November 25, 2015
3000 community keystone Jeff Kell dies
Jeff Kell, the man who founded the keystone of 3000 help, advice and support that is the 3000-L mailing list, died on Nov. 25 of liver cancer and complications from damage induced by a diabetic coma. He'd battled that illness in hospitals and hospice since 2014. Kell was 57.
"It is a very sad day when a good wizard passes on," said coworker and colleague Richard Gambrell at the University of Tennesee at Chattanoona. "Jeff had a gentle soul and brilliant mind."
Kell was the rare IT professional who could count upon 40 years of experience running HP 3000s, developing for MPE, and especially contributing to the state of the art of networking for the server. He created the ultimate network for the 3000's community by establishing HP3000-L, a LISTSERV mailing list now populated with several hundred thousand messages that trace the business computer's rise, decline, and then revival, rife with enduring high tech value and a thread of humor and humanity.
Kell's obituary notes that he came by his passion for scuba early, having worked for a short time at the Chattanooga Aquarium where he fed the sharks. A key contributor to the development of LISTSERV, Kell was instrumental in UTC’s earning the LISTSERV 25th Anniversary plaque, which lists UTC as the 10th University to deploy LISTSERV.
Kell also served as a volunteer to chair SIG-MPE, SIG-SYSMAN, as well as a 3000 networking SIG, but it's nearly impossible to sum up the range of experience he shared. In the photo at the top of this post, he's switching off the last N-Class system at the university where he worked. Almost 40 years of MPE service flowed off those university 3000s. In the photo above, from the HP3000 Reunion, he's updating attendees on how networking protocols have changed.
In the mid-1980s he was a pioneer in developing Internet Relay Chat, creating a language that made BITNET Relay possible. Relay was the predecessor to IRC. "Jeff was the main force behind RELAY, the Bitnet message and file transfer program," Gambrell said. "It inspired the creation of IRC."
My partner Abby and I are personally indebted to Kell's work, even though we've never owned or managed a 3000. The 3000-L and its rich chest of information was my assurance, as well as insurance, that the fledgling 3000 NewsWire could grow into the world of the 3000. In the postings from that list, I saw a written, living thread of wisdom and advice from experts on "the L," as its readers came to call the mailing list and newsgroup Kell started. Countless stories of ours began as tips from the L, or connections to people posting there who knew mission-critical techniques. At one point we hired columnists to summarize the best of each month's L discussions in net.digest. In the era where the Internet and the Web rose up, Kell was a beacon for people who needed help at digital speed.
He was a humble and soft-spoken man, with a wry sense of humor, but showed passion while defending the value of technical knowledge -- especially details on a product better-loved by its users than the management at its vendor. Kell would say that all he did was set up another Listserver on a university computer, one devoted to becoming crucial to UTC's success. Chattanooga is one of the best-networked towns of its size in the world. Kell did much more than that for his community, tending to the work that helped the L blossom in the 3000's renaissance.
Kell looked forward to an HP which would value the 3000 as much as the HP 9000. In 1997 he kicked off a meeting with HP to promote a campaign called Proposition 3000: Common hardware across both HP 3000s and HP 9000s, sold from an Open Systems Division, with MPE/iX or HP-UX as an option, both with robust APIs to make ISV porting of applications to MPE/iX "as trivial as any other Unix platform."
HP should be stressing the strengths of MPE/iX, "and not its weaknesses," he said. "We don't have to be told anymore what the 3000 can't do, because a lot of the things we were told it can't do, it now can. If we take the limitations of the Posix shell and remove them, we have Proposition 3000," Kell said to HP managers. "I would encourage you to vote yes for this investment in the future."
More than 16 years later, when MPE's fate had been left to experts outside of HP's labs, Kell offered one solution on how to keep the server running beyond MPE's Jan 1, 2028 rollover dating gateway.
