INFO-VAX Wed, 09 May 2007 Volume 2007 : Issue 254 Contents: Re: A jihad on the Hobbyist programme ? Re: Anyone know why the Alpha market is so so quiet? Re: Anyone know why the Alpha market is so so quiet? Re: Anyone know why the Alpha market is so so quiet? Re: Digital *is* a Software Company? (Was Re: Sun Studio 11 (Solaris IDE)) Re: Digital *is* a Software Company? (Was Re: Sun Studio 11 (Solaris IDE)) RE: Digital *is* a Software Company? (Was Re: Sun Studio 11 (Solaris IDE)) Re: Digital *is* a Software Company? (Was Re: Sun Studio 11 (Solaris IDE)) Re: Digital *is* a Software Company? (Was Re: Sun Studio 11 (Solaris IDE)) IDE)) Re: Digital *is* a Software Company? (Was Re: Sun Studio 11 (Solaris IDE)) IDE)) For Sue - HP Proliant Server Problems OpenVMS, Oracle, C/C++ software engineer position in colorado springs Re: PID for detached process. Re: PID for detached process. Re: PID for detached process. Re: PID for detached process. Seeking Terminal Server Manager (TSM) manuals Re: Semi-OT: 4K Disk Blocks? Re: Semi-OT: 4K Disk Blocks? Re: Semi-OT: 4K Disk Blocks? Re: Semi-OT: 4K Disk Blocks? Re: SET CONF SMTP/QUEUES Re: SET CONF SMTP/QUEUES Re: Sun Studio 11 (Solaris IDE) Re: Sun Studio 11 (Solaris IDE) Re: Sun Studio 11 (Solaris IDE) Re: Sun Studio 11 (Solaris IDE) Re: SYSMAN problem ---------------------------------------------------------------------- Date: 9 May 2007 10:48:43 -0700 From: Wilm Boerhout Subject: Re: A jihad on the Hobbyist programme ? Message-ID: <1178732923.234048.216360@q75g2000hsh.googlegroups.com> On 9 mei, 07:16, Wilm Boerhout wrote: > May I should worry first about getting .NET 2.0 to run under Wine. Indeed I should. "wine dotnetfx.exe" fails gracefully after reporting it can only be installed when it finds IE 5.01 already installed. I couldn't find an IE 5.01 kit, but downloaded an IE 6 SP 1 kit. It crashed horribly. Anyone ever install *any* Internet Explorer under wine 0.9.36? /Wilm ------------------------------ Date: Wed, 09 May 2007 11:57:09 -0400 From: JF Mezei Subject: Re: Anyone know why the Alpha market is so so quiet? Message-ID: <60cdd$4641ef64$cef8887a$10961@TEKSAVVY.COM> Larry Kilgallen wrote: > And for others, impossible due to the lack of at least two compilers. This was also the case for VAX to Alpha transition where many products didn't make it to Alpha, forcing customers to either stay on VAX or migrate to another vendor's product. HP's IDC study that says that HP will lose 30% of its installed enterprise base due to this unwated IA64 migration must have been factored in and HP must have accepted that fact and will probably have large parties if in the end, it will have only lost 25% of its installed base. In terms of the quiet period, I had heard through the grapevine that there had been a huge rush of purchases of Alpha systems since just before the first "end date" (which was one reason for HP to extend sales since they were raking in the money). So now, it is understandable that there might be a lull. Also, do not forget that Cerner has a deal where cerner customers can contiune to purchase Alpha systems for the foreseable future. ------------------------------ Date: 9 May 2007 10:35:04 -0700 From: jbigboote Subject: Re: Anyone know why the Alpha market is so so quiet? Message-ID: <1178732104.427768.224880@y5g2000hsa.googlegroups.com> On May 9, 10:57 am, JF Mezei wrote: > Larry Kilgallen wrote: ... > > Also, do not forget that Cerner has a deal where cerner customers can > contiune to purchase Alpha systems for the foreseable future. Too bad Cerner decided to drop VMS and replace it with HP-UX. If I'm an existing Cerner client looking to port from VMS, and my choices are AIX or HP-UX, why would I choose HP-UX? Maybe because I already have Alpha servers, which can run HP-UX, but sooner or later you'll have to move to Itanium, and then you'll have to port again, no? And if I'm going to buy new hardware, I might as well go with IBM, who for better or worse have done fairly well with AIX and the POWER platform, and appear poised to continue doing so for the foreseeable future. I administer both AIX and VMS systems, and I can say that in the last 18 or so months getting support for VMS software and Alpha hardware has become noticeably more irksome. ------------------------------ Date: 9 May 2007 19:54:32 +0200 From: peter@langstoeger.at (Peter 'EPLAN' LANGSTOeGER) Subject: Re: Anyone know why the Alpha market is so so quiet? Message-ID: <464226f8$1@news.langstoeger.at> In article <1178732104.427768.224880@y5g2000hsa.googlegroups.com>, jbigboote writes: >Too bad Cerner decided to drop VMS and replace it with HP-UX. If I'm >an existing Cerner client looking to port from VMS, and my choices are >AIX or HP-UX, why would I choose HP-UX? Maybe because I already have >Alpha servers, which can run HP-UX, Try again. Alphas run DEC-OSF1^WDigital-UNIX^WTru64-UNIX (and *BSD, Linux) but not HPsUX. -- Peter "EPLAN" LANGSTOEGER Network and OpenVMS system specialist E-mail peter@langstoeger.at A-1030 VIENNA AUSTRIA I'm not a pessimist, I'm a realist ------------------------------ Date: Wed, 9 May 2007 07:56:18 -0400 From: "Neil Rieck" Subject: Re: Digital *is* a Software Company? (Was Re: Sun Studio 11 (Solaris IDE)) Message-ID: <4641aa83$0$20568$88260bb3@free.teranews.com> "Richard Maher" wrote in message news:f1rkak$jp$1@news-01.bur.connect.com.au... > [...snip...] > > In defense of VMS Engineering, I'd just like to say: - > "We invested a lot of time and energy into creating something better than > VMSINSTAL." > > As for all this "auto-installing" stuff, just ask yourself if most SUN and > Windows developers aren't kicking themselves that they don't have PCSI :-) > > On a serious note though, I just wish VMS Engineering would stop trying to > "replicate" what SUN or Microsoft or Whoever are doing and instead tried > harder to "integrate" these outstanding products with VMS's strengths. I > mean how many IDEs do you need on VMS for Java if everyone is using their > favourite IDE on their favourite platform and, at best, just shipping the > JAR or CLASS files to the VMS/JVM for execution? Rather than playing > perennial catchup, shouldn't HP be applauding SUN's efforts and then > saying > that we can make it more resilient (if not run faster) on VMS? Take > Solaris/Java, Linux/Java, Windows/.NET - Surely VMS should embrace them > all > as potential front-end solutions? (And, most importantly, stop diverting > precious resources into competing with them?) > > Me? I like VMS! I revel in it's existing and rich development environment. > I > love its clustering and reliability, its Run-Time Libraries and language > neutral environment. I love its Lock Manager and its 4 Processor Modes; > yes > its beauty and its terror; the wide-brown land for me! > > Cheers Richard Maher > > PS. Have I missed something and everybody thinks VMS now has a place back > on > the desktop workstation market? Maybe we should start manufacturing VTxxx > again? NintendoVMSboy? VBox? Satirically, that Linux penguin with > diving-flippers and a strapped-on rubber bill is looking more and more > like > logo material :-( VMStunes with Itanium inside? (Could be a bit bigger > than > the Apple one) Who needs a rudder eh? Let's just keep "saling" around out > here and see what happens. > Then we are in agreement because I think OpenVMS is the superior OS as well. I do code development every day with EDT but need to point out that a younger developer in our group copies source code to Windows, edits, then copies back because "no one wants to learn EDT after they have first been exposed to a GUI based editor". So in this light if we don't attract younger developers with GUI based tools, then the future of OpenVMS is doomed. Also, the majority of our software is large but not object oriented. It has been my experience that the only way to properly do large OO programming is with an IDE. OpenVMS engineering supports NetBeans for OpenVMS program development and I think this is a good first step. With regards to your PC comments, x86 was crappy and everyone outside that world knows it. But Intel and AMD are morphing these products into something closer to what Alpha already is. Some day very soon (single years) there will be no real difference between x86-64 and enterprise processors like Alpha, PA-RISC, and Itanium. Some bean counter will then pull the plug on Itanium and we'll all be trapped on another dead technology. p.s. with regard to your new thread name "Digital *is* a Software Company", I used to think that the technical contributions of Digital and Compaq could survive in an HP world. Now that HP sales is promoting HP-UX as a replacement for OpenVMS I can see that HP is treating DEC and Compaq like rejected step children. Neil Rieck Kitchener/Waterloo/Cambridge, Ontario, Canada. http://www3.sympatico.ca/n.rieck/ -- Posted via a free Usenet account from http://www.teranews.com ------------------------------ Date: 9 May 2007 07:42:13 -0500 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: Digital *is* a Software Company? (Was Re: Sun Studio 11 (Solaris IDE)) Message-ID: In article <4641aa83$0$20568$88260bb3@free.teranews.com>, "Neil Rieck" writes: > > Then we are in agreement because I think OpenVMS is the superior OS as well. > I do code development every day with EDT but need to point out that a > younger developer in our group copies source code to Windows, edits, then > copies back because "no one wants to learn EDT after they have first been > exposed to a GUI based editor". So in this light if we don't attract younger > developers with GUI based tools, then the future of OpenVMS is doomed. Well, I do the opposite because nothing beats the power of my personal interface to Eve. But MMS/GENERATE is way behind what most IDEs offer in terms of build tools. I use IBM's Rational Rose for one of my projects, IBM is in the progress of moving it to the open Eclipse base. I understand Eclipse is putting Borland out of the IDE development business. I don't know Netbeans well enough to compare them (none of my VMS systems were fast enough for Netbeans when I had the time to look at it), but I agree VMS needs to go in that direction. ------------------------------ Date: Wed, 9 May 2007 09:26:25 -0400 From: "Main, Kerry" Subject: RE: Digital *is* a Software Company? (Was Re: Sun Studio 11 (Solaris IDE)) Message-ID: -----Original Message----- > From: Neil Rieck [mailto:n.rieck@sympatico.ca] > Sent: May 9, 2007 7:56 AM > To: Info-VAX@Mvb.Saic.Com > Subject: Re: Digital *is* a Software Company? (Was Re: Sun Studio 11 > (Solaris IDE)) >=20 >=20 > "Richard Maher" wrote in message > news:f1rkak$jp$1@news-01.bur.connect.com.au... > > >=20 > [...snip...] >=20 > > > > In defense of VMS Engineering, I'd just like to say: - > > "We invested a lot of time and energy into creating something better > than > > VMSINSTAL." > > [snip ...] >=20 > Then we are in agreement because I think OpenVMS is the superior OS as > well. > I do code development every day with EDT but need to point out that a > younger developer in our group copies source code to Windows, edits, > then > copies back because "no one wants to learn EDT after they have first > been > exposed to a GUI based editor". So in this light if we don't attract > younger > developers with GUI based tools, then the future of OpenVMS is doomed. >=20 And one wonders where all of the scalability and security issues are coming from these days? > Also, the majority of our software is large but not object oriented. > It has > been my experience that the only way to properly do large OO > programming is > with an IDE.=20 [snip...] Much, much more to OO than just the IDE .. it is a major, major culture change that most shops do not realize until they really get into it. I would suggest the first question should be to understand the hype and really determine "is OO the best strategy for addressing the needs of my business?".=20 Yes, there are benefits of OO in specific areas.=20 However, the re-training, code re-writing, integration with current code (never use the term legacy for proven code) and support complexity are also challenges that many Customers trying to make this transition to OO strategies have flagged. Some major projects have also run into big issues with these challenges, so the grass is not always green on the other side. Remember there are many ways to accomplish the same goal and don't get caught up with the industry and media hype that OO is where everyone needs to migrate to. Yes, OO currently has the "cool" and "in" label, but at one point in the past, AI was also "cool" and we know it did not take over the world either. Regards Kerry Main Senior Consultant HP Services Canada Voice: 613-592-4660 Fax: 613-591-4477 kerryDOTmainAThpDOTcom (remove the DOT's and AT)=20 OpenVMS - the secure, multi-site OS that just works. ------------------------------ Date: Wed, 09 May 2007 06:58:11 -0700 From: "Tom Linden" Subject: Re: Digital *is* a Software Company? (Was Re: Sun Studio 11 (Solaris IDE)) Message-ID: On Wed, 09 May 2007 04:56:18 -0700, Neil Rieck wrote: > > "Richard Maher" wrote in