February 24, 2017
Friday Fine-Tune: Opening Up MPE's Shell
Way back in the middle 1990s HP added the Posix shell to the HP 3000. The improvement meant customers who had Unix and MPE running in the same shop could train operators and managers with a single set of commands. Posix was a plus, making the 3000 appear more Unix-like (which seemed important at the time).
Over the years, however, Posix has been a feature waiting be discovered for most 3000 managers and operators. The computer's operating system was renamed from MPE/XL to MPE/iX just for this added Posix feature. But enough history; Posix is still on the 3000 and remains a powerful interface tool, an alternative to the CI interface that HP created for the system. You can even call Posix commands from the CI, a nifty piece of engineering when it can be done.
That's not always possible, though. A customer wanted to know how to "expand wildcard shells" using Posix. He tried from the CI and had this story to relate.
ls: File or directory is not found
So how do I do this? I need to be able to tell tar to archive all of the reels of a STD STORE set via a regexp. It does not work in tar, and it apparently does not in ls—so I speculate that there is something special about the innovation of Posix utilities from the CI that I am not aware of. What is it?
Jeff Vance, the 3000 CI guru at while at HP, replied "Wildcards on most (all) Unix systems, including Posix implementations, are done by the shell, not the individual programs or in-lined shell commands, like ls in your example. A solution is to run the shell and execute ll from within.The magic Posix shell command to do the expansion:
An interesting footnote if you've read this far: The Posix shell for the 3000 is one part of the operating system that was not built by HP. The shell was licensed by HP from MKS, and Hewlett-Packard paid royalties to MKS so Posix could work inside of MPE/iX. That was an issue that posed a potential snag for source code licensing from HP. But the outside license issues never ended up blocking emulation or source-license arrangements. Managers have used Posix on the 3000 as a way to get familiar with commands in Unix systems. In the great majority of instances, these commands are the same.
Follow the 3000 NewsWire on Twitter
for immediate feeds of our latest news
and more twitter.com/3000newswire.
February 22, 2017
Simulator knows what day it is, or was
The SIMH project has created a software release that mimics the HP 3000 Classic CISC hardware. The software makes it possible to emulate HP 3000 servers that go back to the 1970s—the same systems HP mothballed in the middle 1980s even before the PA-RISC products of the past two decades.
So while SIMH won't give anyone an emulated HP 3000 that can run MPE/iX, the package somehow seems to know its way around the calendar. Even after MPE V has long since gone obsolete, the SIMH combo using MPE V from trailing-edge.com adjusts the year to match the current layout. As it turns out, the year 1989 has the same days of the week falling on the same calendar dates as 2017. It offers some hope of getting MPE/iX rewired so its CALENDAR intrinsic works beyond the end of 2027.
An emulator that virtualizes the ultimate generation HP 3000s is the domain of Stromays Charon HPA. SIMH is more of a hobbyist's dreamland, or as one serious veteran called it, "my version of toy trains."
Glen Cole fired up SIMH and reported that "the only user input below was 'hp3000 mpe-auto' ... Neat how it auto-magically knew that 1989 had the same calendar layout as 2017." He did a SHOWTIME to verify the date.
$ hp3000 mpe-auto
HP 3000 simulator V4.0-0 Beta git commit id: f9cfae0c
Logging to file "mpe-auto.log"
Listening on port 1054
LP: creating new file
Cold load complete, P: 177664 (PSHR Q)
Press <CR> to start MPE.
WHICH OPTION <WARMSTART/COOLSTART>? COOLSTART
ANY CHANGES? NO
MON, FEB 20, 1989, 10:35 PM? (Y/N)Y
LOG FILE NUMBER 5 ON
Extract the MPE V/R zip file into a directory along with the HP3000.exe out of the beta, then drag mpe-auto.sim onto HP3000.
COBOL, COBOLII, BASIC, RPG, SPL, FORTRAN are included along with FOS. You can have Reflection connect via telnet on port 1054 to get an actual HP terminal session going that can run FORMSPEC etc.
February 20, 2017
Harris School Solutions buys K-12 ISV QSS
Harris School Solutions (HSS) has announced its acquisition of Quintessential School Systems (QSS). The latter is an HP 3000 vendor whose products have been running California K-12 schools since 1990. The purchase for an undisclosed amount includes a transfer of QSS Chief Operating Officer Duane Percox to the post of Product Owner. The company's QSS/OASIS is capable of going beyond single school districts; it supports multi-district agencies, such as County Offices of Education, and also community colleges.
Scott Schollenberger, EVP of HSS' Financial Solutions unit said of QSS/OASIS, "We see this product as a way to bolster what we offer now, while opening even more doors for HSS in the future.”
Similarly, QSS expressed its excitement over joining with HSS. “Harris School Solutions is an outstanding organization," Percox said in a press release, "not just because of its products and services, but also because of the people who offer them. The people within the company are the real deal, so I’m thrilled to be working with them. Together, we’re going to offer our same great products and services, but to many, many more schools across North America.”
A company press release says QSS OASIS will now be available more widely. QSS has always had a very large share of its customers in California school systems. Selling into a school system in California demands a familiarity of some very unique requirements. Harris brings the QSS software into the rest of the US.
The QSS saga includes a long-term migration campaign on behalf of its HP3000 users. When HP cut its 3000 plans short in 2001, finding a replacement platform with no such trap door was paramount to QSS. Well before the solution was established as a commercial choice, QSS was sent down a path toward Linux. The company calls this Version L, with the migrations coming away from Version H. This past year, the majority of QSS sites crossed over from the 3000 to Linux use.
QSS launched the Linux version of its application suite at Lodi Unified School District in 2008, accessing MS SQL. According to the QSS website, various other customers are scheduled to make the transition from the HP 3000 to Linux during 2017.Calendar year 2016 saw the highest number of conversions of school districts (SD) and County Offices of Education (COE), Eleven organizations cut over to Linux hosts and either SQL Server of PostgreSQL for a database. Those migrated this year include La Habra SD, Mariposa COE, Nevada COE, San Luis Obispo COE and Stanislaus COE, Amador COE, Kern COE, Mendocino COE, Orange USD, Visalia USD, Novato USD
During 2015, five more, schools migrated: Glenn COE, Colusa COE, Modesto City SD, Marin COE, and Santa Clara COE made the switch to version L.
Reports for 2014 covered seven migrations, including the first QSS site making the move from MPE to Linux. Corona-Norco USD was the first QSS customer to make the transition from Version H to Version L in January, 2014. Their HP 3000 was replaced by a Linux application server accessing data from MS SQL databases.
El Dorado COE migrated from Version H to Version L in November 2014 over the Thanksgiving break. EDCOE is running a monolithic system with the Linux application server and PostgreSQL server on the same virtual machine. EDCOE originally planned to use SQL Server as its database server, but opted to use PostgreSQL based on the results of their evaluations. Sac COE replaced their 3000 with a Linux application server using PostgreSQL as the database. San Benito COE switched over the Labor Day 2014 weekend, accessing data from MS SQL. San Ramon Valley USD made the leap over a 4th of July weekend replacing their 3000 system with a Linux and MS SQL combination. Folsom Cordova USD replaced their HP 3000 system with a Linux application server accessing data from MS SQL databases. Merced County Office of Education made the transition to Version L with PostgreSQL as the choice of database.
QSS/OASIS is a suite made up of modules Base Financial (GL, AP, AR, Budget, PO's), Purchasing, Budget Development, Stores Inventory, Fixed Assets, Base Personnel, Position Control, and Payroll, plus a Financial Companion for interfacing to the School/3000 software. School/3000 is an integrated admin system for HP 3000s distributed by QSS that includes GL, AP, AR, payroll, retirement, position control, human resources, stores warehousing, and fixed asset inventory.
February 15, 2017
Wayback Wed: An Emulator's Partners Enter
Four years ago this month, the software that will continue to propel MPE/iX into the next decade earned its first partner. The support for the Stromasys Charon emulator first showed up from Minisoft, the vendor who announced an iPad-ready version of Javelin when Apple's tablet empire was new. Charon got a version of Javelin while the Stromasys product was just making its way into production status.
The promise of an emulator slowed down migrations in 2012. Freeware was showing up during that year that was tuned to Charon's HPA model. Keven Miller created a free utility to transfer Store to Disk files to the virtualized 3000 in the HPA. Minisoft broke the commercial software company ice with a product license created especially for the emulator. For $49, managers could now buy a Javelin to work inside the freeware version's 1-2 user license.
It was a small and initial development to show a marketplace was emerging for the sustaining aspect of the 3000. Freeware Charon (the A-202) was replaced by professional installation and proof of concept within a year. That change elevated the success rate for deployments. Software licensing became the only serious issue to resolve for a Charon site. For nearly all vendors, even though they didn't rework software itself, the licensing became an easy transfer. Software from one 4GL vendor remains an exception, but that company has vexed 3000 sites throughout three different ownerships.For decades now, Minisoft has been selling a terminal emulator as well as ODBC middleware to link 3000 databases. Years ago, the MS92 terminal emulator became Javelin, rewritten to use Java. Javelin got its Charon groove on in 2013.
"We have a special Javelin 2-user HP700/92 Terminal Emulator that is customized to work with the Stromasys CHARON MPE Emulator," said the company's Danny Greenup. A press release announced the first license to be crafted for an emulator since's HP announced its MPE/iX licensing strategy for the likes of Charon in 2004.
Minisoft has enhanced its Javelin HP700/92 Terminal Emulator to work in concert with the Stromasys CHARON MPE Emulator by adding support for raw connections to the TELNET type and support for SSH tunneling. With the communications set to TELNET(raw)+SSH, the console ports are accessible from outside the Fedora (Linux) system to a user with SSH logon privileges.
The cost of this special 2-user version of Javelin is $49. In addition to HP 700/92 terminal emulation, Javelin support access to legacy host computers requiring IBM 3270, IBM 5250, and DIGITAL VT320/420 terminal emulation. All Minisoft Terminal Emulators include scripting, SSH/SSL connectivity and network file transfer.
As of this year, we've seen 15 years of migration and decommissioning 3000s, all of them projects that sparked engines of IT spending and vendor revenues. Some sparks have been as small as $49. It's been an amazing example of dexterity, faith, and hope as your community has pivoted its business and operational practices. 2013 was not the first year companies sold software and services to spark a model of 3000 sustenance. But four years ago, one vendor saw that emulation was going to provide growth in the 3000's ecosystem, too.
February 13, 2017
ODBC treasure might be in your system
Solving HP 3000 challenges can sometimes be as simple as tracking the tools in your hand. Tim O'Neill, a 3000 manager never shy about asking for help, checked in on the 3000 mailing list needing help for his databases.
We would like to export all the data in a format that could be imported by Microsoft Access. Data relationships would be redefined after import. It would be nice to export, with relationships defined, that would run on Windows.
Minisoft's Doug Greenup peered over O'Neill's shoulder, as it were, sitting at his console.
Actually you own our ODBC driver which could be used for the requirement you outline. You were on support until 2009, so you have a version that handles this.The website support.minisoft.com has extensive documentation on our ODBC tool. You could also renew your support and get the most current ODBC version, along with access to our technical team to assist you.
The HP 3000 community is full of databases that need access to the world of Windows. Sometimes those 3000 servers have lightly-used tools to make the connections. As is customary for a budget-sensitive group, O'Neill's collegues on the mailing list had ideas on how to do that export without buying anything.Lars Appel, helping out from a perspective of supporting the Eloquence database (an IMAGE workalike) said a migration tool from the Marxmeier labs might do the job.
If ODBC is not an option, you might also try the DBEXPORT program that is available for migrations to Eloquence. By default, it writes output to "comma separated" text files (unless you use binary mode), which will likely work for importing to MS Access. The link page includes a link to the download location, including source code (so customizations would be possible).
However, using ODBC is likely better for transferring structure info.
Appel's sum-up takes note of the advantages of buying tools and maintaining support for them. Paid solutions do more, do it faster, and include access to support pros to explain how the hidden treasure can solve problems. Minisoft is still selling its ODBC solution, too.
February 10, 2017
3000 support branches into multiple types
Hardware support for HP 3000 sites comes in differing levels this year. At the top is the system administration and MPE support that production machines demand. It's crucial, but many 3000 sites try to self-maintain their MPE/iX. The next level down comes in application and utility support. One step below is support of the hardware hosting the system. Finally there's peripheral support for anything that's not inside HP's servers.
Comprehensive support is a collaborative effort in many cases. Physical hardware support is often a regional affair. For example, Essential.com is located in Pennsylvania. Its website says "We’re central to several major East Coast cities including Washington DC, Baltimore, Philadelphia, and New York City." That seems like a clue that the middle or western parts of the US aren't covered as completely. Whether that's true depends on what you need. Peripheral support for larger storage devices is available at more providers.
For example, Ray Legault at Boeing says his organization on the West Coast has used Essential. "They were okay," he said. "They mainly replaced DLT8000 and SCSI drives for our HVD10." There are no more HP hardware hosts for MPE/iX at Legault's branch of Boeing. The Stromasys Charon emulator drives the production computing at Boeing. It uses standard Intel hardware, boxes with ubiquitous options for support.
Legault employs a different support provider for its software and MPE administration. This is a common combo in the 2017 world of 3000 management. For example, the Pivital Solutions arrangement to care for 3000s combines long-term software experience -- they've been providing support since 1995 -- with hardware partners. A manager needs a provider who vets partners and keeps up with expertise.
"A reliable network is an everyday battle," said Pivital's Steve Suraci. "It used to be one primary and one secondary company to cover the entire continental US and they did it well. Not so much anymore. Almost every contract takes an effort to vet out a reliable resource and a backup." If a 3000 manager's plans don't include a backup to their hardware support providers, that can be a problem during a downtime crisis.Another experienced software company connects with clients from hardware support referrals. "Most of our customers still come through a third party hardware company," Allegro's Steve Cooper said, "but many contract with us directly. Some of those are self-maintainers, some rely on their applications vendor for hardware support, and some are just rolling the dice. We try to be flexible, meeting our customers' needs. As a result, we see a bit of everything."
In some cases a hardware support provider will service those peripherals and offer support for the Series 9xx servers, as well as A-Class and N-Class boxes. Checking out the pedigree and track record on the HP iron can be problematic. Every company has reference accounts, but the references should include some detail on what kind of hardware support issues got resolved. Some companies are stronger in peripheral hardware support than deep knowledge of HP 3000s.
As the community moves into a mature phase with only installed base sites, outside companies new to HP 3000s are thinking of entering this space. Savvy software background is one of the best ways to vet a potential provider of hardware support.
February 08, 2017
3000 hardware support resources requested
We're developing a listing of companies and consultants who do HP 3000 hardware support here in 2017. Recently some customers have been searching for resources to help keep HP's 3000 hardware lively and healthy. It's sometimes surprising to learn where HP's 3000s remain active and productive. Archival systems are at one level, and production boxes at another. Everything that's a working machine needs an expert to call upon.
Self-maintainers are abundant in the market by now, but spots like the Ecometry web and catalog shops and manufacturers the world over still need HP's iron to boot up and run as expected. Even if you self-maintain you need a resource for parts. It won't impress your top management to learn your parts resource is eBay.
Obviously the hardware support arm of Pivital Solutions is our first recommendation for North American HP 3000s. Steve Suraci says that hardware service in 2017 demands a network of providers, coordinated and managed by a go-to, first-call company.