"Well, by 2027, we may be used to employing mm/dd/yy with a 27 on the end, and you could always go back to 1927. And the programs that only did two-digit years would be all set. Did you convert all of 'em for Y2K? Did you keep the old source?" Kell's listserver is the keeper of all 3000 lore, history, and wisdom, a database that can be searched from a Web interface -- even though he started the resource before commonplace use of what we were calling the World Wide Web.
Some might dismiss that resource as a museum of old tech. Others were using it this week, to connect newer-age tape devices to old-school 3000s. He retired the last of UTC's 3000 at the end of 2013 (in the photo above). His own help to the community members on tech specifics and the state of this year's networking will outlive him, thanks to his work setting this keystone for the community's exchange.
Pivital Solutions: Your complete
HP e3000 resource
November 24, 2015
The Wide World of Connecting Storage
IO used to be more complex for IT. Sure, the array of choices for disk is vast today. But in the era when 3000s used to think they were lucky to get SCSI plugged into them, configuring disk connections was not simple. HP-IB protocol, built to link HP's instruments, was simple, used for all HP devices, and slow. But it was integrated and seamless compared to the SCSI of single-ended, fast/wide, and Ultra Fast.
Such was the case for one 3000 manager seeking advice from his colleagues. You never think about these things on a 3000 until the hardware breaks. Or backups fail. Or storage media gets rare. Aging hardware is one of several issues that require expertise, even if a 3000 runs the ultimate 7.5 version of MPE/iX. Our manager hunted for his help on the longest-running 3000 classroom in the world, the HP3000-L mailing list.
A single-CPU A-Class was moving away from DDS technology, the DDS-3 that was first launched in the '90s. There are other options for 3000 tape backup. But these options include single-ended, fast/wide, and other cable and termination combinations. DLT technology, introduced more recently but still a 1990s choice, runs with HP 3000s. It helps to get the ends right, though, if DLT is to have a new beginning on an old-school 3000.
"Until now they have done their backup on DDS," a manager talking to the 3000 newsgroup explained. "Lately they had a failure on the DDS drive, and have realized that it is getting difficult to get new tapes. They have decided to move to DLT8000, model C6378A, and have bought two of them. One is supposed to go live on the 3000, and the other to be stored as a spare device."
The DLT is hooked to the Ultra Wide SCSI interface on the A-Class. But ODE/Mapper doesn't recognize the device."
There was an error, and no DLT joy. Soon enough, one veteran consultant said, "You will have trouble connecting a fast wide SCSI device to an ultra-wide SCSI controller." It wasn't a rookie mistake, but the veterans who still prowl 3000-L had a solution and even a link to an inexpensive fix. So it goes, here in the fifth decade of HP 3000 mission-critical service. Answers are everywhere.
November 23, 2015
Virtualized clouds may shift due to Dell
Although the merger isn't yet complete, EMC will become part of Dell in the year to come. Those are two impact players in the HP enterprise arena, fierce HP rivals as well as providers of gear in HP shops both migrated and homesteading. The biggest impact on HP 3000 customers might come not from either of these companies, though, but from a subsidiary. VMware, which is powering a significant number of virtualized environments, is 80 percent owned by EMC.
That makes Dell the primary owner of the most popular virtualization provider in the industry. In the wake of the merger announcement, consultants, developers and vendors from the community have looked to the future of Dell's VMware ownership. Even a possible impact on cloud computing has come up for discussion.
"Whoever owns VMware next could control and own the future of the cloud," goes the proposition of the new VMware ownership. VMware has certainly promoted its new efforts into cloud computing. But that doesn't make the vendor a controlling force in cloud computing.
The three pillars of cloud computing, according to cloud ERP provider Plex, are elasticity, efficiency, and cloud as a service. VMware is only a backbone for such cloud offerings. The actual cloud applications use a range of backbones. The most common one is Xen, used by Amazon Web Services. HP dropped out of the public cloud business earlier this year, facing losses while going up against Amazon and others.
However, corporate enterprise IT may include clouds on VMware. A VMware-based one might run on an internal security zone not visible to the Internet. Another style can be based on OpenStack, visible to the Internet.