message > news:f1rkak$jp$1@news-01.bur.connect.com.au... >> > > [...snip...] > >> >> In defense of VMS Engineering, I'd just like to say: - >> "We invested a lot of time and energy into creating something better >> than >> VMSINSTAL." >> >> As for all this "auto-installing" stuff, just ask yourself if most SUN >> and >> Windows developers aren't kicking themselves that they don't have PCSI >> :-) >> >> On a serious note though, I just wish VMS Engineering would stop trying >> to >> "replicate" what SUN or Microsoft or Whoever are doing and instead tried >> harder to "integrate" these outstanding products with VMS's strengths. I >> mean how many IDEs do you need on VMS for Java if everyone is using >> their >> favourite IDE on their favourite platform and, at best, just shipping >> the >> JAR or CLASS files to the VMS/JVM for execution? Rather than playing >> perennial catchup, shouldn't HP be applauding SUN's efforts and then >> saying >> that we can make it more resilient (if not run faster) on VMS? Take >> Solaris/Java, Linux/Java, Windows/.NET - Surely VMS should embrace them >> all >> as potential front-end solutions? (And, most importantly, stop diverting >> precious resources into competing with them?) >> >> Me? I like VMS! I revel in it's existing and rich development >> environment. I >> love its clustering and reliability, its Run-Time Libraries and language >> neutral environment. I love its Lock Manager and its 4 Processor Modes; >> yes >> its beauty and its terror; the wide-brown land for me! >> >> Cheers Richard Maher >> >> PS. Have I missed something and everybody thinks VMS now has a place >> back on >> the desktop workstation market? Maybe we should start manufacturing >> VTxxx >> again? NintendoVMSboy? VBox? Satirically, that Linux penguin with >> diving-flippers and a strapped-on rubber bill is looking more and more >> like >> logo material :-( VMStunes with Itanium inside? (Could be a bit bigger >> than >> the Apple one) Who needs a rudder eh? Let's just keep "saling" around >> out >> here and see what happens. >> > > Then we are in agreement because I think OpenVMS is the superior OS as > well. I do code development every day with EDT but need to point out > that a younger developer in our group copies source code to Windows, > edits, then copies back because "no one wants to learn EDT after they > have first been exposed to a GUI based editor". So in this light if we > don't attract younger developers with GUI based tools, then the future > of OpenVMS is doomed. Well, there is nothing wrong with that. In my case Emacs has been my editor of choice for the last 30 years, and i can't get a decent one for VMS. But thanks to WASD, I can use Xemacs on W2k/XP and open any file on my cluster under https. It is a waste of effort for Digital (That is what I call VMS Engineering) to reinvent these tools on VMS when they work perfectly fine > > Also, the majority of our software is large but not object oriented. It > has been my experience that the only way to properly do large OO > programming is with an IDE. OpenVMS engineering supports NetBeans for > OpenVMS program development and I think this is a good first step. There is no compelling reason to go to OO, at least not the way it is done. > > With regards to your PC comments, x86 was crappy and everyone outside > that world knows it. But Intel and AMD are morphing these products into > something closer to what Alpha already is. Some day very soon (single > years) there will be no real difference between x86-64 and enterprise > processors like Alpha, PA-RISC, and Itanium. Some bean counter will then > pull the plug on Itanium and we'll all be trapped on another dead > technology. The transition of Digital from a Hardware to a Software company is a far superior business model, and it has always baffled me that Digital management is incapable of grasping that. > > p.s. with regard to your new thread name "Digital *is* a Software > Company", I used to think that the technical contributions of Digital > and Compaq could survive in an HP world. Now that HP sales is promoting > HP-UX as a replacement for OpenVMS I can see that HP is treating DEC and > Compaq like rejected step children. > > Neil Rieck > Kitchener/Waterloo/Cambridge, > Ontario, Canada. > http://www3.sympatico.ca/n.rieck/ > > -- Using Opera's revolutionary e-mail client: http://www.opera.com/mail/ ------------------------------ Date: Wed, 09 May 2007 15:38:14 +0200 From: Martin Krischik Subject: Re: Digital *is* a Software Company? (Was Re: Sun Studio 11 (Solaris IDE)) IDE)) Message-ID: <4641cec6$1@news.post.ch> Bob Koehler schrieb: > In article <4641aa83$0$20568$88260bb3@free.teranews.com>, "Neil Rieck" writes: >> Then we are in agreement because I think OpenVMS is the superior OS as well. >> I do code development every day with EDT but need to point out that a >> younger developer in our group copies source code to Windows, edits, then >> copies back because "no one wants to learn EDT after they have first been >> exposed to a GUI based editor". So in this light if we don't attract younger >> developers with GUI based tools, then the future of OpenVMS is doomed. > > Well, I do the opposite because nothing beats the power of my > personal interface to Eve. But MMS/GENERATE is way behind what most > IDEs offer in terms of build tools. And how many years did you work on your "personal interface to Eve" until it was as perfect as it is. Or more importantly: how powerfully is Eve without it! It is a sad truth that "the younger" are often presented with an empty sys$login to start with it. And then an old timer tells them to start eve and forgets to mention /INTERFACE=DECWINDOWS. Yes, I know, real VMS Gurus never use "/INTERFACE=DECWINDOWS" but the newer generation might be grateful for: XEVE == "SPAWN /NOWAIT EDIT /INTERFACE=DECWINDOWS" XLSE*DIT == "SPAWN /NOWAIT LSEDIT /INTERFACE=DECWINDOWS" Martin -- Martin Krischik ------------------------------ Date: Wed, 09 May 2007 13:26:29 -0400 From: Stephen Hoffman Subject: Re: Digital *is* a Software Company? (Was Re: Sun Studio 11 (Solaris IDE)) IDE)) Message-ID: Tom Linden wrote: > On Wed, 09 May 2007 04:56:18 -0700, Neil Rieck > wrote: >> Also, the majority of our software is large but not object oriented. >> It has been my experience that the only way to properly do large OO >> programming is with an IDE. OpenVMS engineering supports NetBeans for >> OpenVMS program development and I think this is a good first step. > > There is no compelling reason to go to OO, at least not the way it is done. It's