June 22, 2016
What's MPE got to do with emulators?
Companies that want to use their MPE/iX applications a long time might count their timelines with two eras: Before Emulator, and After Emulator. The B.E. period left the MPE/iX user locked to Hewlett-Packard hardware and waiting for upgrades to HP boxes. The A.E. era uses virtualization via Charon to permit many beefy Intel boxes do the MPE/iX work. But what does MPE/iX code have to do with the magic of Charon? Not much, which is a good thing.
There's a stubborn story we hear about how the gem of MPE's source code is at the heart of what Charon does. What a virtualization engine like the Stromasys product delivers is a new capability for Intel hardware. An Intel box can pretend to be a PA-RISC processor, thanks to the software engineered by the creators of similar products for the Digital market.
But Charon doesn't rely on MPE/iX secrets to do this magic. It's like thinking a jockey is the being who's running a 2-mile racetrack course. He's the rider, and the horse in this metaphor is Charon. The basic design of Charon products, like the ones that virtualize the Sun Sparc systems and the PDP systems of DEC, creates the expertise for booting up Intels like they're 3000s. Nobody expects the ancestry of the jockey to play a role in making the horse faster. We don't sit in the grandstands to watch jockeys hoof it around the track.Source code for MPE/iX was a big part of the push for emulation that started in 2002. The thought at the time, though, was that MPE itself could be taken on as a third party tool. That somehow something could pretend to be MPE, instead of pretending to be the PA-RISC hardware. Emulation is a broad term. Virtualization is more accurate, and these days, almost as well known thanks to titans like VMware.
OpenMPE fought the good battle to get HP to release source code for MPE/iX. It won, but the group never got a copy of the source for itself. The reasons run to finances and organization of the board of directors, struggles we've documented starting six years ago. Then 2011 dawned as one of the dimmest years for the 3000, just dumped off HP's support and with no apparent future for fresher hardware. The only advocacy group was reeling from lawsuits among its leaders. For the last two elections of its board, the number of nominees was the same as the number of directors. Few people could volunteer with gusto. Some of the best were already there, or had already done their bit.
It's a good thing indeed that MPE/iX source code plays no role in the technology of Charon. The code is not open source, as OpenMPE wanted. (That's the whole reason it's called OpenMPE, instead of FreePA-RISC.) What the community got was a license from HP, not freeware. If Charon needed MPE/iX gems, it would mean HP retains control over a product that will let newer hardware take over when the battleship-grade HP iron hits the scrapyard. Nobody wants Hewlett-Packard to have a say in MPE/iX host hardware futures. The vendor's had enough impact on those. In the days of A.E. the virtualization miracle needs nothing from MPE/iX but a stable copy of the OS to carry in its saddle. The horsepower comes from an independent company, the kind that's kept the MPE/iX journey on course.
Follow the 3000 NewsWire on Twitter
for immediate feeds of our latest news
and more twitter.com/3000newswire.
June 15, 2016
Throwback of mid-June marks much change
Amid the midpoint of June, we have reported a lot of change in that month of the 3000 community's calendar. In the blog's first year of 2005, this report said HP's Unix was named in about a third of migrations.
HP-UX gains in later results (2005)
These revised percentage totals keep Windows in the lead. But with 71 companies reporting their migration plans or accomplishments to us, HP-UX has managed to poke above the 30 percent mark, to just about one-third of the target platform choices.
And there remains in the community a vibrant devotion to migrating to Windows. Linux was less than 10 percent back then. How enterprise tastes have changed.
MPE-Education.com becomes the hub for 3000 training as of this week, since HP has called off its training courses for the platform. Many companies still have years of HP 3000 use in front of them.
Paul Edwards and Frank Alden Smith revitalized HP's 3000 training materials and put the education experience online at $1,750 a seat. The market didn't materialize for the noble, useful service.
So much to see, so far to go (2007)
On a rack in one of the Mandalay Bay's wide lobbies at the Encompass show — lobbies so wide that a semi truck can pass unfettered — a stand of adhesive badges sparkles. The array of ribbons stamped with silver letters lays out the known future for an HP customer or prospect.
To no one's surprise, no "MPE/iX" ribbons. This is a conference which looks toward a new future with HP, instead of the past, or MPE's ongoing tomorrow without the vendor. 3000 community members are coming here to make plans for something new from HP—or hear from vendors and experts about how to make better use of something else from Hewlett-Packard.
The new Las Vegas digs for the annual user group show "improved its curb appeal," said the user group president. A sprawling show in a Vegas casino resort still showed off HP-UX training. "Windows on HP" suggested the vendor was scrabbling to keep customers on its platform.
"We did a lot of work in that area," said HP's Jim Hawkins at the Tech Forum. "For a lot of patches that have been languishing in beta test status, we've been able to move them into General Release status so they can be downloaded from the HP ITRC, which makes them freely available."
Indeed, those patches remain free if a 3000 customer knows how to ask for them. Help from an independent support vendor remains a good way to stay in touch with what HP might've forgotten—or which of those patches you ought to avoid.
But while a 3000 issues list logs many HP decisions, some key items remain unresolved.The issue with the broadest potential impact on homesteading customers appears to be resources for the HP 3000 hardware emulator project.
HP didn't release test suites it used to develop MPE/iX, for example. It would be three more years until Stromasys released the Charon emulator. This was the year HP started to change its mind about helping out.
The new Windows-based MBF Scheduler grew up in MB Foster’s labs, nourished by the experience of engagements with several sites migrating from the 3000.The 3000’s depth of scheduling was integrated into the environment from the early days of system delivery. The cloned feature set reminds migrators of what they’ve learned to rely upon.
MBF Scheduler is still the Windows job scheduler that accommodates MPE procedures best. Experience from "true, red-blooded sites" gave the software its feature set.
At one end there's a 20-year-old 927 server still working in a production setting. At the other end, the most powerful 3000s built by HP are now less than $10,000, at least in a spare-parts or hot DR offering with your own licenses.
Prices for N-Class servers have been quoted below $4,000 this year. That 927 may still be working. That's what the indie support companies make possible.
In a Wall Street Journal interview, new CEO Meg Whitman tossed off a message that HP-UX is on its way to the Intel Xeon processor line.
To answer the question: not. The heir apparent to the MPE enterprise-class datacenter will be on Itanium chips for the forseeable future.
Even among the potential allies for the Stromasys emulator, uncertainty is afoot. In a conversation with a reseller last week about the product, he was not sure that IMAGE was a part of the solution. People approach the Charon emulator from their best-known persepective, and in most cases that’s MPE/iX and its database. Good news: Charon doesn’t emulate any of that software. It simply uses what Hewlett-Packard created and installed on everyone's 3000.
This remains a misunderstood point among 3000 customers with very old hardware. The MPE/iX operating system runs the same on Charon as it does on HP's iron.
June 13, 2016
2016 Advice: Emulate Your 3000 System
No kidding, the above strategy is bona fide. It will be online, with time for your questions, next Wednesday at 2 PM Eastern.
MB Foster has a novel webinar scheduled next week, and no, that's not an hour about writing a bestseller. The Web meeting on June 22 will walk through four different HP 3000 emulation options. All of them will mitigate risk, protect investments, and reduce year over year costs. In the end, every one of them should use MPE/iX apps, if they are bona fide emulations. Why else would you be emulating? The webinar promises a tour of how to replace the 3000 hardware, it seems.
As hardware emulation goes — and that's the most popular agent of change — there's only one supplier that we know about. Over the last three years Stromasys has enlisted HP 3000 advocates and experts and customers to embrace the Charon software. We're told that each new customer seems to draw out another.
There are other ways to consider emulation, however. Some of them have been around a long time, if preservation of in-house MPE/iX apps is the goal. AMXW was a sort of emulator: Automated Migration to UniX and Windows. It's a shell that runs atop those two platforms, plus Linux and IBM's Unix, connecting to commodity databases and surround-code tools while preserving the 3000's app code.
"MPE speciﬁcs, such as JCL batch jobs, ﬁle equations, JCW, UDCs, command ﬁles and variables are all supported — allowing the MPE environment to run as is on the new platform." Okay, this is probably a migration solution. You're not supposed to need to change your apps, though. HP's 3000 hardware gets dropped, too.
The two other options? We'll be online to see what they are. Registration is online at the MB Foster website, as always.
You can't say that emulation is the right choice for everybody who needs to change things. Cloud-based ERP and manufacturing is on the horizon from Kenandy, for example, a company with ASK MANMAN roots. Terry Floyd of the The Support Group says Kenandy is MANMAN done better, because the software seems simpler. He's developed and managed MANMAN installs since the 3000 was very new. Floyd goes to work migrating Disston Tools off MANMAN starting next month.
We agree that any range of emulation options must mitigate risk, protect investments, and reduce costs. Risk is in the eye of the manager; we've said that since 2002, when the Transition Era started. Foster says moving away is too risky and costly for customers who have data on HP 3000s.
"Hewlett Packard said it was obsolete 10 years ago," today's teaser email began, "so why are people still running production environments on the HP 3000? We asked the same question. Much of the time, it is too risky and too costly for them to move."
Emulation has its costs, too. The Stromasys option starts at $9,000 for a permanent license — the kind nearly all of the Charon customers buy, says Doug Smith — and then there's efficient and powerful Intel hardware to buy. Although Charon has been demoed many times on a laptop, a computer with a lid which closes is not the sort to run your commercial computing.
But compared to the expense of hiring out for advice on replacement software (you oughta do that) and implementing a package on locally-hosted servers that behaves differently than employees expect (identical functionality is rare unless Charon HPA's installed) and retraining everybody (IT and users) about the new environment — Charon can be very effective as an emulation. Less costly, probably, than anything but staying on the HP hardware that's at least 13 years old today. At least.
The other elements in the equation are investment protection. That's actually what an independent support company does today (yup, Pivital Solutions) to keep hardware that's in place running. Plus all of the company's experience keeping MPE/iX on its toes.
Companies emulate because they recognize value in the original investment. The unmistakeable value lies in the data. Every kind of emulation protects that asset. Foster says the apps are crucial too.
An emulator retains the value of the application long term, while removing the risk of running on old hardware. Being in an emulation environment also stabilizes the development of surround code, reduces disruption to the business, and avoids the need to re-train employees.
June 10, 2016
What A Newer MPE/iX Could Bring
What would HP 3000 owners do with a new MPE/iX release, anyway? On some IT planning books, the frozen status of the operating system counts as a demerit in 2016. Even still, enterprise system managers in other HP-sold environments face a nearly-glacial pace of OS upgrades today. Even while paying for HP’s support, the VMS system managers are looking at a lull.
HP says it still cares about OpenVMS, but that OS has been moving to a third party. Support from a system maker still looks newer and shiny to some companies than the independent support managers available from third parties like Pivital. As it turns out, though, it’s that frozen-as-stable nature of MPE/iX which makes third party support just as good as HP’s—back when you could get support from HP.
“MPE's so solid,” Doug Smith said in a recent interview, “and these applications have been out there forever. There’s not a huge concern out there in the community about needing to have a new release of MPE.” Smith leads the way for Charon emulator installs at 3000 sites.
OpenVMS roadmaps were updated this week. The map shows how slow OS updating can proceed.HP’s more current Poulson Itanium-based Integrity servers now can run OpenVMS, thanks to a springtime release of OpenVMS 8.4.2. There will still be Kittson-based Integrity servers outside the OpenVMS reach, though. These incremental VMS releases are proving that a third party can assume engineering duty for an OS. Linux showed the way for such duty long ago. That OS, however, was never a trade secret inside a system vendor’s labs.
The most cautious 3000 manager didn’t take updates of MPE/iX, in the years HP released them, unless there were essentials inside the new release. That decision point is no longer an issue with 3000 sites. Instead, MPE/iX is getting its newer-gen speed engineering through the Charon solution. Whenever there is a new Intel chipset that can run Linux, the speed of MPE/iX gets a boost.
A third-party OS lab won’t be the crucial element in driving MPE/iX faster. Charon emulates hardware that is not going to change: PA-RISC and the classic 3000 peripherals. VMS Software Inc. is revising an operating system. There’s much more testing needed to do this revision. It’s the cost of those new OS releases.
The newest OpenVMS will arrive in August, according to the VMS Software roadmap. One major advantage the new release brings will be a modern OpenSSL protocol version. It took awhile, and ultimately a third party, to make it so. Until VMS Software got its hands on VMS, the enterprise OS was working with the 0.9.8 SSL release. After more than seven extra years of HP labs support than MPE/iX had received, VMS was just two minor increments newer than the SSL the 3000s can still run: 0.9.6.
If vendor support for an OS is supposed to be so important, we asked up at the beginning, then why is an enterprise HP system so far behind current protocols as OpenVMS? Rethinking the impact of vendor support led many 3000 sites to independent support arrangements for MPE/iX. With the indie MPE/iX support and static OS status proven as a stable combo, it’s the hardware performance that can make strides. The MPE/iX community doesn’t need an OS lab to boost performance. Support for SSL security needs to be moved along, yes. The 3000 community, however, long ago learned to lean on environments like Unix and Linux for highly-secured functions.
Meanwhile, faster hardware support for OpenVMS turns out to be a feature that MPE/iX gained first. VMS Software says it's now working on an Intel-based release of the OS, with a target shipment sometime in 2018. By that date, the virtualized hardware for MPE/iX will have had two additional years of speed upgrades from Intel. MPE/iX already runs on the x86 family in virtualized mode. Integrity is tied to a chip that's now in maintenance mode at Intel. With the 3000 virtualized hardware speeding up, and the OS hosted in a Linux cradle which sports the latest in security protocol support—remind me again what MPE/iX 8.0 would've brought us?
June 06, 2016
SLEEPER, Awakened for the Masses
The contributed program SLEEPER is probably the software cited most often as proof of the riches of the Contributed Software Library. First created outside of the IT shops of Boeing (according to its first writer Ray Legault) the program was among the most classic of solutions for straightforward jobsteam management. You wouldn't mistake SLEEPER for something professional like Maestro, or even the free MasterOp. But SLEEPER was contributed to the 3000 community, not cast into the free wilds like MasterOp was after its commercial career ended. It was meant to be shared.
The trouble was, SLEEPER disappeared from the community's shelves when Interex died. The CSL tapes (and eventually CDs) went off the grid, another skid-mark left when the user group careened into the void in 2005. It's been 11 years, though, and it's finally time to at least make SLEEPER ready for a wake-up call. We've got the two simple source files to share.
Nobody has liability anymore for HP 3000 contributed software. SLEEPER was never released with support or a license; it was simply part of being an Interex member at a certain level. And let us take a minute to recall that Interex folded owing millions of dollars to members and vendors. After 11 years, it's time to make this software a community resource once more. "Hey, I know a guy who can get you that" has been the means to share the utility over the last decade.
Surprisingly, it's just a well-packed 184K of SPL and FORTRAN code. MPE magic never took up much space. That's one of the reasons it was magic. SLEEPER is also a fine example of how 3000 managers helped one another.We've thrown caution to the winds (not exactly a new event here) and loaded those two files onto our website servers for distribution. It comes with no warranty, of course. Like always with the CSL gems, using it carries the same risk that any contributed utility or application did.
The SPL is some of the better looking old-style SPL that I've seen. The FORTRAN is very nice looking.
The SPL program runs in the batch job. The FORTRAN program maintains a configuration file that is presumably used by the SPL program.
So let the jobs of MPE/iX, including those running on standard HP hardware as well as those on virtualized 3000s, be managed a bit more easily. I await your comments on rousing the talents of this classic tool.
June 01, 2016
Recovering Your Lost Logins
In the world of 2016, losing a password can be a time-bending experience. Apple's ID logins for its iCloud services, iTunes, and App Store only bear five attempts at most before they close out access to the phone. From there, you're on to a lengthy call to Apple. The smartphone is a tiny computer. The larger ones are not as strident about refusing repeated attempts, including the HP 3000.
They are secure, though. In these times when the servers can go dark for awhile as cold-start archive systems, though, a login can get misplaced. We're not supposed to write these things down, after all. What do you do when you have a question like the one below?
I restarted a 3000 9x7 after a few years in mothballs to run an old in-house app. I was able to boot up and login as OPERATOR.SYS, but cannot remember or find the password for MANAGER.SYS. Is there anyway to reset, clear, or overwrite the password file? I know the old machine is a very secure one, but now I am hoping there is a way around it.
What follows are a couple of suggestions to get back into the manager's driver's seat. However, you'll need to at least have a ticket to ride in OPERATOR.SYS.
Gilles Schipper shares what he knows.
There's not an easy way around the security. But since you can log on as OPERATOR.SYS, you should be able to store off the system directory to tape, as follows:
Now that you have the directory on tape, you should be able to move around with FCOPY (and its ;char;hex options) to find the password for MANAGER.SYS. And that's all assuming you haven't implemented the directory encryption feature available with HP's security product Security Monitor/iX.