"Dell owning 80 percent of VMware is a huge deal," says Gavin Scott, a developer and a veteran of decades on MPE/iX and former SIG-Java chairman. "But it's not because of clouds. It might actually be bad for VMware because it will push Dell's competitors to look at other solutions. VMware is crazy expensive, so customers may be quite happy to be led to other vendors' doorsteps."
"VMware is like Oracle," Scott told us. "The most expensive way to solve the problem. But it also has the most features and functionality and is a 'safe' choice."
November 20, 2015
Multi-threading traces years of MPE service
Yesterday we explored the prospects of multi-threading for HP 3000 sites. It's an aspect of application and software design that can benefit from virtualization. In years past, when much of the 3000 application base was being created, separate hardware CPUs drove this multi-threading. Stan Sieler of Allegro, one of the authors of the textbook on Precision Architecture RISC "Beyond RISC," told us that multi-threading is likely to have made its way into 3000 software via Unix.
It's a concept, through, that's been possible for MPE ever since its beginning. The MP in MPE stands for Multiprogramming, Sieler reminded me, and that "Multi-threading is a form of multiprogramming or multiprocessing."
Generally, but not always (as words are often abused), “threads” are related to a single process. E.g., my video compression program might work on several parts of the video simultaneously with three or four threads. On some computers, two separate threads of a single process cannot execute at the same time … on others, they can.
On most computers nowadays, threads are implemented at the operating system level. On older systems, threading was sometimes implemented above the operating system, relying on user code to switch threads. (I’ll skip co-routines, which few systems have now, but the Burroughs MCP did.)
Multi-programming is the concept where two (or more) processes (or “programs”) appear to run at the same time, but in reality each gets a short time to run, and then the CPU pays attention to the other process, then back to the first one… or “time slicing.”
On the 3000, few programs use multi-threading, but it is available. It came about the same time as Posix did, perhaps one release later (I can’t recall). In general, if you show me a 3000 program that uses threading, I’ll bet it’s written in C and originated in the Unix/Linux world.
Essentially all computers nowdays have multi-programming. The original HP 3000 (pre-CX) did, too. (The HP 2100 (running RTE) had, IIRC, no multi-programming.)
November 19, 2015
TBT: HP rides into the cloud first on 3000s
In the month of November 17 years ago, Hewlett-Packard drove itself into cloud computing with HP 3000s. It wasn't called cloud computing in 1998. Resolving Y2K was still more than a year away. It was a year with a healthy dose of blue skies for the computer, including the lab manager's plan to put MPE/iX on the company's favored IA-64 chips.
However, HP was positioning the 3000 as a solution for a world that wasn't purchasing as many servers as before. It was a situation much like what HP faces today. New 3000 sales were tough to come by, just like Integrity sales of today. Thanks to HP's efforts, customers were moving off 3000s in favor of Unix and Windows and NT. Today they're all moving away from servers of all kinds, leaving the hardware to offsite management and administration. The Cloud.
The 3000's entry to cloud computing arrived in the form of an acquisition. The 3000 division bought Open Skies, a 38-person software firm which had airlines for clients. Not many major clients for the time. Westjet. Ryan Air. But these were lean airlines that wanted to track miles flown and customers served without developing and maintaining a software application. HP had called the concept Apps on Tap earlier in the year. The 3000's CSY division bought Open Skies to show the way, creating an application that could be tapped.
Roy Breslawski made a shift away from CSY marketing manager to Open Skies marketing manager. Breslawski, like his GM Harry Sterling, took the MPE mission seriously enough to disregard the accepted wisdom about the 3000. (Legacy platform. Fading fast. Jobs there a stepping stone.) Instead, Breslawski set up business with an earnest belief about the product's growth prospects.
The Open Skies deal was sparked by the needs of a much bigger airline, though. British Air was tired of being undercut by smaller operators like Ryan Air and EasyJet, so BA set up Go, a low-end carrier. Go wanted Open Skies to host and manage the HP 3000s handling their reservations.