a tool. OO can be nasty, or it can be useful. There are some OO IDE environments that allow faster coding. Yes, there are trade-offs in image size and overhead from tasks such as garbage collection, etc. OO and IDE can often help hit deployment -- and revenues -- quicker than non-OO. Aim for the target, and pick the tools that allow you to best and most quickly and most efficiently hit the target. >> With regards to your PC comments, x86 was crappy and everyone outside >> that world knows it. But Intel and AMD are morphing these products >> into something closer to what Alpha already is. Some day very soon >> (single years) there will be no real difference between x86-64 and >> enterprise processors like Alpha, PA-RISC, and Itanium. Some bean >> counter will then pull the plug on Itanium and we'll all be trapped on >> another dead technology. Architectures historically come and go. Though x86-64 isn't known as the most elegant of designs, it's ahead of x86-32. It'll be interesting to see if (when?) there's a future generation that drops more of the legacy cruft, or moves it off into some sort of emulation. And x86-64 most certainly works. Itanium has various advantages over Alpha and over x86-32 and x86-64, and in cases where OpenVMS is in use, it's the platform. (BTW, I dug up the various RAS features lists a while back, and posted a summary at the new web site. RAS features can exist within the processor, within in the platform, and within the OS. And whether or not particular RAS features are even a factor depends on your requirements and your environment, and both can vary widely.) All RAS aside, the real question is one of the costs of upgrading and the cost of staying on a static environment, and the cost of porting to a new environment. These are not simple questions. Until you know those costs and what benefits you might exist, there's no real fodder here for discussion. Computer architectures are only good or bad in terms of costs and revenues -- the technology itself is certainly interesting to discuss and to debate, though there's been no shortage of dead elegance left alongside the road to newer products. > The transition of Digital from a Hardware to a Software company is a far > superior business model, and it has always baffled me that Digital > management is incapable of grasping that. The question isn't hardware or software, it's the business model and where you get current and where you expect future revenue from. Folks in the upper echelon of a company have access to details and information that other folks don't have. And some idea of what the trends for these current and future revenue streams might be. Tom's business is software. Other folks sell software, or services, or hardware -- and a mixture of these areas can be required to most effectively sell the more profitable pieces of the puzzle. An x86-64 or Itanium box is basically an expensive doorstop without the operating system software and the requisite tools and application(s). In Tom's case, it likely doesn't make sense to sell a server with his products. There are cases and products where customers want a packaged solution. --- Main, Kerry wrote: > And one wonders where all of the scalability and security issues are > coming from these days? That varies. I've seen various really slick and highly scalable and quite secure environments -- there's almost always more than one way to solve a problem. And security is certainly one aspect and one consideration, but there are others that are often equally important; computers and applications are complex systems, and there are costs and considerations all over the place, from managing the system and ECO kits to security to anti-malware tools, etc. Then there are the discussions of ROI -- computers are cool and fun, but we're here and using this gear to resolve a business problem, or to produce revenue, or both. > Much, much more to OO than just the IDE .. it is a major, major > culture change that most shops do not realize until they really > get into it. > > I would suggest the first question should be to understand the hype > and really determine "is OO the best strategy for addressing the needs > of my business?". Quite true. Backing it up further, you have to look at whether the current applications meet the requirements, and if there are ways to reduce costs or speed production, or to gain a competitive advantage. It's the design, the application and then the stack. Further down, the discussion of the implementation language. > Yes, there are benefits of OO in specific areas. Big time. On some boxes, parallel processing basically falls out of the design. > However, the re-training, code re-writing, integration with current > code (never use the term legacy for proven code) and support > complexity are also challenges that many Customers trying to make > this transition to OO strategies have flagged. Some major projects > have also run into big issues with these challenges, so the grass > is not always green on the other side. OO helps in some areas, and what you can do in a good IDE is quite powerful. There are also some bad IDEs. There are cases when I hike up the IDE and crawl in underneath it, and there are cases where the IDE massively increases my development and debugging speed, or lets me do far more in the same time. Different tools for different jobs. > Remember there are many ways to accomplish the same goal and don't get > caught up with the industry and media hype that OO is where everyone > needs to migrate to. Yes, OO currently has the "cool" and "in" label, > but at one point in the past, AI was also "cool" and we know it did > not take over the world either. OO is in far wider use than AI ever managed. There are techniques from AI that can be useful, though I use OO techniques far more often. (OO techniques in old and non-OO code, and OO techniques in new code.) Different tools and techniques for different jobs, and preferably an appropriate tool for the particular task. OO via IDE is entirely different than working with traditional techniques, but that doesn't make it better or worse. I happen to like OO IDE and I happen to like traditional -- make use of the best of both, as appropriate. It's certainly not an exclusive decision. -- www.HoffmanLabs.com Services for OpenVMS ------------------------------ Date: Wed, 9 May 2007 10:41:41 -0600 From: "Michael D. Ober" Subject: For Sue - HP Proliant Server Problems Message-ID: <4641f9c6$0$3572$815e3792@news.qwest.net> Sue, Since January, we have been attempting to purchase