VMS vs. U*ix - any horror stories?

[Paul Jerrom]

Hi all,
Nice to have this contact with other like-minded VMS-ites.
I am looking after a site that has multiple environments - different OSs, hardware, programming environments etc. Now, unfortunately, the IT Manager has decided that he wants a unified environment - single or reduced number of OSs, reduced languages, uniform hardware etc. I know, I know, it's all my fault and I should have him better trained, and shouldn't have let him mingle with other IT Managers...
Now it looks like there will be a VMS vs. U*ix shoot-off, with the winner taking all. Or at least keeping a job for a while...
Now, I don't need any help with espousing the benefits of VMS, and I can dish as much FUD about U*ix as I receive about VMS [I've been using VMS 20 years, I was in Digital pre-sales for 5 years, and a VMS Ambassador for 3 yrs]. However, as I am continually winning the arguments wrt performance, availability, security, TCO etc, the battle is moving towards the "well, no one develops for VMS nowadays". And this is where I need some help...
I know about the Singapore Stock Exchange, but can anyone help me with examples of a) customers new to VMS or b) applications being ported to VMS or c) companies developing in-house VMS software or better still d) examples of companies who have moved from VMS elsewhere and then come back again. Especially any examples which have the pleasant side effect of knocking Linux, or which emanate from this side of the world.

Thanks guys,

Paul Jerrom (from sunny New Zealand)

[Martin P.J. Zinser]

Hello Paul,

one good source are the success stories at

http://h71000.www7.hp.com/success-stories.html

Then, do not look far, just posted here very recently, Jan described his really long running cluster here in the forum at

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=550510

Last but not least, the company I am working for is running the worlds largest derivatives exchange as well as a couple of other exchanges very successfully on OpenVMS clusters. As far as I am aware we do not intend to move our core systems away from VMS ;-), actually we did commission a new cluster last year to run the central counter party system for Stock trading in Germany.

Greetings, Martin

[Mike Naime]

Paul:

I am covered under one of the success stories that Martin provided the link for. I think that we would also be an Example "C" because we wrote the Application "In-house".

Our App was written to run on both the AIX and VMS hardware/OS. The thought process being that the company did not want to be stuck with only one hardware vendor. Our clients have a choice on what they want to purchase.

If we host the system in house. It goes on VMS. (this is the success story) If I have to host an AIX system, (A migration from a client site) it currently requires twice the disk space that I would give a VMS system for the AIX "Cluster" because it cannot truly access/share the same volume at the same time. I have heard that this issue is being fixed in future releases of AIX. (It may already be there, I don't know)

WHQ is working with the Nashua NH Performance Lab folks to port our App to Itanium. Reports from 1 year ago were that it booted on Itanium, and promply crashed after 15 minutes! :-)

I suppose that my "new install" clients would be examples of Customers new to VMS.

Mike Naime

[Willem Grooters]

You want a 'horror story':

Years ago, I happened to work at a company that was 'in transition' from VMS to Unix.
I don;t know why - but it's likely to be FUD spread by some company at management level. I don't remember their name - doesn't matter anyway, they don't exist anymore.

During that time, the advantages of VMS became clear. Outage-time ran up due to several reasons, but only on the newly installed Unix machines - where the VAXen silently did their job. Furthermore, they had not anticipated the huge cost of transferring their software to the new platform - altough they should use "standrad software" this had to be heavily modified to meet their (legal!) standards.
All that time, the one system-group I worked at (yes, the VMS one), had time left to do good things on the user side (using VT terminals...), where the Unix 'guru's' had to concentrate on their machines. Failure of some of these introduced massive panic - to get the spare system up-and-running, as soon as possible, to minimize production loss.
Well, this failed several times, causing really HUGE damages - so high, insurence became impossible...

Luckily, I left before the transition was completed. After some years I met some ex-collegues, that updated me on the status. It has never come right again. As stated: the company ceased...

You want a success story? Then read the "Celebration" threads: a cluster that is up and running for 7 years now - and has been COMPLETELY renewed within that time. In that, there is a link to the original posting (at www.openvms.org) as well.
NO Unix site can achive that!!!

[John Eerenberg]

Not really horror stories, but two reality stories. I'll appologize in advance for not providing too many details since my employeer may frown on that.

1) In 2003, moved several VMS and Oracle apps to Unix and Oracle. Under VMS, the app was poorly written and ran slowly. Straight forward to fix but not done since the port to unix was pending.
The port to unix included a rewrite; it is still pretty slow. Costs went up for hardware and labor to support the system. Not as flexible or easy/safe to make changes. Former VMS developers complain about the hoops they go through now on unix. I presume some level of productivity was lost. On the plus side, everyone has better job security (tongue in cheek).