Those systems came to be owned by HP and configured in a separate datacenter. That commitment led Open Skies to ask HP for help in meeting manpower arrangements, which developed into discussions about HP taking over the growing company.
November 18, 2015
Application threading a gate for performance
Many an HP 3000 app was designed in an era when threads were expensive. Multi-threading is another way of describing multiprocessing. It's the M in MPE. But few HP 3000 programs use multi-threading. Multi-processing uses multiple processors. These 2-way and greater 3000s could cost upwards of $200,000 over the last complete decade of sales in the 1990s. Since this was the MPE/iX value model, the cost reflected the combo of hardware and system software, during an era with user-count licenses for the OS driving up the capital cost of 3000 computing.
For any customer who had but one CPU propeller to push along their ship of software, a single-threaded app made good sense. But the single threading programs of MPE/iX are a gating device for engaging the full horsepower of virtualization. Dave Clements of Stromasys mentioned the common threading architecture for MPE/iX apps while we talked about VMware's connection with the Charon product. This is a common reason why every 3000 customer's Charon performance is one of those "it depends" solutions.
A user of Charon can sometimes get along with a relatively slow CPU clock speed for the Charon host hardware. At the Conax Technologies datacenter, a 2.7GHz Intel host is standing in for a Series 928 HP 3000. Virtualized CPU power is almost as fast as the original hardware there, according to the system manager — and then any application process that reaches out to the disk screams along, the manager added. But there's not a lot of multi-threading in the 3000 app world.
"We run into a lot of applications that are not multi-threaded," Clements said. "It makes a difference. We see that a lot in database applications. There's not a lot we can do about single-threaded applications," he added, in order to take advantage of the multi-threading abilities in newer and faster host CPUs. What makes Charon an effective emulator is, in part, its ability to excellerate multiple threading of processes. It's the same kind of lift as if the newer Intel chip designs were to give power upgrades to the PA-RISC CPUs. This is the promise of virtualization. Multithreaded apps get more from it.
Stromasys customers and prospects have not been reporting that speed is a barrier to their adoption of the product. Charon has the potential to run 3000 programs even faster if those apps have been written to use multiple threads. "Every customer poses the potential for a unique solution," Clements said. Other aspects can be changed, he said — things that are easier to update than application code which was probably first conceived before the Web was born.
November 17, 2015
Putting PDF Into a 3000's Data Flow
HP 3000 experts know of a wide array of techniques to create PDF files from the server's data, then move them via FTP to a Windows server. While the simplest answer for getting reports into PDF format and then out via Windows is probably Hillary Software's byRequest, there are other commercial solutions. There's also several bolt-together techniques if you've got very limited budget to homestead.
Bob McGregor reports:
We use txt2pdfPRO by Sanface. We had a job that would run and check a pseudo device for spoolfile output, and if the pri > 0, would run the sf2html process, convert to PDF and then FTP to a Windows server. The process would then delete spoolfiles=0 on the pseudo device the next day. Setup took a bit... but once done, worked well.
Lars Appel, author of the Samba/iX file sharing tool, adds:
I wonder if it might make sense to configure a "dummy" network printer on MPE/iX and have it send spooler output to a little socket listener on the Windows system (similar to the FakeLP example from the 3000-L archive) and then invoke GhostPCL on the Windows side for generating the PDF output.
The "dummy" network printer would let the MPE spooler take care of the PCL conversion and also perform the "file transfer" automagically. The GhostPCL software is probably easier to get (or build / update) on Windows than on MPE (okay, I admit that it did also build on MPE long ago.)
November 16, 2015
Webinars set courses for future operations
The next three days each contain a webinar that can help a 3000 manager decide how to best use their IT resources. One of the presentations covers a new cloud-based ERP migration solution, explained in detail, while the other two come from a long-time provider of data solutions for HP 3000s.
