix in our case. YMMV, of course. > I dont know how long ago this incident occurred, some of them occured more than a decade ago, the last one or two just last year, IIRC. Last time it took about half a day to get back to operation. > but if the > VAXStations clustered over the network died, then it was probably > because they were booted over the network from the main cluster system > disk. Maybe, but my observation simply was that such a VMS cluster is more vulnerable than Unix when exposed to the same bad network conditions. > Had they been booted from internal system disks, (and no > VOTES) they would have simply hung until the re-connection interval > timed out. If the connection was restored within this time then > they would simply have continued work. > > It really pisses me off when.... (No! I won't start!) I don't know what reason you have to choose rude language. ------------------------------ Date: Thu, 14 Jun 2007 09:29:03 +0000 (UTC) From: m.kraemer@gsi.de (Michael Kraemer) Subject: Re: Question for the Group Message-ID: In article <00A69111.F9A6C965@SendSpamHere.ORG>, VAXman- @SendSpamHere.ORG writes: > > Such things happened frequently more than a decade ago. Somebody pulled > the electric cord from the wall socket and there was darkness all around. > Similar incidents occured several times as recent as this year. Which > raises some doubts about the claimed superiority of the electric lamp. now this is an extremely relevant advice thank you. Should I forward it to the guys in charge next time the system goes down ? ------------------------------ Date: Thu, 14 Jun 2007 05:39:22 -0400 From: JF Mezei Subject: Re: Question for the Group Message-ID: <798d4$46710cd0$cef8887a$633@TEKSAVVY.COM> Michael Kraemer wrote: > The guys in charge back then were quite some bigots, > so I"m sure they did a lot to make their beloved system look > better than the rest. So it might well be a weakness of the OS. When I setup a disaster recovery site with only 2 vaxes, I had to spend some time explaining why it was impossible to have an automatic failover. And I hadn't even looked at all the possible failure modes, but showed them enough modes to convince them that it was required to have the cluster freeze to prevent data corruption. One day, while I was away, the ethernet transponder for the production node failed. The backup node immediatly froze, as planned. And users had their LAT sessions disconnected from the production node. But nobody had expected that production node to continue to receive/process SWIFT messages via its synchronous port (which was undamaged). When the backup node was brought up in production mode some minutes later, it couldn't log into SWIFT because the session was still active on the primary synchronous line. They disconneced it and logged in. However, during that time, transactions had continued to be received by the production node. So the backup node didn't have a record of those. (they were recoverable from SWIFT so it wasn't the end of the world). However, it really showed the perils of data inconsistancy with disaster recovery schemes. In the post mortem, I had explained that if we had had a 3rd vax, then the production vax would have been the one to be frozen due to quorum loss and the remaining 2 would have continued to function, the backup vax could have been converted to production and no transactions would have been omitted. The fact that VMS nodes go down or crash voluntarily in cluster problems is deep down a desirable thing, even though one swears with many bad words when it happens. Depending on how your data is spread, you can design very robust clusters with a minimum of 3 nodes and judicial use of the voting scheme. If you do not do the voting scheme carefully, you end up with clusters losing quorum at the mere etherner problem and causing nodes to voluntarily crash once connectivity is restored to ensure that such nodes starts off with a clean slate and no "dirty" data sctructures for devices, locks etc. The fact that VMS nodes hang/crash during such cluster events is due to VMS engineers having looked at the real issues and having concluded that this is absolutely necessary to preserve cluster integrity. And that is very very very important when you have real data on those disks. ------------------------------ Date: Thu, 14 Jun 2007 10:04:26 +0000 (UTC) From: m.kraemer@gsi.de (Michael Kraemer) Subject: Re: Question for the Group Message-ID: In article <6bbcc$466fbacc$cef8887a$16700@TEKSAVVY.COM>, JF Mezei writes: > > We just want HP to at least try. We are not asking to subsidize