"We continue to support both MPE and the underlying HP 3000 hardware as one of the select few remaining support companies with access to HP's original MPE/iX source code," Suraci said. "We maintain 7x24x365 phone support for those requiring a total Service Level Agreement. In New England, we support our hardware agreements with our own local technicians. Outside of New England, we support our customers through a network of contracted technicians that have agreed in writing with us to support our customers SLA. In many cases, we will maintain parts on site to help facilitate quicker times to recovery."
That network of technicians covers regional areas. A physical visit is often essential to getting a hardware problem resolved. There are YouTube video services that might be used, or even a FaceTime call or Skype connection that might be a how-to experience. That's a rare solution in your market. The problem with offsite hardware support is liability. Once anybody other than a technician contracted — in writing — troubleshoots and replaces components, the liability lies with the person handling the physical hardware.
We want to build a thorough list of resources, even while the Stromasys Charon emulator continues to replace HP's iron for MPE/iX. Vendors, send an email to us if you've got current clients. Be sure to provide an email and web address, plus a phone number, so we can contact you to follow up. Customers, if you use a hardware support company, tell us who it is. We'd all be happy to hear how it's worked out for you, too. To be fair to everybody, we'll want to use your company name in any references. Share your wealth.
February 06, 2017
German A-Class sells for $162 per CPU
Yesterday afternoon the seller of the A-Class twin-processor model A500 closed his auction of the server. After seven days the bidding rose from an opening bid of $1.07 to $323.59, not including shipping. Some lucky bidder who's been using eBay for stocking up on computers, terminals and servers now owns a system that sold for $37,000 new: A greater than 99 percent discount.
One way to sum this up is to watch nearly all of the hardware value of an A-Class—a device that represented the ultimate line of HP's MPE/iX hardware design—evaporate over 15 years. However, the computer sells in today's US market for at least $1,300. That preserves almost 4 percent of original pricing.
However, another way to calculate this turn of events relies on return on investment. These servers are clearly in their 15th year of service. Dividing that original price by its incredible term of service gives you a cost of about $200 a month for hardware which will run a business and doesn't require replacement. The enduring benefit of MPE/iX was its astounding value. This discouraged hardware replacements, a problem HP could not solve.
Half-empty or half-full? HP's 3000 iron keeps dropping in cost. The components are aging, of course. Finding a handful of systems to part-out for spares could keep such a 15-year-old server running. Intel hardware, of much newer vintage, provides an unlimited lifespan if you're using the PA-RISC emulator from Stromasys.
eBay can be a resource for HP's MPE/iX hardware, but my, a manager must be cautious. A hardware resource that's a company rather than an individual seller—or better yet, a coordinated hardware-software support enterprise partnership—is more prudent. At $162 per processor, eBay might be worth a gamble. But getting money for a server returned may not be as simple as for a disappointing collection of sports cards: one of the other purchases the new owner of the German A-Class made last week.
February 03, 2017
Fine-tune Friday: Care and feeding of UDCs
Mercury Insurance is a long-time HP 3000 shop still running a server in production. Last week Reggie Monroe reached out for a refresher on administration of HP 3000 User Defined Commands (UDCs). These are the HP3000's equivalent of scripting in Unix environments. UDCs are a better version of Command Files, according to Jon Diercks and his MPE/iX System Administration Handbook. UDCs are catalogued, Diercks says, so they can be loaded for individual user accounts.
There's a superior PowerPoint slide deck online at the 3K Associates website that covers how to create and use UDCs. But the Diercks book (no longer in print, but available online) is more concise on the use of UDCs. It's also only available as an $80 book today on the used market; put yours in a safe place. Monroe's question asked about "a command to list all users, and the logon UDC associated with them, if one is set."
The initial answer was the command HELP SHOWCATALOG,ALL. This brings an administrator to
But Alan Yeo pointed out that the MPE/iX command only locates system-level UDCs.
You don't actually get what you think you asked for, so whilst :showcatalog ;user=@.@ sounds very hopeful, in fact it only shows the system level UDCs not account ones. As far as I'm aware the only place you can find them all is in the BULDJOB2 file in PUB.SYS. You do have a BULDJOB2 file don't you? And it's up to date?
And here's where Vesoft's utility does a job the 3000's OS cannot. VEAUDIT LISTUDC @.@ finds UDCs of all kinds.We have chronicled much of MPEX during the 21 years of the NewsWire's publication. The utility was even the sole subject of the Inside Vesoft column back in the era when HP was starting to lock down 3000 futures. In 2002 Steve Hammond illustrated the distinction of UDC administration under VEAUDIT. It becomes important because security on a 3000 includes management of the UDC catalogs. And yes, there's a tool for the security, too.
VEsoft’s Vladimir Volokh told me he had been asked to find out if any users on a system had the VEsoft utility GOD in one of their UDCs and if it had the lockword embedded in the UDC. He gave me a series of two commands that did the trick and they had some added value to boot. Once I saw the commands, I was impressed with the simple elegance, but like a good programmer, I had to deconstruct it, break it down and reassemble the whole thing. If you’d like to play along, you need: MPEX, VEAUDIT (both available from, who else, VEsoft) and a healthy programmer’s curiosity (you’re going to have to provide that yourself).
The details of the exercise show "GOD.PUB.VESOFT’ [is found using VEAUDIT] and we have accomplished our mission. But wait. What are those CIERR907 files? Those are files in the list that don’t exist! But they are UDCs that have been set! Looks like you can do some housecleaning and those UDCs can be un-set. How about that — you got some value added, you killed two birds with one stone, (insert your favorite cliche here). Time to play system manager again."
When you add a third party tool to your administrator's box, you can make a purge of such files foolproof. MPE/iX cannot select to show a complete set files by attributes such as program capability. Or for that matter, by last accessed time, or file size, or file security. It's a long list of things that MPE makes an administrator do on their own. Missing something might be the path to looking foolish.
VEAudit and MPEX will root out UDCs and do a foolproof purge, including file names. VEAudit will list all of the UDCs on a server, regardless of user -- not just the ones associated with the user who's logged in and looking for UDCs. The list VEAudit creates can be inverted so the filename is the first item on each line. Then MPEX will go to work to do a PURGE. Not MPE's, but a user-defined purge that looks for attributes, then warns you about which ones you want to delete, or would rather not.
January 30, 2017
German A-Class for sale at $1.07 + shipping
Managers running MPE/iX in Europe can get a backup HP 3000 server today on auction at $1.07 plus shipping costs. Considering this is a two-CPU server which you can essentially tuck under your arm, (as HP product manager Dave Snow does at left, when the server was unveiled in 2001) the shipping costs might permit even North American 3000 managers to bid.
Dennis Grevenstein posted his notice of the sale last night on the HP3000 mailing list. The eBay auction starts at a minimum of 1 Euro, or about $1.07 at today's exchange rate. As of this evening there were no starting bids posted. The listing starts in German, but not too far down the page Grevenstein has English translations on the details.
Other than servers which have been given away for the cost of shipping, this is the lowest price we've ever seen for an HP 3000—especially for a model that's nearly portable and was built after 2001. Many 9x7-9x9 servers have been offered for outrageous discounts, especially considering their original pricing. This is an ultimate-generation HP 3000. Earlier this month, a single-processor A500 was being offered for $1,200 in North America.
The description notes that the German server is an rp2740 "with a slightly different firmware. It will also run HP-UX or Linux without problems. There is one hard disk with MPE/iX 7.5 on it and a spare disk." The note reminded one 3000 veteran about the performance drag HP that saddled the RISC processor with as a result of that firmware. The eBay listing is straightforward about how much the HP of 15 years ago hobbled the A-Class.The A500-200 server "naturally runs MPE/iX. It has two 650MHz PA8700 CPUs (clocked at 200MHz when running MPE/iX)."
Even though those CPUs are now selling for about 54 cents apiece, with two disks and 3GB of memory included, the down-clocking of the processors still bugs Patrick Santucci. "I'm still irked at HP for underclocking and crippling the 3000 like this," he said on the mailing list discussion.
Grevenstein said in this email offer, "I need to free up some space. International shipping is possible. Please ask for shipping costs." A-Class servers continue to run production-grade IT datacenters around the world.
January 27, 2017
Friday Fine-Tune: Memory and disk behavior
By Jeff Kubler
Along with the relationship between your CPU measurements and overall performance, memory and disk make up the other two components of your HP 3000 performance picture. Main memory is the scratch pad for all the work that the CPU performs. Every item of data that the CPU needs to perform calculations on or updating to must be brought into main memory.
The CPU must manage memory. It must cycle through the memory pages, marking some as Overlay Candidates (this means that new data from disk may be placed here), noting that some are in continued use, and swapping others out to virtual or what is called transient storage. Swapping to disk occurs when data is in continued use but a higher priority process needs room for its data.
To accommodate this higher priority process and its need for memory space, the Memory Manager will swap the memory for the lower priority process out to disk. The more activity the Memory Manager performs, the more CPU it takes to do this. Therefore it is the percentage of CPU used to manage memory that we use as a measurement.Page Faults per Second
A Page Fault occurs each time a memory object is not found in memory. The threshold for the number of Page Faults per second that can be incurred before a memory problem is indicated varies with the size and the power of the CPU. Larger machines can handle more Page Faults per second while a smaller box will encounter problems with far fewer. We have found that the number of Page Faults per second a system can endure without problem rises with the relative performance rating of the machine.
An exceptional number of Page Faults should never be used as the sole indicator of memory problems but when observed should be tested with the memory manager percentage. If both agree, you have a memory shortage. There are some strange things that I have observed with Page Faults, so it does not stand alone as an indicator of memory shortage.
The number of Page Faults per second and the amount of CPU needed to manage Memory are always evaluated in conjunction with each other. That is to say the high Page Fault Rate will not be considered a problem if the Memory Manager Percentage is not above 4 percent.
The Disk Environment is referred to as Secondary Storage. This is where all the data needed for system use is stored. Since Main Memory is not large enough to store all of the data that will be needed by all the processes, there must be a location for this larger pool of data. Even though the Disk Environment does not have the significance it once had, this area can still be a bottleneck. As the CPU speeds increase, bottlenecks become more significant.
Several different factors can affect the Disk Environment. One of these is data locality. Data locality pertains to two different types. There is data locality within Image datasets and data locality across the disk itself.
Data locality across disk: This refers to the location of separate pieces of files (called extents). When files are placed on the disk, they can be placed in contiguous sectors or sections of files, or they can be placed in non-contiguous locations or even on many different disks. When files are not in contiguous locations they are said to be fragmented. The advantage of contiguous location is that greater efficiencies are allowed in retrieving data. When files need to be read, the head movement of the disk drive is minimal if files are in contiguous locations. The head moves to the location and the retrieval begins.
As the disk fills up the system cannot find one contiguous location to build any new file. Therefore, the system breaks the file up into extents and places the file wherever it can. A system reload will put files back into contiguous location (usually back on the location of the files file label) or products such as Lund Performance Solutions De-Frag/X can be used to put the files back into contiguous location.
Operating systems allocate disk space in chunks as they create and expand files and transient disk space (swap areas, etc.). When files are purged, these chunks are released for reuse. Over time the disc space may end up fragmented into many small pieces, which can slow the performance and the reliability of the system.
To observe and correct MPE fragmentation, you can use the De-Frag/X product from Lund Performance Software or use the Contigvol command of MPE/iX's Volutil program. The latter creates contiguous free disk space on a volume. Contigvol work about as well as VINIT CONDense did -- that is, it's stable and reliable, but requires multiple passes to get the best results.
Data locality within IMAGE datasets is the other area of major concern. There there are two different types of datasets to be concerned with, detail datasets and automatic or master sets.
The Detail Datasets
This type of set holds the day to day data input. Detail sets begin with nothing in them. When records are added 1 is added to something called the high-water-mark, a number that tells how many records have been in the set, and the record is placed in the set.
The problem is that IMAGE automatically reuses space that is given up when a record is deleted. This space is often called the delete chain. New records are placed in the most recent location available on the "delete chain." This means that new records are not in the same physical locality as the rest of the records and may be far removed from the other records.
The ideal state for a detail database is one where the detail entries are sorted by the key field. This allows the data to be retrieved in the smallest amount of IOs making efficient use of the MPE systems pre-fetching of data. When this is not the case we can measure the dataset lack of efficiency with something called the Elongation factor. This is simply a measure of how many more IOs the user must perform to retrieve desired data.
The Master Datasets
These have unique identifiers (field names). There are two types of master sets, a manual master and an automatic master set. Manual masters have user-entered master entries while automatic masters have automatic entries placed in them to accommodate access to detail records. The issue of importance to performance here is something called the hashing algorithm. This is the method used by the database to calculate the location of the next record placed in the database. The intent is to cause the master set to be as equally distributed as possible.
The hashing algorithm uses the size of the set in its calculation. A poor size or a size that is not large enough will result in an unequally distributed database. A poor size is most easily described as one that does not consist of a prime number. This means that when the hashing algorithm calculates a location there is a higher potential that a record will already exist in that location. When this happens a secondary position must be calculated. When secondaries are placed in another block within the database, another IO must occur to retrieve needed data. Since IO to disk is the slowest type of access, we want to avoid this at all costs.
January 25, 2017
Migrate, emulate: Wednesday show's for you
Thursday, at 2 PM EST (11 PST, 8 PM CET) there's an MB Foster webinar show covering emulation options. For the 3000 owner and manager who hasn't yet moved off HP's 3000 iron, no what matter where you're headed, there's something in this 60 minutes for you.
Last summer's version of the webinar walked its viewers through Foster's eZ-MPE, Ordat's TI2/SQL, Marxmeier's Eloquence database suite, and the Stromasys 3000 hardware emulator Charon. Only the last product delivers no changes to software and frees you from HP's aging boxes. But the other three offer ways to mimic parts of the 3000's heart and soul.
eZ-MPE is the newest of the emulate-to-migrate products. Introduced in 2013, it's a suite of software to accommodate the data infrastructure and scripting needs of today's HP 3000 sites. The Thursday show includes a demonstration of the MB Foster product.
TI2/SQL gives TurboIMAGE users (pretty much everybody who's still running a 3000) an avenue into SQL databases like SQLServer. And Eloquence replaces the IMAGE database wholesale, using an SQL-based data platform with deep work-alikes for IMAGE intrinsics and features.
It should be an interesting show. The distinctions between the first three products and Charon will be obvious by the end of the presentation, so stick around to the finale. That wrap-up is also the portion of the webinar for free-form questions. It's getting rare to have a place to ask those in a semi-public setting. I hope to hear from you during the webinar. MB Foster's got a means to listen and watch these shows after their airing. But the Q&A part is live-only.Knowing the computing processes of HP 3000 managers for more than 35 years gives MB Foster the insight to build a complete ecosystem eZ-MPE, said the company’s sales and marketing chief Chris Whitehead.
“What we’re really doing here is we’re mimicking the environment that everybody’s accustomed to using,” Whitehead said. “To get all those nuances, you must have all the specific capabilities already there. With all HP 3000 sites they have some similarities. They have UDCs, file systems, KSAM that’s involved with MPE files. They all have an IMAGE database.”
Whitehead says the biggest nuance of eZ-MPE is its focus on custom code and surround code, “to transition to a supportable platform with the least amount of risk. The value of MBF eZ-MPE is its collective ability to mimic the HP 3000 environment — but aiming the customer at the advantages of the Windows environment."