a RAID 5 subsystem for a Proliant ML110G4 server. The configuration was to be 3 x 500GB internal SATA Drives with a E200B Raid controller plus the cables required to connect them. The first time, the wrong drives were sent with no cables (Case #3218708870 & 3218583632). HP reluctantly agreed to replace the drives with the correct drives. The cable that was sent was the wrong cable. Today, we received "replacement cables", but instead of a single cable that connects to the E200B controller (part # 411508-B21) with multiple SATA drives (Part #404469-B21), they were four distinct SATA cables. Note that all purchases, including the server, have been directly from HP. The cables we have received so far are Part #413957-001 (Controller to multiple drives) and four cables Part #430542-001 (single controller to single drive). In addition, since the server is in a remote site (Tulsa, OK), I asked how much it would cost to have HP install this subsystem as was told that HP would not install it - it was user installable. I have no one in Tulsa who can install or maintain hardware, which is why the server has a 3 year next business day CarePaq. Mike Ober Sr Systems Engineer, Wakefield & Associates, Inc. 303-537-2900 x6715 mailto:mdo@wakeassoc.com ------------------------------ Date: 9 May 2007 09:01:32 -0700 From: roger tucker Subject: OpenVMS, Oracle, C/C++ software engineer position in colorado springs Message-ID: <1178726492.051453.261010@q75g2000hsh.googlegroups.com> Senior Engineer - SW SYSTEMS Ref XX59364 Country USA State Colorado City Colorado Springs Business Unit Operations & Technology Speciality Engineering Job Description DBA position within the IT Advanced Communication Engineering Services (ACES) Integrated Intelligent Network (IIN) organization. The primary platform (INCP) is a real time call processing system written primarily on the Alpha OpenVMS platform, using C,C++,ProC, and PL/SQL languages. INCP processes over 280 million calls and 600 million provisioning updates each day. Requirements Installation and maintenance of Oracle10G on Sun Solaris, Red Hat LINUX and more importantly ALPHA OpenVMS. Database Administrative tasks, including database design, data migration, application maintenance and disaster recovery. Diagnosis, troubleshooting and resolution of a broad spectrum of database problems. Solid understanding of relational databases and writing SQL Queries and stored procedures. Strong skills in analysis, design, development and testing. Ability to design, document and model databases. Proven analytical skills. Attention to detail. Excellent verbal and written communication skills. Team player with ability to multi task and produce quality work under high pressure. Ability to take initiative and successfully complete projects independently. Highly organized with the ability to handle multiple projects simultaneously. In addition to new product development and DBA maintenance, the INCP group is responsible for all production maintenance and 7X24 4th level support ? pager rotation is required. Bachelor's Degree in Computer Science, or equivalent 9-10 years job- related experience. Java would be a plus. Apply at www.verizonbusiness.com. http://www.verizonbusiness.jobs/Search/Search.html ------------------------------ Date: 9 May 2007 04:47:08 -0700 From: BaxterD@tessco.com Subject: Re: PID for detached process. Message-ID: <1178711228.002106.52140@h2g2000hsg.googlegroups.com> On May 8, 11:41 pm, Kilgal...@SpamCop.net (Larry Kilgallen) wrote: > In article , bri...@encompasserve.org writes: > > In article <1178632977.412050.112...@e51g2000hsg.googlegroups.com>, Baxt...@tessco.com writes: > >> Does anyone know if there is a symbol which keeps the PID of the > >> last detached process created. > > >> I am thinking of a equivalent to the $ENTRY symbol which holds the > >> entry number of the last job you submitted. > > >> Basically I would like to get the PID of a process and store it for > >> future reference. > > > You can capture it from the status line returned by the RUN command. > > It's ugly as sin > > For something less ugly, use a compiled language to write your own > version of the RUN command that creats the desired symbol. Just in case anyone is interested, this is how I currently do it. $ Pipe Show Sys /Proc="(DET) QUOMOVR" | (Read Sys$Input Line ; - Read Sys$Input Line ; Read Sys$Input Line ; - PID = f$extract(0,8,Line) ; Define/Job/NoLog PID &PID ) $ PID = f$trnlnm("PID","Lnm$Job") It is similar to Dave Dachera's approach. I like it because it creates no unnecessary output files (needint to be cleaned up). Dave ------------------------------ Date: Wed, 09 May 2007 11:53:36 GMT From: VAXman- @SendSpamHere.ORG Subject: Re: PID for detached process. Message-ID: <00A67597.9D0C7AB2@SendSpamHere.ORG> In article <1178711228.002106.52140@h2g2000hsg.googlegroups.com>, BaxterD@tessco.com writes: > > >On May 8, 11:41 pm, Kilgal...@SpamCop.net (Larry Kilgallen) wrote: >> In article , bri...@encompasserve.org writes: >> > In article <1178632977.412050.112...@e51g2000hsg.googlegroups.com>, Baxt...@tessco.com writes: >> >> Does anyone know if there is a symbol which keeps the PID of the >> >> last detached process created. >> >> >> I am thinking of a equivalent to the $ENTRY symbol which holds the >> >> entry number of the last job you submitted. >> >> >> Basically I would like to get the PID of a process and store it for >> >> future reference. >> >> > You can capture it from the status line returned by the RUN command. >> > It's ugly as sin >> >> For something less ugly, use a compiled language to write your own >> version of the RUN command that creats the desired symbol. > >Just in case anyone is interested, this is how I currently do it. > >$ Pipe Show Sys /Proc="(DET) QUOMOVR" | (Read Sys$Input Line ; - > Read Sys$Input Line ; Read Sys$Input Line ; - > PID = f$extract(0,8,Line) ; Define/Job/NoLog PID >&PID ) >$ PID = f$trnlnm("PID","Lnm$Job") > >It is similar to Dave Dachera's approach. I like it because it >creates no unnecessary output files (needint to be cleaned up). > >Dave You could install SYMBOL and then it's just: $ PIPE RUN/DETACH program | SYMBOL/SET PID ; PID == PID - "%RUN-S-PROC_ID, IDENTIFICATION OF CREATED PROCESS IS " $ SHOW SYMBOL PID PID == "20201A32" -- 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: 9 May 2007 07:36:42 -0500 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: PID for detached process. Message-ID: In article , Kilgallen@SpamCop.net (Larry Kilgallen) writes: > For something less ugly, use a compiled language to write your own > version of the RUN command that creats the desired symbol. I really think that process names and previous posts with examples of how to use them to get the PID directly addresses the OP's needs in a straightforward manner using documented interfaces. While it would be possible to do the same by writing a program, I'm left wondering how many UNCHECKED_CONVERSIONs it would take to do it in Ada? (The last time I wored with VMS routines form Ada every routine seemed to have its own type for a 32 bit integer. 