On Nov. 17 (Tuesday) Kenandy demos its cloud-based, Salesforce-driven ERP stack. It's a new performance of the overview show broadcast at the end of September. Kenandy has enough features to replace more than a few MPE/iX apps, for any sites which are looking for replacement solutions on the way to migration. Registration is here on the Web, and the program starts at 1 PM Central Time, US.
Over the following two days, MB Foster airs a pair of Q&A, webinar-driven broadcasts about best practices for data management. The company is serving customers beyond MPE/iX sites now, from the experience of carrying out a migration as well as the integration of its software and practices in non-3000 customer sites.
Wednesday Nov 18th's Webinar covers Data Migrations Best Practices. IT operations generate opportunity and challenges to organize data into useable information for the business. The Webinar will deliver practical methodologies to help you prevent costly disruptions and solve challenges. "A data migration project may not be your specialty," says CEO Birket Foster. "We are offering an opportunity to learn from our successes and minimize the business impacts of data migration, through best practices." The Webinar begins at 1 PM Central US, and registration is here on the Web.
Thursday Nov. 19th's Webinar (a 1 PM Central start time; register here) from MB Foster explains the strategy and experience needed to employ Operational Data Stores in a datacenter. An ODS requires integration, Foster says.
"Essentially you’re changing what and why you deliver information, and where that information resides for end-users decision support and reporting," he says. "You would also change ongoing management and operations of the environment."
The meeting will deliver insights into MB Foster’s ODS and DataMart services, its technology, and best practices including:
1. What an Operational Data Store and DataMart are
2. How actionable data can be delivered, quickly
3. Why investing in an ODS and DataMarts are smart choices
November 13, 2015
Quotes On A Happening, 5,111 Days Ago
My career has not changed significantly, but I no longer believe anything HP tells me. They could say the sky is blue, and I'd seek a second opinion. They lied to our face once, I won't give them a chance to do it again. — Terry Simpkins, TE Connectivity
It was very difficult to reinvent and took several years. HP's decision almost killed our company. But we survived and are stronger as a result — Doug Greenup, Minisoft
I had received the news prior to the public announcement. I was very angry with HP after being told by Hewlett-Packard at HP World that there was a long future for the system. — Paul Edwards, Interex director
We felt like we were supporting legacy products already, because most of our MANMAN customers were off of applications software support anyway, so it didn't change our plans much. — Terry Floyd, The Support Group
When I joined the conference call, in which management announced to CSY staff that they were pulling the plug on MPE and the 3000, I remember the date and the hour. My feeling was one of relief that they were going to stop pretending that the 3000 had a future. It might have had a future, but not with the level that management was investing in R&D at the time. — V.N., HP 3000 labs
I remember heaving a big sigh and realizing that, in the aftermath of the Compaq takeover, HP would not keep two proprietary platforms. Between a 71,000-unit installed base (HP 3000) and a 700,000-unit installed base (VMS), the choice was quite obvious. To this day, VMS still exists. — Christian Lheureux, Appic
I was working for a company called Hewlett-Packard at the time. I don't know what's become of them; I think they still sell ink. Last I knew, they sold personal computers too, but they weren't sure about that. — Walter Murray, California Dept. of Corrections and Rehabilitation
This really scared a lot of people at the company where I was working, but I kept telling them we had third party support, and not to worry. The directors decided to leverage our 350-plus programs with a migration to an HP 9000. We secured a used 9000, only to have them reverse their decision and opt instead for a newer 3000. — Connie Sellitto, Cat Fanciers’ Association
We were well into MPE/iX and the Posix environment, and there appeared to be some real solidarity given its Internet capabilities. The 2001 announcement was a knife in the back of our long-term planning, from which we never fully recovered. — Jeff Kell, founder of the 3000-L mailing list
I was working a long-term consulting contract managing HP 3000s and several datacenters for the US government. The job that pays the bills these days has nothing to do with HP 3000s — and thankfully very little to do with HP at all. — Chris Bartram, founding 3000newswire.com webmaster.
Share your memory of the day below. Or email the Newswire.