There's one other way. It's a good thing, because the alternatives do not include a rescue call to HP. There's nothing like Apple's password recovery support left for the HP 3000, as the vendor has long ago left the field. However, a good support contract for an MPE/iX server can be just the thing, and we know where you can get one.
May 27, 2016
A Weekend Memorial to the Future's Past
Here in the US we start our Memorial Day holiday weekend today. Plenty of IT experts are taking a few days off. I reported the start of the HP 3000 emulation era over a Memorial Day weekend, five years ago. We'll take our long weekend to celebrate grandkids and a cookout, and see you back here next week.
In the meantime, here's that first report, a three-parter, showing that Stromasys set and met its development schedule, one that gave the 3000 homesteaders a future beyond the lifespan of HP's MPE/iX hardware. One year later, the software, called Zelus at the time, had a formal debut at a Training Day. Now as Charon it's preserving MPE/iX applications.
During that 2011 springtime, Stromasys offered screen shots of the PA-RISC emulator as evidence the software could serve as a virtual platform for the 3000’s OS. The screen above shows the beginning of the boot sequence (click for detailed view). HP provided internals boot-up documentation to assist in the software's design.
A product journey toward a 3000 hardware emulator took another significant step this spring, as the Zelus cross-platform software booted MPE/iX on an Intel server.
CTO Dr. Robert Boers of Stromasys reported that the OS has come up on a version of the emulator that will managed, eventually, by Linux. Although the test screens that Boers sent were hosted by Windows, the "fairly preliminary version" will be released on an open source OS. "Windows is a little passé," Boers said. "But we now have a first prototype."
Stromasys said it has now been able to use Zelus to tap PA-RISC hardware diagnostics to get the bugs out. "The way we had to debug this was just looking at the code instruction by instruction," Boers said, "to figure out what it does. That took us a long time." Compared to the emulators for the DEC market, "this is by far the most complex emulator."
The accomplishment means that Zelus can do enough to create an MPE/iX image in memory and log to the files. For MPE that was complex, Boers said, while examining and transferring bits and pieces of 32-bit and 64-bit code. Linking to the Processor Dependent Code (PDC) calls that check for 3000 hardware held the project up. One decimal in a table — which turned out to be 666 — "kept us from booting for three months," Boers said. "It's an infamous number that turned out to be a coincidence when we found it."
The pilot milestone comes about one quarter later than the company estimated last year. Pilot versions of the emulator were scheduled to be in beta test by now, with a full release available by mid-year. Boers said the complexity and construction of HP's MPE boot code taxed the tech skills of a company which has built thriving DEC Alpha and VAX and hardware emulators.
"It was a tough one to write," he said of the 3000 effort that began in earnest last year but reaches through HP's licensing delays back to 2004. "It's a pretty deviously complex system. The big problem is that large parts of the operating system are still running in 32-bit mode. MPE's basically an emulated operating environment. We were debugging an emulator running on an emulator."
Hewlett-Packard said in the 1990s that MPE/iX was going to get its full 64-bit version when HP revised it for the Itanium processors. When the vendor cancelled its product futures, the OS remained in emulation mode.
Zelus product delivery to a limited number of sites will take some time, "because it's been such a long project and it's a matter of pride. This has been just a proof of concept. We started trying to build a 918, but then we decided to build something really good, so it now is [software that emulates] an A400."
May 23, 2016
Moving off a 3000, or just some MPE/iX app?
Wednesday afternoon MB Foster leads another of its webinars about migration advice. The company is the community leader in data migration, data migration projects, data migration service. You're moving, they're the folks to contact. On Wednesday at 2 PM EST they're reaching out to explain the methodology the company uses to process departures from the 3000 world.
The options on exits "have not changed much over the last decade," the company's email teaser says. "They include; Stay, Rehost, Replace/Buy, Rebuild. The best choice for you depends on growth expectations, corporate standards, risk and cost." The other determining aspect is how much exiting a migration prospect must do immediately. Several of the current generation of migrators have gone to the app-by-app model.
The largest single migration of educational 3000s, 36 of them at the SBCTC, was pulled off in some pieces. This usually follows a methodology of getting a key app onto another platform in a lift and shift. Rewrites have become rare. Later on the lifted app can be replaced. Sometimes, as is the case at SBCTC, the whole migration platform shifts. Eloquence database to Oracle was the shift there. Another higher-ed site, at Idaho State University, moved its apps a few at a time over several years.
It's always worth mentioning the choice that MB Foster notes: a choice to stay on the HP 3000. But you won't even have to do that if all you need to accomplish is an update of hardware. Choosing Stromasys and the Charon emulator is also a move off the HP 3000: the Hewlett-Packard servers and disks get left behind. New PC hardware and a Linux control center take the place of the HP iron.The pressure to move an app depends on the need to bring it forward with new technology and more wide-ranging interfaces. When Secure FTP was a crucial transfer mechanism for data, the 3000 was often a special-needs case. HP never finished FTP for MPE/iX so it worked fully to industry standards. By now, though, FTP is yesterday's transfer technology. Managed File Transfer came of age several years ago. A 3000 app that was moved to embrace SFTP was hitting a target that moved already.
The point to remember is that the instincts that got a 3000 IT manager to their current post are still sound. Change only what brings needed value and business-competitive features. Foster's strategy says that "In 2006 there was no pressing need to migrate the [MPE/iX] application. A decade later, the pressure to ‘get off’ the HP 3000 has surely increased."
Where the pressure has become higher, the most crucial play involves moving data. More than 20 years ago, IBM was making a play to migrate 3000 systems to AS/400s. Computerworld called me to ask how I believed that'd go. "Not so good," I said at the time, "because there's all that IMAGE data that's essential to a 3000 customer. A tough move today." Migrations are easier now thanks to all the tech advances. It's business choices, though, not technology promises, that propel any migration.
May 18, 2016
Tape drive changer a powerhouse for MPE
Autochanging tape drives used to be the stuff of science fiction among 3000 managers, but those days passed by before HP cut off making Classic 3000 MPE V systems. Just because an autochanger is a 3000 storage option does not make it automatic to program, however.
A question posed to the community by Ideal Computer Services Ryan Melander reached for help on programmatically controlling such autochangers -- to select a slot, and load the tape and come to ready. "I am trying to configure an old DDS3 auto-changer, one that I don't believe will unload and load the next tape," he said.
Gilles Schipper noted that the command ad ldevno id=hpc1557a path=?? Mode=autoreply configures the device, and to advance tape after use, employ the command (from Devctrl.mpexl.telesup) ldev eject=enable load=online
Denys Beauchemin mentioned HP's pass-through SCSI driver as a tool to drive the device's robot. The software was built by HP's labs and labeled as "not for the faint of heart" by engineers, but can assert a programmatic control over autochangers. Some third party programs such as Orbit's Software Backup+/iX can also do this work.
If ever there was a theme song for an autochanger at work, it would be a tune called Powerhouse. Children of the Fifties and Sixties will know it as soon as they hear it, if they've ever watched a Warner Brothers cartoon.Some programming ideas came from Beauchemin, the engineer who developed at HiComp for the HiBack 3000 solution. "For the next tape to be brought online automatically, I seem to remember there had to be a special setting with the dip switches."
As for being able to control the robot itself, you definitely need to have the [HP] SCSI pass-through driver configured and loaded, and then you need a program to actually issue the IOCTL calls to the robot with the properly formatted SCSI commands. There was such a program a long time ago from a vendor, but that's all gone now.
Perhaps the high-test flutes and heavy octane horns of Powerhouse -- used in Duck Dodgers and the 24th and a Half Century -- can be put up on the MP3 player while fitting the driver to MPE. ("Oh drat these computers -- they're so naughty and so complex," says Marvin the Martian in one installment. "I could just pinch them.")
From our archives we found this advice from the late Jack Connor of Abtech, pointing to similar complex answers about controlling DDS changers.
Typically, there's a second SCSI port/address assigned for the transport control which allows the selection of specific tape. For MPE, stacker mode is typically selected, which tells the drive to just mount the next tape in line when requested. I don't know if the DDS autoloaders have a network connection available like the C7145NA DLT autoloaders do; with that device's web interface you can reload any tape, bypass a bad tape, and so on.
Back in 2011, John Pitman checked in to report that a much simpler solution to his changer's control needs popped up. "On re-examining my code for HPDEVCONTROL, I found I had catered for 1- and 2-digit device numbers in the string passed, but I had configured the drive as dev 777. This produced a string dev number of 77, which doesn't exist as a tape drive. Once I fixed this, it works like a treat."
May 16, 2016
A Spring When The Web Was New to You
Twenty years ago this month we were paying special attention to the Web. We called it the World Wide Web in May 1996, the www that does not precede Internet addresses anymore. But on the pages of the 3000 NewsWire released in this week of May, a notable integration of IMAGE and the Internet got its spotlight. We've put that issue online for the first time. The Web was so new to us that our first 10 issues were never coded into HTML. Now you can read and download the issue, and it's even searchable within the limits of Adobe's OCR.
As an application for higher education, IRIS was serving colleges in 1996 using MPE/iX. The colleges wanted this new Web thing, popular among its professors and students, to work with the 3000 applications. Thus was born IRISLink.
IRISLink is not a product that Software Research Northwest will sell to the general market. But SRN's Wayne Holt suspects that a generic version of something like it is probably being built in the basement of more than one third-party vendor for rollout at this summer's HP World meeting.
"The message traffic on the HP 3000-L Internet list shows that a lot of sites prefer the COBOL lI/IMAGE model over writing piles of new code in a nonbusiness oriented language," Holt said. "But people are telling them that won't fly in the world of the Web and - take a deep breath here - the time has come to dump their existing well-developed COBOL lI/IMAGE infrastructure on the HP 3000. Not so."
The integrators on this project made themselves big names in the next few years. David Greer convinced Holt at a face-to-face meeting at a Texas user conference where "I listened to him share his vision of what the Web would someday be in terms of a standard for access to resources and information." Chris Bartram was providing a freeware version of email software that used Internet open systems standards. Take that, DeskManager.
It was far from accepted wisdom in 1996 that the WWW would become useful to corporate and business-related organizations. Even in that year, though, the drag of COBOL II's age could be felt pulling away 3000 users from the server. An HP survey we noted on the FlashPaper pages of that issue "asks customers to give HP a 1-5 rating (5 as most important) on enhancements to COBOL II that might keep you from moving to another language." There wasn't another language to move toward, other than the 4GLs and C, and those languages represented a scant portion of 3000 programs. Without the language improvements, some 3000 customers would have to move on.
HP was struggling to find enough engineers to do everything that was being proposed in a wild time of Internet growth and innovation. We complained of this in an editorial. HP would tell a customer who needed something new in 1996, "Help me build a business case for that." As in, let's be sure you'd buy it before we build it. Puh-leeze, I wrote.
One business case - a need for a product - doesn't eliminate another. Some customers need COBOL 97 support, the speed of the Merced chip and the ability to run Java native on their HP 3000s. Maybe they need the COBOL support first, Java after that, and the Merced by decade's end. There are others who need the same things but in a different priority. If CSY draws its input only from customers, they pit one set of priorities against another. I doubt this is the intent of being Customer Focused - but it's what happens when every development needs a Business Case.
HP itself was still pulling away from legacy technology: systems running corporate IT that didn't even have an HP badge on them.
For many years mainframes from IBM and Amdahl have been among the most business-critical servers in the company, and on May 17 HP will replace those systems with its own. According to reports from the Reuters news service. HP 3000 systems as well as HP's Unix systems will take the place of those mainframes. That comment came from HP's CEO Lew Platt, interviewed while on business in Asia. The mention of the HP 3000 by the company's CEO begins to fulfill at least one Proposition 3000 proposal - a higher profile for the computer system within HP's own operations.
Proposition 3000, of course, was the advocacy push launched to put the 3000 in a frame like the propositions on the California ballots of that era. Changes in the infrastructure, voted in by the constituency. Computerworld "asserted that HP had been "put on notice" during the SIGMPE meeting at San Jose in late March where the Proposition was first presented to HP management."
May 06, 2016
Options for STORE You May Not Know
STORE is the default backup tool for every HP 3000, but this bedrock of backup has options which might exceed expectations for a subsystem utility. Gilles Schipper, the support guru of three-plus decades on the 3000, told us about a few STORE sweet spots.
To start, using the :MAXTAPEBUF option can cut a four-hour backup to three hours or less. Schipper says that increasing the buffer size default to 32K, from the usual 16K, speeds up the backup when STORE sees MAXTAPEBUF. "That's a pretty good payback for one option."
Backups don't need to be specified with an @.@.@ command to be complete. "People should really be using the forward slash," he says, "because it's easy to accidentally omit the Posix file structure if you're not careful constructing your fileset backup." The slash is so much better that a backup specified by HP's TurboStore will replace any @.@.@ operation with "./" Combining @.@.@ with exclusions can lead to omitting files which should have been in a backup.
Schipper says that including a directory on a backup is smart, but private volumes in use on a 3000 need more than :directory as an option.
"If you're using private volumes and the directory option, you'll only get the system volume directory — and you will not get the private volume directories," he says. The backup must explicitly specify the volumes through the ONVS= command, using the long name of the private volume.
The partdb option on a STORE command ensure that any 3000 databases which are incomplete will get backed up. Without partdb, if a root file "doesn't have its corresponding data sets, the root file won't get stored. It's silly, but partdb ensures they get stored." A privileged file can also look like a partial database, so partdb brings those files into a STORE backup.
A 3000 with HP's TurboStore, rather than just the default STORE, can take advantage of the :online command. "It will give you zero downtime, if you have that version from HP," Schipper says. But that begins to drift away from the no-cost STORE options available to any HP 3000 administrator or owner.
May 04, 2016
CPR for a Non-Responsive Console
On my HP 3000, after a short power blip, the console is now non-responsive. I can connect to the system's GSP port and the session is connected, but nothing is displayed. Neither <ctrl> A or <ctrl> B works. I type away, but get no response. I can then connect via VT-MGR and take the console :console !hpldevin and I receive all the console messages.
So, the messages are being sent (since I see them on the VT connection), but neither the physical console or the GSP gets any console messages. What can I try?
Gilles Schipper says
I believe a START NORECOVERY reboot is in order here. Since <ctrl> A <ctrl> B do not work, you will need to power-recycle the machine to effect a reboot. Presumably you would want to do this after gracefully stopping all jobs and asking online users to log off, if possible.
Depending upon which patch level your level of MPE is on, the :SHUTDOWN RESTART MPE command may also work from a logged-on session with at least OP capability.
Mark Ranft adds
If you haven't rebooted, I've seen similar issues. From the VT console can you try to do 'abortio 20' until it says no I/O to abort. A WHILE loop may make this easier. I've had luck with this in the past. But since Ctrl-B doesn't work, you may be out of luck.
Robert Thwaites notes
These are the simplest things to try first
Among the commonest issues: forgetting to do an x-on after a <ctrl>S (x-off) to stop output, so you can look at the line you are interested in. One time I saw another issue where someone had pressed <space> on the console and hadn't pressed <return>.
May 02, 2016
New encrypted hardware solves aged issues
Security standards have advanced in IT, while HP's 3000 hardware has not. Encryption resolves a key need for data security that's a part of the HIPAA regulations. The 3000 components won't allow for full disk encryption. There's another approach. A replacement hardware solution for MPE/iX -- which is still being used in the insurance industry -- has been on the market for more than four years. In fact, the hardware is all around us.
Encryption solutions for an older 3000 hardware's data are available. FluentEdge Technologies has sold a PCI-ready solution for Ecometry sites for more than five years. A built-in full-disk approach is only an option with a fresher OS, though. We don't mean the environment powering the application; that's still MPE/iX. The control of the hardware is where such new hardware can be put into play.
Virtualizing with the Charon HPA software offers several advantages over relying on HP's hardware. Component failures are a matter of when, not if, in 15-year-old hardware. If the 3000 isn't an A- or N-Class, it's even older. Shrink-wrapping replacement drives won't look as good to security auditors as a full disk encryption of recent-model components. Newer drives include broader options.
The virtualization of the MPE/iX hardware can become an encryption strategy. Alternative methods that rely on legal defenses don't exist like they once did. A security expert friend of mine tells a story about using lawyers instead of encryption. It's a story from a different time: the era when 3000 hardware was not so old.
The certified CISSP Steve Hardwick was once involved in a HIPAA audit. After the presenting the audit results to the CIO, the next question to be resolved was remediation -- bringing the systems into compliance. The CIO’s response, Hardwick said, took advantage of an older version of HIPAA instead of newer hardware.