VMS. But > we are asking for VMS to be given the freedom to use its own money to > market itself. We are asking that VMS be allowed to push press releases > through the official HP channels. We are asking for VMS to be mentioned > by HP corporate manegement as a product with lots of potential etc. > > If HP really does give it a good try and it still doesn't work, then we > can all know what will happen next. But if HP doesn't try, or gives it a > dishonest try (designed to fail, but also designed so they can tell us > they tried) then it is not a fair treatment. > The board could have thrown La Carly out when she started to want to > distract from her poor performance by getting married to Compaq. If HP > wasn't interested in getting all the baggage that came with Compaq, they > shouldn't have bought Compaq. Well, the tiny footprint of VMS certainly doesn't make it baggage, just a minor nuisance. If you look at HP-UX, which is also part of BCS, I presume, the HP-UXers would have as much reason to bitch as the VMSers. Both are just a small fraction of a business unit which is another small fraction of HP total. Compare that with the situation of, say, early 1990s, when HP still was a high-tech company: HP "Snake" workstations, HP VUE desktop, PA-RISC boxes entering the data centres, etc. If I were an HP-UX bigot, I would walk away and weep. But such things happen when a company decides to go for the mainstream rather than for technical excellence. And the bottom line for the share holders doesn't look that bad, unfortunately. HP has about $100B annual revenue, which puts them ahead of IBM by some 10%. Their profit margin, 8% on average, isn't that far below IBMs (10+x% last time I checked). > Consider how Apple was able to bring its Macintosh back to life. It > still isn't too late to put VMS back on a growth curve. I don't think this is easily repeatable in your case. What happened with the Mac was more or less a whole-body transplantion (including the brain, i.e. Jobs Himself). They supplanted the aging 68k CPU with a more powerful PPC (but that happened already under Spindler, I guess). A similar move for VMS' hardware you would probably applaude, I presume. But they also put an entirely new Unix OS under the hood, relegating the aging classic one to some compatibility box. I guess you don't want to see the same happen to VMS. Plus they added some cool apps Jobs knew the people would like, i.e. Garageband and all that i-stuff. All that took surely a lot of money and endeavour, which was justified since the Mac was Apples main product. If their last u-turn towards intel will pan out for the Mac in the long run remains to be seen. Maybe it will simply be sucked up by the wintel world and Apple just contineus selling iPods and other lifestyle toys. ------------------------------ Date: Thu, 14 Jun 2007 10:23:11 GMT From: VAXman- @SendSpamHere.ORG Subject: Re: Question for the Group Message-ID: <00A691D4.F218F2D3@SendSpamHere.ORG> In article , m.kraemer@gsi.de (Michael Kraemer) writes: > > >In article <00A69111.F9A6C965@SendSpamHere.ORG>, VAXman- @SendSpamHere.ORG writes: >> >> Such things happened frequently more than a decade ago. Somebody pulled >> the electric cord from the wall socket and there was darkness all around. >> Similar incidents occured several times as recent as this year. Which >> raises some doubts about the claimed superiority of the electric lamp. > >now this is an extremely relevant advice thank you. >Should I forward it to the guys in charge next time the system >goes down ? Why not? Then, perhaps, they'll have a better understanding as to why their cluster goes down when they yank its lifeline. It has nothing to do with inferiority. If they insist on pulling this lifeline, you might have them try increasing the value or RECNXINTERVAL. This is sort of the "how long can your cluster hold it breath" parameter. -- VAXman- A Bored Certified VMS Kernel Mode Hacker VAXman(at)TMESIS(dot)COM "Well my son, life is like a beanstalk, isn't it?" ------------------------------ Date: Thu, 14 Jun 2007 11:40:07 +0100 From: Anton Shterenlikht Subject: Re: Question for the Group Message-ID: <20070614104007.GB27339@mech-aslap33.men.bris.ac.uk> On Tue, Jun 12, 2007 at 02:08:56AM -0700, IanMiller wrote: > To answer the original question - I think the complaints in this > newsgroup have two negative effects > > 2. any new people may have heard of VMS and read this newsgroup to > find out more would gain a very negative picture. There is very little > positive news here. that's true in my case to some extent > > My impression