On the subject of those other solutions in MB Foster's perspective, some well-established migration products have received a new label. This is an emulation-to-migration path that lands a 3000 customer in the world of Windows. Eloquence, the database that doesn't run under MPE/iX but has a TurboIMAGE Compatibility Mode, handles data. The Marxmeier product has always been sold as a migration tool. For years the ads on this blog called it "Image migration at its best." Users have testified to the strong value of Eloquence.
Another third party tool, resold and supported by MB Foster, got a mention in last summer's webinar and earned a label as an emulation solution. Ti2SQL, software that moves IMAGE data to SQL databases, was released by Ordat in the early years of the migration era. In 2003, Expeditors International included ORDAT’s Ti2SQL in Expeditors' rollout away from the 3000 because the software emulates IMAGE inside a relational database. The end result produced CLI calls native to a Unix-based database.
"Ti2SQL uses CLI," said MB Foster's Chris Whitehead. "Think of it as going to a complete native environment, while leveraging/using all of the business logic developed on/for the HP. Additionally, Ti2SQL allows someone to go to an off-path server and database, such as AIX and DB2."
January 23, 2017
SD cards take a hand in 3000 storage
One of the most unpredictable hardware devices in HP's 3000 iron is its SCSI drives. Out in the user community one enterprising manager is trying to link the server to microSD cards. John Zoltak checked in with other users last week about the project.
Zoltak was simply trying to copy one 917LX disk to a new disk on the server's SCSI bus. A 4GB drive is standard on a 917, so just about any microSD card would match that storage. A bit of open source wizardry props up SCSI2SD, a combination of hardware and software. You can purchase an SCSI2SD card on eBay and in other Web outposts.
Zoltak didn't begin there, however. He was searching for an offline diagnostic tool to do the disk copying. "I want to copy the system volume sets, so using VOLUTIL is not an option. And at this point just how does anyone get the diagnostics passwords? My other choice is to attach the 3000 disks to a PC and copy there."
That other choice leads the way to SCSI2SD. Using PC-based disks, of course, is one of the serious advantages to using a Stromasys Charon emulator for 3000 work. The 9x7s are so old they don't have a Charon equivalent, but the strategy is the same.
As for ODE, once you locate a diagnostics password (an exercise left to the 3000 customers who have a support provider) you must be prepared to wait on DISCCOPY. "There is a DISCCOPY in the ODE," Craig Lalley reports. "Hook up both drives and you should be able to copy the drive if it is copyable. I must warn you, it is slow, like all night all day sort of slow. But it is free."Stan Sieler of Allegro took a hand in this quest to more reliable disc via cloning.
My recollection of using ODE is that portions of it might not allow reading/writing to non-HP disk drives. I know that's true for some diagnostics, just can't remember which ones. If you use ODE, try to somehow check that it copied the entire disk. At some point, these HP diagnostics didn't handle large disks.I don't recall what "large" meant, but I recall bugging HP about it for some time until they released a (perhaps final?) version of the diagnostics that did support larger disks.
A few days later, Zoltak announced he was working on getting his 3000 to boot off an SD device.
I'm attempting to install to a device called SCSI2SD V6 and run the 3000 off of an SD card. So far, a few glitches. I'm in discussion with the developer and working through issues. I'll let everyone know how it goes.
Sieler said he'd tried out SCSI2SD a while ago. "I found that it worked, but that it was much slower than it should have been. I worked with the developer, but with no real success."
Things have gotten better with the device, perhaps, according to Zoltak.
I'm using the V6 version which should be faster. It supports sync transfers up to 20MB/sec. I can't yet get MPE to install on it. It hangs when I try.
So far I've hit a few repeatable problems using an Adaptec 2940 simple setup on a PC and have been emailing [developer] Michael McMaster.
January 20, 2017
RAIDing Your Storage to Homestead
By Gilles Schipper
One straightforward way to improve the value of an HP 3000 is replacement. That is, finding a better disk storage hardware component—replacing what shipped with your 3000 with a corresponding disk subsystem, one that offers the superb protection of hardware RAID (Redundant Array of Inexpensive Disks).
I would also recommend this replacement for those who utilize HP 3000 Mirror/iX software—since Mirror/iX does not protect the MPEXL_SYSTEM_VOLUME_SET volume set . It also forces one to utilize user volume sets even if the situation does not otherwise warrant it. I believe there is a place for user volume sets, but only in very specific and limited circumstances.
The main advantage of hardware RAID is that it offers excellent protection from disk failures and resulting data loss and time lost due to data recovery requirements. Chances are good that you are if you are a candidate to benefit from hardware RAID, your existing disk technology is relatively old and prone to failure as a result of years of use.
Let's face it, unlike a good wine, older HP 3000 hardware and its associated peripherals do not get better with age. Quite the contrary.
However, even if you choose a RAID technology that is also relatively old or obsolete, the nature of the technology itself affords you a great measure of protection from disk failures and corresponding data loss and downtime.
The inexpensive choices include the Nike Mod10 or Mod20 and the HP Autoraid 12H. A higher-end RAID solution is in HP’s VA7000 family of products.
January 18, 2017
Recruiter opens book on college opportunity
You don't see many requests for HP 3000 expertise by now, at least not in a public setting. But a boutique placement agency posted a request for COBOL experience on the 3000 mailing list this week. The notice doesn't deliver many details, but it stands out in a job market where opportunities have been few.
Under the covers, where consultants and developers serve 3000 shops both on the move as well as homesteading, gigs lurk. One veteran knows another and they'll contract for a period together. Most of these engagements involve finding someone familiar with a piece of 3000 software.
MRG Search and Placement has a website but there's no public listing of available positions. It's just as classic old-school as a lot of the talent that could fill those jobs. The message in public was simply "HP3000 skills needed in an University setting," and went on to mention COBOL was involved. The language usually is, considering the vast percentage of in-house apps written in the language.
Jon Culotta runs MRG, which is upfront about keeping 3000 customers and talent connected. "Established in 1976, MRG started its niche recruiting in the HP 3000 arena. That core market is still served today." The company's heartland is Amisys healthcare software talent, but a university might only be involved if it was a health organization operated by a school.
The job is on-site and contract. Culotta's email is firstname.lastname@example.org.
January 16, 2017
Older hardware, current support, new prices
HP's 3000 hardware is still being offered for sale. Hewlett-Packard Enterprise wants none of this 2017 action. Independent hardware brokers sell HP 3000s today, and by the looks of the pricing the transactions might be simply for parts. How could anyone operate a company while they rely on a $975 server?
The price is one data point on a wide spectrum of a sweeping array of servers, all offered on the 3000 mailing list this week. At the tip-top of the spectrum was a $3,175 system, first introduced early in the 1990s. At the very bottom was the faithful Series 918LX, priced at $675 including a DDS-3 tape drive. The newest computers came in at that $975 price.
The range of power ran from the 918 to the Series 989KS/650. It was a $290,000 system sold new in the late 1990s. The one on offer this week from the broker carried a price tag that was discounted $288,625.
Antiques? Some, perhaps, but not all. Series 918 and 928 servers from HP—both on the list—are running production systems today. Roy Brown, a consultant and developer in the UK and a member of the 3000 list, is running two Series 918s. One much newer server is holding archives at a migrated shop in Texas. While using the old, or very old HP iron one smart customer keeps support current for such boxes. Even when they're not on the critical path for computing.HP's sales ended in the fall of 2001 for those 918s and 928s. In that year the servers were sold for $3,700 at Phoenix/3000, the used hardware outlet operated by the North American HP 3000 distributor. In 15 years' time those boxes have held on to about 20 percent of their price.
The hardware is only one part of the ecosystem that's gotten inexpensive. We've heard of simple support agreements that are just $140 a month. At Republic Title of Texas, Ray Shahan said he's got an N-Class system hosting archived data. Shahan's company has a current support contract for this archival 3000.
It's been over a decade since that 3000 went into archive mode, so long ago Shahan said he's not sure anymore what the actual model is for the HP server. Independent support is around now to keep track of such details.
The original sales prices for those older systems "might be too depressing to hear," according to Terry Simpkins at TE Connectivity. Simpkins is among those 3000 veterans who remember when something like a $311,000 Series 997-500 included MPE/iX license fees charged by the number of users. HP placed value in its databases for the 3000, too. Non-3000 servers were less costly, until you added in the software HP included with MPE/iX.
Today's prices don't suffer under the valuation of included software. Transferrable 3000 licenses remain an audit-worthy strategy. Management rigor won't be stout for licensing software on a $675 backup server, though.
Moving onward to new prices will remind 3000 migrators of the old HP midrange pricing. For example, an LTO-5 tape duplicator—an device useful for anyone keeping archives of older enterprise data—costs $12,000 from TapeMaster today. That's an entry-level 1:1 unit that simply replaces older tape with new. Someday that duplicator will be discounted by 96 percent. It will be sold as scrap or for parts much sooner than a 3000. It won't be working in 2033, 15 years from now. The A-Class servers for sale this week for $1,200 are already 15 years old and are still working in shops like Republic Title.
It's not easy to say for certain it's depressing to see a $311,000 server go on the market for $3,175. The 9x7 line was rolled out before Bill Clinton took office. That a 9x7 is worth anything is a tribute to the stubborn economics of the 3000 line. As Clinton liked to say while winning office, it's the economy, stupid.
January 13, 2017
Emulation review will air out all options
On January 26 MB Foster is airing the 2017 edition of its emulation webinar. The 40-minute show will walk 3000 managers through four emulation options. Last year's show had four very different products. Three will address the MPE/iX environment: how to get your applications onto the Windows OS. One will give you emulated hardware. In the first edition of the webinar, Birket Foster called the Charon emulator for 3000 hardware emulation "flawless."
The other three solutions — unless the lineup changes from last year's show — are all based in software methods to replicate databases and surrounding code. They are
- Ti2SQL middleware from Ordat (a holder of an MPE/iX source code license, along with Pivital Solutions)
- Marxmeier's Eloquence database environment
- MB Foster's eZ-MPE
The MB Foster environment emulation solution has been working for at least one customer. We introduced it in 2013. Here's our story from that year for reference. We'll all look forward to the update at 11 AM PST.
MB Foster is announcing a hybrid of solutions aimed at making migrations off the 3000 easier. The company is calling its offering MBF eZ-MPE, and it’s aiming customers at the native benefits of working in Windows once they make their transition. MBF eZ-MPE is a solution for HP 3000 sites that have a keen interest in transitioning to a Windows environment, while they preserve their company’s competitive advantage and legacy applications.
Knowing the computing processes of HP 3000 managers for more than 35 years gives MB Foster the insight to build a complete ecosystem, said the company’s sales and marketing chief Chris Whitehead.
“What we’re really doing here is we’re mimicking the environment that everybody’s accustomed to using,” Whitehead said. “To get all those nuances, you must have all the specific capabilities already there. With all HP 3000 sites they have some similarities. They have UDCs, file systems, KSAM that’s involved with MPE files. They all have an IMAGE database.”
Whitehead says the biggest nuance of eZ-MPE is its focus on custom code and surround code, “to transition to a supportable platform with the least amount of risk. The value of MBF eZ-MPE is its collective ability to mimic the HP 3000 environment — but aiming the customer at the advantages of the Windows environment.
January 11, 2017
Finding the knowledge HP once shared free
In 2008 was the debate, and I don't mean between our now-outgoing President and his rival. The debate was in your community about future knowledge. Where could you expect to find HP 3000 and MPE/iX manuals in the coming years? It didn't turn out to be where it was planned and proposed, but a manager of a homestead 3000 does have options today.
For many years, MMM Support hosted the full range of HP's manuals for hardware and software. As of this morning the website is offline, but it's probably a configuration error and not a sign of a company's demise. You'll find plenty of links on our blog to the hpmmmsupport.com site. The manuals are in PDF format and you don't experience any pop-up or page-takeover ads like you see in YouTube.
The newer player in the hosted HP manual arena is TeamNA Consulting. While it's a newer site, the venture is led by one of the older (in history) resources. Neil Armstrong, one of the tech wizards at Robelle, is the NA in TeamNA. Armstrong started with the HP 3000 more than 34 years ago—an era where MPE IV was still a common OS for the servers. Plenty of experience there, and plenty of manuals available too. More manuals than HP will share with the world today. The extra information is hardware documentation.
This wasn't the future the 3000 community expected in the days when official 3000 support from HP was nearly gone. Today that support is well-filled by companies like Pivital, a bedrock upon which homesteading and 3000 emulation rests.In the waning years of HP's support, when its Jazz 3000 server was HP's exclusive repository of what the community learned, such independent companies were supposed to hold the tech history of the 3000. Speedware and Client Systems paid for licenses to HP's technical content. The document licenses went beyond the Jazz whitepapers and jobstream scripts created by the likes of Jeff Vance in the HP labs. Those licenses at Speedware and Client Systems were supposed to ensure 3000 manuals remained available to the homesteading community.
Even though the two companies made good on promises to preserve the Jazz content, including programs, the manuals escaped re-hosting there. It was an oversight or perhaps a over-reach on the part of the companies; logging and making access to hundreds of manuals is a big job. Business focus changed as well. Those Jazz links at Speedware (now Fresche Legacy and absorbed with IBM work) are tucked away under hpmigrations.com. Not exactly the place where you'd look for homesteading tools.
This kind of confusion was not supposed (there's that word again) to matter so much. HP said it would keep its manuals online through 2015. A very long time for a corporation where those promises emerged from a division that was being closed down.
The website docs.hp.com lands you on a mostly-useless landing page at HP, Inc. That's the half of Hewlett-Packard with scant link to anything related to MPE/iX. A Google search on hpe.com today unearthed those HP-hosted manuals. Well, some. At this moment they're a collection of 7.x documents, 269 of them, plus a tracer-file for the Jazz content that goes nowhere. That link above is 436 characters long, something that looks like it's going change based on how HP Enterprise keeps rearranging its business. But there it is, for now, keeping HP's promise two years later than the 2015 plan.
As for manual hosting from the companies with continuing business and with knowledge of MPE/iX, the TeamNA and MMM websites are far better Web addresses. Today. Armstrong is like me, a half-generation younger than the most senior wizards in 3000 lore. He's got more years in the future with MPE/iX, probably. Knowing where to get answers and relying on experience can keep us in the 3000 knowledge game.
It's a intern-style assignment to download the 321 manuals off the HP site for homesteading reference. This assignment seems like a good idea. It's certainly easier than locating (and storing) those blue HP binders full of paper—which were the only bibles before PDF was our tabula rasa for knowledge.
January 06, 2017
Friday Fine-Tune: Logging, IP logins, SNMP
Due to a disk crash, I had to reload my HP 3000 system recently. I’ve just discovered that system logging has been suspended. How do I resume system logging?
Paul Christidis replies:
The reason for the suspension of logging was most likely due to a duplicate log file name. When the SLT was created the then-current log number was recorded, and when you restarted the system from your most recent SLT it tried to open the sequentially next log file. Said file already existed.
- MOVE the existing log files to a hold area
- Determine what logfile the system resumed on
- Perform a series of SWITCHLOG commands until the logfile number advances to one more than the highest number in the hold area
- Then move the held logfiles back to the pub.sys group — replacing the ones created by the series of ‘switchlog’ commands.
Is there a way to see the IP address associated with a particular login?