The CIO said that "after consulting with legal counsel, we are taking no action to mitigate the deficiencies found in the audit. They have informed us it will be cheaper to litigate than spend funds on security changes.”
That was in the early days of HIPAA. In those days, that regulation lacked teeth. To rectify this, the US government passed the Health Information Technology for Economic and Clinical Health (HITECH) Act for enforcement. It defines what's a breach of information. plus the responses that organizations must take after a breach. Very quickly, the cost of not putting security controls in place changed, especially due to the enforcement defined by the act.
Bringing 3000 applications into line with regulations like HIPAA and HITECH usually includes securing the healthcare data. Full disk encryption is an option if the drives are controlled at the host level by an OS other than MPE/iX. At the host level, Linux is the controlling environment in a virtualized environment. Drives in Charon, for example, are disk images in OS instances such as RedHat. Choosing virtualization can supply something to pass an audit. It's not exactly brand-new hardware, but it can be generations newer and leave the old and reliable app software in place.
April 27, 2016
The Remains of Any Need for CPUNAMEing
Earlier this week a reseller in the 3000 market offered an N-Class 3000 with an upgraded CPU. The server's HPCPUNAME, however, reflects a slower model of the system. These names of systems are actually a part of the 3000 process that Hewlett-Packard regards as ongoing business. There are oh-so-many names of 3000 systems. In some cases, what your software sees in the name is not what you get.
HP's support for the 3000 terminated at the end of 2010, but there are a few HP services that continue today. Name changing is among them. Steve Suraci of Pivital, which sold 3000s as an official reseller until HP stopped that sort of business, says it's a lock that HP didn't do an N-Class upgrade that doesn't have a corresponding HPCPUNAME change. "For sure HP didn’t do the upgrade," he said. Outside the lines upgrades can be a way to skip software upgrade fees.
The only thing pertinent to the current 3000 community is HP's CPU rename service. When HP did this with support engineers at a time plus materials engagement, a software-to-hardware blessing changed the HPCPUNAME and HPSUSUAN numbers for replacement 3000s. When a CPU board dies, or a system needs to be updated at a fundamental level, Hewlett-Packard still owns the only software that can transform replacement hardware into your hardware — complete with reinstated numbers that allow third-party programs to run unfettered.
This service is still available, if you insist, from HP. A server that reports it's a 550 MHz N-Class, while it's actually a 750 MHz system, could use this kind of correction. What's important, though, is whether an N-Class will be fast enough. Reporting a different HPCPUNAME can keep third party software from running. That issue can be corrected by calling the software vendor, who'd probably be glad to hear from an MPE customer.
Of course, there's always the possibility the caller is a former customer, running software which is legal but not on support. Getting an HPCPUNAME change to make the software match the hardware then becomes a courtesy—or a service extended in exchange for renewing support. "You will know me by my commitments," says one sage, and keeping a support commitment seems like a good idea if you're using third party software products for MPE/iX in 2016.
Modern budgets for homesteaders, however, sometimes don't have room to pay for application and toolset support. HP can't hold a hostage anymore by waiting for MPE/iX support ransom. It ended all support business in 2011, business the independents picked up nicely. They don't have the use of SS_CONFIG (for the system up to 900 Series vintage) or SS_UPDATE (for the ultimate models of 3000s) can cost a customer on a time and materials basis, and HP's last stated plan said it would offer these reconfigurations of stable storage for an undetermined period.
HP long ago considered its customer communication to the post-2010 community complete. "We at HP believe we have responded to and addressed all of the HP e3000 end-of-life requests our customers and partners have made in recent years," one document stated while there was a e3000 webpage at HP. Such pages are long gone now, farmed away to the archives of 404-land and the Internet Wayback Machine.
The portion of HP which continues to touch the 3000 community in perpetuity is licensing operations. Software License Transfers between HP's 3000 systems sold on the used market are still being offered through HP’s SLT organization. This SLT operation serves all of HP’s licensed products, not just the HP 3000. Transferring a license officially is still an essential step to some customers, even when they're buying a used 3000 like an N-Class. That makes full disclosure of the HPCPUNAME important.
In 2008 while those webpages were still alive, HP was candid enough to admit that only a portion of its customers would make the effort to have 3000 MPE/iX software licenses transferred. There was no support eligibility that HP could offer after 2011 in exchange for this compliance with license requirements. Customers now need to supply their own reasons to do things by the book.
April 25, 2016
Proving concepts leads to hardware exits
They've been called straw men, and more lately proof of concept projects. These assessment steps have often represented significant change at HP 3000 sites. Few migrations got the green light to proceed with the raw change and full-on expense without demos of replacement apps. Even when the change was limited to applications only, with no platform replacement, testing with production data was the most secure choice.
That's why the strategy sounded familiar when Stromasys hosted its first webinar in years. The company calls its assessment engagement to test Charon a proof of concept. Led by Global Accounts Manager Ray LeBrun and system engineer Darrell Wright, the talk included a note on how essential the PoC step has been to success with the Charon virtualized system.
"We're pretty confidant that if we engage in a PoC with you, then we're 99-plus percent sure Charon will work for you," LeBrun said. "We will not engage if we're not confident this is the right solution for you."
Stromasys works with a site's production data to prove the concept of giving HP's 3000 hardware an exit date. MPE/iX and the applications, and of course the data, stay in place. However, LeBrun said Charon has also been "a bridge to allow you to get to a migration. We have folks who say, "I'm only going to use that  application for another two years. Well, two more years oftentimes becomes three, four, and five years."
The technology concept behind virtualization is well known by now. People are so familiar with it that LeBrun said the vendor gets asked regularly when HP-UX Integrity server virtualization via Charon is coming. The question came up in the webinar, too."It's not in our roadmap anywhere," LeBrun said of a Charon built to give HP's Unix hardware an exit date. "Even if it was, I probably couldn't say anything about it, but it'd certainly be viable if there was a business case for it."
A business case for migrations — either off HP's hardware, or away from an operating environment — usually gets built at a customer site only after the technology has been given an all-clear. Stromays Proof of Concepts are paid engagements, because they include vendor staff hours spent onsite, services to transfer data to the Intel-based Charon system, and training a customer's IT staff to use the software.
A five-day onsite engagement, followed by a 30-day test period, makes up a PoC. "We basically do a live install of your application [on Charon] in your environment," LeBrun said. "You're not testing for look and feel. You're testing your application with your data."
April 18, 2016
A Parts Supply Non-Problem for HP's 3000
One part of Hewlett-Packard's end-game fantasy about the 3000 pointed to parts. This was a server the vendor wouldn't build after 2003, HP warned. You could not be sure your server and its essentials could be serviced -- where would the parts come from? For the last decade and more, HP's 3000 parts have come from everywhere. About the only hardware services supplier constricted by the halt in HP manufacturing of parts was -- wait for it -- HP.
While practicing the careful shrink-wrapping of HP-built replacement motherboards, disks, IO buses and power supplies, the market has shared and sold ample hardware to replace 3000 systems. One reseller reported on the 3000-L he has hundreds of HP 3000 terminals on hand and was ready to send them to the scrapper. There might be sites where HP's tubes are essential for production operations, but I hope not. The scrap heap looks like the next stop for those 700/92s.
On the other hand, there are a few consumable items that make HP's hardware hum. One is essential to smooth operation of a service processor. You can get a replacement part for this processor at your grocery store.Craig Lalley reported this week that a customer needed their Guardian Service Processor looked after inside the HP 3000 hardware. The GSP is mighty useful for an A-Class or N-Class customer. However, it carries a battery inside that dies, like all batteries do.
(I still have an HP 14B calculator here, given to me as a 1989 HP 50th anniversary memento, which fires up each time I press the On key. I used to think it was solar-powered. How could any batteries last 27 years? Ah, the HP of old. Perhaps those batteries came by way of a NASA supplier.)
And the GSP's battery? Lalley says it's a CR2032, a part HP installed in a commercial server that once sold for tens of thousands of dollars. Or to put it another way, without ever owning an HP 3000, I've got a replacement part for an essential 3000 subassembly sitting right here in my hutch. Right next to that 14B with its 1989 batteries.
There are good reasons why HP's hardware for MPE/iX may not be a wise very-long-term component for production computing. But a lack of parts never was a good reason back in 2002 -- and it's just a tall tale today, even if the parts are aged like fine wine.
April 15, 2016
Use GSP and More for 3000 ID, control
There's been plenty of change in the 3000 manager's life over the last 15 years. Some of the change is a little easier by exploring the use of the Guardian Service Processor. It can help in changing the location of HP 3000 consoles from one part of a shop to another. There's even a story out there that says the identity of the HP A- and N-Class hardware resides in the GSP.
Kent Wallace, while a 3000 manager for Idaho-Oregon healthcare delivery system Primary Health, needed to move his 3000 console.
I must move the 3000 console another 10 feet farther from the rack (it's an N-Class HP 3000/N4000-100-22). What are the 3-pin positions on the wire that I need to extend its RS-232 cable?
Our blog contributing editor Gilles Schipper chipped in with a solution offering even farther movement:
If you want to extend the range of the console to anywhere on the planet (at least where there’s Internet access) you could consider the HP Secure Web Console to replace the physical console. Depending upon the condition of your physical console, this solution may also save a bit of wear and tear on your eyeballs, instead of adjusting those pins.
Schipper wrote us a great article on setting up such a web console. The GSP is quite the tool.
Former HP support engineer Lars Appel offered another take on Schipper's strategy:
While Gilles is right about the possibility of using the web console, it would probably be easier to use the already built-in dedicated LAN port of the N-Class systems that gives access to the GSP by telnet.
I prefer the “telnet console” over the “web console” because it gives more freedom in the choice of terminal emulator — whereas the web console typically lacks features like “easy cut and paste” or special key mappings (e.g. German language ;-) or something similar.
This prompted Schipper to clarify his suggestion:
Lars is absolutely right about the built-in “secure-web-console” that comes with all N-Class and all but the earliest A-Class e3000s.
And, yes, the built-in is definitely more functional, allowing cut-and-paste as well as telnet access, whereas the external variety has only Java access to it via a web browser and no cut-and-paste.
So, if one has a choice, the built-in is definitely superior and available with only proper configuration. However, the external secure web console is available for all HP 3000s, and would still be most useful where is internal secure web console is not an option.
The GSP is a powerful part of an HP 3000 that runs whenever the server is plugged in. It's the maintenance control console commanding the ultimate class of the server to reboot, do memory dumps and even fully power down the 3000. Consultant Craig Lalley has noted the GSP has one fewer feature than its Unix counterpart, though.
"On HP-UX it is possible to reset the GSP from the host OS," he said. "I have not found a way from MPE."
From time to time a reset may be required for diagnostics services on A-Class and N-Class servers. If your 3000 gets loving care from a consultant or service provider outside your computer room, you may need a paper clip to keep up service levels.
The GSP can also reveal the 3000's speedometer, as profiled near the bottom of an Allegro webpage.
The gap between controlling the 3000 and HP's HP-UX Integrity GSPs is a common shortfall of HP designs. Even though the 3000's MPE/iX includes a Posix interface, HP didn't engineer enough Unix into the 3000 to enable some of the administration that HP-UX users enjoy. (That lack of Unix can sometimes be a good thing, however, when a security breach opens up in the Unix world.)
But when a 3000 needs a GSP reset, pressing a recessed button on the 3000's back will do the trick if a telnet command doesn't work. You can telnet to the IP address of the GSP, log in and do the reset. But you can also get someone to press the physical reset button at the back of the machine. It's recessed into the cabinet so you may need a magic paper clip bent just-so.
Lalley calls the GSP, which HP introduced with its final generation of 3000s, one of the most useful things in the A-Class and N-Class boxes.
The GSP is a small computer that is always powered on when the plug has power. With it, it is possible to telnet to and BE the console. While multiple admins can telnet in and watch, only one has the keyboard.
It is possible to reboot, memory dumps and even fully power down the HP 3000 from the GSP. Use the command PC OFF.
It is probably the best feature of the N-Class and A-class boxes. The problem is sometimes it needs to be reset, usually with a paper clip. Since the GSP is a different CPU, this reset can be done during business hours.
April 13, 2016
Stromasys returns to webinar lineup April 19
Emulation and virtualization vendor Stromasys returns to the web airwaves next week with an event that examines the return on investment for its Charon family of software products, including its HP 3000 model HPA. The event begins at 11 AM EST for US viewers, 5 PM CEST for Euro IT managers.
The company last did a webinar for the 3000 market in 2012. At the time, Charon HPA was still a nascent product, without the portfolio of success stories and case studies it's built in the 3000 market. The previous show explained the concepts and demonstrated administration and management of the software. The April 19 event will reach for more strategic perspectives, but will include a technical angle, too.
Worldwide Strategic and Global Accounts manager Ray LeBrun will be joined by Systems Engineer Darrell Wright on the one-hour webinar. Doug Smith is the HP 3000 product manager for the company. Smith is managing a limited-time discount offer on Stromasys services for Proof of Concept and full integration of Charon HPA.
The company promises the webinar will offer
An overview of legacy system pain points and the difficulties businesses like yours may encounter as they determine how to move forward with their legacy systems. Learn how your organization can improve the ROI of your legacy systems while also minimizing risks of unplanned downtime.
Stromasys says that space is limited for the webinar. The event includes information for Digital and Sun systems as well as the 3000. Registration is online, which yields a confirming email that provides login instructions for the GoToWebinar event.
April 11, 2016
Yes, Virginia, there is MPE at the Terminal
One of the first HP 3000 migration success locations might have hung on to an MPE/iX app since its story was introduced by the vendor. A lively discussion popped up last week when Don Seay asked on the 3000-L mailing list about running Speedware on the Stromasys Charon HPA virtualizer software. The chatter included updates on the work to cross the 2027 hurdle for MPE/iX use, as well as reports on the speediest settings for Charon.
Seay was emailing from an address at VIT.org, the legacy location of Virginia International Terminals. It's the port authority for all shipping in Norfolk, Newport News and environs. A shiny website handles just about all of the data requests at portofvirginia.org. But there's still data being fed to VIT.org, and Seay's request seems to hint that an application continues to work there. We're checking in with him.
Taking a full-on approach to a migration is a typical opening strategy, but there are sometimes good technical reasons why apps remain on 3000 hardware. This didn't seem likely when we first heard about the 3000 and VIT in 2002. HP was promoting the practices and concept of retiring 3000s during that time, the first full year after Hewlett-Packard's announcement it would leave the 3000 marketplace.
VIT’s assistant IT director at the time, Clark Farabaugh, said at HP World 2002's migration roundtable the decision to shift to HP’s Unix servers “has changed our shop, for better or worse.” That summer, IT began to migrate at VIT. The organization took delivery of a HP 9000 rp8400 server to replace its HP 3000s, and Farabaugh said “we were the first ones on board.” We took note of the report of 13 years ago.
Even though Farabaugh described TurboIMAGE as “the fastest database I’ve ever seen,” it was moving toward using Oracle on its migrated system. Three IT staffers — VIT said it was doing all the work in-house — were trained in database admin for Oracle.
The applications running at VIT handle shipments through a terminal with 7,000 international longshoremen at work, and a desire to Web-enable the apps led VIT away from the 3000. The IT director said the migration project will take 12 to 18 months to complete using the 45-person IT staff, taking apps from Speedware on the 3000 to Speedware on HP-UX.
A baker's dozen years later, it's a surprise to see that perhaps something on MPE has continued to lift bits of the cargo of terminal data at VIT. A datacenter manager and IT execs do need to keep everything online, though. If a 3000 survived the HP-UX shift for this long, it's bound to have a better reason than "we don't really want to change things."
Does MPE/iX still exist in the field? Some in your community feel that asking about that is akin to asking about Santa Claus. But yes, at Virginia, it seems there's still MPE at the Terminal. If we're reading the messages right, it's doing more than a fairy, too.
April 08, 2016
Hardware's emulation puts software at ease
In the earliest days of the 3000's Transition Era, advocates for MPE/iX formed the OpenMPE user group. But the first campaign for these engineers (and a few businesspeople) was for the emulation of MPE itself. The ideal was that if MPE/iX source code could be turned over to the community -- since HP had no real interest in the future of the 3000 -- then the OS and its subsystems would be pushed onto newer hardware.
The ideal was open source for MPE/iX. That campaign assumed plenty of change was in the future of 3000-based software. The reality that formed about compatibility of software is illustrated in the everyday experience of Charon users.
One checked in this month with a summary of how smooth his software slipped into the Charon HPA environment. The emulation that paid off was virtualizing the RISC hardware. The caliber of the solution made things easy for Jeff Elmer.
The report was sparked by a question about whether the Speedware 4GL suite was in production in a Charon site.