8-) ) ------------------------------ Date: 9 May 2007 09:20:23 -0500 From: briggs@encompasserve.org Subject: Re: PID for detached process. Message-ID: In article <46410E0D.A7352D14@spam.comcast.net>, David J Dachtera writes: > briggs@encompasserve.org wrote: >> You can capture it from the status line returned by the RUN command. >> It's ugly as sin, but here's a proof of concept: >> >> $ define sys$output gork.dat >> $ run sys$system:loginout.exe - [... ugliness snipped ...] > Here's another approach to try... > > $ pipe - I had tried that at first but ran into command element length restrictions. > run sys$system:loginout.exe - > /detach - > /input=eisner$dra3:[decuserve_user.briggs]nothing.com - > /output=eisner$dra3:[decuserve_user.briggs]nothing.log | - > @tee sys$command | - > (read sys$pipe status_line ; - > pid = f$extract ( f$length(status_line)-8, 8, status_line) ; - > def/job subpid &pid) > $ pid = f$trnlnm( subpid ) > $ deas/job subpid The bit with tee is a nice touch. > See HELP PIPE Description Pipelines_and_TEEs Using_TEEs_and_SYS$PIPE for an > example of a TEE.COM. This may only work on V7.3-2 and later due to the command > line length. Ahh. That explains it. Eisner is still on 7.2 ------------------------------ Date: Wed, 09 May 2007 10:19:38 -0700 From: Alan Frisbie Subject: Seeking Terminal Server Manager (TSM) manuals Message-ID: <1178731178.688027@smirk> I am looking for a machine-readable manual for the old DEC Terminal Server Manager product. The original part number was AA-JF99C-TE. And as long as I'm asking for unobtainum, I might as also ask for an electronic copy of the TSM Software Installation Guide (AA-JF00C-TE). :-) Alan ------------------------------ Date: 9 May 2007 07:47:37 -0500 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: Semi-OT: 4K Disk Blocks? Message-ID: In article , JONESD@ecr6.ohio-state.edu (David Jones) writes: > > But if the 'B' in the LBN fields of the on-disk structures are still 512 byte > blocks, your volume size is still limited to 1TB. Once upon a time they were smaller and lots of folks were claiming VMS would have to add partition support to deal with larger disks. Instead of adding that level of complication, the file system was modified to handle larger disks. No reason why it couldn't be done again. There's a perfectly good example of how to handle larger fields in an upward compatable manner in RMS when you look at NAM vs. NAML, and such. Even though VMS was originally written largeley in Macro-32 the basic data structure work that was done gives it internally much of the rich power of the then not coined "object oriented" approach. ------------------------------ Date: 9 May 2007 07:48:43 -0500 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: Semi-OT: 4K Disk Blocks? Message-ID: In article <46411ab6$0$90269$14726298@news.sunsite.dk>, =?ISO-8859-1?Q?Arne_Vajh=F8j?= writes: > > My guess is that there will be a lot of broken code out there with > hardcoded 512. As there was for pagesize prior to Alpha. Gee, none of my code broke. ------------------------------ Date: 9 May 2007 10:05:36 -0500 From: brooks@cuebid.zko.hp.nospam (Rob Brooks) Subject: Re: Semi-OT: 4K Disk Blocks? Message-ID: JF Mezei writes: > Newbie question: > > When the application requests to read 16k of data, does this generate 1 > SCSI command to read 32 sequential blocks ? Or does it result in 32 SCSI > commands to read each 512 byte block ? A single transfer (SCSI, MSCP, IDE, etc. . .) can be up to ucb$l_maxbcnt bytes at a time. For any device, you can get this value using $GETDVI with the DVI$_DEVICE_MAX_IO_SIZE item code. This item first appeared in V8.2 (I think), and was unofficially backported to V7.3-2 some time ago. -- Rob Brooks MSL -- Nashua brooks!cuebid.zko.hp.com ------------------------------ Date: Wed, 09 May 2007 12:30:27 -0400 From: Stephen Hoffman Subject: Re: Semi-OT: 4K Disk Blocks? Message-ID: JF Mezei wrote: > Newbie question: Folks that have been here a decade don't get to call newbie all that often, eh? My own newbie-isms have been on display elsewhere... > When the application requests to read 16k of data, does this generate 1 > SCSI command to read 32 sequential blocks ? Or does it result in 32 SCSI > commands to read each 512 byte block ? That's an open-ended question. The application might request 16 KiB, and could end up generating zero I/O requests, one I/O request, or some number of I/O requests to service the transfer. For most transfers not serviced from host or controller cache, I'd expect one 16 KiB data transfer, and one or maybe two I/O SCSI commands. A READ, potentially followed by a SCSI status command. The disk itself might or might not hit the platters to read the data; it's possible for an I/O to be completed out of drive cache. > If applications already send SCSI command that request chunks of data > that are larger than a physical block, then would it be correct to state > that it wouldn't change much in terms of IO load at the OS level ? I'd expect the aggregate might increase slightly, in terms of data that previously fit in 0.5 KiB and now sits in 4 KiB. This will be very slight increase in the data getting shuffled, and might or might not be noticed if the system is already implementing read-ahead. This is the same sort of situation around the wasted space in the last cluster; that's about half the cluster factor times the number of files on the disk, on average. It's there, but it's not something that generally becomes overtly visible. The bus-level I/O isn't where I'd expect a sector size change becomes centrally visible to an application. The I/O stack is a fairly involved place. I've previously posted up a quick intro to the OpenVMS I/O stack, and I posted up another article built from that previous reply of mine in this thread. -- www.HoffmanLabs.com Services for OpenVMS ------------------------------ Date: Wed, 9 May 2007 08:52:01 +0000 (UTC) From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) Subject: Re: SET CONF SMTP/QUEUES Message-ID: JF