Any user with SM can do the following, for example:
HPSTDIN_NETWORK_ADDR = 172.16.0.30
The command :listf ci.pub.sys,8 will list all sessions and will show their associated IP address.
I’ve got an older model HP 3000 and I'd like to start monitoring it with SNMP for things like CPU utilized, jobs running or whatever other cool stat I can SNMP-grab. The problem I have is I can’t find the MIBs for it anywhere.
Andreas Schmidt replies:
First of all, I do not recommend the use of SNMP on the 3000, for performance but also security reasons. SNMP is not the securest protocol, as you know. Nevertheless, here are some hints:
• In the group NET.SYS you will find the SNMPUDC. This should be set in any case for MANAGER.SYS or on system level.
• Having set this, a SNMPCONTROL STATUS will show you the status of the SNMP subsystem.
• SNMPCONTROL START / STOP are self-explaining.
• The MIBs specific for MPE can be found in the document HP SNMP/XL User’s Guide
January 04, 2017
Future Vision: Too complex for the impatient
Seeing the future clearly is not simple, and planning for our tomorrows is a crucial mission for most HP 3000 owners and allies. Changes easily cloud the vision of any futurist—people who dream up scenarios and strategies instead of writing science fiction.
Or as Yoda said, "Difficult to tell; always in motion is the future."
Economics makes every future vision more compelling. A friend who just became a city council member reminded me of this when she talked about taxis and hotel checkouts. These things are the equivalent of COBOL and batch job streaming—just to remind you this post is an IT report. Disruption surrounds them. COBOL, batch, hotels, and taxis still keep our world on its feet. Nearly all of us reach for a legacy solution when we're finished sitting in the bathroom, too.
The new council member forwarded a futurist's article on Facebook—where so many get their news today, alas—an article that pegged so many bits of the economy that are supposed to be going the way of MPE V. (I think we can all agree it's really over for the OS that powered 3000s before PA-RISC.) The Facebook article says we need only to look at Kodak in 1998 when it "had 170,000 employees and sold 85 percent of all photo paper worldwide. Within just a few years, their business model disappeared and they went bankrupt." The timing is wrong, just like the timeframe predicted for total migration of the 3000 base. Was: 2008. Now in 2017: still incomplete.
The futurism you hear predicts things like "What happened to Kodak will happen in a lot of industries in the next 10 years — and most people won't see it coming. Did you think in 1998 that three years later you would never take pictures on film again?" Nobody did, because it wasn't true in 2001 that film disappeared. Neither had MPE disappeared by 2006. These predictions get mangled as they are retold. This year's IT skills must include patience to see the future's interlocking parts—a skill that a 3000 owner and manager can call upon right now. Since it's 2017, in one decade we'll be facing the final year of the date-handling in MPE that works as HP designed it. I'll only be 70 and will be looking for the story on who will fix the ultimate HP 3000 bug.
I love reading futurist predictions. They have to concoct a perfect world to make sense, and the timing is almost always wrong. Kodak took another 14 years after 1998 to file for bankruptcy. But after I disagreed with my friend, she reached for her own success at using disruptive tech to make her point. Even an anecdotal report is better than retelling abstracted stories. The danger with anecdotes is that they can be outliers. We heard them called corner cases in support calls with HP. You don't hear the phrase "corner case" during an independent support call. The independent legacy support company is accountable to a customer in the intense way a hotel operator commits to a guest. A guest is essential to keeping a hotel open. A lodger at an Airbnb is not keeping the doors open, or keeping jobs alive for a staff of housekeepers. There can be unexpected results to disrupting legacies. People demand things change back from a future vision. Ask voters in the US how that turned out last year.
You can call the OS running Amazon an environment, but Linux doesn't much care if you succeed with it or not. Investing in your success was what brought companies to HP's 3000. It's too much to hope for benevolence from a corporation. However, if we can all stop peeling the paint off of future visions, if only we can stick to the details and know that change doesn't come easily, or quickly, we'll be okay. They're still building hotels in spite of Airbnb, just like you're still maintaining COBOL code and modifying those jobstreams first written in the previous century.
It helps to get the facts right. AirbnB isn't a hotel company at all, and faces laws to curtail its business in US states including New York. It has few provisions for safety and fraud that can stand the test of a court matter. Watch out for auto-driving cars, auto industry. Another slice of folly is that this industry is headed for the scrapyard by the time MPE/iX gets to the end of its CALENDAR function. Auto-drive car tech is more decade away if it can evade the non-auto-drive cars that will litter the roads for decades.
Onward the bright future goes, with tech saving the day by saving lives and shutting down medicine as we know it. Who needs so many doctors when you have a Tricorder X? Revised rules for that tech-doctor device contest say the Tricorder X won't have to detect tubercolosis, hepatitis A, or stroke. "Goodbye, medical establishment," so long as you don't need those conditions detected. 3D-printed houses might be built, but who will assemble them: robots that cost no more than today's tradesman labor? You can get a 3D selfie today, and a gun's parts printed 3D. We were promised code that writes itself, weren't we, when object-oriented computing and Java swept in?
A sweep of futurism helped HP put away its 3000 business. The lives that are changed and jobs lost are not a concern of the futurist. Then another change enveloped the futurist who was certain that selling systems was a secure spot. This year there are rumors Hewlett-Packard could sell off its servers business. That one is a piece of data like those ever-present reports of HP splitting up. They were just rumors for years. Then it came true. Economics, not technology, made that come true.
Nothing is impervious to change, and to celebrate the marvel of technology upending legacy leads us astray. The future is a blend, not nonsense like "Facebook now has a pattern recognition software that can recognize faces better than humans." Or, "In 2030, computers will become more intelligent than humans." How many faces, and how many humans? I'm still waiting on the flying cars I was promised at the World's Fair of 1964.
My council member says that while in Amsterdam last spring she was struck by the stark difference between ornate 16th Century architecture downtown and the simple square box apartment buildings in the suburbs. "I asked our Airbnb host about it and suggested this: There has not been a reduction in human creative intelligence. It's just that in the 1500s all that creative energy was being put into architecture, and today it's being put into the digital world. Our host, a bright young Dutch digital engineer, smiled and said he agreed with me." As every good host does.
Then Uber arrived for the ride to the airport, I presume, using a car that the company wasn't invested in, driven by a person who was working a 12-hour day pitted against a fleet of freelancers that keep Uber's business model thriving for the corporation. "And no money changes hands" was my friend's punchline, overlooking the part of the Dutch economy using ATMs and currency, or the fact that you tip your housekeeper in currency unless you don't pay one.
The futurists want you to be wary. If you don't prepare for the future, "you're going down with Kodak, the cable companies, landline phone makers, Macy's, video rental places, printed books and tape backup media." Or you can find a life keeping yourself in the present, the happiness of the now. Making good things last longer is resourceful and sometimes inventive work. If the last 15 years have taught our community anything, it's that the future arrives slowly and looks nothing like we expect. Even my council member knows the value of legacy, asking "If we close down all our paper mills, who will make our toilet paper?"
January 02, 2017
Where's your backup media in the new year?
Here in the opening days of the New Year it's time to resolve your way to a cleaner 2017. People in the US and the UK voted for changes starting this year and they'll get some, including unexpected ones. You don't want unexpected change on your homesteading HP 3000 system, though. One of the simplest means to forestall a crisis is getting fresh media for your backups. MPE/iX system backups are no better than the media they employ.
Not long ago, a 3000 manager was looking for fresh DLT tape for his backups. Tape remains a part of the backup regimen at some shops, never more true than at a site still using HP's 3000 hardware. DDS drive verification should be among your new year's examinations.
New tape media is available for purchase. New tape all the way back to DDS-1 is on the Data Tech Store website. As a minimum 2017 homestead resolution, write a fresh backup onto new tape.
Disk backup will pull your homestead practices out of the 1990s. As DLT technology fades, cheap high capacity Serial ATA discs took their turn as the method of choice for large backups. Store to disk should be the next generation of MPE/iX backup. Using an SCSI to SATA converter, newer drives can capture backups from 3000s. HP's SCSI storage devices for 3000s are at least a decade old by now. SATA disks work well for smaller systems where Model 20 HP backup units are overkill.
The age of media can be offset by more recent design. Although it's slower and has lower capacity, tape is a seasoned technology. On the other hand, disk has the advantage of being engineered more recently. Pencils versus rollerball pens is a similar consideration. You know exactly how long a pencil can be used. Pens are more indelible but expire unexpectedly.
MPE/iX servers created using the Charon emulator from Stromasys can even employ SSD disks for backups. Verifying any media, new or old, should be on a manager's to-do list for 2017. It's even better to craft a regimen that rotates fresh media, whether you're relying on tape or storing to disk.
If your management style takes incremental steps into change, then using classic backup technology alongside newer host options might work. For example, even while using the Charon emulator, an external DAT device can be plugged in to keep backups. A few years back, Paul Taffel reported on DLT tape options for the Stromasys hosts that use what us old-timers call PCs. Charon will boot up on something as modest as a laptop, he pointed out.
I had a USB-connected external DAT 72 drive and plugged it into my laptop. It is very simple to hook this HP DAT drive up to any PC (server or laptop) running Charon HPA. The drive can read and write older DDS-3 and DDS-4 tapes, and is a very cost-effective solution. I picked one up new for $300. There's also the old-way, which involves adding a SCSI controller card to the server PC, and then connecting a SCSI tape drive.
Independent software support has embraced the mission of taking 3000 backups onto the Charon emulator. Keven Miller of 3k Ranger wrote a utility a few years back that will
- Convert MPE STD (Store-to-Disk) files to/from HPA/3000 Tape Image files
- Create an MPE STD file
- Convert the STD to a tape image file
- Transfer the image to your Charon HPA emulated system.
- Link the image to a tape device
- Put the tape online in your VM MPE
- VSTORE or RESTORE from the STD
You gotta love automation for backup processes, especially while making changes for the new year.
December 21, 2016
Wayback Wed: A Dark Day for Emulation
The future looked dim for hosting MPE/iX on virtual hardware in December of 2009. Your market had little news about the forthcoming Charon HPA 3000 emulator. That software was only in alpha testing. This was the month that Strobe Data announced it was curtailing development of its 3000 emulator. Your community headed into 2010 with the hope of a Stromasys success and HP's promise to announce the new independent holders of MPE/iX source licenses.
Licensing source for an OS that only runs on aging HP hardware has value, indeed. Support customers benefit from outside licenses. It's well worth asking if your support vendor has such a license. But as a model to extend the lifespan of MPE/iX in production, source won't do the work that an emulator does: create new boxes.
Strobe hoped to do that using new hardware. The company started as a venture to emulate Digital computers as well as the HP 1000 real time machines. Many roadblocks stood in the way of a successful 3000 emulator launch in 2009. Strobe's founder Willard West intended to sweep away some obstacles by obtaining new PA-RISC processors. The chips were to be integrated on cards that would go into high-end Windows servers.
But development takes money. The resources for non-Digital development at Strobe did not materialize. It would take two more years for the ultimate winner in 3000 emulation, Stromasys, to bring out a product that needed no special HP hardware—just a special OS to run, MPE/iX.
An economic lull at the end of 2009--HP was reporting declines in all of its businesses except services --set the 3000/PA-RISC emulation work onto Strobe's back burner. The rate of hardware aging made a profound difference to Strobe, a small concern compared to Stromasys.
"We are just trying to survive the lull in government orders right now," the company's Alan Tibbetts said during the dark of that December. "The trouble is that the sales of our [Digital] PDP-11 line are down. The PDP-11s became unreliable more quickly and we have sold a bunch of them in the past, but the easy ones have already been captured." The month was a moment like the epic one in The Empire Strikes Back. Yoda watches Luke fly off Degobah, his training unfinished. "That boy was our last hope," he said. "Now matters are worse."
"No," Obi Wan replies. "There is another."Stromasys announced in the summer of 2009 it was putting its PA-RISC emulation solution into alpha testing in the fall. We reported the Stromasys product "won't rely on hardware components, going to an all-software solution that provides cross-platform virtualization. The emulator will permit MPE/iX to boot up and run on Intel's Xeon-x86 processor family as well as AMD's PC chips." A stalled IT economy looked like it just claimed the leader in emulator work.
Tibbetts said that Strobe has leaned itself up in order to weather the lull and it continues to meet with customers to secure new emulator sales in the 1000 and PDP markets. He added that he's traveling to New York State this week to install an emulation product at BAE Systems, which is testing US military jet engines using 1985-era minicomputers.
The sidetracking of emulator work at Strobe can be viewed in more than one perspective. HP 3000 community members have long wondered if competing emulator solutions could survive in the MPE/iX marketplace. The market has a strong inventory of used hardware, much of which could be considered an upgrade for owners of older 3000s. Companies have already left the market who might have been emulator customers—had HP made technology licensing available sooner to the vendors' R&D teams.
Stromasys bridged that gap, finding new 3000 clients from companies who were not on obvious maps. Two years later the first steps of a public Charon showing appeared on the trail. Watching an emulation company run short of funding didn't spook Stromasys—it also had Digital emulation customers. It had a different concept, through, as well as a broader set of resources to make the design a reality.
December 16, 2016
Friday Fine-Tune: Login recovery strategies
We have changed passwords on MANAGER.SYS—and now we cannot locate the changed passwords, thanks to some staff reductions that made the new passwords unavailable. Any ideas on how to recover them?
John Stevens says
If you're logged on as OPERATOR.SYS, do LISTACCT on all the accounts that may have SM capability, then logon as the MGR/MANAGER of those accounts: TELESUP, and SUPPORT, and there are others; LISTACCT will find them. Login as those users (unless you don't have those passwords either) and LISTUSER MANAGER.SYS. Vesoft's MPEX might help ease some of this as well.
Duane Percox of QSS adds a simpler approach:
Using your favorite editor or other utility search for the string: "ALTUSER MANAGER SYS"
You will notice: PAS=<the pwd> which is your clue.
Steve Ritenour suggests that a logon to the TELSUP account will unlock the passwords.
Some 3000 managers believe the subject itself should be filed in a place not easily found. "These responses are all well and good," said Bruce Collins of Softvoyage, "but shouldn't we be thinking twice about posting this kind of information (i.e. how to hack an HP 3000) to the 3000 newsgroup?"
Bill Lancaster disagreed. Secrecy about password recovery is not really a secret, he said.
A Google search with the right words will yield far more dangerous information about the 3000 that anything in this thread. The genesis of this information being on public networks came through BBS’s in the 1980s. I’m afraid the barn door is already open.
Can the password for MANAGER.SYS be reset?
Gilles Schipper says
Not easily. If 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 look around with fcopy (and the ;char and ;hex options) to find passwords for manager.sys.
John Stevens says
If you're logged on as OPERATOR.SYS, do LISTACCT on all the accounts that may have SM capability, then logon as the MGR/MANAGER of those accounts: TELESUP, SUPPORT, (I can't think of others right now, but LISTACCT will find them). Login as those users (unless you don't have those passwords either), and LISTUSER MANAGER.SYS MPEX might help ease some of this as well.
December 14, 2016
HP: Still a font of talent after all these years
It's Wayback Wednesday, but the 3000's history recall has fresh entries from the current day. A lot of HP 3000 sites turned away from Hewlett-Packard's offerings over the last 15 years. But more than a few have not, even after three CEO ousters and a split up of the company into consumer and enterprise parts. There's still something in the split-off parts to admire. A new book chronicles lasting HP lessons to the industry players who are lapping HP today.