I can say that since what is emulated is the PA-RISC hardware and not MPE, it seems unlikely that there would be any software incompatibilities. Everything we use (multiple third-party tools plus in-house COBOL/IMAGE software systems) just worked. It really was true that no one would have noticed a difference unless we told them.
The single item that we had to modify was in our backup job stream. We had a tape rewind command in the job that was no longer needed and which the emulator at that point (in 2013) did not understand. The "fix" took less than 60 seconds when I removed that clause from the job.
In summary, I would expect Speedware to work without incident but I couldn't speak to what combination would provide optimal performance (that is, which class HP 3000 should be emulated or what physical hardware should be under it). We spent a long time testing the emulator without charge before we proceeded with the purchase. I would think the possibility exists that Stromasys would extend a similar courtesy to you so that you could find out first hand with your data in your environment.
In fact, there's a Proof of Concept arrangement that Stromasys uses today to introduce its product for this kind of evaluation.
April 06, 2016
Stromasys reports aim at speed, and help
The fine art and craft of tuning an Intel-based server to mimic HP's 3000 hardware has evolved. The Charon HPA emulator has been in production shops for more than three years. In the beginning, the software's demands on hardware were outlined in a table of preferred servers. Or in calls to a product manager. The latter has always produced more robust performance than the former. A recent string of messages on the 3000-L showed why. They also showed that a 3000 jobset that ran three times faster, after "setting power management to dynamic."
Performance tips on the L about selecting and tuning for the best hardware have included the following advice
Set other settings for performance
System Isochronous Mode enabled
Hyper-thread off or 1
Intel Speed Step enabled
If this set of instructions doesn't make much sense to a prospective user, it illustrates why Charon HPA is a fully-guided product by now. Customers and prospects buy services from Stromasys to deploy this solution. There's no other way. Downloadable freeware copies left the marketplace last year.
Emulating a legacy hardware server to run enterprise-grade applications is not a hobbyist's mission. Stromasys product manager Doug Smith says the customers have been better served with engineering-driven integration insights. He's got success statistics to prove it.The nuances of installing and integrating Charon for the success include networking deployment advice. Ray Legault, a systems manager at Boeing, shared these insights when another manager asked him about the impact of networks on Charon.
I did not perform any tests over the network. My actual servers are 1,800 miles away. In Linux, I make sure my ETH1 is set correctly.
ethtool -K eth1 rx off tx off sg off gso off gro off txvlan off rxvlan off
To avoid having to redo every time you reboot, just add this line to the bottom of the ethtool file:
post-up ethtool -K eth1 gso off gro off
So it looks clear that knowing Linux's ethtool will be an essential skill of integrating Charon, too. Expert services are now crucial for the product. That's why it's become a solution for the serious user, one trying to eliminate the need for 15-year-old HP hardware.
April 04, 2016
Working to Set MPE's Future to Forever
When a 3000 manager asked about running Speedware on the Stromasys Charon HPA emulator, the question evolved quickly. In just a few hours, MPE experts were talking about how long the OS could keep running. The detour of the 2027 CALENDAR intrinsic came up. It turns out the community experts are already working on that.
Jeff Elmer of Dairylea Cooperative, whose success story with Charon was part of our 2014 reporting, told the readers of the 3000-L that he's pleased with the way the Stromasys product cut out HP's MPE/iX hardware. The words "run MPE forever" were part of his message.
We used HP's 3000 hardware for 30 years. We've been using the HPA3000 emulator in production since December 2013. Our users would have never known the difference if we had not told them.
We had a 969KS 100 and went to a 2-CPU A-Class on the emulator. Performance is essentially identical but all concerns about "ancient" hardware went away. (Our RAID array hard drives were older than our web developers). Charon is running on a 1U "off the shelf" Proliant server under the Red Hat Linux environment (if we didn't have a DLT8000 and a DDS tape drive attached to it, all that it would take up in the rack would be the 1U). We run our disaster recovery version of the emulator in another location under VMware on OmniCube hardware, although we have never used it for anything other than testing.
"Based on our experiences we would recommend it to anybody," Elmer said. "You could run MPE forever with this setup and over time your performance would only improve as you put newer, faster hardware under it." Whoa, forever? It's the promise of virtualized servers that emulate antique hardware. But MPE/iX has that calendar problem that'll rear up at the end of 2027, right? Not so fast there, said one MPE expert.When Denys Beauchemein said that forever really meant December 31, 2027, Robelle's Neil Armstrong begged to differ. "That doesn’t stop the system from running, and a lot of issues can be handled at the application level quite easily," he said.
There are people working on such things at the OS level. I’ve been reducing the dependency on CALENDAR in all our software as well. By reducing the number of calls to CALENDAR, this helps mitigate the impact, of course, and adding options to change the result of a call to CALENDAR directly after a call are being considered.
It is an interesting problem.
Beauchemein retorted with the viewpoint of the IT manager who needs to be away from MPE/iX, since it's old.
That is fantastic news. Now I need to find my abacus and see if I can get them to refurbish it along with my slide rules.
Has anyone here even booted a 3000 for December 31, 2027 and see what goes on at the virtual stroke of midnight? Unfortunately, I do not have one handy.
All that Armstrong could offer in reply about a Jan. 1, 2028 bootup was "Yes, it’s been done. Nothing catastrophic happens."
He is right, nothing catastrophic happens. But what does happen?
SETCLOCK allows a manager to revise such a future date and time, but only up to December 31, 2027. It won't accept a higher date; it reports "out of range" if you try. MPE/iX continues to run. After midnight SHOWTIME will give the wrong result, year 1900. But if your application doesn't care about a time stamp (which is unlikely in a business computer) this doesn't matter. SHOWTIME will show Jan. 1, 1900. If days of the week matter, by setting the system date to Y2K, the day of the week will align with the correct day for 2028.
One of the sharpest of the MPE minds, Vladimir Volokh -- who co-created MPEX -- has given us a deep dive on that set of CALENDAR problems. He reports he's done that 2027 boot 10 years ago. In our conversations, his opinion has been that someone in the community will find a way to reroute MPE/iX to a future in 2028 and beyond, relying on current dates.
Dates don't vex MPEX, Vladimir added when we talked about this a few years ago. MPEX can do operations with dates—because unlike COBOL, or even the first language of MPE, SPL, MPEX has a DATE datatype. "If you have MPEX," he said, and here we could hear a wink, "and who doesn’t—DATETOCALENDAR is a function in MPEX."
Last week's exchange on 3000-L shows there's already work underway on getting beyond 2028. The thing about an abacus is that it does still work. Ours hasn't needed to be refurbished since we bought it in 1989. Unlike an abacus, MPE/iX has a purpose for being an everyday tool when the software has close ties to specialized business logic. It has a better reason to keep working forever—however you define that timeframe.
March 30, 2016
Big G anniversary recalls era of 3000 crunch
This month marked the 150th anniversary of General Mills, the benevolent cereal giant that started its business just after the Civil War milling flour. The maker of Wheaties, Gold Medal Flour and Play Doh, the company known as the Big G got a rousing eight minutes of celebration on the CBS Morning News this weekend. When the report turned to Wheaties, it triggered a memory of one special era for the HP 3000. MPE/iX once managed a giant boxcar-load of operations for the food company, a firm so large it acquired fellow 3000 customer Pillsbury in a 2000 deal that teamed century-old rivals to make the world's fourth-largest food company.
Powerhouse was an essential part of the Pillsbury legacy, but the reach of the 3000 was even deeper at General Mills. Mark Ranft, who operates the Pro 3K consultancy, said his time at the Big G covered the years when core corporate functions were controlled by a fleet of 3000s.General Mills was glad to point the way to lifting the 3000 into a higher rank than Unix. In the period where The Unix Hater's Handbook was making the rounds, IT Manager Mike Meinz booted out HP-UX from General Mills' datacenters after a brief fling. In language of the era, Computerworld said that General Mills "tried Unix, but it did not inhale."
"There is a panacea of thought that you have to have Unix," Meinz said in the article. "You don't have to have Unix."
General Mills went so far as to pull an HP 9000 out of the IT lineup and move its warehousing application over to its HP 3000s. The company was just into the process of converting those Classic 3000s to PA-RISC models. The vendor was taking steps to position the 3000 as a less-proprietary choice. "Not only is the HP 3000 open," Meinz said in the ad, "but it's an excellent, easy-to-use transaction-processing system for business-critical operations."
The headline that provided too-rare coverage of the 3000 in Computerworld enjoyed a joke at the expense of Unix. "Cheerio to Unix, cereal giant says," noting that the 9000 was chosen at first because it was the only platform that could host a preferred warehouse system. General Mills bought the source code for the application and did the porting. "What followed became a testimonial to MPE's portability," the article said. Meinz said he had anticipated the porting project would take six months, but it only took two. And much of that time was spent developing enhancements rather than actually porting it."
March 23, 2016
Putting a CPUNAME on HPSUSAN's profile
The MPE is a most unique creature of the computer ecosystem. This is software that does not have its own license, specifically. According to HP, the ownership of any MPE/iX version is determined by ownership of an Hewlett-Packard 3000 server, one built to boot up MPE/iX. When a copy of MPE is moved onto a Charon virtualized server, it must come from one that's been assigned to one of HP's 3000s.
We reached out for clarity about this when a major manufacturer was looking into replacing HP's 3000 iron with Charon licenses on Intel systems. After the MPE/iX software is turned off on any replaced 3000 hardware, does its hardware-based license then expire? The operating system license, according to HP's MPE Technical Consultant Cathlene Mc Rae, is related to the HPSUSAN of the original HP hardware.
So wait a minute. Are these HPSUSAN numbers of 3000s considered de-licensed, even if they're going to be used on the Charon emulator? Mc Rae explained.
The HPSUSAN number is different from the MPE/iX license, although there is a relation between the two. The ability to use MPE/iX on the emulator is a result of completing a Software License Transfer. The original MPE/iX license on the HP e3000 would then no longer exist.
In the hardware world of HP 3000s, HPSUSAN takes the original serial and model numbers on the system. It remains the same, as long as the customer owns the system. This combination was used to ID the hardware and enable diagnostics for the correct system.
However, that transferred license for the MPE/iX installation on the Charon emulator -- available via a $432 Software License Transfer Fee -- won't be getting a new HPSUSAN number during the process. HPSUSAN gets re-used, and so it leads us to see what HPSUSAN stands for, and how the HPCPUNAME is a key in emulator installations.The U in HPSUSAN stands for Unique, as in System Unique Serially Assigned Number. Mc Rae said that HPSUSAN is one of a kind for HP-built 3000 systems. But SUSAN doesn't designate an MPE/iX license, even though MPE is licensed via hardware ownership.
Mc Rae explained to us, and to any Charon prospective user, "MPE hardware and software was created before the technology of virtual systems and emulators, in the 1970s. Licenses were based on hardware ownership."
This sounds familiar. HP once compared the licensing of MPE/iX to license plates issued for a car. They could not be separated, these numbers and the car that was the HP 3000 iron. (Let's just put aside the common practice of those metal-plate days, when they'd give you a new number after your plate was older than 8 years in Texas.)
In 1999, HP was busy suing Hardware House and a few other resellers over the resellers' separation of HPSUSANs from HP's 3000 hardware cars. The House was taking other PA-RISC servers and pressing valid HPSUSAN numbers onto the non-3000 iron. People went to jail. Lo-jacks were ordered for ankles.
Thanks to the passage of 17 years' time, an HPSUSAN number can now move to a USB thumb drive plugged into a Charon Intel- or AMD-based server. Those license plates can travel to a newer model of car. The emulator's HPCPUNAME, however, can only be designated as an A-Class or N-Class system, according to HP's knowledge. That'll likely be a reason to contact all software vendors whose products operate on the replaced HP 3000 iron.
You see, vendors use a combo of HPSUSAN and HPCPUNAME to control licensing. Products such as Infor's MANMAN or PowerHouse not only want to read HPSUSAN -- which you can move to Charon -- but also HPCPUNAME. If you're moving off a Series 979, for example, "979-100" isn't an emulated system under Charon. No 979-100 for HPCPUNAME. You've got to get license permission from your software vendors to enable an A-Class or N-Class HPCPUNAME.
The HPCPUNAME on the Charon system may not be set to 979, Mc Rae said. "Based on the Charon HPA/3000 family, it is assumed that the HPCPUNAME will be set to an A-Class or N-Class CPUNAME," she said. "For example: HPCPUNAME = SERIES e3000/A500-200-50. As far as I know, Charon can only emulate A- and N-Class systems." That's true: a Series 9xx model isn't on the HPA/3000 product list.
The silver lining in this cloud is that you're only doing this contacting and CPUNAME-changing once. Moving to an A-Class or faster CPU from a 9x9 system is the last time you'll be changing from an unsupported CPUNAME to something included in the Charon product line.
In short, independent software vendors are going to have to be contacted, if they've licensed their products with the HPCPUNAME-HPSUSAN combo on a Series 9xx. Contacting your software vendors about a system upgrade is a fair business practice. But it's more than the right thing to do. Series 9xx users headed to the emulator look like they need that refresh to boot up their indie software.
March 21, 2016
Free software worth the time to track it down
It's entertaining and heartening to discover someone who's new to the HP 3000 and MPE. Fresh users tend to run in the hobbyist lanes of the IT race these days. Sometimes, however, they can ask questions that uncover values for the existing managers of the MPE server.
That's been the case with Michael Kerpan. He's just discovered the new freeware simh emulator engine for creating MPE V Classic HP 3000s. Kerpan is just pursuing this as a hobby project. "I'm not retired, but I'm also not in the IT business at the moment," said, "though I do maintain my SF club's library catalog server, which is a Linux box."
On the HP 3000 front, his box is a Windows server running simh, but Kerpan wants more than just the stock MPE V Fundamental Operating System to use. Kerpan specifically asked about the old Interex Contributed Software Library. The CSL started out as a swap-tape built from reel tapes that attendees at conferences brought along. Drop off the programs you wrote on your reel -- or eventually, DAT tape -- and pick up a compilation of such contributed software when the conference adjourned.
The CSL dropped off the radar of the 3000 community once Interex went bankrupt. The collection of programs wasn't even listed in the organization's bankruptcy assets. In some places out in the community CSL tapes still exist, but trading them hasn't been a compelling pasttime. However, MPE contributed software, now called open source and freeware, still exists. Knowing where to track it down is often worth the effort, if managing a 3000 is still your job.The biggest sources of contributed freeware live on three servers with public access: 3k Associates, Client Systems, and Fresche Legacy (formerly Speedware). The first company is an Internet-savvy company from as far back as the early 1990s; its founder Chris Bartram wrote the first non-HP email application, NetMail/3000. The 3k website has dozens of programs.
In terms of number of programs, the Client Systems-Speedware servers are right in the running. They're the licensed sites for HP's former Jazz public server content. About two dozen or so programs, plus some UDCs and scripts, are available for downloading. Without a doubt the Speedware server is better organized and uses a cleaner interface. HP forced these companies to install a vast EULA on the front end of these collections. You must click through it each time you access. Annoying, but just about toothless by now.
Other locations for freeware include Allegro Consultants and Beechglen. Those companies include products they've built to sell and use in their own labs work. Some of the for-sale titles have gone free now there's less market for 3000 software. There's also freeware available through AICS Research, the QCTerm free terminal emulator. A 3000 Classic emulator in software, attached to an emulator of the HP 2392 terminal, represents the ultimate in live-forever status for the legacy of MPE. It's something like the mighty state that Gandalf occupied once his body -- that hardware -- was gone forever.
Live-forever is also a promise of the Stromasys Charon emulator for the final generation of up-to-date 3000 boxes. As for the CSL, it's bound to live forever, but in locations as mythic as anything in Lord of the Rings. Perhaps we'll see a return of the king of contributed software.
March 18, 2016
Big files get zipped, moved on HP 3000s
A computer manager who's new to the HP 3000 is looking for CSL files this week. The Contributed Software Library is just an oasis to this IT veteran, something shimmering in his future that holds a highly useful thicket of utilities and more.
Someone in the 3000 community is bound to connect our new user with this CSL, for one reason: he's looking for MPE V programs to supplement his discovery of the emulated Classic HP 3000, simh. That's the MPE V-ready version of a virtual HP 3000: what amounts to a CISC skin for a 3000 on top of the simh code. Whenever the newbie connects with a CSL resource, if they've got their files on a 3000 they're bound to need to send about 24MB to him. That's going to require zipping them.
The act of zipping to compress for a transfer is an essential in 3000 management. Although the code for compressing files on HP 3000s is more than a decade old, like a lot of things on the system, it continues to work as expected.
Tracy Johnson, who manages the Invent3K server operated by OpenMPE, noted he's using the MPE/iX Posix shell's compress and uncompress. "It creates a file that ends in capital Z. Seems the compressed format is compatible with both GNU-zip and Winzip programs or any other Unix/Linux machine."