is that the majority of posts here appear to be off > topic and/or negative. I don't read newsgroups dedicated to other > operating systems to know if this is typical for this sort of > newsgroup. If anyone here does read newsgroups for other OS then it > would be interesting to know their thoughts. I read two FBSD mailing lists (there are maybe 30 or more lists in total). On these two OT are rare. The negative and gloomy posts have a very different content, mostly those are from very new people who struggle to understand the basics or get some errors during install. On VMS list, in contrast, most people are very experienced so this sort of frustration does not occur. The other difference, of course, is the pace of development of FBSD and ported applications, which means there is something new to discuss every week. -- Anton Shterenlikht Room 2.6, Queen's Building Mech Eng Dept Bristol University University Walk, Bristol BS8 1TR, UK Tel: +44 (0)117 928 8233 Fax: +44 (0)117 929 4423 ------------------------------ Date: Thu, 14 Jun 2007 11:48:00 +0100 From: "Richard Brodie" Subject: Re: Question for the Group Message-ID: "Michael Kraemer" wrote in message news:f4r18s$t91$1@lnx107.hrz.tu-darmstadt.de... > Maybe, but my observation simply was that such a VMS cluster > is more vulnerable than Unix when exposed to the same bad network conditions. Yes, that's a given. A fully redundant network is highly recommended for a cluster. Configurations where you scatter a load of client workstations about the building, and run a dodgy network between them - not really a good fit. ------------------------------ Date: 14 Jun 2007 08:37:49 -0500 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: Question for the Group Message-ID: In article , Michael Kraemer writes: > Bob Koehler schrieb: > >>> >>>This story is as old as 1988. >>>BTW, a possible reason why VMS wasn't affected might by >>>they didn't have TCP back then. Sometimes it seems to be >>>an advantage to be behind. >> >> >> We certainly did have IP back then. > > I think UCX came later, 1989 ? Probably, but I hardly think that stopped us from having a better IP stack earlier. And we were no where near alone. > > If it's particularly easy to obtain such passwords, > it is an inherent weakness of that platform. > But I assume they have fixed that by now. Long since fixed. But at the same time MS-DOS didn't implement passwords and the installation passowrds of some UNIX might have been as easy to obtain. ------------------------------ Date: 14 Jun 2007 13:46:28 GMT From: bill@cs.uofs.edu (Bill Gunshannon) Subject: Re: Question for the Group Message-ID: <5dcv5kF33ir71U1@mid.individual.net> In article , koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes: > In article , Michael Kraemer writes: >> Bob Koehler schrieb: >> >>>> >>>>This story is as old as 1988. >>>>BTW, a possible reason why VMS wasn't affected might by >>>>they didn't have TCP back then. Sometimes it seems to be >>>>an advantage to be behind. >>> >>> >>> We certainly did have IP back then. >> >> I think UCX came later, 1989 ? > > Probably, but I hardly think that stopped us from having a better IP > stack earlier. And we were no where near alone. > >> >> If it's particularly easy to obtain such passwords, >> it is an inherent weakness of that platform. >> But I assume they have fixed that by now. > > Long since fixed. But at the same time MS-DOS didn't implement > passwords and the installation passowrds of some UNIX might have > been as easy to obtain. I have never seen any version of Unix with an "installation passowrd". Every version I have ever installed prompted for a root pasword at some point in the install procedure. Actually, I don't remember any version of Unix I have ever used that had default passwords for any account. bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves bill@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ Date: Thu, 14 Jun 2007 12:00:24 -0500 From: Ron Johnson Subject: Re: Question for the Group Message-ID: On 06/14/07 08:37, Bob Koehler wrote: > In article , Michael Kraemer writes: [snip] > >> If it's particularly easy to obtain such passwords, >> it is an inherent weakness of that platform. >> But I assume they have fixed that by now. > > Long since fixed. But at the same time MS-DOS didn't implement > passwords and the installation passowrds of some UNIX might have > been as easy to obtain. And Windows still doesn't make you use a password or make your company or family use anything other than the default (Administrator) account. -- Ron Johnson, Jr. Jefferson LA USA