Among the former: thousands of HP employees who've spent decades serving the HP customer. From engineering desk to conference presentation room, too many people to count or name have lifted the level of service. We heard from one today, Guy Paul, who once managed HP 3000s for the vendor and now is working on network storage for HP Enterprise. When asked what's remained stellar about the company where he's worked for 32 years, Paul pointed at people.
In a great book review and summary at the MIT Technology Review, the HP Way is among four lessons Hewlett-Packard's departed leaders still offer for top movers of our current day.
Make sure “culture” is about values, not practices. HP’s founders created what became known as the HP Way in several ways. Examples include insisting that the company enter markets only where it could make a meaningful contribution of valuable technology; asking employees to take pay cuts in tough times to avoid layoffs; and fostering understanding and collaboration between all corners of the company. “Management by walking around,” they called it.
But as the years passed, many employees came to equate the HP Way with particular traditions, such as the daily doughnut breaks meant to encourage conversation, or the right of top performers to earn full product-and-loss authority over their own product groups. That last one became a huge problem for [former CEO] John Young, because building computer platforms requires development of hardware, software, and other technologies that are all interdependent.
The future leaders of today’s tech giants should be prepared for similar grumbling if they have gotten too many employees accustomed to such perks as on-site massages, laundry service, and climbing walls. Dropbox said in a filing this year that it spent $25,000 in perks on every employee.
McKinney took note of the software-hardware interdependence of the mid-1980s Hewlett-Packard. His story about the era when the 3000 was growing fastest includes references to the HP 150, PC software created to enhance the value of such hardware, and a multi-division company that was ready to roll out something way ahead of its time called NewWave for PCs.
He praised HP in that oral history interview and can help us see how people like Guy Paul were attracted to—and stayed with—the HP that was built upon the Way.
When HP got in the minicomputer business...there was the HP 1000 and the HP 9000 and the HP 3000 and the HP 250 and then it kind of got all sorted out and they said, “Oh, we need [to have] one architecture and we need to be able to market [a product line].” One of the interesting parts about HP is it's just a very creative place and somehow it gets rationalized in time and [inter-divisional] doesn't become a general problem.
The part of HP that was split off, PCs, took its first steps in HP as a product to sell to 3000 customers. McKinney explained that 3000 begat PCs at HP.
In the beginning of this period there was still a hope that we could build a proprietary architecture [PC] product. Now obviously, how you sell it was one of the issues. Well I think in the beginning the [market for our PC] was major accounts who were buying the HP 3000. This is a little bit like the saying: “when the only tool that you have is a hammer, then everything looks like a nail.”
The 3000 was that hammer in an era where some top talent worked at Hewlett-Packard. It's refreshing to see that the subtitle of the new book is "Why Strategic Leadership Matters." The answer: you want to be around for decades making a difference and growing by 20 percent a year from 1958-1998. The HP of the Way did that and built the 3000, too.
December 09, 2016
Friday Fine-Tune: Driving Filesystem Checks
In the middle of a full backup, the HP 3000 at James Byrne's shop came to a system halt at 3 AM. It was the kind of halt that puts up those puzzling abort messages not even HP has fully documented. For example, about SA 1458, Robelle's Neil Armstrong said, "My experience with SA 1458 is that it is a catch-all abort. You need to look at the subsystem information and the only way to truly know the root cause would be to get a dump and analyze it." He referenced a webpage that breaks down the process of doing 3000 system failure analysis, too.
When a halt occurs during a backup, there's always the chance the 3000's filesystem has been injured. "I'd say run FSCHECK.MPEXL.TELESUP and check your filesystem," said Keven Miller of 3k Ranger. He added that a former HP support expert, Lars Appel, "instructed me that System Abort messages are in subsystem 98. From the MPE Error Messages Volume 2, Chapter 4, System Aborts, 1458 MESSAGE means A critical process is being terminated due to a trap."
Sure enough, power interruptions at Byrne's shop introduced damage to an Image database.
We reached the point last summer were we were toying with going off-grid simply to avoid the repeated power interruptions. If this sort of thing is causing damage then we will have to consider it. And it seems to; we now have a broken backward chain in one of our Image databases. A thing that I cannot ever recall. Coincidence? We are doing a backup, and then I will be using Adager to go in and take a look.
FSCHECK is an included tool on the 3000. It's simply there to validate extents and scan the table cache for missing files. Better tools include not only the legendary Adager, but independent support suppliers for the 3000 owner. People who know the fast commands of tools like CSTM. "What does your support provider say?" asked one support vendor. Self-support can be backed up by 3000-L questions. Some of the advice about the halt even ran to looking at memory issues. A provider can help eliminate these possibilities.Capricious power service from his utility — a government service that's been privatized — has extracted a price on Byrne's Series 918.
I am simply trying to find out if there is any way of examining whether or not we actually have a failing drive. We have spares but if there really is no need then I would rather not take the system down again after such a short interval. It has been a bad fall for our poor old 918. The system HDD was toasted by a whipsaw set of power outages on October 11; now our data disc is suspected of being ready to let go as well.
While running FSCHECK, Byrne was advised to use the commands
Check all Dev=all
FSCOUNT / 10000
He also received instruction from Mark Ranft on how to scan logs using CSTM to find disk errors.
Sign on as Manager.SYS. Do a LISTF LOG####,2 to find the start and ending Log files. Alter the log file number range in the commands and enter the commands in LOGTOOL.
list log=3404/3477 type=111 "device class"="hard disc",da,ca,"bus converter" out=LogOut1
list log=3404/3477 TYPE=111;'MGR CODE'= 241,242,900,901,951 out=LogOut2
The results are written to two files LOGOUT1 and LOGOUT2. I had this set up to run weekly on my systems. And if the files had errors in them, the job would email the results to me for review. You will see errors due to SCSI or FC resets on every boot, so check the timestamps to tell if the boot caused the error.
At Byrne's shop the Series 918 was recovered after Adager did its repairs. "During the Adager repair an infrequently occurring high-pitched but low volume sound was noted," he said. "This appeared to emanate from the 3000. We have not heard it since the Adager repair completed. A suspicion arose that we might have a disk about to lose a bearing."
Once the recovery was complete and the disk replaced, the trademark wisecracks of long-time 3000 vets began to arise. The sound during Adager's repairs "would have been the sound of the chains being dragged around the disk to put them back straight," said Alan Yeo, "and possibly the sound of a very tiny virtual Alfredo welding a few broken links back together."
Byrne noted that the problems with the Series 918 disks "started I was working on installing the 9.6 version of PostgreSQL on a new FreeBSD host. I wonder if the HP 3000 is throwing a temper tantrum? Naaah. Cannot be."
December 07, 2016
Taking Steps into Open Source with a Plan
A significant number of HP 3000 shops have employed Linux as a replacement over the last 15 years. (Yes, it's been that long that the MPE/iX community has been migrating or homesteading their systems). Over that time, open source software has become so mainstream that an architecture meeting often includes a line like, "Well, what can open source do for us here?"
If open sourcing a commercial datacenter sounds enticing—think of the size of the community you join, for example—it's wise to remember a commercial open source is the way to success. Downloading and testing is always essential, but adding open source has its best prospects when there's a commercial, paid support aspect to the choice.
This week we reported on one HP 3000 site where the system is making a slow exit. Harte & Lyne is still using a Series 918 with MPE/iX 7.5. The operations are being supplanted by what manager James Byrne calls FOSS: Free and Open Source Software. He's got his reservations about doing much more in that direction, though. Byrne said a more commercial—though not vendor-specific—approach to new architecture is in order.
HP was advising this to its enterprise computing customers as far back as 2006. Linux in the datacenter was a lot more exotic in that year, a time when HP was still selling support for the 3000. That vendor-based support is all gone by now, right down to the demise of docs.hp.com webpages where advice and training materials once lived. If you need 3000 support, third parties like Pivital Solutions are the best way to go forward, even if you're going away slowly.
An HP exec of 2006 said it only made sense to look for a supported FOSS design. David Claypool said
The rational thing to do is to choose something from a commercial company, whether implementations available and supported by a Linux distribution or non-affiliated Xen implementations like those from XenSource, Virtual Iron, and now Oracle.
Working together in such alliances was part of what FOSS was all about at the beginning. It would be another four years before Oracle would hire the departing CEO of HP, Mark Hurd, to run Oracle's software business. In 2006 all was pretty collegial between Oracle and HP.Campbell said, in a reply to the story we ran about choosing open source software
Certainly, it's possible and may even be prudent for some to download and run the bits from a raw open source project. But it's incumbent upon the adopter to understand the commitment to self-reliance that's being made if it's being used in any operational or revenue-producing capacity.
Linux was free until users understood they still needed a support provider to contact when things went awry. Support is the enduring part of any software relationship, and it's something critical for everyone who's using computers to drive an enterprise. Even HP 3000 shops need someone to call when the bits get out of alignment.
December 05, 2016
COBOL's Continuing Value for 3000s
The venerable workhorse of COBOL is often maligned during IT strategy meetings. The language usually has to make do with a lot less educational opportunity these days in universities, too. It is verbose and legacy and not ever going to captivate a ping pong table discussion about which platform development platform is best. There are perhaps only 1 million people in the world still trained in using it.
However, COBOL brings a single, enduring asset to the 3000s and other mainframe-caliber servers where it runs. COBOL is a standard, one that's exploited and extended and supported by many vendors, not to mention carried through decades of use. This is no one-man show.
You cannot say that about Powerhouse, or any other fourth generation language. If there's a standard out there for C-Sharp, or the wonders of Visual Basic, it is controlled by a single vendor. Powerhouse users are in a pickle. A single company, Unicom, controls the fate of all users employing the 4GL, and the vendor is jerking its leash on users. When Unicom said it was canceling the license of one customer, James Byrne at Harte & Lyne, then Byrne had a response.
"I am ignoring that," he said. "One cannot cancel a contract without cause."
He went on to point at what sets COBOL apart while he's choosing foundational software. "Not that we use it," Byrne said, "but there is a reason that COBOL is still around. The people who do not understand why are at the root of many of the problems with FOSS." That's open source software he's referencing, something that a vendor cannot cancel, but drifting toward commercial prejudices anyway."Actually, I have come to the conclusion that a great deal of the FOSS environment has exactly the same problem as [Ruby on Rails]," Bryne said. "Too many people are looking forward to their next contracting gig and padding their resumes when they introduce changes into projects. Too few are wondering about who has to maintain the trash that they leave behind."
COBOL won't regain its footing at Harte & Lyne, where a 3000 still supports the logistics vendor. But those ping pong developer tools are going out of vogue.
I have pretty much given up on Ruby-on-Rails, having slowly reached the conclusion that the maintainers are more interested in being fashionable than producing a stable and useful product suitable for enterprise deployment.
We will never return to proprietary software, but we are making some significant changes in our approach. We are presently migrating off of Linux onto FreeBSD—and we are moving away from the tech-du-jour crowd onto something more grounded in the commercial world.
November 30, 2016
This just in: Generalissimo Cobol is not dead
A favorite running gag on the most antique Saturday Night Live shows was Chevy Chase reading the fake news. With each broadcast he'd repeat this joke: "Generalissimo Franciso Franco is still dead." Same result, week after week. A situation with a lot in common with COBOL's current fortunes. Despite what people think they know about it, the Common Business Oriented Language still props up a vast swath of the business data across the world.
Nothing has changed with COBOL's fortunes since we last visited this topic with a podcast in 2014. During that year's spring, the NPR Planet Money radio team posted a show that blamed COBOL for the slow pace of money-changing in clearinghouse transfers. The mistaken report was like fingering English for the outcome of the Presidential election. Yes, the COBOL code in banks turns the IT cranks. The result is not the fault of the tool, but how it has been used. Yes, English was used in the 2016 campaign. [Insert joke about Twittering here.]
Our COBOL correspondent Bruce Hobbs pulled this story back into the light this week. He pointed to an article on the HackerRank blog, examining COBOL's not-dead-yet status once again. If you like numbers, the article included these above. Its still a language that supports 80 percent of all point of sale transactions and routes health care to 60 million patients a day.
To be fair, one of the sources of that graphic is the company still selling COBOL, Microfocus. But Gartner is also cited, an impartial consulting giant. In the NPR show the reporters interviewed an exec from Fiserv, a vendor who might have known better; they made some of their fortunes selling Spectrum/3000 for credit unions.
In the HackerRank piece, the author quotes an article from 20 years ago that surveyed why COBOL has held on so long.
COBOL does the 4 essential business tasks better than most modern languages today:
The capability for heterogenous “record-structure” data
The capability for decimal arithmetic
The capability for convenient report generation
The capability for accessing and manipulating masses of data (typically made up of heterogenous data structure).
“COBOL is either good or adequate in all 4 (except for database access and GUI construction, they were designed into the language from the outset), whereas the COBOL replacement languages, like Visual Basic and Java are good at few if any of them,” notable software engineer Robert L. Glass says.
Since the COBOL community's average age is 55, and the estimated 2 million IT pros who knew the language in 2004 declined by 50 percent since then, the technology that's as old as any Chevy Chase joke on SNL will see a new spotlight soon. Even in replacement, old skills will be in demand. Technology reporter Ritika Trikha looked to the future in her article.
As taboo as COBOL might be in the ping pong rooms of modern startup-driven culture today, its influence and irreplaceability will result in a spotlight on the dinosaur language again. Businesses must figure out who will maintain their mainframes when COBOL programmers retire in the near future.
November 25, 2016
Friday Fine-tune: Adding disks and IP blocks
Is it possible to add a disk drive "on the fly" without doing a reload?
Jeff Kell replied:
You generally have to shut the system down to install and cable the disk to avoid electrical/interface problems. The usual approach is to use SYSGEN to configure the new device on the path where it will reside, keep the new configuration, shutdown the system, install the disk and do a START NORECOVERY.
Once the disk is recognized by the system, you can add it to your running configuration as follows (assuming the new drive will be LDEV 5 in the system volume set):
> newvol mpexl_system_volume_set:member5 5 90 90 (DISC,SPOOL)
[For details, see "Volume Management", HP Part No. 32650-90045 or "Performing System Management Tasks" HP Part No. 32650-90004.]
This will add the volume to the system volume set, but it also has some side effects. Since the new volume is "empty" and the disk space allocation routines attempt to "balance" loads across drives, all of your new files and transient space will be allocated on the new drive until it's capacity approaches that of the other volumes. This will create an I/O bottleneck on that drive, at least initially.
You could selectively :RESTORE certain accounts (or the whole system) to try and balance the allocation. You could also perform an INSTALL and a :RESTORE for better efficiency, but at the cost of a great deal of time. There are also certain third-party utilities that will balance disk utilization across members of a volume set. These utilities work online on a running system and don't require any downtime.
The network configuration of our HP 3000 was originally set up with one block of IP addresses. Now I need to add another block of addresses. Where do I add these in NMMGR?
You can add an IP address using NMMGR the following way:
- After typing NMMGR, select "Open Directory" .
- Then select "Update Dir."
- Now select the "Add" option (F5.)
- You are placed in a screen where you can enter the IP Address of the machine. The type is generally set to 1( IP).
- Now press the "Save Data" (F6) option, back out of NMMGR, and you are done.