Lars Appel, who ported the Samba file sharing tool to MPE, offers a comprehensive answer. He points to an HP 3000 Web starter software kit that resides on a development server, open to the public.
Lars explained in a post a few years ago:
You can pick up the InfoIP zip/unzip programs (in a tar file) at my area in the Infocorp server. The link in that webpage that contains the zip/unzip programs is
Transfer it to the 3000 in bytestream or (fixed) binary format and then unpack with :/bin/tar "-xvzopf FILENAME". Place the two programs where you like; I typically have them in /usr/local/bin or (with uppercase filename) in a group or directory that is part of my HPPATH settings.
The webpage also contains a tar.Z file with /usr/local/bin/gzip
(gzip -d decompresses; creating a symbolic link gunzip is also useful)
March 14, 2016
Upgrade bargains on 3000s remain in play
"We still have units that are licensed and salable," said Pivital Solutions' president Steve Suraci. "We still have customers occasionally looking to upgrade."
Prices for even the largest of HP's MPE system line are being quoted below $10,000, and in some locations, a deeper discount than that. Like the goods sold in the basement at the legendary Filene's, the word cheaper comes to mind—because the pedigree of each 3000 system's MPE license is sometimes the most important element. (Healthy disks are pivotal, too.)
Bonafide machines have valid HPSUSANs. It's essential for moving MPE apps and utilities during an upgrade. In the scruffiest days of the 3000 resale history, HPSUSANs were being slapped onto HP's L-Class hardware with rogue software, making a 3000 out of a cheap 9000. People went to jail over that episode from the end of the 1990s.
But even a valid HPSUSAN is not the same thing as a proof of license. A continuous chain of ownership paper trail makes for a fully-licensed system. Such a license can be important to the customers who care about keeping auditors happy. That level of validation isn't required for a support contract, though, since HP's long been out of of the MPE/iX business.
March 11, 2016
New 3000 simulator looks back, not ahead
Community members on the 3000-L newsgroup have been examining a new entry in the emulation of HP hardware. However, this simulator creates a 3000 under Windows that only runs MPE V. The MPE version of SIMH — a "highly portable, multi-system simulator" — is a Classic 3000 simulation, not something able to run PA-RISC applications or software.
Some 3000 users are embracing this software though, maybe in no small part because it's free. It's been more than 15 years since HP supported MPE V and the CISC-based systems that launched the 3000 line starting in 1972. One of the experts in PA-RISC and MPE/iX computing, Stan Sieler, briefed us on what this freeware simulator can do, and what it cannot — in addition to not running MPE/iX.
Currently only Charon from Stromasys runs PA-RISC. Thus, the SIMH runs only the Classic HP 3000. At the moment, it’s an old version of MPE V (Q-MIT, release E.01.00)
And, the machine probably has no networking support. It probably has some kind of serial datacomm support, but I haven’t looked at that yet (all my use has been via the simulated console, LDEV 20).
I’ve put several hundred CM programs on the “machine” to see which will load and run. Many won’t, because they use newer features (e.g., FLABELINFO intrinsic which came out on the T-MIT with the Mighty Mouse).
So, you ask, can you put a newer version of MPE V on the emulated 3000?
The answer is, I don’t know. If I recall correctly, the machine isn’t emulating (yet) the “Extended Instruction Set,” but the authors claim MPE has a run-time emulator for them, so perhaps that won’t be a problem.
It comes with a version of MPE V, if you download the two packages that the release notes file mentions.
It’s fast. On my Mac, it runs CPU bound stuff about twice as fast as a 400 MHz HP 3000 would.
This is classic software running on classic hardware, so it's strictly for the hobbyist. Or someone who still has MPE V apps running their company. The software is downloadable from Trailing Edge in a pre-compiled .exe file.
The discussion has already generated 40 messages on the 3000-L, easily the biggest discussion of the year.
March 09, 2016
Powerhouse MPE futures clouded in silence
Users of Powerhouse have a mailing list, much like the one that HP 3000 owners and managers have enjoyed for decades. Powerhouse-L has traffic on several flavors of the Cognos creations. It's that range of product platforms that gives readers a chance to compare.
The news for MPE users of Powerhouse is that there is no news. This isn't a fresh take on the future of Powerhouse, Quiz, and the other products like Axiant which remain in use in the homesteading marketplace. Ken Langendock, a consultant in the market, asked on Powerhouse-L about the future of the products. He added that getting a response from Unicom Global, the owner of the products, has been difficult for his boss. It's not a good sign when a customer cannot get multiple calls returned.
Langendock was plain-spoken about which Powerhouse he needed an update for. A webcast from the vendor on March 4 didn't include Powerhouse futures by the 30-minute mark, so he pulled the plug on his viewing. He said MySQL support was high on his list of needs.
"The HP version to me is dead," he said. "I expect nothing more to happen on that version." That tracks with the reality of 2016 management from Unicom. When the software changed hands at the start of 2014, hope for changes to licensing and features rose in the MPE user base, but not for very long. Unicom owns scores of products by now, using a model that runs smoothly for Infor, owners of MANMAN.
Ownership of a product includes a paid support option, but not much will change for the 3000 world. An update from Bob Deskin, longtime Powerhouse product manager who consulted with Unicom in 2014, made no mention of MPE/iX after he reported his contact with Unicom.
We've reached out to Russ Guzzo, who heads the company's communications efforts and led the integration of Powerhouse into the company, but didn't get a reply to our question about, as Landendock put it, "doing something with Powerhouse."
"They assure me that they are continuing to develop and support PowerHouse," Deskin said. "I’m told that they are in catch-up mode because of the volume of support issues that have been reported."
They asked me to pass on the following concerning current releases which tells me that they are indeed continuing to develop and support PowerHouse.
For PowerHouse 4GL and Web, the current Windows version is 8.41U, and for Axiant 4GL, 3.4u. They support Windows Server 2012, MS SQL 2012, and Oracle 12c.
For PowerHouse 4GL and Web on Unix and Linux, the current version is 8.43G build 3350.101.It supports RHEL 7 and Oracle 12c. This is still a G version because they couldn’t find anyone to beta test the U version. It is a fully branded UNICOM version and is compatible with the Windows 8.41U and Axiant 3.4U even though it might seem otherwise. They still plan to have an 8.43U version soon.
The OpenVMS current version is still 8.40G and while 8.40U is still planned, there are major support issues on this platform.
If you'd like more information, here are the contacts:
UNICOM client services department at: Powerhouse.email@example.com or call 818-838-0606
For technical questions (may require an active support agreement):
Powerhouse.firstname.lastname@example.org, or call within the US toll free 800-866-6224; outside the US 973-526-3888
I’ve always found that providing support to UNICOM customers has always been their top priority.
March 04, 2016
3000-L breaks silence with DTC primer
On the verge of four weeks without a new message, the 3000-L mailing list and newsgroup delivered a primer on the control of DTCs on MPE/iX networks. A longtime contributor to the community set up a question about using these venerable devices that connected HP's terminals and other devices to a 3000.
I have multiple HP 3000s sharing one DTC. My problem is, which one controls the DTC? In the event of a power cycle, there is a race between them for which 3000 will download the new configuration. I need system A, my A-Class, to be in charge. However, systems B or C are most likely to download first, leaving me with the manual step of unplugging the network from B and C, power-cycling the DTC, and then waiting for the A-Class to download the configuration. Is there anything that can be done to just leave the A-Class in charge?
Tracy Johnson replied, "It's always been a crap shoot. I'm of the opinion the first HP 3000 to notice the DTC needs downloading will do the job. Which usually means the less busy machine." As it was just the first answer on the newsgroup, there was still a need to do unplugging.
"For example," he added, "if you have a test machine on the side nobody's using, I would simply unplug the test machine's network cable until the DTC was reset. Since there will be no other machine in control, the regular machine will do the download. When complete, I'd plug the network cable back into the test machine. However if both machines are in production and you have users active on both, you'll have to make a decision."
Lalley came back with an answer of his own, at least to eliminate the travel from chair to server.
"On the DTC's I guess it could be possible to do it without leaving your chair, if you TELNET to the console on B and CDTCCNTRL (has to be run from the physical console) to stop the DTC subsystem, then restart it."
Mark Ranft at Pro3k had another idea.
Back in the good old days, we used DTC Switching - a feature that allowed a DTC prompt to appear at a terminal. The user entered 'C HPA' to connect to host HPA. I believe you could set a default system. Initially DTC Switching was only available if you configured and downloaded the DTC using OpenView DTC Manager software running on a PC. Later HP set up NMMGR to allow DTC Switching.
The Communicator for that release may have more details, but I only entered the MAC of the DTC on the one system that was supposed to download the DTC. The other systems would have the DTC name, but not the MAC. This allowed the system to have printers and other devices accessible on the DTC without concerns about which one downloaded the configuration.
Of course, the location of that MPE Communicator took all of 41 minutes to dig up. Barry Lake pointed at HP documentation (now hosted outside HP's baffling website) that covers Enhanced Host-Based DTC Management Functionality in Chapter 3. Plus, a manual on Configuring and Managing Host-Based X.25 Links.
Not too shabby for a mailing list that's more than 20 years old, but still manned by dozens of experts. It's the kind of expertise a good third party support provider offers—as usual, better than HP's today.
February 26, 2016
21 days of radio silence on the 3000-L
The slowing current of 3000 communication showed a fresh signal by the end of this month. As we write it's been 21 days since a message of any kind on the 3000-L MPE newsgroup. The resource that carried 45 messages during last February has 10 for the current month. All of this month's traffic was wrapped around finding resources: Brian Edminster of Applied Technologies and Vesoft support. Both were located.
However, the three weeks without a new message is new territory for the community's log of technical help and outreach by cohorts. Among those who were posting during 2015, several told us they're on the mailing list-newsgroup out of habit — rather than needing details for their datacenter's 3000s.
"I’m still on the list out of inertia, nostalgia and mild interest," said Dave Heasman, a UK IT manager. "My employer got rid of their 3000s and me in 2008. Bought a series of packages to replace a big bespoke brokerage/investment system."
Robert Mills said he "remained a member of the list, mainly as a lurker, to keep appraised of what was happening in the 3000 community. Except for three requests in September 2012, December 2014, and February 2015, I've only posted to the list when I felt that the 3000 knowledge I had would help somebody solve a problem." Mills said he retired when his company went insolvent in 2009, but he's kept his hand in IT.
"I have been involved with the GnuCOBOL (formerly OpenCOBOL) Project on SourceForge since October 2014, and decided to write a macro preprocessor that emulated the functionality available on the 3000," he said. "The preprocessor, CobolMac, is now in its 5th version (B.04) and has received good reviews by its users."
Others who contacted us said they haven't worked on the 3000 since the days that HP sold support for MPE/iX. "I have been a BizTalk developer full time since 2008," said Kent Wallace. "I needed to work, and this was the direction the world was going." The 3000-L still has more than 500 subscribers on its mailing list rolls, but much of the messaging comes from consultants and vendor experts, supplying answers to questions and tips. A total of 45 messages have passed through the list since the start of 2016. The IT pros like Wallace have taken the path to other platforms, first to HP-UX, then to Windows.
"I left my previous employer in Boise and I moved to a Microsoft shop, whose mainframe was HP-UX," Wallace said. "However, in 2015 we migrated off HP-UX and onto SQL Server on Microsoft Server 2008. We do health insurance and the purchased software, Trizzeto, was moved to MS SQL servers."
Another registered user keeps up with the community, but he can imagine a future where he'd be back on the MPE/iX front lines. "We're totally out of the 3000 business," said Ted Johnson of Wake Forest University, adding a sad-face emoji. "But I love seeing the 3000-L posts and keeping up. Who knows — maybe they'll get rid of me one of these days, and I'll end up back on a 3000."
The 3000-L, hosted at the University of Tennessee at Chattanooga where the late Jeff Kell launched it the early 1990s, holds more than two decades of traffic. 10 years ago the list was big enough to measure a signal-to noise-ratio, but by now it's almost entirely signal. When John Burke was a monthly columnist for the NewsWire who summarized its content in net.digest, he rounded up the following help in just one month's communications. For a 3000 owner managing a homesteading shop, the 3000-L's tips still carry some value.
• Do you want to know when a particular account or group was created? LISTACCT and LISTGROUP are no help. But “listfile /ACCOUNTNAME,3” for the account or “listfile /ACCOUNTNAME /GROUPNAME,3” for the group tell all. And then some.
• The number of sectors reported by the REPORT command for a group or groups is sometimes inaccurate, sometimes very inaccurate. Running the program FSCHECK.MPEXL.TELESUP and issuing the SYNCACCOUNTING command will fix this problem.
• In case you were wondering, despite many requests for the enhancement, TurboStore will NOT append store sets to tape. Well, it might if you use the proper incantations, but it is unsupported and highly dangerous because under certain circumstances you could overwrite a previous backup without knowing.
• Speaking of things you cannot do that you might like to do, the ALLOW command is not persistent across sign-ons unless you use the extremely dangerous “ALLOW @.@; commands” version. This is another example of an enhancement that has been requested for years, but now will never happen. Fortunately, there are a number of options, for sale and free (MPEX, CSL, etc.).
• CI integer variables are signed 32-bit entities. So be careful if you are doing some wild arithmetic in your CI scripts.
• Here is a little trick when using Apache’s indexing (for example to keep track of documentation) to index file displays. You can override the default ascending sort by name by appending “?N=D” to the url. Instructions on changing Apache’s default behavior are available on the Web.
• If you are trying to program VPlus applications and are interested in working examples programmed in your favorite language, look in the group HP32209.HPPL89 (which should be on every FOS tape). This group contains source code for the ENTRY program in a variety of languages including COBOL, Fortran, Basic and Pascal.
• To see the firmware (aka PDC) Revision of a system (CPU): Run cstm, and at the cstm > prompt, type ‘map’ and note the Dev Num of a CPU and then type ‘sel dev DEV_NUM’ (e.g., ‘sel dev 41’) and then type ‘info’ and then type ‘il’ and look at the output for the ‘PDC Firmware Revision’. Easy, huh? Thanks to Guy Paul of HP for this tip.
• SPFXFER will allow you to write to disk (undocumented “feature/bug”). But don’t do it, because SPFXFER cannot read the disk file it creates! Doing this could lead to a big oops.
• While it would certainly be a nice to have, MPE/iX CI scripts have no provision for inline comments. Sorry, don’t even bother trying.
February 24, 2016
Bringing a First 3000 Love Back to Life
Stories of HP 3000 longevity are legend. Less than 10 years ago, Paul Edwards could report on a Dallas-area customer who was running a Series 70 system in production. Paul was circumspect about who the lucky company was — lucky because they were still leveraging a system HP stopped selling in the late 1980s.
We heard from a longtime 3000 lover in Buffalo recently who wants to turn back the calendar on his Series 42 system. By his system, we mean that literally: Matthew Bellittiere took personal possession of the same system which he learned MPE upon in the early 1980s. The 42 was a server that considered a DDS tape drive an upgrade. Reel to reel was the standard backup peripheral for any computer HP first sold during the early half of the 1980s. HP gave the Series 42 its debut in 1983.
Bellittiere waited awhile to rekindle his old flame. About 20 years ago, he took the Series 42 into his home, but only this month is he working on getting it up to speed. A system that is 30-plus years old, that hasn't been started in 10 years: some might think this is scrap, or worse. But listening to his request, we hear a man who's finding a long ago sweetheart, rescued from the mists of time.
This HP Series 42 is the first HP mini mainframe that I started on around 30+ years ago. I arranged many updates over its active life. Some of the updates include increasing the memory by exchanging the 1/2 meg cards with 1-meg boards. I upgraded to the HP670H disc drives, and also to the DDS tape drive. In 1996 the company upgraded to a Series 947, and HP did not want the 42 back. It was going to scrap, so I requested it and it was given to me. I have had it ever since with plans to get it up and running.
I had to ask: Is the Series 42 project a hobby, or a work system? "Yes," Bellittiere admitted, "it is more of a project for me." But he needs the help of MPE V experts in our community to bring his old flame back to life.
Bellittiere understands there are special procedures required for a server whose discs are its newest parts (circa 1990, so 25 years old already). "My first question: does it need any special treatment before powering up?" he asks. "I think any internal memory will have been lost long ago. It has been at least 8 to 10 years since being powered up."
The components that die soonest in a 3000 are usually the power supply and the internal battery, although the disks are often not far behind. "I am not sure of the power up routine — can you help with some ideas?" I said we knew some 3000 experts with MPE V, CISC-generation hardware savvy. He replied with some hopeful praise aimed at his community.