November 12, 2015
TBT: HP translates brags about fresh e3000
On a November afternoon fifteen years ago, users and vendors met in an Amsterdam conference center to celebrate integration. A handful of companies had melded their HP 3000 applications with the Internet. "All of the users I spoke with were already doing some kind of e-something, whether elementary or quite advanced,” said Adager CEO Rene Woc. One showed off how Java had helped create an interface for a company that was selling parts for power looms. Their customers were all over the world.
The users' presentations were especially notable because they were offered in five languages. Simultaneous translations were paid for by the HP 3000 division, the only time in more than 30 years of conferences I've been able to pick up a wireless headset and hear technical reports translated. Not into everyday C-level language, but into French, Spanish, German, Dutch and English. HP set up two rooms with a total of 10 translators. The vendor was working to encourage 3000 managers to speak the language of the Web. HP collected $365 per attendee to help defray the cost; 90 customers and partners attended from 14 countries.
Users wanted their 3000s to be better connected because they didn't want their systems left behind as IT expansion ramped up. Everyone had escaped Y2K worries by November of 2000. The dot-com boom hadn't gone bust, and in some segments like e-commerce, Web interfaces were bringing genuine innovation for interfaces.
The surge was less certain for companies which had limited their 3000 communications to data swaps over internal LANs. Some were using an intranet, employing the Web technology without exposing the 3000's data to the outside. Others like Lindauer Dornier used the Enhydra Web application server and Java/iX to send the power loom manufacturer's parts data to its customers across the world.
The HP 3000 at the heart of Dornier's operations was plugged in when Windows NT proved too slow. The Windows product that became Windows Server a few years later got dumped in favor of MPE/iX. The meeting "had a lot of flavor of the old days," said HP's Sally Blackwell. The emphasis was not on sponsorships. It was an exchange of information, with HP's help."
HP 3000 Division Product Marketing Manager Loretta Li-Sevilla made the trek from the HP 3000 headquarters, telling customers that “the 3000 is a rock solid foundation for an Internet future. With the 3000 as your platform of choice, that future is unlimited.” There was another 12 months of future remaining with an unlimited flavor.
November 11, 2015
Protecting a 3000 by Eliminating Its Services
Here on this day when we celebrate people who have served in the armed forces, a question emerged about enabling HP 3000 JINETD services. Or disabling them, to make a 3000 more powerful and secure. (Yes, it seems to defy the logic about more services being better, one that we can hear in national defense debates. We didn't have such debates at Signal Corps training for the Second Battalion.) The solution to the 3000 service problem included advice on how to trim back risk as well as performance drains on a 3000.
Grigor Terterian said he was having a Series 979 freeze up, because JINETD was receiving a call "for echo udb." Mark Ranft and Denys Beauchemin said the fastest repair would be to comment out echo in the inetdcnf file. Ranft got specific with an example.
:print inetdcnf.net # Internet server configuration database # #echo stream tcp nowait MANAGER.SYS internal #echo dgram udp nowait MANAGER.SYS internal #daytime stream tcp nowait MANAGER.SYS internal #daytime dgram udp nowait MANAGER.SYS internal #time stream tcp nowait MANAGER.SYS internal #time dgram udp nowait MANAGER.SYS internal #discard stream tcp nowait MANAGER.SYS internal #discard dgram udp nowait MANAGER.SYS internal #chargen stream tcp nowait MANAGER.SYS internal #chargen dgram udp nowait MANAGER.SYS internal telnet stream tcp nowait MANAGER.SYS internal #bootps dgram udp wait MANAGER.SYS /SYS/NET/BOOTPD bootpd #tftp dgram udp wait NET.SYS /SYS/NET/TFTPD tftpd ftp stream tcp nowait MANAGER.SYS /SYS/ARPA/FTPSRVR ftpsrvr
In the example above, only telnet and ftp services are enabled, Ranft said. This led Art Bahrs, a Certified Security Professional, to add that the services you leave on are the ones that can cause trouble, if you don't need them enabled.