November 23, 2016
Mailing news from the HP 3000: an old skill
Internal mail hosts remain a crucial tool in datacenters, even some running MPE/iX. "You still host your own email?" is not a question you'd only pose to a crazy manager. An organization's security standards can be so high that no outside mail server will be trusted. In the earliest days of email, 3k Associates built and sold a beautiful native MPE mailing system, Netmail/3000. It's a smart mailserver, meaning it doesn't require that an organization's e-mail be piped through an Internet provider's mail server for final delivery. Then in the late 1990s, HP's lab started the long process of porting sendmail to MPE/iX.
Now some 3000 sites are looking at how to replace their 3000-based mailing software as they migrate. One of them contacted us this week to ask about an alternative to sendmail. Linux is their migration target, after a history using the 3000 that goes back to the days of HP Deskmanager. Tim O'Neill shared a story while asking about an alternative to sendmail.
I saw that FreeBSD Unix has its version of sendmail. Seeing reference to FreeBSD made me recall a story about FreeBSD running on an old HP 3000, maybe a Series 70 or an early Spectrum system. I think I have read that FreeBSD is at some sites still running in production mode, as MPE and MPE/iX are. It also made me wonder what the installed base of FreeBSD might be — and how that compares to the installed base of MPE and MPE/iX on old hardware and on Charon hardware.
FreeBSD, like MPE/iX, has some surprisingly large companies using it. You might have heard about one of them called Netflix. Of course the Charon HPA emulator from Stromasys makes every remaining product and archival 3000 a candidate for the kind of longevity we see in FreeBSD.
Sendmail has a colorful history. The Unix Hater's Handbook devoted a full chapter to the software's vulnerabilities; sendmail comes from the Unix heritage, after all. By 2003, HP was still patching sendmail to shut down security breaches, although the breaching wasn't nearly as serious on MPE/iX as on Unix variants including Linux. Sendmail's open source capabilities are now under the banner of ProofPoint, the company that purchased the sendmail resources in 2013.
Sendmail's worldwide release was last updated in 2014. HP announced it was testing sendmail to place in the Fundamental Operating System in November, 2001—a month that's famous in the 3000's history for other reasons. But the software moved along to an 8.13.1 release in FOS. It's only one major release behind the worldwide open source version, now advanced to an 8.14 release. Sendmail also includes encryption.Sendmail has included encryption facilities since 8.11. That's where security capabilities descend onto the requirements. Encrypting mail is a common feature in commercial hosting solutions. Sendmail/iX sends mail created by and triggered from HP 3000 applications, given enough technical know-how.
There's a robust webpage about the 3000 mail solution that was started by Mark Bixby. He's the engineer responsible for lighting the fire of open source flames at HP. Keven Miller of 3K Ranger has updated and maintained the page and its knowledge about Sendmail/iX. The software itself is in your 3000's SENDMAIL account in a version-specific group named vuuff.
November 21, 2016
Middleware rescues a Quiz captive, again
Old friends can help with new challenges to homesteading. Minisoft's ODBC software stepped in again to get a 3000 customer away from the pricing schemes of Powerhouse and its Quiz reporting software. That ODBC link between MPE/iX and Windows databases and tools turned out to be the essential component in pulling a client away from Powerhouse, according to the Support Group's Sue Kiezel.
Kiezel said her support client got an eye-popping quote to move Quiz from a 9x9 to an A-500 server: $27,000. Like some sites who are learning about the new regime of pricing from Unicom, this Support Group customer was returning to Quiz support after years of no improvements to the former Cognos product. Quiz made its way onto many MANMAN installations in the during the 1980s on the ERP suite. Getting out from under that legacy required a reasonable tool to connect the data with more modern reporting.
Enter the Minisoft ODBC software. The middleware connected with SQL Server to build a reporting database, data that was used to create the Excel spreadsheets everyone wants to use. As we've seen before, Windows-based reporting solutions like Crystal Reports can carry the 3000's data into departments better than Quiz did.
"SQL Server has turned into a beautiful database," Kiezel says. "You don't need a database administrator for it. Because of this kind of connection, my users no longer need paper for their reports. The middleware opens it up for MANMAN uses, and Excel can make joined tables for reports. Instead of just sending out a paper report, I'm sending out a spreadsheet, with the first three sheets of them working like a visual dashboard."
Visual Basic does analytics in this kind of report solution, too. "We are now in the modern world," Kiezel says. The bonus? Finding an expert to tune up these reports is a $50-75 hourly charge, instead of the $200 hourly that a Powerhouse consultant will charge to beautify and enhance Quiz. There are features and solutions that are worth the extra cost you'll sometimes encounter in MPE/iX. But reporting doesn't turn out to be worth the extra expense in licenses and expertise — not when there's middleware to open up reporting options.
November 18, 2016
Friday Fine-Tune: Tricks with command files
I'm working on a command file on my HP 3000. Is there any way to have it copy part of itself into a separate (temporary) file?
Jeff Vance replies:
MPE does not support the Unix concept of ‘here’ files, where input data for the command can reside in the same file as the command, except in the case of jobs. But even in a job, you may not include inline data for a script or UDC invoked by that job.
The SPOOKHELP script may be of some use. This single script contains the help text for all of the SPOOK commands plus the code to search for and display that text once HELP xyzzy is entered.
How can we execute a command after a user enters the :bye command in MPE?
Olav Kappert replies:
It is possible to execute many commands after the bye has been entered. Simply create a UDC (maybe a cmd file) called bye.
The contents of the UDC for the command bye is up to you. This would be useful if you want to do statistics before the session terminates.
John Pittman adds:
Don’t let them do a bye. We don’t allow any users access to OP system prompt at all. They get a logon no break UDC that runs a menu, and when they end the menu, they get logged off.
Inside that UDC at exit time, we build a string giving user, connection point (LDEV or IP of their PC) connect time, CPU date etc and append it to a log file. Then we know when anybody last used the system, how many users are using different connections, or when different user names are using the same connection point.
November 16, 2016
Noteworthy dates drive views of the future
This week on the 3000 newsgroup, Alan Yeo of ScreenJet picked up the remembrance torch to note the anniversary of 2001's 3000 business shut-off at HP. About your resilient computer he added, "In some ways it seems to have survived in some places in better shape than the HP that announced they were killing it!"
We agree and noted as much in the Nov. 14 NewsWire article. I promised to make not such a big deal about the history of the event; instead I tied it to recent advice about a hybrid of local and cloud-based ERP alternatives.
That event brought some benefit along with all of its carnage. Canceling the HP business operations for the 3000 (never an end-of-life; vendors don't get to define that) also sparked the completion of the first PA-RISC hardware emulator from Stromasys. The software continues to assure us all that the aging HP hardware won't be our only option over the next 11 years or so. Remember, on Jan. 1 2028, at 0000 hours, the dates stop working. Not MPE altogether, however.
A fix for that date issue might become a project for some remaining support company which has an MPE/iX source license. As you might infer from a date in this month's political events, stranger things have already happened.
November 14, 2016
The best wishes for your long life: a Plan B
Congratulations to us all. This is the 15th anniversary of the "we're killing off the 3000" announcement from HP. The end-game hasn't played out like HP expected. In 2001 the company's management didn't see three CEO resignations coming over those 15 years, or the company being forced to split itself to stay relevant to enterprise IT. Those two events are related. Yes, the 3000 got its pink-slip notice at the HP of 2001. So did the overstuffed, unwieldy Hewlett-Packard. The company that lurched toward every business while stepping back from others. It took 14 years almost to the day, but HP is half the size it was: HP Enterprise is the severed sibling from 2001's family.
Inside the 3000's division during that year, no one was talking about emulating the 3000 PA-RISC hardware that the company would stop building in 2003. That's now a reality, a new development since the 10-year anniversary of this sobering date. Hewlett-Packard was going to lead four customers out of every five away from MPE/iX, delivering them to the Unix alternative of HP-UX. Windows was going to get new customers out of the upheaval, too. No one figured three of every four departing companies would choose a non-HP environment.
Here on this date in 2016, the idea of an environment as a crucial strategy is feeling outdated. IT directors always cared about applications. Now they're told they don't have to worry about environments. The cloud computing providers will do that for them. Except when they cannot provide the cloud. Behold (above) the map of Internet outage from last month on an ugly day.
The Support Group's Terry Floyd offered a Plan B strategy to the manufacturing customers of CAMUS last week. More than 30 companies using HP 3000s and MANMAN are in the CAMUS user group. Floyd's company is delivering a fresh alternative to help MANMAN sites move on from the 3000. But he also supports homesteading sites. With a foot in both worlds, he recommends staying safe by having a Plan B, even while you employ cloud computing for your future."I'm still a little bit paranoid about the cloud being out there," Floyd said on the 90-minute RUG conference call. (Keep in mind, he's bringing a traditional manufacturing site's IT onto the Kenandy ERP cloud solution, so he's being extra-careful.) One of the Support Group services runs manufacturing datacenters for some clients.
If any of you are thinking about cloud apps, you should think about a hybrid app. You'd have some stuff in-house on your own boxes, and some stuff out there on the cloud. For instance, we're doing EDI [for a client]. It's pretty much local. We'll be able to receive and send stuff even if the Internet went away for a day. It would kill us not to be able to do EDI. Even hours of Internet downtime would kill us in some situations.
Think about what you might consider really critical to your company—and think about putting some of that stuff in-house. Having shipping on a local server, for example a SQL Server, we'd be able to ship whether the Internet's up or down.
"Sometimes the Internet goes away for different people for different reasons," he said, and it's so very true. DDoS attacks are becoming a too-regular event for the world's Internet. When Twitter, Netfix, Amazon, Tumblr, Reddit and Pinterest can be taken offline at once, as they were on that map of Oct. 22, everyone needs to manage the risk. A Plan B once meant staying on the HP 3000 in spite of HP's community exit. Today it means keeping some computing local, no matter what your enivronment.
November 11, 2016
Friday FineTune: Internal disk plus VA array
FiberChannel Storage Area Networks and shared tape libraries became popular in the years after HP stopped making its MPE/iX hardware. The HP 3000 supports SAN from the XP series of RAID devices to the VA7100 disk arrays. But how much should you rely on a RAID or SAN device? Internal storage devices might seem to be yesterday's tools, but the modest drive inside an HP 3000 can still be very useful — even if a company has invested in the FiberChannel storage solution of the VA7100.
Moving to the VA solution has great benefits, as reported in a story about using a VA7100 array with the 3000s. But booting directly from a VA array — well, you'll need an N-Class server (native FiberChannel installed) or a very expensive HP A5814A-003 Fiber/SCSI router (if you can find such a thing) to employ VA in the 9x9s.
The Crossroads SA-40 Fiber/SCSI switch will link a VA array to 9xx 3000s. It just won't let you boot your MPE/iX system from any of its drives. Craig Lalley of EchoTech recommends the affordable Mod 20 arrays for boot capability.
Internal drives remain as important as the VA arrays for a Series 9xx HP 3000, or even to the XP line of HP arrays. Even important enough to even duplicate them.
A second bootable disk inside your 3000 can take some forethought, but can be essential to smooth recovery of an LDEV 1 failure. James Killam of HP once reported to the 3000 mailing list, "Keeping a bootable image of MPE on one of the internal drives... saved me once at 2 AM when we lost total connectivity to the XP array the system was attached to and we had some serious troubleshooting to do."
Donna Hofmeister, former OpenMPE board member, has noted:
I did one internal disc and it was expressly for memory dumps. A second internal disc with a bootable image would be wonderful insurance. It would take some planning to be able to manage it all, but there’s no reason why you can’t have two bootable discs. I’ll point out the obvious: if LDEV 1 is internal, and you have a multi-disc system volume, and the remaining system storage is on a disc array — uh, what’s the point? If LDEV 1 fails, you’re toast.
Hofmeister said she's had a drive fail on a VA array. "That array worked perfectly and switched over to the spare without a blink. Given that I had two systems sharing this array, I was more than pleased with how well it worked."
Drive failures are among the most likely of hardware problems using HP's 3000 devices. A second internal drive in a system can make a big difference in recovery time. The other way around drive failures is to make your way onto emulated PA-RISC systems from Stromasys.
November 09, 2016
A Response to Being Stunned: No Tribute
Citizens of the US woke up this morning to a turn of political events described everywhere as stunning. There's nothing anyone can do to change that today, but in the event of a stunning relicense quote for Powerhouse products, you can respond with software that preserves your reporting administration. Some customers using HP 3000s can stun right back by leaving Powerhouse, using software from Minisoft to pave their data's way.
Steve Suraci of Pivital Solutions told us that one of the HP 3000 GrowthPower clients he supports has opened up one of those stunning relicense bids. In trying to get their software back onto support with the vendor, the customer received only an offer to relicense the full version of Powerhouse. "The most current product doesn’t even run on the 3000," Suraci said, explaining the folly of the return to support tribute being demanded by Unicom this year.
It's easy to think of back-support fees, levied in a market the size of the 3000's, as tribute: money demanded for nothing in return except a promise of help. A small promise indeed for software like a Powerhouse suite that hasn't had one MPE bit improved in more than 7 years.
The demand made even less sense considering what the customer was using. Quiz, the reporting end of the GrowthPower application, was the only Powerhouse software running on the 3000. "They originally acquired the product embedded in their ERP application," Suraci said. "They ended up purchasing the Minisoft ODBC and recreating the necessary reports using SQL tools like Crystal Reports, SQL Server, and Access."
Minisoft's products have never had an acquiring entity like Unicom take over and then demand such tributes from 3000 sites. Returning to support is a noble practice, something a manager with integrity does. However, this is a good deed that can be punished by ignoble companies. Support returns are a tradition that can trigger back-support fees. You don't have to pay them, but then your data has to live software else to get its support. The situation mirrors the dilemma of more than half of those who voted in the US yesterday. They don't want their President-elect, but they want to be citizens, too. It'll be awhile to see how much tribute the new President will demand. HP 3000 data is in a luckier situation.Data hosted on HP 3000s goes wherever the tools like ODBC can take it. The reports then flow from non-MPE software like SQL Server or Crystal Reports. That GrowthPower site, new owners of the Minisoft ODBC software, are a small division of a very large corporation. Tribute is not about the size of the IT budget, al it's the integrity that stings when a company gets stunned. Responses to being stunned include resistance, as well as reaching for alternatives.
The alternatives are more affordable for the Powerhouse customer who's only got reporting to replace. The full-development installation of Powerhouse faces a much bigger problem when they get their demand for tribute. They have to move their data's house to another country, if you will—transforming their application's platform. One solution for that is the Core Migration software and services. The transformation is a mighty task, though—a bit like thinking you can become an expat after a lifetime in a country you thought you understood.
While the United States was very new, the young country was put on notice by France. Pay us a stunning sum, said the French, and we won't attack you. The demand sparked a classic cry of resistance from the era of Founding Fathers: "Millions for defense, but not one penny for tribute," said Robert Goodloe Harper. When a 3000 site finds a means to defend its investment in MPE, they're resisting all demands for tribute. When you're lucky, your resistance can be as straightforward as ODBC.
November 07, 2016
Work of 3000s Helps Preserve Democracy
Tomorrow is a very special day in America. In a land called the United States we're going to elect a President to unite us. The kind of future we work toward will be chosen on that day. I'd like it to be the same kind of future the HP 3000 community has always worked toward.