"I am glad that you are still out there. I would not know who else could help me."
The Series 42 was a noble steed, one of the genuine workhorses of the 3000 line. HP used it like a team of draft horses in its labs. I took a tour of the company's disk drive manufacturing plant in the late 1980s — in the days when HP still built some of the world's most dependable drives in the industry, in Boise, Idaho. A wall stacked with Series 42s was doing burn-in testing for the 7973 drives that were already a mainstay in 3000 shops. At five-plus years already, the Series 42s looked like tiny tugboats, computing craft like ships whose decks was peeling but whose hulls were still buoyant.
I hope there's an MPE lover out there who's got advice for Bellittiere. The wisecracks are easy enough about boat anchors or semiautomatic target practice. People said the same thing about F-150 pickup trucks for awhile, too. My son Nick bought his first F-150 right after cut its lines on the 3000, when that truck was 22 years old, an age Nick hadn't yet achieved himself. People live and work in our world who find old tech a delight. Send your help and advice to Matthew via email, or at 716-536-3298. Let him see a colon prompt from a server introduced before fax machines were common office tools.
February 19, 2016
How hot plug disks can replace DDS offsite
I need to find an alternative to DLT and DDS tapes for offsite storage. Sure, there's DS2100 and Jamaica drives. But a few $35 300GB Ultra SCSI drives would hold a lot more data with less points of failure. I'll set up a BACKUP_VOLUME_SET and use the internal disks to do store-to-disk backups of the system.
I've always used my A-Class and N-Class systems with fiber-attached disk. Are the internal disk drives hot-pluggable?
Jim Hawkins, IO maven for HP 3000 systems at HP, replies with details.
There are multiple layers of changes for actual hot plugs or swaps to work.
- You need the disk HDD to handle this electrically.
- You need HDD physical carrier and physical interface to comply.
- You need the system physical interface and receptacle to comply.
- You need your Host System Bus Adapter (HBA) to electrically support this.
- You need the OS to be aware enough of the HBA to not get flustered by absence of the device and deal with any notifications from the HBA of the activity.
Given that the N-Class disk cage has a screw-based cover and the HDD carriers have no quick release levers (as compared with HASS/Jamaica or VA7400) I would state definitively that there is no hot-plug intention.
At the same time, the SCSI bus is pretty low power and low voltage, so it would be generally not-too-unsafe to experiment. But you're also close to AC inputs and they are not low power.Hawkins explored the not-too-unsafe scenario with theoretical possibilities.
Might you be able to pull/push a drive where you've closed the volume? Likely it would work, but there may be all kinds of noise and stress on the SCSI bus which may not be well handled. However, I think each disk is on its own HBA channel which isn't shared with anything else, and so unlikely to abort someone else's IO.
This takes us to the last issue: mechanical wear.
These connectors were likely intended for more or less permanent mating of two components. Very likely they have a limited number of cycles that they are specified to hold-up. I've seen connectors that are specified for fewer than 25 cycles before you lose gold contact material. This is okay for normal HDD where one might replace one or two per slot in a system lifetime, but not sufficient if you're doing nightly back-ups and swaps. Connectors, where there is an expectation of a high number of pull/replace cycles, have special designs.
Now a little good news here is that the N-Class was still pretty much old-school HP design, so likely they didn't pick up something cheap that saved them .2 cents per unit on gold plating. No idea though if the HDD connector is a 10-, 100-, 1000-cycle part. Your system, your risk.
Consultant Mark Ranft points out that the HP design for 3000s seems to make the servers and components good candidates for exceptional mechanical wear tolerances.
It is especially helpful to understand the concept of mechanical wear on the connectors. HP always had excellent and innovative hardware engineering on their HP 3000 servers. Remember, you can drop them off a building and still self-test them.
The Unix N-Class appears to allow hot-pluggable drives.
The actual power supply and the fans are in the front of the N-Class. The power receptacles in the back have internal cords that lead to the front.
February 12, 2016
How to get specific about IP access for PCs
I want to give a 3000 a static IP, so I can permit a user to access the HP 3000 from that PC with that static IP. Is there a way to force a particular user ID to use a specific IP address?
Tracy Johnson replies:
A simple logon UDC should suffice:
IF HPREMIPADDR = "aaa.bbb.ccc.ddd" then
ECHO Evil message here.
Bob Schlosser adds:
You can set up a logon UDC that checks that the var HPLOCIPADDR is equal to the device (PC) that you want them to use. Something like this:
IF "!HPLOCIPADDR" <> "123.456.789.321" change "123.456.789.321" to
your IP address
Using this, we verify that the user is on the correct (assigned) IP address, and log them off if not.
Chris Bartram, who's created e-mail solutions for the 3000 at 3K Associates, and hosted Web servers since early in the 1990s, adds:
The following is an excerpt from system UDCs I use on my HP 3000s that might give you some ideas.
The "VALIDATEIPADDR" call in this UDC calls another command file that actually does a validation of the logging-on user based on data in a control file to determine if he/she is allowed to log onto the system from the specific host/IP address they are coming from.
The variables the UDC sets will work whether the logging on user is coming in via Telnet or NSVT (or hardwired or via a modem).
The TELLOPs also leave a nice log on the system console (and log file) of the login, including where they came from, and what protocol was used to access the system.
setvar _network_node ''
if bound(hpstdin_network_node) then
setvar _network_node '!hpstdin_network_node'
setvar _na ''
setvar _at 'HARDWIRED'
if bound(hpstdin_network_addr) then
setvar _na '!hpstdin_network_addr'
elseif bound(hpremipaddr) then
setvar _na '!hpremipaddr'
if bound(hplocport) then
if !hplocport=23 then
setvar _at 'TELNET'
IF BOUND(HPSTDIN_ACCESS_TYPE) THEN
SETVAR _AT "!HPSTDIN_ACCESS_TYPE"
IF BOUND(HPSTDIN_TRANSPORT_TYPE) THEN
SETVAR _TP "!HPSTDIN_TRANSPORT_TYPE"
IF "!_AT"="TELNET" THEN
SETVAR _TP "TCP/IP"
SETVAR _TP "SERIAL"
IF BOUND(HPVT_CLIENT_VENDOR) THEN
SETVAR _VND " (!HPVT_CLIENT_VENDOR)"
SETVAR _VND " "
TELLOP LOGON VIA !_AT USING !_TP !_VND
setvar _node ups(ltrim(rtrim("!_network_node")))
setvar _addr ups(ltrim(rtrim("!_na")))
if '!_node'<>'' then
tellop !_at, IP: "!_addr" Node: "!_node"
tellop !_at, IP: "!_addr"
if !cierror<>0 then
echo ** NODE/IP CONTROL FILE CORRUPT **
February 10, 2016
Linux box feeds Series 918 for daily needs
HP never designed a smaller PA-RISC 3000 than the Series 918. The server that was released in the middle 1990s helps untold 3000 sites keep MPE/iX in the production mix. While surveying the customer base to learn about the 2016 state of the server, James Byrne of Hart & Lyne reported that a 918 at the company processes data FTP'd from a Linux system. The reason for sticking with MPE/iX, Byrne said, is the state of today's toolset for Unix and Linux. We'll let him explain
By James B. Byrne
Our firm has been running its business applications on HP3000s since 1982/3. First on a time-share service, and then on our own equipment. Our first in-house HP 3000 was a Series 37 ("Mighty Mouse") running MPE IV, I believe. Anyway, that is what my little brown MPE software pocket guide tells me.
We subsequently transitioned to a Series 42 and MPE V, and then a 52, and then to a Series 925 and MPE/XL, which soon became MPE/iX. Then through a 935 to our present host, a Series 918LX running MPE/iX 7.5.
And in all that time we ran the same code with the same database. We still can produce reports of transactions going back to 1984.
Presently the HP 3000 runs the greater part of our online transactions and handles all of our billings and payables. Due to changes in our business model, our main business operational application is now provided by a service bureau. Twice each working day a separate process, written using the Ruby on Rails framework, scans the PostgreSQL database, extracts all unbilled items, and produces a transaction file that is then forwarded via FTP to the HP 3000. Once the transaction file is transferred, the same FTP process triggers a job on the HP 3000 to process that file into invoices.
Our intent is to move off of the HP 3000 and onto Linux, moving away from proprietary solutions to open source computing. This includes bringing our operational software back in-house and off of the service bureau. We are actively developing software in pursuit of this strategy. However, the progress toward a final departure from the HP 3000 has not been as rapid as we had hoped.
There are many reasons for this but the main one is the primitive nature of the tools in common use by the Unix-Linux community. These have improved greatly over the past decade, but they are still nowhere near the effectiveness of efficiency of software I used on the HP 3000 in the 1980s.
There are exceptions, of course. Git as a version control manager is head and shoulders above anything I was exposed to on the HP 3000, or any other platform of my personal knowledge, whatever may have existed elsewhere. Likewise Perl, Bash and Ruby are far superior to MPE's native command scripting language. And the sheer variety of software tools available for Linux dwarfs by several orders of magnitude that which was ever provided for the HP 3000. Even if you could afford the 3000 tools.
But for online transaction processing and speed of development, not to mention stability and reliability, nothing in the *nix world that I have encountered even approaches the HP 3000. PostgreSQL is certainly a more then adequate replacement for IMAGE/SQL, but the open source rapid development tools are a different story.
February 08, 2016
Newer XP storage works for HP 3000s
HP 3000 systems often use antique disks for storage and as boot drives. No device HP has integrated in a 3000 server is newer than 2003 in age, and even some later-generation disk arrays have design dates that throw back to more than 10 years ago. (We're looking at you, XP12000.)
Thankfully, there's newer storage available to HP 3000 sites. The XP20000 and XP24000 can be integrated with HP 3000s. We know of a couple of support and resale companies which do this work. Pivital Solutions continues to support 3000 sites, including integration like this. Newer storage can assure more confidence in the HP-built versions of the 3000. Older hardware gets dinged during datacenter audits, after all.
Other companies that don't write 3000 support contracts are able to connect these XP arrays. One of the other providers calls these newest StorageWorks devices "amazing storage devices for HP 3000 servers." HP put out an end-of-life notice for these arrays' XP12000 predecessor more than two years ago.HP's 3000s are not listed among the supported systems in the quick HP datasheets for these arrays. But there's nothing like field reports to give a manager the accurate state of 3000 storage. "We just sold and integrated two XP20000 and XP24000 systems over the holidays," reports one reseller.
It's not ridiculous to want to attach these storage beasts to a datacenter that includes HP 3000s. The XP24000 maxes out at 1.98 petabytes, can handle up to 64,000 LDEVs, but can start as small as 1 TB. The XP20000 goes as high as 413 terabytes. HP's list isn't complete, but it says the following operating systems and hosts can share in the XP24000/20000 storage:
- IBM AIX
- Sun Solaris
- Linux - IA2, Red Hat
The chassis on the XP20000 is listed at under $14,000 for a refurbished model on one website. As a shared device across a heterogenous datacenter, these XP units are the newest edition of the StorageWorks lineup. Moving parts are the highest risk elements in homesteading a 3000, like any other system. In the right datacenter, these can help keep MPE/iX systems in the mix.
February 05, 2016
Number, Please: Finding the 3000 Set
When I started in this line of work in 1984, writing about the Hewlett-Packard community, I had a directory. Literally, a perfect-bound directory of HP staff that worked in the company headquarters and labs in California. HP shared it with me as HP Chronicle editor, updating it every year. When someone's number at HP came up missing, you'd call up company HQ and ask for the division operator. It was the 411 of the middle 1980s. It's obvious the 3000 world needs something similar today.
As it turns out, the community does have it. The most dynamic directory resource is the 3000-L, still in use this month to locate information about contacting experts. What makes it powerful is the wetware behind the bits. Knowing which of the 3000-L posters are customers, rather than consultants, is one example of the power of that wetware.
As the week began, Bob from Ideal Computer was searching for Brian Edminster, he of Applied Technologies. Bob slipped a message under the door of 3000-L, then got an answer back about a current email address. I followed up today, just to make sure Bob got something useful. Brian's on the lookout for consulting opportunities, as well as longer engagements.
Yesterday Al Nizzardini was seeking an email address for Vesoft. A couple of replies on the L misinformed Al that Vesoft doesn't use email. That might have been true 10 years ago, but the address email@example.com lands in the offices of Vladimir Volokh and his team. Vladimir far prefers to use the phone, but he's old-school enough to enjoy an in-person visit, too.
In another update, 3K Associates and Chris Bartram are now at 3kAssociates.com. Bartram, one of the very first of the 3000 community to set up shop in the Internet, sold his two-character domain name 3k.com for a tidy sum. "We continue to sell and support our entire like of HP 3000-based software products from 3kAssociates.com," he reported on the L.
The community is moving outward like a starburst, all getting older and ever more dispersed. But the 3000-L remains a good place to post a "number, please" request on how to find someone who knows the 3000. If you're like me and archive all those messages (more than 9,000 over the last 10 years), you can search your email client and find the latest communique from someone like Brian. Many of us don't include phone numbers in our email signatures, but the email is always there.
And if you don't have 10-plus years of emails archived, the server at the University of Tennessee at Chattanooga does. It hosts the L, as it has for two decades, and you can search it via the Web.
If you're counting up such things, more than 450 people still subscribe to 3000-L today. There's also a directory of vendors and consultants here on the NewsWire site. If you're offering services or solutions to the community, I'd be glad to take down your number.
February 03, 2016
MPE site sizes up Linux distro for Charon
When we interviewed one HP 3000 manager who's homesteading, James Byrne had a question about the kind of Linux that's used as a platform for Charon on the 3000. Byrne's heart rests in the ongoing lifespan of MPE apps, a thing Charon can help make possible. There's a matter of spending additional money on a proprietary solution, though, no matter how stable it is.
There's another issue worth looking at in his organization, Hart & Lynne. The Canadian logistics company has Linux wired extensively into its datacenter. Having been burned with an HP pullout from MPE, the solutions that go forward there have to meet strict open source requirements to run in the datacenter there. Nobody wants to be caught in the vendor-controlled blind alley again.
Bynre's got a problem about about something called KVM, and how genuine open source Linux needs to adhere to that product. Byrne described KVM as a Linux-kernel-based virtualization system and is therefore Open Source software.
Doug Smith, the HP 3000 Director of Business Development at Stromasys, said KVM isn't a part of the Charon installation set. "KVM is part of the Linux kernel, the part that allows Linux within itself to create virtual machines—kind of like a hypervisor. This is not utilized by our software."
KVM users have strong feelings about hard-line open source licensing. Byrne's issue is that VMware's software—which isn't required for every Charon install—looks like it might be operating outside the General Public License that many open source solutions utilize.Byrne says that "Charon-HPA runs on ESXi vmkernel, which VMWare claims is not derived from Linux." Then he explains why that's a problem for his adoption of Charon.
VMware is presently being sued by Linux developers for violations of the GPLv2 with respect to the Linux kernel. It is alleged that VMware is in fact using GPL code but are not providing the source for their derived vmkernel, as is required by the terms of the GPLv2.
VMWare is thus attempting to benefit from Open Source projects through misappropriation of public goods for private profit, and attempting to assert proprietary rights over the work of others. In short, they are not a company we wish to deal with, either directly or by proxy.
(Below, VMware's overview of the architecture of VMware's ESXi architecture.)
Regardless of what happens between VMware and those Linux developers, VMware doesn't have to be deployed as part of Charon HPA, according to a Stromasys product manager. VMware is a commonly used component, but it's not mandatory.
This alliance of Linux and MPE was considered beyond a dream back in the days when the HP lab for MPE was closing. A fully open sourced OS acting as a cradle for a legacy OS first created in the proprietary era? Cats and dogs living together. It says something nice about the flexibility of Linux, a trait that's a byproduct of its open source development community.
But the alliance also says something about MPE/iX and its continuing value. Stromasys believes as much, investing in R&D not even HP could get into its budget to give MPE/iX a way to boot up on Intel hardware. Extend the value of your apps with fresh hardware, the vendor says. To this day, even HP-UX won't jumpstart on Intel systems—unless they're Itanium servers. X86-Xeon won't work with HP's Unix.
That enduring value of MPE and the 3000's PA-RISC architecture is something Byrne sees clearly after decades of managing 3000s. "The real problem with the HP 3000 is that it just works," he said, "and so every other issue gets precedence above migration."
February 01, 2016
Loyalist, laggard, loser: who are you now?
When 2016 arrived on our calendar, we looked for signs of the 3000's present and its future. A survey of frequent 3000-L contributors was answered by about half of those we polled. Among that group we found half of these IT pros — selected to be sure they owned 3000s, not just consulted on them — have plans for MPE/iX in their companies in 2016 and beyond.