2) In 2004, moved an app from VMS to AIX. Old hardware was a VAX 7800 with 6 CPUS and memory (twice for the redundant system in another city). New hardware is a 12 processor AIX box, unix, and suitable/adequate memory (twice for the redundant system). The port was to include new enhancements. The actual port did not include new functionality. Very costly in terms of both hardware and labor. Performance is better then the old VAX but some performance problems exist (which doesn't make sense when going from CI based MHz VAXen to GHz AIX with new IO subsystems). Hardware costs are estimated to be roughly 2-3x the VAX. System Admin support is a full three times the old VMS staff. Same 3X for developers and application support. This was planned for and understood up front but I never understood why . . . until now.

In both cases the unix staff and developers are very good and skilled people. This appears to be reality in our field of business.

[Jan van den Ende]

Paul,

hope this suits you

(it DID happen some 10 years ago, but I guess the main line would still be valid).

At a previous client (sorry, no name) the computer centre was running some IBM MVS, some BULL UNIX, and a 2-node (8800's) VAX-VMS cluster. During the nearby construction activities the main powerline was dug up (~08:00). It was more-or-less anticipated (previous experience), and an emergency cable was rolled out and in place in some 10 minutes. In the mean time 3 brands of "sys-admins", and the head of the computer centre had all found their way to the computer room.
The head told us "BULL holds the most important apps, when they are through with their inital power surges, they signal IBM to go ahead. VMS 'only' runs the architect's CAD software. Better tell those 40 odd people to find something else to do today that doesn't require the system." Meanwhile power was back on, the bootstrap sequence was already displaying, and by the time our BULL friends started their power-up activities, our console reported database recovery completed, so at about 8:30 every(-VMS-)body was back at work, and we left the computer room telling the head we were fine.
At ~ 16:00 the head came to us:
H: "Well, it may be getting late today, but you can go reboot"
We: "Reboot? Why?"
H: "Because the IBM's are up now, and the Bull guys thinks they won't be troubled if only you are rebooting"
W: "But WHY should we reboot now"
H: "Well, then the architects can work again tomorrow".
W: "But they have been working all day".
H: "How can, I know you crashed as well, and tou left the computerroom when power came back up?"

We were back up and recovered in less the 15 minutes, while the Unix guys took over SEVEN hours to get "back" to a consistent situation (yes, full restore!). And then they were lucky that the only mutations they missed were bach-input while the input was still available and cound be re-run. Normally at about 8:00 all kind of individual transaction mutations start going, so they were very lucky. Had the outage happened some later...

But then, of course, every U*ix promotor will be telling that nowadays they have very good disaster recovery etc.
Try and get a demo of a U*ix system with full applic activity and a simulated hard crash. And if they "of course" have no-break, then simulate a power-supply failure, or pull the cable form (a) disk(s).
Try the same on VMS, and observe the resilience and/or recovery difference.

I sure hope you succeed in winning your shoot-off!!

Jan

[Willem Grooters]

Try and get a demo of a U*ix system with full applic activity and a simulated hard crash. And if they "of course" have no-break, then simulate a power-supply failure, or pull the cable form (a) disk(s).
Try the same on VMS, and observe the resilience and/or recovery difference.


I've seen the demo.
On last year's ENSA@WORK in Amsterdam there has indeed been a demo with combined NAS and SAN storage, "fully redundant" to show the capacity of data-failover for Windows and Unix (that is: HP-Ux, Linux and Solaris) clusters. Access problems were manullay caused by pushing a (big red) button, so one side simply 'broke down'. The operator (= HP representative) said, on my question of automated fail-over, that that wasn't possible: it ALWAYS required operator intervention, it had to do with data replication and checks before control could be handled over to another machine. I must have looked surprised, he looked at me, noticed my (VMS) badge and said: "If you're running VMS, Sir, then I know _you_ won't have to worry, but here we cannot be so sure and need to check first".

[Andreas Fassl]

Paul,

I'm sending this forum thread to Sue Skonetski, she's the editor of the VMS Technical Journal, she has LOTS of success stories and she will help you in your task.

With kind regards

Andreas Fassl

Von "Skonetski, Susan"
Datum Montag, Dezember 29, 2003 4:39 pm
An "Skonetski, Susan"
Betreff OpenVMS: When Continuous Availability Really Matters
Dear Folks,

This is from 2001 but it is a good reference paper. Alan Muir, thanks
for forwarding.

I just stumbled on the white paper entitled "OpenVMS: When Continuous
Availability Really Matters" by accident (while Google searching for
information on OpenVMS support in Veritas) at the ITPapers.com website
(http://www.itpapers.com/abstract.aspx?&sortby=compd&scid=260&docid=4010
4). It was a great read!