Give a man a fish, and he eats for a day. Hit him with a fish, and he goes away for good! ------------------------------ Date: Thu, 14 Jun 2007 07:40:21 -0700 From: Trefor Subject: RPC and clnttcp_create Message-ID: <1181832021.531805.48610@q19g2000prn.googlegroups.com> We have a OpenVMS system with multiple IP addresses. I need to make a RPC client call, but to force the call to be made from one of the IP addresses. So - my Plan 1 was to A. create a socket, B. bind it to one the desired local IP address, and C. use the clnttcp_create and set the sockp parameter to my socket. clnttcp_create does not fail, but ..., I get an error "ENOTCONN" from clnt_call within the generated code for the RPC call. So Plan 2 A. create a socket, B. bind it to one the desired local IP address, C. connect to the remote system, and D. use the clnttcp_create. But the connect fails - EADDRNOTAVAIL and I've failed to get past this... Note: If I use use clnttcp_create and set the sockp parameter to RPC_ANYSOCK then the RPC call works, but not from the desired IP address on the client. So. How do I configure a socket to make it useable in a call to clnttcp_create? I've tried to look for some example code but failed.. Any help greatly appreciated.... Trefor ------------------------------ Date: Thu, 14 Jun 2007 14:48:38 GMT From: "Jamez" Subject: SSL FTP Message-ID: Is there any software that can do SSL FTP from OpenVMS? ------------------------------ Date: Thu, 14 Jun 2007 11:45:38 -0400 From: norm.raphael@metso.com Subject: Re: SSL FTP Message-ID: I was looking for FTPS a short time ago on OpenVMS and could not find any implementation. Nor HP TCPIP nor Multinet TCPIP nor FTPware TCPIP has it, although Process Software said that Multinet had looked at it. If you get a yes, pls let me know, although we've got to sending it to another box to send FTPS. "Jamez" wrote on 06/14/2007 10:48:38 AM: > Is there any software that can do SSL FTP from OpenVMS? > > ------------------------------ Date: Thu, 14 Jun 2007 15:18:02 GMT From: VAXman- @SendSpamHere.ORG Subject: TCPIP Services SFTP Message-ID: <00A691FE.22B769DC@SendSpamHere.ORG> When I use sftp to transfer from my Powerbook or my Linux box to VMS, I cannot get the session to terminate when I am done. I will type 'exit' at the sftp> prompt and the VMS session hangs. This doesn't happen if I sftp between the Powerbook and linux machines. I find myself entering a CTRL-Z and then doind a 'ps' to find the sftp process and I then 'kill' it off. Is there some magical incantation to get the TCPIP Services sftp to terminate the connection? -- VAXman- A Bored Certified VMS Kernel Mode Hacker VAXman(at)TMESIS(dot)COM "Well my son, life is like a beanstalk, isn't it?" ------------------------------ Date: Thu, 14 Jun 2007 12:27:01 -0400 From: Chuck Aaron Subject: vms 8.3 to Red Hat Linux AS4 Message-ID: Group, Is there software or a way to convert VMS 8.3's sysuaf.dat and rightslist.dat files onto a windows server running the same application using RHL AS4 so these two files can be sync'd up and utilized? Thanks in Advance. ------------------------------ Date: Thu, 14 Jun 2007 11:52:32 -0500 From: Ron Johnson Subject: Re: vms 8.3 to Red Hat Linux AS4 Message-ID: On 06/14/07 11:27, Chuck Aaron wrote: > Group, > > Is there software or a way to convert VMS 8.3's sysuaf.dat and > rightslist.dat > files onto a windows server running the same application using RHL AS4 so > these two files can be sync'd up and utilized? How can a Windows Server run RHAS4? -- Ron Johnson, Jr. Jefferson LA USA Give a man a fish, and he eats for a day. Hit him with a fish, and he goes away for good! ------------------------------ Date: Thu, 14 Jun 2007 13:09:05 -0400 From: Chuck Aaron Subject: Re: vms 8.3 to Red Hat Linux AS4 Message-ID: Sorry, I meant to say a Dell running RHAS4. Chuck Ron Johnson wrote: > On 06/14/07 11:27, Chuck Aaron wrote: >> Group, >> >> Is there software or a way to convert VMS 8.3's sysuaf.dat and >> rightslist.dat >> files onto a windows server running the same application using RHL >> AS4 so >> these two files can be sync'd up and utilized? > > How can a Windows Server run RHAS4? > ------------------------------ Date: Thu, 14 Jun 2007 12:36:18 -0500 From: Ron Johnson Subject: Re: vms 8.3 to Red Hat Linux AS4 Message-ID: I see you've been well conditioned by The Evil That Is Redmond. Anyway... SYSUAF.DAT is simply an indexed file. Standard VMS languages can read it, and CONVERT can make a text copy that Perl or Python can convert to a properly formatted /etc/passwd file. You can probably do the same to RIGHTSLIST.DAT, but