If you're still using HP 3000s here, getting on to 15 years after HP announced the system's "end of life," then who are you? Among your own kind, you're possibly a loyalist, devoted to tech that's still better than the alternatives to your company. After all, almost 5 percent of every Mac user runs their systems on Snow Leopard, an OS released six years ago and decommissioned by Apple in 2013. Some experts in the community say it runs faster on the newest Macs than any other OS release, though.
The glove on this page came from a Mac conference of 2006, when Snow Leopard was three years away. Maxtor was sure we'd be losing files unless we backed up to their disks. They gave us a set of three instead of a pair of gloves. The way things turned out, Maxtor lost its company status that year, purchased by Seagate. The Maxtor brand went dark in 2009, the year Snow Leopard made its debut. The OS got a small update this month, though, to keep the door open to a newer OS X.
Your 3000 loyalty may label you a laggard. That's one way to describe somebody who's among the last to migrate somewhere when anybody who's savvy has already departed. Tough word, that one. It can inspire some dread and maybe shame about holding out. Or holding on. If the vantage point and the capabilities of MPE/iX in 2016 suit you, though, laggard is just a way to segregate you from someone else's visions.
The implication and suggestion is that laggard would mean loser. Nobody will actually use that word while identifying advocates for old tech. It surely doesn't fit when your applications are solid and cannot be replaced by a migration project priced at more than a full year's IT budget. There's also the matter of keeping IT headcount lean. The most expensive parts of running a datacenter are the people. That's why cloud solutions are getting airtime in boardroom planning. MPE demands fewer heads.
"We're still using our HP 3000s," said Frank Gribbin, running the servers for the law firm of Potter, Anderson. "It's just too useful a tool to do without."To be accurate, some of the datacenter managers who shared their 2016 status said they're on their 3000s because they're sticking to something they'd rather replace. HP poisoned the well for proprietary systems, one said, and extending MPE/iX use at Harte & Lyne Limited would be "simply shoveling money out the door that is better spent on moving off the platform entirely."
That said, in a moment of optimism this user of the Powerhouse platform on the 3000 said "Keeping the HP 3000 and Cognos software going on Charon has its attractions. Frankly, I never expected an MPE/iX emulator to see the light of day, and for that reason alone I am interested in seeing it work."
Byrne might think of himself as a laggard, and look like a loyalist -- but his loyalty is to the product, rather than its makers. He shares his tech strategy and his insights at length, though, and it's pretty clear he's only lagging because there's nothing better that fits the company's needs and resource capacity. Sometimes that's budget, and sometimes it's people. Watching somebody wire MPE/iX into a significant Linux shop shows he's not lagging, but looking forward. And back at his MPE, regretting the loss of HP loyalty. It makes everyone who's endured from 2001 onward a loser, or at least a victim of a loss.
There will be losses out there in the MPE world in 2016, right here and in some 3000 sites, too. Enduring them is the opposite of being a loser. And if you're lagging at a leap into a migration, there are probably reasons that satisfy your flight plans.
January 27, 2016
Keeping up lets you receive what you give
We've been checking in on how companies are keeping their MPE/iX servers up to date. One element is consistent in successful updating: continuing maintenance contracts for the software that's in production or development use. It's the heart of a healthy body of IT resources.
In one recent story we followed up on Reflection, the Attachmate HP 3000 terminal emulator product. Things have changed in PC desktop environments, since Microsoft has been hawking its Windows 10 update automatically. To get the latest Reflection version from Attachmate, keeping up on support is required. It's a paid enterprise to work on making changes to software like Reflection to support new environments such as Windows 10. Not many software solutions update themselves, said Birket Foster.
"Even free, open source software has programmers that are paid," he said when we checked up on Reflection updates. MB Foster has sold many copies of the product over the last 25 years. "Even for open source, there's some support and other positions also being compensated if these volunteers are working for a university or a large company like HP."
Foster says yes, there is an upgrade fee to bring Reflection up to date. "For customers that have been using the software for 10 years, they might want to remember that there is a cost to keeping the software in sync with the Microsoft changes," he said. "Continuous development is required and the programmers need to be paid."
One alternative to Reflection terminal emulation is Minisoft 92, from the company of the same name. CEO Doug Greenup said his product's got Windows 10 support, but even more interesting is the fact that it's got as many as 25 sites using the Charon emulator. Moving from HP's 3000 iron to Charon is a complimentary relicense at Minisoft, without a fee — so long as there's a current support contract.
Greenup says, "In the past couple of years some of our customers have moved to the Stromasys platform. Off the top of my head it's in the 20-25 range."
As long as they are current on support we allow our customers a no-cost license transfer on their software. They just need to provide us with the new CPU information for licensing purposes. These transfers are handled electronically and so are very quick and easy to implement.
Long ago when paper was the primary medium for sharing stories, one editor used an apt metaphor to describe ongoing support. In the publishing business of that day, that support was a subscription, an annual payment to ensure the resources remain available. "Subscriptions are the meat on the bone of any magazine," said Sy Safransky of The Sun. Support is the meat on the bone for software customers, especially for products that are meant to keep MPE/iX on duty with so little attention required.
January 25, 2016
VMware solution assists Win10's 3000 debut
Windows 10 is making its way into HP 3000 shops. Earlier today a manager had loaded up Win10 and then discovered that Reflection, the terminal emulator built for HP 3000 access, wasn't working anymore.
"My Attachmate Reflections v184.108.40.206 does not work — it has an error when trying to start," said George Forsythe. He wanted to know about any available updates for the former WRQ product. It's not a former product, but Reflection for HP, as it's known today, is a Micro Focus product. Last year Micro Focus bought Attachmate, the company that purchased WRQ.
The short answer is version 14.1.543 (SP4), according to Craig Lalley. It's a matter of an update, but a mission-critical connection might demand a faster solution. One well-known program that aids Windows migration of 3000-attached desktops was mentioned by Neil Armstrong, developer of the Robelle data utility Suprtool. VMware can have your back if you're taken a PC onto Win10 and something critical like the 3000 connection stops running, he said.
This is why I've "virtualized" some key environments that are used for development. If something like this comes up, you're not stuck with a critical problem at a key moment.
Supported software is sometimes built with customized routines to use desktop OS modules. That means it can stop working when a desktop environment changes. There's profound changes in Windows 10. Forsythe reports the AICS freeware terminal emulator QCTerm, built for the 3000, still works on Win10, even while his not-quite-fresh Reflection didn't.
Armstrong said the reliance on using VMware to preserve stable desktops comes with a cost. You can't ignore updates to the virtualization engine.
Once something like [a desktop OS release] is stable and set up, you just turn off all updates and back it up. Of course, the weak point then becomes if VMware doesn't work with whatever OS update is currently going on. But there seems to be enough resources and typically there is a solution on hand, as long as you keep that software up to date.
Micro Focus is maintaining Reflection, but one 3000-L member reports the upgrades are no longer free. Older versions of Reflection work with Win10, according to Steve Cooper of Allegro, "with only a few nuisances that can be worked around."
Cooper was using version 10.0.5 of Reflection. When we last checked, that's software more than a decade old. Apparently the extra value of later releases is offset by their compatibility challenges. There's a lesson in there about older software, like QCTerm and elderly Reflection — and MPE/iX — being a more stable solution, even in the face of change.
And if Windows 10 is software that's too new to behave well on a PC connected to a 3000, there's a way to stay on a prior release and stop the "upgrade to Windows" reminders. Paul Edwards, consultant, board member and OpenMPE volunteer, offered this advice.
For those of us who really want to stay on Win 7 for a while and not be reminded to upgrade to Win 10, there is a tool available from www.ultimateoutsider.com/downloads. It is GWX Control Panel. The control panel has a status page to tell you whether the “Get Windows 10” app is running, whether it is enabled, whether the Win 10 files have been downloaded to your PC, and if so, how much room the files are taking up on your computer. If the files are there, the control panel can remove them for you. This is much better than modifying the registry.
I have installed and run the GWX Control Panel. I had it delete the Win 10 logo, folders, and files (6 GB). I had no problems with my PC afterwards. And no Win 10 reminder.
January 22, 2016
A 3000, awaiting replacement, still at work
If the above headline sounds like your homesteading situation, then you're an interim homesteader. Or a wannabe migrator, which can amount to the same thing if the pain of retaining a 3000 and MPE is low. In the hospital they ask you to rate your pain on a scale of 0-10. Nobody says 0, unless they're deep into morphine. There's usually some.
At Cerro Wire, the pain level must be not more than a 2, but the 3000 is being targeted for replacement. As part of our survey of the 3000 managers who speak up on the 3000-L, we got a report back from Herb Statham. He's led the 3000 computing at the manufacturer based in Alabama, with operations elsewhere in the US, too. Statham notes that the MPE server at Cerro continues to work. It's something like staying on your job even after you've been laid off, because they can't find a replacement yet.
Uncommon for an employee. Commonplace among interim homesteading systems. Statham, who was hiring for 3000 operations as recently as 2014 -- and had a contract 3000 expert at work until October — reports that Intel-based systems are preferred now at Cerro.
We are still running an A500 box at Cerro Wire. The game’s afoot to replace our current business applications with ones that are Intel- and Microsoft-based. I do not know when the final decision will be made, but the HP 3000 just keeps chirping along. I am trying to get “semi-retired” to only work two or three days a week, until the “new and better” system is in place.
Intel had prospects earlier at Cerro, in a different capacity. Statham was public about a 3000 emulator's chances there, even before the Stromasys Charon software had a big footprint. Cerro was going to be a classic 3000 manufacturer pushing their MPE apps into a long-running role. Leaving the HP hardware behind looked to be important, but other apps on other platforms were already working there.Some IT managers call this situation "floating." So long as the MPE applications don't fall short, their cost of ownership and low need for attention keeps them running. A turn-off date at the start of 2016 becomes a midyear close-out, and then that depends on how soon replacement apps on Windows get integrated. Any nagging pains about relying on an environment now in its fifth decade of useful life are offset by the Tylenol of low costs and stability.
It works for companies that don't see massive growth coming soon. At Cerro, which is a Berkshire Hathaway company, business has been good. Back in 2014, just before the help-wanted call went out, the pressure to migrate was low.
In profile stories from 2014, we heard this report.
Statham has no pressure from Cerro management to replace the applications that are successful at running the company. With ample spare parts, independent support and storage consulting, and his own source in hand, he needs only the green light from Dell to move forward. Specifics on pricing and performance are still in play from Stromasys, at least from his vantage point. A 1.5 version of CHARON HPA/3000 was announced late last year, promising increased performance. But meeting the speed needs of an A-Class would be no challenge for the CHARON lineup.
This veteran of 3000 deployment and management has little desire to send his company toward an application replacement that might end up with Cerro "spending millions of dollars." There are many years left for MPE/iX, and his company is an all-HP shop, with the exception of a couple of Dell monitors on Statham's desk. He can see a long future for the app the company has fine-tuned to its business.
The CALENDAR intrinsic roadblock is the only thing he can forecast by now. He's not sure how HP might react to an independent fix for that issue, a date challenge that's still 13 years away. (Of course, now it's 11-plus years until the December 31, 2017 deadline)
"If we could ever get this 2027 thing out of the way, you could run your applications indefinitely, so long as you’ve got someone to support them," he says. "My only concern is HP themselves, in the event that someone said they had a patch to the operating system. You wouldn't have to worry about the year, because there was some type of workaround."
There's a number of ideas in there, from relying on MPE doing its job 11 more years (not out of the range of possibility) to seeing an independent lab develop a 2027 workaround (also not impossible, so long as community experts don't do more than semi-retire) to HP getting in the way of this kind of lifespan extension. There's zero pain to the MPE's creator in letting the OS keep working. It doesn't require much pay by now. That's the sort of thing that makes some migrations wannabes, or at least keeps them floating in the future.
January 21, 2016
Taking a Charge at Transition's Costs
Changing your IT infrastructure might become more critical in 2016. Hardware is older, especially the hardware HP built and sold to run MPE/iX servers. One solution is to migrate to a new OS environment. Another refresh for IT might come from emulating the PA-RISC servers with Intel-based servers. But in either case, some software will have to come along, a move to help contain transition costs.
License transfer practices come into focus during these projects. While moving from MPE/iX to another OS, most shops would like to keep what's been working, if the software's got prospects to grow along with IT's needs. In some cases that's possible, because the vendor has put in its work to adopt a new platform. A couple of middleware providers have done this. MB Foster and Minisoft both reached out to HP-UX users coming out of their MPE/IX environments. Minisoft's Doug Greenup reported this week that Summit Information Systems Spectrum users — whose vendor is now Fiserv, post-migration — headed to HP-UX when leaving the 3000 credit union application. Their target was Eloquence, the database designed to embrace IMAGE applications into an SQL world.
"We have quite a few Eloquence customers," he said, "more then 100. Many of the Fiserv Credit Union customers moved from MPE to HP-UX and use our ODBC driver for Eloquence." Minisoft's also got an ODBC for IMAGE product. That's an example of a cross-platform development strategy, something to keep costs under control. When your existing vendor does a version of your product for a migration target, that's fortunate. It's even more fortunate when you're not expected to re-license the product.
Last week the Minisoft ODBC for IMAGE product became the target of a competitive upgrade campaign. MB Foster says it will let a Minisoft ODBC customer switch to UDALink for MPE/iX for the price of a support contract. We took note of that campaign, a classic move from the days when new MPE/iX software was being sold in a market active enough to support multiple vendors for a single product like middleware. Going into competition, and retaining customers in the face of it, smacks of moxie in a market that's quiet and stable by now. It helps if your product has feature differences, so a 64-Bit ODBC driver, and the ability to use Suprtool's Self Describing (SD) Files, are getting touted in that offer.
Both vendors say they support Windows 10 with their middleware. No matter how much grief the new Microsoft environment is causing, it's still a certain part of IT futures. Windows 10 support is essential to keeping a 3000 current with the latest PC clients tapping IMAGE/SQL data.
Vendors in the 3000 market had to go where new system sales were happening, though. For MB Foster, its HP-UX version of UDALink is preserving investments at a site where the biggest single group of 3000s was migrated. At the same time, this site is using Minisoft's middleware on HP-UX, too. The situation at the college group looks like a lesson in preventing extra costs in a transition. Migration has plenty of prerequisite costs.There are no more MPE/iX computers in service at the Washington Community College Consortium, a group of 34 organizations once run by 3000s. The college collective turned off its 3000s in 2012, giving over the services to Unix systems built by HP. Minisoft's middleware works there for the Unix servers. So does the MB Foster software. The educational organization didn't have to re-develop its hundreds of reports it built over years of 3000 services.
"They use UDALink Reporter for UNIX (formally DataExpress) for reporting on their HP-UX systems," said MB Foster's Chris Whitehead. "We converted DataExpress from the MPE version to the Unix version, so that the college could continue to use the hundreds of host-based reports they created over the years. They also are using it well for bulk extracts, offloading data."
The cost of the software for a fresh platform was not a show-stopper there, which seems to be the case for 3000 vendors who serve with middleware and key utilities. Database management, data extract software, middleware: much of it can be moved to something like the Stromasys Charon emulator, or even a new OS, for little to no charge. "Preservation and continued use of reports created on the 3000 was key to the conversion of UDALink Reporter on MPE to UDALink Reporter on UX," Whitehead said.
Minisoft's product made a move to contain costs, too. "Some of our MPE customers have migrated over the years to the HP-UX platform, running with Eloquence," Greenup said. "So long as a customer is on a current product support contract we allow them to "swap" or transfer licenses at no additional cost."
Even where there are what amounts to administrative charges — the equivalent of HP's license $432 transfer fee to move an MPE/iX instance from one kind of iron to another — many vendors make these minimal. Everybody's budget is important, but I find it interesting to see IT managers squeezing costs so hard that a $125 per month support fee, or $1,000 one-time to administer a license change, gets scrutiny. Yes, that's about $125, or less than my U-Verse TV bill.
This is corporate IT we're talking about here. The minimums shouldn't be that low. There's money needed to be spent on Windows turnover, yes. But a corporate server needs a budget bigger than a TV bill.
Redevelopment costs a great deal more than that, of course. One month's work might not be enough to bring hundreds of reports into an HP-UX environment; it might take much longer. In a case like that, the $1,000 looks like a better option than three months of an IT pro's time, at around $25,000.
Both Minisoft and MB Foster help to enable Charon HPA emulation projects, too. Those kinds of transfers don't require a new license, something that the new-ish owners of Powerhouse can't embrace yet. Containing costs with support-based license transfers is a forward-looking move. Forward is a direction you'd expect vendors to focus upon. IT changes often. It shouldn't be more expensive than absolutely necessary.