This computer is called a business server because it works to meet the needs of business. A business relationship is at the heart of manufacturing concerns, insurance organizations, e-commerce companies and more. Business is at the heart of good relations with others in our world. MPE/iX software has always been a part of good relations. Much it serves the processes of business like invoicing. Going Forward Together might as well be a way to say Make Relations Through Documents. Business documents are the bedrock of your community.
In the earliest part of our 21st Century, Wirt Atmar was holding a seat as the conscience of this community. The founder of vendor AICS Research railed at HP's plunder of loyal customers, then proposed a Plan B to resist needless change. It was a time of high passions. The most crass and base expressions of the IT pros in our world were on display in the 3000-L listserver in that era. But since this is a republic with freedom of expression, although that trolling was revolting, it was tolerated. Much of that era's tone seems gentle compared to what's assaulted our ears and our spirits since this year began.
Back in 2004, Atmar was teaching his community how affordable Web-based lecture software could give minds a common ground. His QCShow product followed QCTerm, and both of those sprang from the makers of QueryCalc. In an HP World demo and lecture, Atmar explained his belief about how an HP 3000 was an alternative to war and atomic armageddon. These are real prospects for an American future. It feels like a disturbing misfit that anyone devoted to MPE, and having built a life's work from it, should vote for anything but a diplomatic leader.
Atmar had a fascinating background, including a stretch of his life when he worked to estimate and calculate the effects of annihilation. Nuclear throw weights -- the number of tons of atomic bomb to destroy various numbers of people and structures -- were his everyday work as a scientist in a government defense contract. He said he hated every day of his life that he had to wake and perform that work.
In contrast, when he created business tools that delivered invoices and orders, he felt his work spoke to the very root of human decency. Invoices, he said, were the everyday diplomacy of enterprises and organizations. I agree to purchase these goods and services, each would say. I agree to make and deliver them as you ordered, replied each sales receipt. A world still sending invoices, he said, ensured that war and revolt was a poor choice. Invoices were an expression of peace and a shining light for democracy and capitalism.
Something approaching half of America has already voted in this year's Presidential election. For those who have not, asking if a leader should respect business partners, find allies, and preserve relationships with respect— these all are a guide for anyone who's ever programmed or managed an HP 3000. Nobody is perfect. Anyone who wants to lead us should respect invoices, contracts and agreements. Tearing up a legacy is a poor start toward the future. Every HP 3000 community member should agree on that, and agreement is a good start toward where we need to go. We don't need to migrate away from working together and moving forward. Rather than looking back, we should take a hand in making history. Vote tomorrow and make some.
November 04, 2016
Reporting software takes over for pricey 4GL
By some estimates the 4GL software from Cognos sat on 7,000 HP 3000s at the 4GL's high water mark. A very serious share of that installed base was using only the reporting tool associated with Powerhouse, Quiz. Manufacturing sites such as those employing MANMAN, plus other applications, relied on Quiz to produce reports for managers and C-level executives. In many instances, these Quiz licenses came without restrictions or separate support agreements.
These are the sites that never had much of a business relationship with Cognos, and none at all by the time IBM bought the 4GL suite in 2007. Some of these sites eventually felt they needed to buy support, though, and some believed maintaining a license was important -- even though they'd become Quiz users when they implemented their application. The majority of Quiz sites stopped paying for support long ago. Like many bits of MPE/iX software, Quiz was frozen in time, a day when a reporting tool could cost thousands to support.
It was a bolt-on module, something that customers could be taught to un-bolt when pricing got outrageous, though. Cognos used to try to tamp down the outrage during the 1990s about license costs. Renegotiations were common, because the default pricing maintained strategies of an era when Windows was not a lower-cost enterprise option.
Steve Suraci of Pivital Solutions supports HP 3000 sites as an exclusive business model. It's an all-3000 vendor. He checked in about the latest shock over Powerhouse re-license prices. In the era of of ownership by Unicom, the licenses have soared agin. But there's another option to letting Powerhouse lock up a company to a 3000 license.
Back in the day, Cognos was always out of their minds regarding pricing. When they were bought by IBM, they got much more realistic and started offering a seat-based licensing model. Then Unicom entered the picture and they lost their minds!
We have a customer that owned the reporting-only version of the product, but was no longer on support. The only option available from Unicom was to re-purchase the full-blown development product in order to upgrade. The most current product doesn’t even run on the HP 3000. Needless to say, we replaced Quiz with something more robust and much less costly.
October 31, 2016
A Scary Kind of October Surprise
James Byrne, a systems manager at Canadian logistics management firm Harte & Lyne, has reported a hair-raising development at his 3000 shop. A straightforward request to relicense Powerhouse from the MPE/iX version of the software to Linux resulted in an eye-popping quote.
The supplier of the software, Byrne said, has told him they want $300,000 to move the 20-seat license. Byrne noted dryly, "I recently had my decision to move our company away entirely from proprietary software validated in a most dramatic way."
It's always possible, when numbers like this surface on a Powerhouse relicense bid, that the wrong person in the Powerhouse business line has responded to a request for a quote. Byrne reported this exchange on the 3000-L mailing list, but didn't want to name the software vendor of Powerhouse. It used to be Cognos, but that stopped being true many years ago.
In a message of nine years ago, the debut of Powerhouse for Linux seemed tied to the fortunes of Powerhouse for HP-UX.
Cognos continues its ongoing commitment to its PowerHouse customers with the upcoming release of PowerHouse 4GL and PowerHouse Web for Linux. This is a direct port of the industry-leading application development tool that is so successful on other UNIX platforms as well as MPE/iX, OpenVMS, and Windows. User-based pricing for PowerHouse 4GL and PowerHouse Web for Linux is the same as for other UNIX versions. Please contact your Cognos Account Representative for availability.
Byrne said the exchange with the current supplier of Powerhouse licenses ended with a termination of the Harte & Lyne license for the software -- just after he was told the annual support fee for the relicensed copy was going to be $60,000 a year.Hopes ran high for awhile when Powerhouse became part of the Unicom Systems software and services business early in 2014. The corporation has a long history of serving the mainframe and IBM communities with products. This legacy business might have made Powerhouse an attractive acquisition more than two years ago.
But the product was acquired from Cognos by IBM just a matter of months after the Linux version of Powerhouse emerged. Even 2007, people wanted to say that Powerhouse was past its use-by date. It became easy to find companies using HP 3000s who were sticking with it, of course. The University of Idaho State was a notable installation, a 3000 site that's moved on to other platforms by now.
Quotes like the one mentioned in Byrne's post are a good reason to try to take a step away from a development platform tied to the rosiest days of 3000 data processing. It's not easy, though -- and doing the right thing by a relicense can generate quite the October surprise of a quote.
October 26, 2016
MPE/iX to private licensees: A new HP way?
Fifteen years ago HP was cutting its 3000 business loose and software vendors scrambled. A few of the bigger ones, like Adager, were looking for a way to buy the MPE/iX assets from Hewlett-Packard. Nothing could be arranged. However, HP recently started posting notices about its patented technology it's trying to license.
The IAM Market (free registration required) has started to hawk the intellectual property of both sides of the HP, a company about to mark the first anniversary of its split-up. Hewlett-Packard Enterprise is offering a range of patents, all designed to let a company use HP technology to serve business users.
HPE Patent Sale – Mission Critical Computing Portfolio
46 issued patents (41 US, 2 JP, 2 GB, and 1 FR) relating to servers and storage products for Mission Critical Computing (MCC). Key applicable areas include High Availability, High Reliability, Replication/Failover, SSD/HDD, System Management.
Except for that SSD element, everything in the portfolio could fall into the realm of HP 3000 and MPE technology. If only such a marketplace existed 15 years ago. More importantly, if only HP was actively licensing its IP back then. Something could have been worked out. Today, at least there's a mechanism for listing patents for sale and finding interested buyers.By the time the 3000 pullout at HP started to set in stone during 2002, the licensing discussion was being steered toward emulator-ready user licenses. Meanwhile, the source code licenses were more than seven years away from being a reality.
For many customers, the only issue that matters is the licensing of MPE to be used on an emulator that mimics HP 3000s running on Intel hardware. Rumors continued to abound that HP is proposing the destruction of a licensed HP 3000 for every MPE license for an emulator that a customer wanted to deploy. The 3000’s longest advocates are howling over that possibility.
“It’s clear that HP’s intention is to limit the MPE user community to the number of licenses at the time of its death,” said Wirt Atmar of AICS Research, “thereafter drawing the population constantly downwards over the years. This is completely opposite to the future that I believe is possible, where MPE would be distributed world-wide, at prices comparable to Linux distributions.”
Emulator licensing makes the Stromasys Charon product a realistic choice for companies keeping their HP 3000s in use, either as permanent archive machines or in production. But the software communnity was reaching for so much more in the months following HP's 2001 notice of a pullout. Perhaps those MPE/iX bits, as well as the related subsystems, are available today.
The only companies who've even seen MPE/iX source, to use as a support and development reference, are the eight licensees of the code. They don't get to sell it or release modified MPE/iX. But it's there as a resource for support providers like Pivital Solutions and software companies as well.
Any dreams of restarting an MPE/iX business would have to begin with IP licensed for commercial use.
Stranger things have happened. The HP WebOS software developed for the Palm smartphones, as well as the heart of ill-fated HP tablet, found its way into LG Smart TVs, after all. At one point, LG was planning a range of consumer devices that would use the HP IP which the vendor sold to LG. LG never got WebOS into its ThinQ line of refrigerators. MPE/iX, should it ever go into the market, might become an item of the IAM Market.
October 19, 2016
Come together to conference with CAMUS
Admit it. It's been a long time since you talked person to person about your HP 3000 with somebody outside your company. User conferences and one-day meetings for 3000 folk used to be as common as leaf piles in October. That's what happens when you live a long time. You can outlive your community and lose touch.
CAMUS, the Computer Aided Manufacturing User Society, has a way to reconnect. At 11 AM Central Time on Thursday, Nov. 10, the Annual User Group meeting of the organization will form around a conference call. Terri Glendon Lanza of CAMUS is organizing the call. It's free.
The agenda, shared by CAMUS member Ed Stein of MagicAire, is 10 minutes of CAMUS announcements, followed by general discussion with the Board of Directors and everyone on the call. It's manufacturing managers who make up CAMUS, but you might have questions about a certain emulator that earned its stripes in the Digital market before arriving to emulate HP's 3000 systems. Both Digital and MPE managers will be at this conference.
Or you may be interested in the new ERP replacement for MANMAN, Kenandy. Experts from the Support Group -- which is installing Kenandy at Disston Tools this year -- will be on the call. You might just want to know something about MPE management that could take only a minute to answer.
Send an email to Terri at email@example.com, or call her at 630.212.4314, to get your conference call-in phone number. The call runs until 12:30 Central Time. You might learn something, or get to show what you know.
October 17, 2016
Rebuilding Systems Faster and Better
I'm looking at how to save as much time as possible in rebuilding an HP 3000's software and directories. My options seem to be using STORE, versus the sysgen tape command "tape store=@.@.@". What's the best way to go here?
Donna Hofmeister of Allegro replies
Unless your system is small (like a 918 with 8-12GB of disc), you don't want to try to do a full backup via sysgen. If you really do a full backup then I prefer this syntax “store /;...” as it is self-documenting and you know that the Posix files will be backed up as well. (On older releases of MPE, @.@.@ did not back up Posix files <eek>)
You want to make sure that you run 'buldacct' periodically (and routinely). You also want to make sure that you are somehow backing up your directory (store /;*t;directory, for example). Between the two, you have belts and suspenders (for recovering your accounting structure).
On older releases of MPE, you want to make sure that the network is shut down prior to making your SLT tape. And it's still a good idea to have the system quiesced when making an SLT, since everything in the sys account (and .pub.sys in particular) will be locked while the tape is being made. Nothing quite like grumpy users to make your day.
Just as a matter of preference, you should normally do a BULDACCT @ at the beginning of the weekly full backup, then the DIRECTORY option along with ;ONVS=MPEXL_SYSTEM_VOLUME_SET, PRODUCTION_SET, etc as a belt and suspenders approach for that day we all hope never comes.
Mark Ranft of Pro 3K outlined the use of sysgen
Here are instructions for a complete backup of basic MPE/iX system via SYSGEN's TAPE option. It is best if everything fits on a single DDS (or DDS-2 or DDS-3 or DDS-4) tape cartridge, but it will ask for a second (additional) tape(s) as needed.
Note: I included a 'second_volume_set' which can be changed or removed
Note 2: The line below is >80 characters, so you have to know how to create a file so that this does not wrap (or make other adjustments.)
Step One - Create an indirect file containing the following as a single line...
Step Two - Create this job
:tellop END OF JSLTALL -------------- EOJ
Feel free to add a BULDACCT to this.
October 10, 2016
Duke diners deliver some wayback news
It's always a great event—since it's so rare now — to see 3000 folk gather in person. Last week an invite over 3000-L and other channels requested the pleasure of the company of anyone in the Bay Area who remembers — or works with — MPE and HP 3000s. The number of lunchtime diners at The Duke of Edinburgh pub was at the intimate level, which is not a surprise. What was interesting was how informed some attendees were.
"Some were finding out about the [Stromasys] emulator," Stan Sieler reported. He was among the few who were still working on MPE tasks. I was surprised that the news of the emulator was just arriving in October 2016, five years after the product's debut in the Bay Area.
In the fall of 2011, about 80 HP 3000 folk gathered at the last HP3000 Reunion. (I won't say final, because reunions tend to hold on until organizers and the ardent alumni lose the ability to travel, drive, and have meals together. We're not young, us 3000 folk, but we're spry.) The story of the Charon HPA product has orbited the MPE solar system for many months. Not everybody looks up at the sky to see the stars, of course.
Those getting wayback news about Charon included one who needed a free hobbyist license. That kind of license went off the market at the end of 2014, when Stromasys transitioned to an all-proof of concept licensing and sales plan. The PoC strategy has yielded a string of green-lit transitions to the non-3000 hardware. Hobbyist/freeware licenses got abused; free software was caught running in commercial settings. Other people might have failed at their no-cost DIY approach. You don't always get news of failures when you never knew about the attempts.
News travels slowly, especially for managers who are not in everyday contact with MPE and 3000s anymore. Sometimes 3000 news has traveled slowly for reasons other than simple oversight, or becoming busy with non-3000 computing.It's been embarrassing to see, year after year, that the events we publish as news just don't stick with all people who rely on 3000s. You can't get everybody in the loop, not on anything. (Okay, all people have heard of Donald and Hillary, if they've heard of the United States.) HP discovered these gaps in the news loop when it started to spread the word that the 3000 was finished. The company that created MPE called the end of its 3000 business an end of life announcement. Almost a decade after that first finale, the computer isn't finished. The HP we knew, that's finished.
The news of an end of life sometimes flows slowly, or too fast like in that End of Life fable above. This afternoon I heard someone register surprise that Arnold Palmer was dead. It's not just a popular soft drink; Arnold was among the greatest golfers ever. Late last night, Abby was pretty sure the actor Rip Torn wasn't alive anymore. Not yet true, and Not Dead Yet could be a regular magazine, sort of a more gruesome What Are They Doing Now?
When HP's news about the end of its 3000 business rolled out -- trumpeted by at least five publications, covered in ABC News, circulated in serious correspondence to the customers paying for HP support -- it took years to become universally-known. That 2001 announcement was still news to customers more than five years later. For other managers who knew, they kept it to themselves, since the computer was still working. What's surprising is that good 3000 news, like Charon HPA, hasn't arrived yet in a few quarters.