Unix's security model is so different that it might not be useful. P.S. - Please don't top-post. On 06/14/07 12:09, Chuck Aaron wrote: > Sorry, > > I meant to say a Dell running RHAS4. > > Chuck > > Ron Johnson wrote: >> On 06/14/07 11:27, Chuck Aaron wrote: >>> Group, >>> >>> Is there software or a way to convert VMS 8.3's sysuaf.dat and >>> rightslist.dat >>> files onto a windows server running the same application using RHL >>> AS4 so >>> these two files can be sync'd up and utilized? >> >> How can a Windows Server run RHAS4? >> -- Ron Johnson, Jr. Jefferson LA USA Give a man a fish, and he eats for a day. Hit him with a fish, and he goes away for good! ------------------------------ Date: Thu, 14 Jun 2007 08:05:16 +0200 From: "P. Sture" Subject: Re: VMS analogue of FBSD and linux hier(7) man pages Message-ID: In article <4670BCDF.3020600@comcast.net>, "Richard B. Gilbert" wrote: > Ron Johnson wrote: > > On 06/13/07 21:26, Richard B. Gilbert wrote: > > [snip] > > > >> same system disk although, in principle, you could have as many as > >> thirteen. A cluster could have one system disk for each node in the > >> cluster or small cluster could share a common system disk (disk > >> loading would make this impractical in a large cluster) imagine > >> fourteen nodes beating up the same system disk, it would run like a dog! > > > > > > Only if the page/swap files, or some other important system files (audit > > logs, etc) were on it, no? > > > > How long would it take to boot thirteen nodes from the same system disk? > You may assume that each node has its own page/swap disk. I haven't done that with VAX systems, but a couple of Alpha 4100s could serve 20+ Alpha PWS systems with page, swap and dump files on their local disks quite nicely. Caveat: before implementing DOSD, a workstation rebooting would initiate a full shadow copy. Prior to that, if all the workstations rebooted at the same time due to a network outage, DECnet/Plus would timeout "somewhere" during startup on one of more systems and not come up fully functional. The initial workaround was either to wander around 3 or 4 floors starting the workstations manually (Yuck), or drop a shadow member of the system disk until all the workstations had booted, and bring it in then. -- Paul Sture ------------------------------ Date: Thu, 14 Jun 2007 11:46:57 GMT From: rdeininger@mindspringdot.com (Robert Deininger) Subject: Re: VMS analogue of FBSD and linux hier(7) man pages Message-ID: In article <5dacfrF334k19U1@mid.individual.net>, bill@cs.uofs.edu wrote: >In article , > koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes: >> In article <5d7l6eF342pf9U1@mid.individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes: >>> >>> I am sure someone will correct me if I am wrong, :-) but I don't >>> think the VMS filesystem would be considered "hierarchical" which >>> would preclude such a listing. >> >> Why not? > >Well, I thought the locations of files was not particularly relevant and >the system would function just fine if the filesystem was kept completely >flat like, say, CPM. I thought LOGICALS were used to identify all the >important things like SYSEXE, etc. Is there anything in VMS that expects >files to be in any particular hierarchy? Or can the system be setup with >the directories in pretty much any location? The first few layers of boot and startup code have hard-wired assumptions about where key files are. The logical names that point to most of the standard directories aren't defined until relatively late, in the STARTUP process (which runs STARTUP.COM and a whole list of subsidiary procedures like SYSTARTUP_VMS.COM). If you stop in SYSBOOT (using the conversational boot flag), and SHOW/STARTUP SET/STARTUP OPA0: ! instead of STARTUP.COM SET WRITESYSPARAMS 0 CONTINUE ...you'll see how little has been done. Most of the familiar environment isn't there yet. ------------------------------ Date: Thu, 14 Jun 2007 11:56:22 GMT From: rdeininger@mindspringdot.com (Robert Deininger) Subject: Re: VMS analogue of FBSD and linux hier(7) man pages Message-ID: In article <4670BCDF.3020600@comcast.net>, rgilbert88@comcast.net wrote: >Ron Johnson wrote: >> On 06/13/07 21:26, Richard B. Gilbert wrote: >> [snip] >> >>> same system disk although, in principle, you could have as many as >>> thirteen. A cluster could have one system disk for each node in the >>> cluster or small cluster could share a common system disk (disk >>> loading would make this impractical in a large cluster) imagine >>> fourteen nodes beating up the same system disk, it would run like a