Mezei wrote: > Phillip Helbig---remove CLOTHES to reply wrote: > > > Apart from the obvious question---what's the purpose of having a generic > > queue with only 1 execution queue---does anyone have this set to a > > number higher than 1? > > In my case, I have define the generic queues on nodes velo and bike to > point to the execution queue on node chain. The TCPIP software still > insists on having an execution queue on VELO and BIKE, but those are > never used. > > Sending SMTP emails on BIKE or VELO results in the symbiont on CHAIN > processing the message. > > Unfortunatly, SET CONF SMTP/QUEUES does not accept a value of 0. It > would be logical to have all workstations have a value of 0 and have any > outbound mail processed by a central SMTP server queue. Why? Why not have every machine be able to send email itself? Ideally, every machine in the cluster should be able to do anything, so that the cluster can continue if a node goes down. You don't want a single point of failure. Even satellites can send and receive email. (In my case, incoming email comes to the cluster alias, and a) it is probably a good idea for it not to hop around and b) I keep it on an ALPHA so that I have more anti-spam features, but those are details. In principle, it shouldn't matter which machine has the cluster alias, except perhaps for performance considerations.) > > SMTP_RECV_MAIL: getpeername failed: vaxc$errno: %SYSTEM-F-NOLINKS, > > maximum network logical links exceeded > > This would have nothing to do with the queues. This is related to DNS > queries. HELP/MESSAGE %SYSTEM-F-NOLINKS discusses only DECNET with NCP > (so it is rather dated). > > What you can do is to limit the number of concurrent receivers. ($TCPIP > SET SERVICE SMTP/LIMIT=5 for instance). Is there any way to see the current limit? (TCPIP SHOW SERVICE doesn't seem to be able to show it.) Do I want to increase or decrease the limit? What will happen in each case? What happens NOW when the "vaxc$errno" message comes? > It might allow you to bypass the stuck queue and get messages processed > by the new second queue while the first one remains in "stuck" state. I > think the original intent of having many queues was for performance > reasons when you have very high load of messages. I have moved up to two queues. Occasionally, both are busy. Presumably, in the past, in the case of incoming mail, the sender saw some sort of delay (or even temporary failure?) if the one queue was busy. Presumably, there is no reason not to have a very large number of queues? David J Dachtera writes: > See the help for SYSCONFIG: > > $ UCX > TCPIP> HELP SYSCONFIG > > ...., especially for "-r socket somaxconn" and related parameters. You may need > to refer to the on-line documentation or the hard-copy documentation as the > on-line HELP does not list all the modifiable parameters. Indeed, the help doesn't say much. Is this related to the SET SERVICE SMTP/LIMIT which JF mentioned? ------------------------------ Date: Wed, 09 May 2007 12:49:32 -0400 From: JF Mezei Subject: Re: SET CONF SMTP/QUEUES Message-ID: Phillip Helbig---remove CLOTHES to reply wrote: > Why? Why not have every machine be able to send email itself? Because you then have multiple log files, you then have multiple configuration files etc. (ok, you still need multiple config files, one for each workstation so that SMTP% will work, but it is less critical). Also, when you run Mozilla on a workstation, it will take every possible resource it can lay its hands on and adding a functional SMTP adds to an already overloaded workstation. > You don't want a single point > of failure. You can still have multiple execution queues on each "server", so if one server goes down, workstations still submit messages to their generic queue that will feed whatavere execution queues are still available. > Is there any way to see the current limit? (TCPIP SHOW SERVICE doesn't > seem to be able to show it.) TCPIP SHOW SERVICE SMTP/FULL , you will see the service limit on the left. The reason I set mine to a low value (2) is that I had noticed that spam came in waves. Having a low limit means that only the first 2 attempts at delivering an add for toe growing magic pills get connected, the rest just bounce right off even before the system can decide the come from a blacklisted IP. However, for a real shop, you couldn't do this because with many users, your mail traffic would likely have many instances of it arriving at the same time as other traffic. In terms of your VAXC error number, I would suspect some TCPIP stack config limit or perhaps the DNS server limit. Is your dns server local on VMS or something provided by your ISP ? ------------------------------ Date: 9 May 2007 07:26:25 -0500 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: Sun Studio 11 (Solaris IDE) Message-ID: In article <4640deea$0$16306$88260bb3@free.teranews.com>, "Neil Rieck" writes: > > "Bob Koehler" wrote in message > news:nUKRx35Zd0e5@eisner.encompasserve.org... >> In article <463fbcfb$0$16288$88260bb3@free.teranews.com>, "Neil Rieck" >> writes: >> >> Sun doesn't do Windows. That's probably good. They only do UNIX. >> That's bad. >> > > I disagree. Solaris is a very popular OS used by industry throughout North > America as well as the world. Also, some Sun software (like Java) works on > many platforms including WIndows. So Sun does not just "do unix". The popularity of Solaris doesn't change the fact that when it comes to making and selling an OS, Sun is a one-tune shop. ------------------------------ Date: 9 May 2007 07:31:43 -0500 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: Sun Studio 11 (Solaris IDE) Message-ID: In article <5aca95F2klvcbU1@mid.individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes: > > Well, first, I would imagine that people here are only concerned with VMS > and would consider VAX dead once you could no longer run VMS on it. People on c.o.v are certainly concerned with VMS on any platform and VAXen running VMS. That doesn't cover everybody with a VAX as a visit to comp.sys.dec will show. > Second, BSD has never been a serious solution for the VAX Prior to RISC architectures, BSD on a VAX was the bulk of UNIX. If Berkely hadn't ported UNIX to VAX there's a good chance the rest of us would never have heard of it. I great many kids I hired in the 80s came out of college with years of experience only in Ultrix on a VAX. The denizens of c.s.d seem to think that's a fine thing to do with a VAX. ------------------------------ Date: 9 May 2007 12:51:43 GMT From: bill@cs.uofs.edu (Bill Gunshannon) Subject: Re: Sun Studio 11 (Solaris IDE) Message-ID: <5aduevF2o7hpdU1@mid.individual.net> In article , koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes: > In article <5aca95F2klvcbU1@mid.individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes: >> >> Well, first, I would imagine that people here are only concerned with VMS >> and would consider VAX dead once you could no longer run VMS on it. > > People on c.o.v are certainly concerned with VMS on any platform and > VAXen running VMS. That doesn't cover everybody with a VAX as > a visit to comp.sys.dec will show. I realize that, but the question was raised here and I am fairly certain the concern was about the effects of eliminating what may still be a significant number of VMS systems. > >> Second, BSD has never been a serious solution for the VAX > > Prior to RISC architectures, BSD on a VAX was the bulk of UNIX. > If Berkely hadn't ported UNIX to VAX there's a good chance the > rest of us would never have heard of it. No doubt, and BSD 4.4 is probably still a good bet on a VAX, but that doesn't change the fact that while in general {Net|Open}BSD have gotten more and more bloated (as they target machines with greater capabilities than the VAX) neither of them has added any of the later hardware leaving an OS that runs only on a very small subset of systems. (Hint: I got rid of all my RA80 disks quite some time ago. {Net|Open}BSD will not run on my 4000's and certainly wouldn't run on the 7000's I just got rid of.) > > I great many kids I hired in the 80s came out of college with years > of experience only in Ultrix on a VAX. The denizens of c.s.d seem > to think that's a fine thing to do with a VAX. Which is fine for hobbyists (especially those who have no problem running software without a license) but not for someone wanting to do serious business on a VAX (and considering how many people we hear from here who are still running VMS on Vaxen for business.....) 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: Wed, 9 May 2007 13:11:22 -0400 From: "Neil Rieck" Subject: Re: Sun Studio 11 (Solaris IDE) Message-ID: <4641f45b$0$8173$88260bb3@free.teranews.com> "Stephen Hoffman" wrote in message news:f1r89s$1aij$1@pyrite.mv.net... > Neil Rieck wrote: > [...snip...] > > There are various IDEs available for various platforms, including the > Sun Studio, NetBeans (of which HP OpenVMS engineering has certainly been > targeting for a while; Sun Studio is itself based on NetBeans), Eclipse, > Visual Suite, Xcode, and others -- these are as much a part of the > operating system as not; compilers were once largely separate pieces, and > now they're tied into the operating system and its APIs and tools. IDEs > are now central to the run-time application environment and vendor and > partner software development operations. Xcode has included over two > hundred megabytes of manuals and source examples and such built in and > periodically updated, and it's not alone in this regard. > > Unlike the compilers of old, many of these IDEs are also now free. > > And everybody I know of has program(s) aimed at the developers. > This is welcome news because the younger generation won't give you a second look if you don't have one. > In addition to Java and Solaris, Sun is also working with NFS (now at > NFSv4.1), has open-sourced most of Solaris, and is just starting to market > its Fortress (currently in R&D) language. Sun Fortress certainly looks > interesting, though it seems a whole lot like writing source code in a > language and a display and an IDE that looks built from the output from > TeX; programming in a fully formula-based programming language that was > once envisioned as TeX output intended for publication. > > There's a whole lot going on in the computer hardware and software > industry, yes. Both in terms of the products and technologies and tools, > and in terms of the marketeering. But this is nothing new. Increasing > integration, increasing competition, increasing marketeering, and product > prices that continue to drop -- OpenVMS I64 is itself down to about US$900 > per core with the FOE package on boxes with one or with two sockets. > True. But remember that my initial post mentioned 64-bit Solaris coming up on a dual-core x86-64 (2.8 GHz) for free. You already know that I prefer OpenVMS over every other OS but, like the story of the "tortoise and the hare", many VMS people are not noticing that other OSs are slowly catching up. Some of these people also refuse to acknowledge that x86-64 is getting better every year as well. As I previously said in another thread, I had been installing both "Red Hat Linux" and "Solaris" on a home PC every spring for the past 4 years, and I really didn't find anything worth mentioning until this year with Solaris-11 and Sun Studio 11. For example, even though I'm building code on an "x86-64 based" PC, I can generate and optimize code for SPARC. This means that developers don't need to be in possession of a SPARC in order to support one. (but how is this any different from running OpenVMS in a virtual environment on top of HP-UX ?) I really hope that HP has got a skunk-works porting OpenVMS to x86-64 because like the VHS vs. BETA thing (BETA was better) I'm thinking that x86-64 will eventually dominate the marketplace. Neil Rieck Kitchener/Waterloo/Cambridge, Ontario, Canada. http://www3.sympatico.ca/n.rieck/ -- Posted via a free Usenet account from http://www.teranews.com ------------------------------ Date: Wed, 9 May 2007 07:44:04 +0000 (UTC) From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) Subject: Re: SYSMAN problem Message-ID: In article <46410A40.C00DC096@spam.comcast.net>, David J Dachtera writes: > > > $ TYPE A::NL: > > > > > > ....and see if it returns a similar result. If so, troubleshoot DECnet. > > > > No DECnet running, all nodes in the same cluster. > > ??? > > That makes the message all the more puzzling. Except for PHONE between nodes, I don't think one needs DECnet at all within a cluster. I am familiar with the message if the SMI server is not running on the target node, but that is not the case. This came about as a result of a dirty reboot caused by spontaneous reboots of a VAX in the cluster, which put an important disk into mount verification. As a result, the other nodes couldn't be shut down cleanly. I'll probably combine a needed reboot for some ALPHA patches with a cluster reboot, so the problem should go away after that. ------------------------------ End of INFO-VAX 2007.254 ************************