In a way, though, that's good news for Stromasys. "They're coming out of the closets," said product manager Doug Smith earlier this year about finding 3000 users who were on nobody's radar. When you're selling a solution that keeps good software in place, it has a better chance of surfacing than reports of the death of a close ally like HP's 3000s.
October 05, 2016
How to fail at mission-critical IT with 3000s
We recently tried to be helpful for a 3000 manager who was desperate to get an MPE/iX server back online in steady, reliable service. Our role was just to feed questions to the volunteer force of experts on 3000-L and then pass back answers to the manager. The experience led us to think about what any company should do to fail at using a 3000 for mission-critical service.
Be assured, following these helpful hints will ensure your 3000 cannot do its work.
- Do your support with someone who'll just help out from time to time. Save your support budget for your other servers that are mission-critical. Let the 3000 fend off errors with volunteer help.
- Let your inventory of spares of the 3000's moving parts take care of themselves. A power supply or a hot-spare CPU board takes up a lot of room; set aside space for more modern computer components. Someone will be able to find something soon enough when trouble comes up.
- When a software or network problem starts to occur, give the situation awhile to work itself out for a few months. Save your support budget for the time when things are crashing because they've gotten serious.
- When your support vendor bills you on your 3000, let that expense take the same place as less-critical services. This isn't a vaccine, after all. It's just support for mission-critical servers.
- Make it clear to your management you're saving money by using the 3000 in a mission-critical role. Reinforce the cost-effective nature of the use of MPE/iX by keeping the software on 15-year-old HP hardware.
- If No. 6 might raise attention you're using MPE/iX, keep the age and support matters internal to datacenter planning. A 9x9 with no support provider is a fine way to ensure the future.
If a 3000's performance is crucial to staying in business, then not doing any of Nos. 1-6 will be a better course for the health of your company as well as your career opportunities. Keeping the investment of MPE/iX applications in top shape is not simple these days. But not having a support expert on the budget, with a monthly retainer, makes success difficult to impossible. The factor of luck shouldn't be a part of your operational formula.
If a 3000 is crucial to a company's success, then a cost as small as $10 a day shouldn't be out of budget. Some 3000 support companies are glad to stay available these days for a nominal fee.
October 03, 2016
Emulation customers got all they wanted
Five years ago this week Stromasys was doing a full technical detail demonstration of its PA-RISC emulation software. Since then, such virtualization has become an everyday choice for interim homesteading (just a few years of use needed) or long-term plans, too.
The software got its debut in front of a sophisticated crowd: HP 3000 veterans at that year's HP 3000 Reunion. In 2011 skeptics were schooled and devotees bowled over.
The rap on emulator choices from out of the past was performance. That's gone away by now, because moving an environment to a quick-growing OS like Ubuntu Linux -- the foundation for the emulator -- gives MPE an accelerating train of processor improvements to leap onto. Itanium won't leap like Intel's Xeon chips will over the year to come with Skylake. Here's a surprise nobody saw coming: the ultimate Itanium chip, Kittson, began development in 2011, and it's still not running in HP's servers. To think, MPE/iX could've had that fate if HP had chosen to port the OS to that chipset.
HP 3000 hardware and MPE experts at the Reunion believed in Charon's emulation future. In 2011 there were more in attendance at the Reunion than could fit in a single-family home. What's still in the years to come is making a home for MANMAN on one Ubuntu-Charon partition of a big Skylake Intel server, and MANMAN's replacement Kenandy on another.Terry Floyd, founder of the Support Group manufacturing and 3000 support firm, posted glowing comments five years ago about the future of Charon in a CAMUS.org report. What he's foretold has come to pass.
It was amazing to learn that within a year, MANMAN (and everything else that runs on MPE/iX 7.5) will be running on Intel/AMD 64-bit machines. MPE Virtualization: what a home run! Dr. Robert Boers, who came all the way from Switzerland to give his speech at the Reunion, showed MPE/iX running on a small Linux PC costing about $600, and MPE/iX is expected to run many times faster than on an HP 3000 A-Class machine. They also had it running on Craig Lalley’s laptop in the same room; he’s been consulting on this project, but now it’s open to any developer with a good reason to download it.
It was non-obvious to me that MPE would need to boot up in 2 or 3 minutes, mainly because all the memory, IO, and disc checking had been done by the underlying OS (Ubuntu Linux in this case), but also because of the PDC rewrite they must have done. No more watching all the dots and 1s, 2s, and 3s etc. going by on the console for 10 or 20 minutes (or longer on large-memory HP 3000 machines).
Later, in a more technical briefing at the Reunion's hotel, Floyd noted that all the right answers flowed from Boers.
It was like Christmas and Boers was Santa Claus (there is a slight resemblance). MPE booted on both the laptop and the little Stromasys server Dr. Boers carried under his arm off his flight from Europe. Fun was had; DEBUG was run; Glance worked in Block Mode! Stan Sieler asked if MPE crashed in all known ways.
September 30, 2016
Earliest birds to eye Charon stick with 3000s
One week ago the 3000 Simulator Project rolled out a new version of software to simulate an MPE V Classic 3000. That news led to a look at the modern emulation product Charon HPA and what has helped make it a success. Diligent engineering and testing of the Stromasys product across the community started just about five years ago. One of the earliest vendors to green-light their software for emulation was a company who's still selling new customers on MPE software: Minisoft.
Founder Doug Greenup called last month to report on some new sales into your market, the one which established his company. He mentioned Minisoft's connection to See's Candies' HP 3000s. See's is using Minisoft's middleware, and the connection between emulation and Minisoft popped up when I found Greenup's earliest report on testing against Charon. Minisoft was the first third party company to announce their products were Charon-ready, including ODBC, JDBC, and OLE DB products. These were the days when PA-RISC emulation was as new as Clarence Birdseye's frozen food was in the 1940s. Greenup's report was so early in the Charon HPA lifespan that the Stromasys software was being helped into the market by independent consultants like Craig Lalley.
Craig [Lalley] gave us access to the Stromasys emulator to test some of our legacy MPE products. The HP 3000 terminal emulators under Windows and Macintosh worked fine connecting up via Telnet. We ran some VPLUS screens with no problems. Connections were reliable and fast. We also tested our middleware drivers, connecting and running queries.
The bottom line is our products worked like they were interacting with an HP 3000. So if any of our customers deploy Stromasys, we are confident our MPE products will work.
Charon HPA needed software vendors who were familiar to the 3000 community to step up and certify. It's satisfying to see that one of the earliest adopters of your market's emulator is still selling software to MPE/iX sites. We'd call those sites 3000 customers, but its possible the HP hardware has been replaced by Charon HPA. Which is precisely why it was good business to step up and demonstrate that the emulator worked just like an HP 3000. Works better, now that HPA is not five years older like those boxes with "HP" on the front.
There's your report. MPE/iX still running at high-profile candy manufacturer. New 3000 software still being sold in a few places. Stromasys now moving toward five years of support from the MPE third party vendors, support that started with Minisoft.
September 28, 2016
Meeting at Building D: the rarest 3000 link-up
Notices were posted this week on the 3000-L mailing list about a rare meeting next Monday, Oct. 3. At opening time 11:30, people who know and remember the 3000 will gather at The Duke of Edinburgh pub. It's a site popular enough with the MPE crowd that it's still called Building D by some seasoned community members. The Duke is on Wolfe Road, just to the west of where the 3000 grew up. As the 3000 group intends to arrive at opening time, it might be able to commandeer the snug (above).
In-person meetings for the 3000 community happen in bars and pubs by now. The last one we heard about this public was SIG-BAR's meeting in London in 2014. Dave Wiseman, a vendor and software maven whose history includes a software project called Millware for 3000s, set up SIG-BAR. The 2014 meeting was announced so far in advance that people were able to plan their summer vacations around a gathering at Dirty Dick's. There's something about English pubs that attracts the 3000 crowd.
The Duke of Edinburgh is within walking distance of a mecca of the 3000 world, now departed: The HP Cupertino campus. Building 48 has been replaced by the rising concrete and steel of the new Apple world headquarters building. There's no word yet if the 3000 friends who meet Monday at Building D will bring their drones to take their tour of the Apple-ized HP campus.
A walk through the HP parking lot and across a cozy margin of poplars used to bring you to the Duke. "It's right across the street from where MPE lived," said Stan Sieler of Allegro while announcing the meeting. As of Monday, MPE's heart will be among the taps and chips of The Duke. Two years ago, Robelle's Bob Green said this about the last in-person meeting at that London pub:
We exchanged notes on the current state of the machine—especially the new emulator—- and discovered what each of us was doing. An amazing number of people are still doing the same thing: helping customers with their IT concerns. But in reality, most of the time was spent swapping war stories from the past, which was great fun.
As for that emulator, Charon HPA is in full swing by now, a certainty of life going forward with MPE/iX systems. For one additional lunchtime, a pub will be emulating the home of the system, even as it continues to move into a virtual existence.
September 26, 2016
3000-L connects again after a silence
As if on cue after our report about its silence, the 3000-L mailing list and newsgroup sprang back to life over the weekend. One problem solved by the 415 readers was how to identify if a store to disk backup is a LZW compressed backup file. A Tracy Johnson report also confirmed that a LISTF,2 can report the time of each LISTF, by writing a specialized job.
Meanwhile, a 15-year-old HP 3000 with network connection troubles got advice from the newsgroup's readers. A Series 969 running MPE/iX 6.0 would not be the first thing you'd choose for interfacing to an internal website. But when a 3000 has data that a user needs over the Web, the server is the place to go.
Trouble started to surface when clients access a webpage which then opens a telnet session with the 3000, grabs the info, and then returns the data to the webpage.
We’ve been getting more and more errors over the last year, culminating in non-stop Could not initialize data in path with TCP, which then blocked anyone accessing us through our webserver. We’ve tried many changes but cannot seem to get past this.
When it locks up, the HP 3000 keeps running but won’t accept any new sessions. Which means our clients can’t run searches. Which is very bad for us. Sometimes we can stopnet and startnet and it will work for a while, but then the errors start again. Eventually, we have to coolstart to be able to have clients log in.
Mark Ranft suggested "If they are already running in the CS queue, here is the likely cause. Is there some new monitoring in place? If so, it may not be behaving well on the network.
What happens is someone uses a telnet or ftp connection to monitor whether the network is up on the 3000. They send the SYN, the hp3000 answers with the SYN ACK, and then the 3000 receives a reset before the connection is complete. This handshake sequence causes this exact error.
Also inetd and other HP3000 networking improved greatly in later versions of MPE/iX. If they must stay on 6.0, they should at least be sure to install the latest patches.
A third party support company served this troubled user until the support vendor folded, "and the only options we found weren't affordable." Getting the 3000 back up will trigger a revisit of that situation. Any server with critical customer data on it—and doesn't have a support vendor—relies on the largesse of the 3000 volunteers of 3000-L. That mailing list did go without a new message for more than a month, a troublesome response time for a critical server.
September 23, 2016
Simulator for Classic 3000 gets third release
A third release of an HP 3000 Series III simulator is now available from the Computer History Simulation Project website. J. David Bryan of the project reports the software which simulates the old MPE V HP 3000 Series III now has a cold dump facility.
Entering the DUMP command simulates pressing the ENABLE and DUMP front panel buttons. The contents of main memory are written to an attached magnetic tape in a format suitable for analyzing with the DPAN4 program provided with MPE. The new SET CPU DUMPDEV and SET CPU DUMPCTL options specify the default device number and control byte for the dump.
Known as the SIMH project, the software is aimed at hobbyists who are using MPE V programs and utilities. Even though a power failure is not a desired event, the simulator has a capability of creating one. This is in addition to yanking the plug out of the laptop or PC running the simulator software.
"The user may simulate a system power loss with the POWER FAIL command and resume powered operation with the POWER RESTORE command," project notes from Bryan state. "The SET CPU ARS/NOARS command determines whether or not MPE automatically restarts when power is restored."
A full set of new features is listed in the release notes that accompany the simulator source files. Aupdated HP User's Guide covers the new commands is provided in Microsoft Word format with the source download. The guide is also available as a PDF file at an MIT website. A preconfigured MPE-V/R disc image available from Bitsavers was not changed for this release.
September 21, 2016
Power outage, or no problems? It's been quiet on the 3000-L. "Yeah, too quiet."
In the classic war movies, or a good western with Indian battles, there's the moment when someone notices the silence on the field. "It's quiet out there, Sarge," says the more innocent hero. "Yeah, too quiet," the non-com replies. That kind of quiet might be the sound we're hearing from the 3000-L mailing list today.
It's been five weeks without a new message on the mailing list and newsgroup devoted to MPE and its servers. Advice and solutions has flowed for two decades and more off a mailing list that still has 498 members subscribed. The number of subscribers has remained steady over the last three years. Like the number of migrations in the market, the exit from the list has slowed to a trickle. So has new traffic, of late.
The silence may not be ominous. In 2016 the 3000-L is used almost exclusively to resolve MPE/iX problems. The hardware posts are limited to the rare announcement of used server prices, messages that the members still howl at if they don't include <PLUG> in the subject. The server hasn't been sold by HP in more than a decade, but its owners still don't like to be bugged by sales messages. They solve problems in a grassroots manner. As a notable ballplayer once said, you can look it up. There might be no problems to solve.
However, no messages at all over 35 days sets a new record for the 3000-L quiet. This 3000 resource was much more lively a decade ago. And 20 years back? Well, HP was still selling enough 3000s in the fall of 1996 to be sending its new marketing manager Kathy Fitzgerald to speak at an Indiana RUG meeting about the new servers. There was also advice on storage compression, because compression-enabled DDS drives were becoming more common.
Good advice: If you can find a DDS tape drive from 1996, you should take it out of service. Your MPE server, no. And evergreen advice from the L is still available online. Jeff Kell, the deceased 3000 guru who started the server on a University of Tennessee at Chattanooga server, built it to last.In 1996 people wrote on the L that they understood most drive and backup software vendors recommend against using both hardware and software compression. "You should try each separately and use the technique that achieves the best result consistently," we reported. Mark Klein of ORBiT Software gave an interesting explanation of what was going on and the compression possibilities.
"Actually, there are cases where multiple levels of compression are useful. But first let me describe the various types of compression available.
"Hardware compression is typically LZW or some variant thereof. This type of compression uses a dictionary of repeating strings that can be dynamically determined on the fly and, as such, doesn't require the dictionary to be stored with the data as with other types of compression. There are other types of hardware compression available, but LZW is the most common found on compressing tape drives. LZW can also be done in software.
"Since the compressibility of the data really depends on the data itself, there are instances where negative compression will be achieved as well as instances where very large files can compress down to almost nothing. In fact, I've seen an instance where a large, multi-Gb database that was mostly empty got compressed into less than 32K using LZW.
"LZW is not effective in trying to compress something already compressed with LZW. This can result in negative compression (the resulting data actually gets larger). For that reason, I wouldn't recommend using LZW software compression on top of LZW hardware compression.
"Another type of compression is called run length compression. This is in essence a combination of a length tag and a string. The length indicates how many times to repeat the following string. For example, a line of 80 blanks would be represented by (80," ").
"Now, using a combination of RLC and LZW one can achieve better levels of compression than with one or the other method. So, if you want to use software compression with a hardware compressing tape drive, I would recommend using RLC compression in software."