dog! >> >> >> Only if the page/swap files, or some other important system files (audit >> logs, etc) were on it, no? >> > >How long would it take to boot thirteen nodes from the same system disk? > You may assume that each node has its own page/swap disk. With modern systems, disks, and networks, OS startup is fast. The main VMS development/build cluster has recently consisted of 19 nodes, a mixture of VAX, Alpha, and Integrity. (This isn't supported, but it works, and people who do their own support do run 3-architecture clusters. We do our own support. :-) The Alpha (nothing older than a DS10, IIRC) and Integrity systems all boot from FibreChannel disks (1 or 2 Gb/sec). There's typically 1 Alpha boot disk and 1 Integrity boot disk. I haven't timed them precisely, but I believe they are all booted in 4 or 5 minutes. The VAXes boot from CI-based disks, and take longer. There are a LOT of disks. They are mostly mounted outside of the startup process, and they are NOT all mounted on every node in 4 or 5 minutes! The Alpha and Integrity systems usually spend more time doing HW- and FW-based diagnostic tests than they do booting VMS. ------------------------------ Date: Thu, 14 Jun 2007 13:16:52 +0000 (UTC) From: moroney@world.std.spaamtrap.com (Michael Moroney) Subject: Re: VMS analogue of FBSD and linux hier(7) man pages Message-ID: Ken Fairfield writes: >At my previous, previous employment, we had 2 Alpha 4100's on one >system disk, and 3 VAX 6000-class machines on another (of course) As an exercise in uselessness, I once created a system disk that was bootable on both VAX and Alpha, just to prove it could be done. >it's a set of trade offs. But I don't believe there is any >architectural limit to the number of nodes using a single system >disk (other than supported cluster sizes and the maximum system >specific root being SYSFF, i.e., 255...). From CLUSTER_CONFIG.COM: $max_root = %XFFFF Of course, CI clusters have a hardware limit of SYS0-SYSF. ------------------------------ Date: 14 Jun 2007 08:42:45 -0500 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: VMS analogue of FBSD and linux hier(7) man pages Message-ID: In article <5dacfrF334k19U1@mid.individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes: > > Well, I thought the locations of files was not particularly relevant and > the system would function just fine if the filesystem was kept completely > flat like, say, CPM. I thought LOGICALS were used to identify all the > important things like SYSEXE, etc. Is there anything in VMS that expects > files to be in any particular hierarchy? Or can the system be setup with > the directories in pretty much any location? Installation and update procedures have explicit knowledge of the layout of VMS. The amount of disk space not consumed by unecessary duplication of files from multiple boot roots on a single disk is not readily separated from the hierarchy. (Look at the confusion that was such a disk under VMS 4). Some logicals are defined by the kernel during boot and you can get in trouble if the kernel can't define them correctly. ------------------------------ Date: 14 Jun 2007 08:49:48 -0500 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: VMS analogue of FBSD and linux hier(7) man pages Message-ID: <1KsxPeDNJ2yu@eisner.encompasserve.org> In article , Rich Alderson writes: > > You are correct, even though someone incorrectly said otherwise. I'll explain > below. Look below, you goofed. > > Yes, in so far as that goes, but it does not address the main point. > > The Unix-related family of operating systems use a single hierarchical > filesystem, as did (soon again to be does) Multics. There is a single system > root directory, under which all other directories are found; other filesystems > may exist on subsidiary disks or partitions of disks, but they are attached to > directories located within the root hierarchy. Tying multiple fileystems into a single directory hierachry does not invalidate systems which have multiple hierachies from being hierarchical. > > VMS started out with a non-hierarchical filesystem (now called ODS-1) which was > identical to that of RSX-11{A,B,C,D,M,S} and similar to those of Tops-10 and > TSS/8 (Tops-10 inspired), and IIRC those of the 18-bit family. (Tops-10 added > "SubFile Directories" eventually, so some hierarchicality, but this was not > very cleanly done. IMAO.) No. VMS 1.x in first ship supported ODS-2 with hiearchical directories. ODS-1 was, and still is, supported on the VAX version to assist in porting RSX-11M applications. Early versions of VMS relied somewhat on RSX and