January 20, 2016
Pricing, Value, and Emulating Classics
Editor's Note: Yesterday we ran a story about the impact of proprietary software lock-in, as reported from a manager's office where HP 3000s still do their work. Amid that story was a quote about predaceous pricing (love that word), the act of outre increases to the cost of emulator MPE server solutions because of upgrade charges. It's blocked several adoptions of Charon HPA, even among managers who love the ideal of non-HP hardware that keeps MPE apps alive. Tim O'Neill wrote the following editorial, prompted by our article. Although companies do need to generate capital to keep supplying software, the matter of how much to charge for a shift to an emulator remains a flash point.
Editorial by Tim O'Neill
James Byrne brings up important point about proprietary software running proprietary hardware: it enabled predatory pricing, both by HP and by third parties.
At this stage, it appears that Charon could be bought affordably, but the problem is the third parties' still seeing the opportunity to gouge existing customers.
This is why businesses become former customers and change to shareware and open source operating systems and databases, e.g. Linux and open database systems like Postgres. There are still costs as a part of such a change. They might need to hire more in-house staff to do what HP and third parties used to do for that one huge cover-all price. It might not be wise to entrust critical applications to shareware, but are customers avoiding doing so?So the huge predatory prices were not without value. This is not to say I defend them.
That said, it is still shameful that at this point, third parties are unwilling to honor their customers' long history of loyalty, by requiring emulator relicensing. These third parties should realize that they might realize longer-term benefit by keeping their customers, not driving them away.
It would be interesting to compute the price and valuation of HP stock since the point just before they announced the death of its MPE business, through the split in 2014. One might be able to say that the company's value has fallen without MPE. It may fall further when OpenVMS is eliminated and when HP-UX is not marketed, not enhanced, not written for any CPU other than HP's own Itanium, and not licensed at prices that are fair to customers.
January 19, 2016
It's becoming an MPE Server, this HP 3000
Hewlett-Packard stopped building 3000s in 2003, cutting off a product line in the belief that users would leave the server. But after thousands of them did just that, thinking there would be no more MPE/iX servers to be purchased, an emulator emerged. After more than four years, it might be changing the concept of what is an HP 3000. Brian Edminster of Applied Technologies wonders what's the future for the system that delivers MPE/iX apps.
"It seems to me that it's almost more accurate to call these beloved hosts 'MPE/iX' systems," he said, "rather than 3000s, since — eventually, at least — no one will be running 'original' HP hardware."
We have asked around the community about how this concept plays out. James Byrne, 3000 manager at logistics provider Hart & Lyne, offers one view on what makes up his idea of a device to use MPE/iX.
There's more at stake at his shop: software migration patterns, a way to ensure what's running on HP-built hardware operates on a fresh MPE/iX server. Pricing for a key 4GL-reporting tool — you'll know which one — got in the way at Hart & Lyne. MPE's the keystone there, but Byrne says his company won't tie itself to a single-vendor system in the future.
I consider our systems to be MPE/iX rather than HP 3000. The hardware does not really matter to us any more, since most of the rest of our critical infrastructure is already running on commodity Intel 64 bit boxes. We simply keep two or three of everything running on different 3000 hosts most of the time, and have them continually cross checking each other. That approach has covered us well in the one or two serious incidents we have experienced these past 15 years since HP gave up on the 3000.
If the Charon emulator was priced in the same range as a used HP 3000, and ran on Linux, and used KVM virtualization, then we would in all probability move to it as an interim step, if only to escape the aging hardware MPE/iX is running on.
I believe those conditions are unlikely to all be met, so we do not consider the emulator as a possibility. We still would have to deal with the issue of Powerhouse licensing fees. The last inquiry we made with respect to Powerhouse provided a price that was startling to say the least. We would even entertain moving to Powerhouse on Linux as an interim step, if the price were not so exorbitant and the product supported PostgreSQL. However, when last we looked Powerhouse only supports proprietary databases, so again it is not even a consideration.
Those examples are representative of why we are never going back to proprietary software: predaceous pricing and technological limitations dictated primarily by marketing. Whatever we write for ourselves in future, we are not going to be held to ransom if we wish to move it from one system to another.
January 18, 2016
The GSP makes the A and N worthwhile
It's a powerful part of an HP 3000 that runs whenever the server is plugged in. The Guardian Service Processor (GSP) is the maintenance control console commanding the ultimate class of the server to reboot, do memory dumps and even fully power down the 3000. Consultant Craig Lalley of EchoTech has noted the GSP has one fewer feature than its Unix counterpart, though.
"On HP-UX it is possible to reset the GSP from the host OS," he said. "I have not found a way from MPE."
From time to time a reset may be required for diagnostics services on A-Class and N-Class servers. If your 3000 gets loving care from a consultant or service provider outside your computer room, you may need a paper clip to keep up service levels.
The GSP can also reveal the 3000's speedometer, as profiled near the bottom of a webpage from Allegro Consultants.
The gap between 3000 and HP's HP-UX Integrity GSPs is a common shortfall of HP designs. Even though the 3000's MPE/iX includes a Posix interface, HP didn't engineer enough Unix into the 3000 to enable some administration that HP-UX users enjoy. (That lack of Unix can sometimes be a good thing when a security breach opens up in the Unix world.)
But when a 3000 needs a GSP reset, pressing a recessed button on the 3000's back will do the trick if a telnet command doesn't work. You can telnet to the IP address of the GSP, log in and do the reset. But you can also get someone to press the physical reset button at the back of the machine. It's recessed into the cabinet so you may need a magic paper clip bent just so.
Lalley calls the GSP, which HP introduced with its final generation of 3000s, one of the most useful things in the A-Class and N-Class boxes.
The GSP is a small computer that is always powered on when the plug has power. With it, it is possible to telnet to and BE the console. While multiple admins can telnet in and watch, only one has the keyboard.
It is possible to reboot, memory dumps and even fully power down the HP 3000 from the GSP. Use the command PC OFF.
It is probably the best feature of the N-Class and A-class boxes. The problem is sometimes it needs to be reset, usually with a paper clip. Since the GSP is a different CPU, this reset can be done during business hours.
January 15, 2016
Competitive upgrading lives on for 3000s
In the 1990s, HP contracted to send its ODBC middleware development to MB Foster. The result was ODBCLink/SE, bundled into MPE/iX from the 5.5 release onward. The software gave the 3000 its first community-wide connection to reporting tools popular on PCs. HP decided that the MB Foster lead in development time was worth licensing, instead of rebuilding inside the 3000 labs. Outside labs had built parts of the 3000's fundamental software before then. But ODBCLink/SE was the first time independent software retained its profile, while it was operating inside of the 3000's FOS. Every 3000 running 5.5 and later now had middleware.
Other ODBC solutions were available in that timeframe. Minisoft still sells and supports its product. That's one reason why MB Foster's running a competitive upgrade offer for users of the Minisoft middleware. The upgrade was announced yesterday. 3000 owners who make the switch from Minisoft for IMAGE ODBC to Foster's software will get a full version of UDALink for the cost of only the annual support payments.
This kind of competitive offer was one of Minisoft's sales tools while it competed with WRQ for terminal emulation seats. There was a period where NS/VT features were not a part of every Reflection package, but were a staple in the Minisoft MS/92.
Foster's ODBC software has been extended to use 64-bit ODBC drivers, embrace Suprtool's Self Describing Files, and more. UDALink was a part of the migration that the Washington State community college consortium pulled off in 2011 when it moved 34 systems to Unix. The vendor has continued to develop to make a state of the art middleware solution.
Almost as notable: seeing MB Foster compete for business like vendors did routinely in the 1990s. The upgrade offer tells us that there are 3000 sites out there still looking to extend their development cycles. UDALink is also built for platforms other than the 3000, but any outreach to capture MPE/iX customers is news here in 2016. Chris Whitehead is fielding the calls and emails for the upgrade offer, which runs through June of this year.
January 14, 2016
HP's 3000 now at $149 until Sunday
Google is happy to trawl the Web for HP 3000 news, a search that I've had in place for the past 10 years. I receive a lot of notices about horsepower of auto engines (the HP) and a few about printers. But today a link showed up that features a computer called the HP 3000, currently selling for $149 plus shipping.
There are a few unique and important qualifiers. To start, this is an HP3000 model with an Intel server, literally a PC powered by an Xeon X3330 CPU at 2.8 MHz. That's a quad-core processor, though, and the box is already loaded with 4GB of memory. (It's a start, but nowhere near enough RAM to power software such as, for instance, the Stromasys Charon HPA emulator.)
In short, this is an HP3000 built by Hewlett-Packard that can run MPE/iX, but does not use PA-RISC. Hewlett-Packard Enterprise has not restricted the use of "3000" to the PA-RISC servers well-loved by the MPE community. Over on the HP Inc. side, there's a large-scale printer also called an HP 3000.
This HP3000 running a Xeon chip has another, less significant qualifier. It's being sold by a New Zealand owner on TradeMe.co.nz, "Where Kiwis Buy and Sell." And the shipping options don't go beyond Auckland, or the North and South Islands.
However, this TradeMe model might be something that could be shipped to the 3000 stalwarts Ken and Jeanette Nutsford. The former chairs of SIGRAPID and SIGCOBOL still live in NZ, when they're not gadding about the globe on their epic cruise calendars. Their total mileage easily runs into the hundreds of thousands. Trans-Pacific flights are embedded in their history. So perhaps the 6,693 miles to the US is not completely out of reach, in a hop. The Nutsfords travel regularly to the US, and this PC looks like it would be cargo-bay ready.
Yes, you could file this article under clickbait. It's an online auction after all, and $149 is only today's price. However, if you consider your systems to be MPE/iX servers by now, rather than the Hewlett-Packard PA-RISC 3000 hardware that hosts that OS, this is technically a server that can run your apps.It will require an installation of the HPA emulator, which at last report started at $9,000 for A-Class power. The combination can be compared to A-Class boxes that sell for under $2,000, but those include few options to increase speed. The A-Class had a 2-CPU model running at 220 MHz. There's genuine, hard limits on RAM.
You don't have to go to New Zealand to get this kind of HP3000, although this one looks ready to boot up and run. This ProLiant blade-caliber box does illustrate how much hardware remains in the world that can run MPE/iX software. If a manager's concern is the reliability of the HP hardware that's at least 12 years old -- the last server was built in 2003 -- this leaps over that hurdle to homesteading.
January 13, 2016
Using Store-To-Disk for Backup Preservation
By Brian Edminster
Second of two parts
Yesterday I outlined some of the powers of the Posix program pax, as well as tar, to move MPE/iX backup files offsite. Here’s a warning. There are some file types that cannot be backed up by tar/pax while also storing their attributes: ;CIR (circular) and ;MSG (message) files (and possibly others. I haven’t tested all possible file types yet. Also, there is an issue with tar that is a fairly well known and has been discussed on the 3000 newsgroup. Occasionally it does not un-tar correctly. It is unclear if and when this was fixed, but I’d love to hear from anybody that might be in the know, or which specific situations to avoid.
Regardless of these limitations, I’ve found a simple way around this. Use store-to-disk to make your backup, then tar to wrap it, so as to preserve the store-to-disk files’ characteristics, before shipping the files off-system. Later, when you retrieve your tar backups and un-tar them, you’ll get your original store-to-disk files back without having to specify the proper ‘;REC= , CODE= , and DISC=’ options on an FTP ‘GET’. I’ve been doing this for several months now on several systems, and I have not had any failures.
If you have a version of STORE that has compression, use it to reduce the size of backup. If not, use the ‘z’ option in the tar/pax archive you create from your store-to-disk backup. Do not use both. They don’t play well together, and you may end up with a larger tar file.
But what about the tar archive size limit of 2GB? There’s an easy way around this as well, as this limit is common on early Unix and Linux systems. Just pipe the output through ‘split’ to create chunks of whatever size you want. Below, there's simple examples for both directions.
Below, Figure 2 is an example of a ‘cksum’ created of the files as they’re stored on the NAS.
As both the hashes and #bytes shown in each file are the same as on the MPE/iX server — we know the backups are transferred correctly. The same technique can be used ‘in reverse’ to verify that when FTP’d back to the FTP server, they’re still intact.
When un-taring this backup, ‘cat’ the pieces together and pipe it through tar. At least, that’s the way it’s supposed to work. Yes, there is a known issue with the MPE/iX Posix shell’s built-in cat command. But I’ve so far been unable to successfully use the external cat command to successfully cat either. Here’s how this should work for a 2-chunk tar backup:
sh>/bin/cat ./CS1STD1.ustar.aa ./CS1STD1.ustar.ab | tar -xfv - *
Unfortunately, for me at least, it always throws an error indicating bad format for the tar files.There is a work-around, however. Note that while ‘cat’ing the tar ‘chunks’ didn’t work using the internal or external cat command, untar with multi-file option does work. Even though it gives a minor error messages, files were returned to proper store-to-disk format, and the recovered store-to-disk backup is intact and has been used to recover the desired files. To do this, use tar like this:
sh>tar -xfv ./CS1STD1.ustar.aa *
Also note that when using tar in this way, it will ask for the name of the 2nd-nth component tar files, as it finishes reading each prior piece. You must give the filename and press return to continue for each. I believe that it should be possible to script this so that it’s fed the filenames, but I haven’t gotten around to doing that yet.
Brian Edminster is president of Applied Technologies, a 3000 consultancy serving MPE/iX sites in contract and ongoing engagements.
January 12, 2016
Backing Up Your 3000 Backup Files
By Brian Edminster
Once store-to-disk backups on the 3000 are regularly being processed, it’s highly desirable to move them offsite — for the same reasons that it’s desirable to rotate tape media to offsite storage. You want to protect against site-wide catastrophic failures. It could be something as simple as fire, flood, or a disgruntled employee, or as unusual as earthquake or act of war.
Regardless of the most pressing reason, it really is important to keep at least some of your backups offsite, so as to facilitate rebuilding / recovering from scratch, either at your own facility, or at a backup/recovery site.
The problem comes in that the MPE/iX file system is far more structured than Unix, Windows, or any other non-MPE/iX file system-based storage mechanisms. While transferring a file off MPE/iX is easy via FTP, sftp/scp, or rsync, retrieving it is problematic, at least if you wish the retrieved files and the original store-to-disk files to be identical (i.e., with the same file characteristics: filecode, recsize, blockfactor, type, and so forth).
What would be optimal is automatic preservation of these attributes, so that a file could be moved to any offsite storage that could communicate with the MPE/iX system. Posix on MPE/iX comes to the rescue.For FTP transfers between late-model MPE/iX systems this retrieval is automatic, because the FTP client and server recognize themselves as MPE/iX systems. For retrieving files from other systems, HP has made that somewhat easier by making its FTP client able to specify ‘;REC= , CODE= , and ;DISC=’ on a ‘GET’:
If you do not specify the ‘buildparms’ for a file being retrieved, it will default to the file-type implied by the FTP transfer mode: ASCII (the default), binary, or byte-stream (often called ‘tenex’ on Unix systems). The respective defaults used are shown below:
What follows is an example of automatic preservation of these attributes, so that a file could be moved to any offsite storage that could communicate with the MPE/iX system. And this is yet again where Posix comes to the rescue, via the venerable ‘tar’ (Tape ARchiver), or ‘pax’ archiving utilities.
‘pax’ is a newer backup tool, designed to be able to read/write with tar format archives, newer ‘ustar’ format (that includes Extended Attributes of files). At the same time it has a more ‘normal/consistent’ command syntax (as Unix/Posix stuff goes, anyway), plus a number of other improvements. Think of it as tar’s younger (and supposedly more handsome) brother.
A little known feature of most ‘late-model’ tar and all pax commands is the ability for it to recognize and utilize Extended Attributes. These will vary with the target implementation platform, but for the tar and pax commands included with releases after v5.5 of MPE/iX this capability is not only present — but contrary to the man command’s output and HP’s Posix Command Line manual, it’s the default! You use the -A switch to turn it off, returning tar to a bytestream-only tool.
While not externally documented, via a little experimentation I’ve determined that the following series of Extended Attributes value-pairs are in the MPE/iX Posix implementation of a tar or pax ‘file header’ for each non-Posix file archived:
MPE.RECORDSIZE= value in bytes
MPE.BLOCKFACTOR= integer value
MPE.RECORDFORMAT= integer value (0=unstructured?)
MPE.CCTL= integer value (0=nocctl)
MPE.ASCII= integer value (0=binary, 1=ascii)
MPE.FILECODE= integer value, absent for ‘0’
MPE.FILELIMIT= value in bytes
MPE.NUMEXTENTS= integer value, may be absent
MPE.NUMUSERLABELS= integer value (0=no user labels), and
MPE.USERLABELS=[binary content of user labels]
Brian Edminster is president of Applied Technologies, a 3000 consultancy serving MPE/iX sites in contract and ongoing engagements.