took advantage of ODS-2 layouts that resembled ODS-1, but they were on ODS-2 disks. > At some point in its history, IIRC v4, VMS adopted a hierarchical model for its > filesystems, on a per-filesystem basis, similar though not identical to the > model of TENEX and Tops-20: There is a root directory for each filesystem, but > mounting a filesystem does not require that it be associated in any way with a > directory on a central filesystem. Not "at some point in its history". First ship was that way. But your right at TOPS-20 not getting it done very well. ------------------------------ Date: 14 Jun 2007 10:23:11 -0500 From: cornelius@encompasserve.org (George Cornelius) Subject: Re: VMS analogue of FBSD and linux hier(7) man pages Message-ID: <+T3s37JdA2dP@eisner.encompasserve.org> In article , "P. Sture" writes: > In article <4670BCDF.3020600@comcast.net>, > "Richard B. Gilbert" wrote: >> How long would it take to boot thirteen nodes from the same system disk? Don't recall, but then my cluster has been up for awhile: View of Cluster from system ID 16604 node: ROSRVX 14-JUN-2007 09:46:59 lqqqqqqqqqqqqqqqqqqqqqqqwqqqqqqqqqklqqqqqqqqqqqqqqqqqqqk x SYSTEMS x MEMBERS xx CLUSTER x tqqqqqqqqwqqqqqqqqqqqqqqnqqqqqqqqqutqqqqqqqqqqqqqqqqqqqu x NODE x SOFTWARE x STATUS xx FORMED x tqqqqqqqqnqqqqqqqqqqqqqqnqqqqqqqqqutqqqqqqqqqqqqqqqqqqqu x ROSRVX x VMS V7.3-2 x MEMBER xx 18-JUL-1998 23:22 x x ROSRVC x VMS V7.3-2 x MEMBER xmqqqqqqqqqqqqqqqqqqqj x ROSRVB x VMS V7.3-2 x MEMBER x x ROSRVZ x VMS V7.3-2 x MEMBER x x ROSRVN x VMS V7.3-2 x MEMBER x x ROSRVY x VMS V7.3-2 x MEMBER x x ROSRVM x VMS V7.3-2 x MEMBER x mqqqqqqqqvqqqqqqqqqqqqqqvqqqqqqqqqj [here the boot disk only serves seven nodes]. The truth is it takes a long time to boot a single node, but that is because STARTUP has to mount a _ton_ of FC based shadow sets and there always seem to be path switches triggered by the mount operations. [...] > Caveat: before implementing DOSD, a workstation rebooting would initiate > a full shadow copy. You did have volume shadowing licenses for the workstations, and were setting up a path the the license database in SYLOGICALS, I assume? Or are you talking about the fact that the writes to SYSDUMP.DMP only go to the original (boot) member, after which a shadow merge is forced in order to get the shadow members back into synchronization? -- George Cornelius cornelius(at)eisner.decus.org cornelius(at)mayo.edu ------------------------------ Date: 14 Jun 2007 10:34:15 -0500 From: cornelius@encompasserve.org (George Cornelius) Subject: Re: VMS analogue of FBSD and linux hier(7) man pages Message-ID: In article , moroney@world.std.spaamtrap.com (Michael Moroney) writes: > Ken Fairfield writes: >>At my previous, previous employment, we had 2 Alpha 4100's on one >>system disk, and 3 VAX 6000-class machines on another (of course) > > As an exercise in uselessness, I once created a system disk that was > bootable on both VAX and Alpha, just to prove it could be done. > >>it's a set of trade offs. But I don't believe there is any >>architectural limit to the number of nodes using a single system >>disk (other than supported cluster sizes and the maximum system >>specific root being SYSFF, i.e., 255...). > > From CLUSTER_CONFIG.COM: > > $max_root = %XFFFF > > Of course, CI clusters have a hardware limit of SYS0-SYSF. Is this actually true? On a VAX you were limited to passing the system root in as the upper four bits of R5, but on Alpha would it not be possible to specify I higher number? I know there is a max possible CI node number (15, I assume), but that is not what we are talking about here. -- George Cornelius cornelius(at)eisner.decus.org cornelius(at)mayo.edu ------------------------------ Date: Thu, 14 Jun 2007 16:58:32 +0000 (UTC) From: david20@alpha2.mdx.ac.uk Subject: Re: VMS analogue of FBSD and linux hier(7) man pages Message-ID: In article , "P. Sture" writes: >In article , david20@alpha2.mdx.ac.uk >wrote: > >A good summary David. > >Just one nitpick, for accuracy: > >> Logical names such as SYSEXE are then used to point at both the >> >> SYSEXE directory in the system-specific root for a particular node and the >> SYSEXE directory in the cluster-common root > >Should read as > >"Logical names such as SYS$SYSTEM are the used to point at both the >SYSEXE directory in the system-specific root for a particular node and >the SYSEXE directory in the cluster-common root" > Yes thanks good catch. David Webb Security team leader CCSS Middlesex University >-- >Paul Sture ------------------------------ Date: Thu, 14 Jun 2007 03:21:23 -0400 From: Bill Todd Subject: Re: What's the implication of this development if any... Message-ID: Ken Robinson wrote: > Headline: Unisys Hires Former HP Exec to Head Up Server Development > (from the Linux Beacon) > > Start of story: > > "Server maker Unisys has hired away a top executive from > Hewlett-Packard's Business Critical Systems division to take over the > management of its own Systems & Technology Group. > > Rich Marcello, who had a long career at the former Digital Equipment > and who stayed in charge of various server units as Digital was eaten > by Compaq in 1998 and then Compaq was eaten by HP in 2001, is now > president of the Unisys Systems & Technology Group, and reports > directly to Unisys president and chief executive officer, Joe McGrath. > Marcello is replacing Leo Daiuto, who has spent 39 years at Unisys > (and its predecessors) and is now retiring." > > The rest of the story can be read at > Good for Rich - I hope he'll be happier at Unisys than he appeared to be at cHumPaq. While he never visibly stuck his neck far enough out supporting VMS to get it chopped off, my impression was that he did what he could - and I wondered whether he finally left HP of his own accord because of accumulating disgust with his employer (I suspect he was well-enough off not to have to worry about finances, and competent enough not to have to worry about other employment should he want it). Unisys has more interesting x86-64 hardware out than most competitors: they have for years now had medium-sized (up to at least 16 sockets, which should now support 64 cores) x86 (and then x86-64) systems that out-performed Itanics with the same core counts (as well as out-performing the IBM X3 'Hurricane'-based x86 competition at anything over 4 sockets, IIRC). That makes them a leader in exactly the technology toward which the market is moving, even if until now they may have been too far out in front to have felt much benefit from it. They and (or perhaps through) NEC are also partnered with Stratus in developing the kind of Tandem-like high-availability hardware-fault-tolerant x86 systems that Stratus pioneered. So they certainly have the hardware underpinnings required to do well - especially in the interim before Intel's presumably-improved Common System Interconnect (or whatever CSI stands for) hits the streets late next year or in 2009. Too bad they don't have VMS to run on that hardware, but Solaris could be an interesting option. - bill ------------------------------ Date: Thu, 14 Jun 2007 05:34:02 -0700 From: Andrew Subject: Re: ZFS and OS X 10.5 (Leopard) Message-ID: <1181824442.283512.191780@i13g2000prf.googlegroups.com> On 13 Jun, 07:19, "P. Sture" wrote: > > > In article , > Ron Johnson wrote: > > > > > On 06/12/07 14:25, Ron Johnson wrote: > > > On 06/12/07 13:22, P. Sture wrote: > > > [snip] > > > >>> And by allowing ZFS to be ported to OS X, they are again expanding > > >>> mindshare: "ZFS on my MacBook Pro is great. For that server I need > > >>> to build at work, it'll be even better than Linux and ext3 (which is > > >>> what I was planning to build)". > > > >> As a point of information, ZFS isn't due on OS X until the Leopard > > >> release later this year. > > > > Well, yes, it *is* a bit of a forward-looking statement, but Apple isn't > > > in the habit of promising the moon and delivering mud. > > > Guess I'd better stop believing rumors! > > >http://www.informationweek.com/story/showArticle.jhtml?articleID=1999... > > =RSSfeed_IWK_News > > > An Apple official on Monday said Sun Microsystems' open-source > > file system would not be in the next version of the Mac > > operating system, contradicting statements made last week > > by Sun's chief executive. > > Yep. I've just watched the Sun video showing that Jonathan Schwartz > (Sun's CEO) announced that ZFS would be _the_ file system on Leopard. > > He was out of order there. It's highly unlikely that a brand new file > system would be anything but an optional extra at this stage. > > For a more up to date report on the state of ZFS within Leaopard, see > > http://www.theregister.co.uk/2007/06/12/sun_apple_zfs_denial/ > > but we'll have to wait to see what actually ships. > > Let's not forget what happened to Spiralog :-) > > PS, and more on topic for this news group, Schwartz did have a positive > message about the penetration of OpenOffice in schools and universities. > > -- > Paul Sture Apple have confirmed that ZFS will be in OS-X but that it will not be the default filesystem. Regards Andrew ------------------------------ End of INFO-VAX 2007.322 ************************