Warm Regards,
Sue


Von "Skonetski, Susan"
Datum Dienstag, Dezember 2, 2003 5:28 pm
An "Skonetski, Susan"
Betreff Excellent OpenVMS Pearl - Vienna Firefighters plan a great future with OpenVMS - OK for external distribution
Dear Folks,

If you would like the nice word document of the attached please let me know.

OK for external use per Peter Ranisch.

Thank you very much Peter.

Warm Regards,
Sue


-----Original Message-----
From: peter r. ranisch - OpenVMS V8.0 on Itanium(R) is on the road [mailto:peter.ranisch@chello.at]
Sent: Tuesday, December 02, 2003 9:02 AM
To: Skonetski, Susan
Subject: Vienna Firefighters plan a great future with OpenVMS


Sue,

Attached is a short overview about the histroy and the future of OpenVMS usage at the Vienna firefighters. You can distribute this as an OpenVMS Pearl. Further information about it could be acchieved by contacting Peter Götzl at gop@rts.co.at

Peter

_____________________

Vienna firefighters use OpenVMS for their call and mission center
The mission planning system for Vienna's professional firefighters was put into service in 1988. The system was in charge of the active missions and did the alarming of 24 remote fire stations throughout Vienna. The Hardware consisted of 2 VAX-11/750 servers with 20 terminals (VT220 and VT240). The external sites were connected via terminal servers and serial lines to the headquarters. The operating system was VMS 3.7 (?), ORACLE was used as a database, and programming was done in C. Two PDP-11s running RSX-11M were used as preprocessors for data gathering on the status of the emergency vehicles. The PDP-11s also displayed the status of these vehicles on 6 monitors. In 1989 the VAX-11/750s were replaced by VAX 11/780s to improve performance. In 1993 the 11/780s were replaced by 2 MicroVAX 3800s. An additional MicroVAX 2000 was used for training and test purposes. VMS was upgraded to version 5.1. In 1998 the PDP-11 was replaced by Windows/NT boxes due to millennium end. In 2000 the MicroVAX 3800 and 2000 were replaced by 3 DS20s which included a RAID 320 controller to be able to do Hardware-based mirroring. OpenVMS 7.1 was used together with ORACLE 8.0 and the program source was ported to C++. The control stations are still terminal-based, but some of the terminals have been replaced with Windows-based PCs and terminal emulators. Since July 2002 the DS20 has been running under OpenVMS 7.3-2. Starting in 2003 a general redesign of the environment will be done: The application software will be completely rewritten in C++ and C#. The 3 DS20s servers will be connected via DCOM and .NET to 50 Windows clients. Redundancy will be application-driven; no clustering will be used. The operating system will be 7.3-2 and ORACLE 9.x. The servers will be connected via two independent LANs, one 100MB-based and the other Gigabit. Each external site will be connected via 3 different independent networks to the headquarters. All systems will be monitored via SNMP. A giant video wall with twelve projection screens will display all mission-relevant information. The project is planned to be completed by 2006.

Von "Skonetski, Susan"
Datum Mittwoch, März 24, 2004 3:24 pm
An "Skonetski, Susan"
Betreff OpenVMS Pearl - HP Awarded $784 Million Services Contract by Department of Veteran Affairs-(OpenVMS Clusters) Please distribute
Dear Distribution lists, this is a public Document. Many thanks to the
OpenVMS Ambassadors working on this account. In particular Mark Z.

Warm Regards,
Sue


-

HP Awarded $784 Million Services Contract by Department of Veteran
Affairs 3/24/2004 7:45:00 AM



PALO ALTO, Calif., Mar 24, 2004 (BUSINESS WIRE) -- HP (HPQ) (HPQ) today
announced that it has been awarded a ten year, $784 million contract by
the Department of Veterans Affairs (VA) for Engineering Support Services
and Maintenance of the VistA Health Information Systems. Under the VistA
Maintenance and Expertise Center (VMEC) contract, HP Services will
provide support and maintenance to the mission critical VistA systems,
helping the VA deliver vital health care data across the VA's 21
networks.

The Veterans Health Administration (VHA) operates 170 medical centers
throughout the United States, Puerto Rico, and the Philippines. More
than 4.5 million people received care in VA health care facilities in
2002 and more than 6.8 million people are enrolled in the VA healthcare
system. The unprecedented growth in medical system workload has
reinforced the VA's need for an adaptive environment that facilitates
the seamless flow and availability of critical data across its networks.


"HP is honored to have been selected by the VA to provide mission
critical support to the clinicians and IRM personnel dedicated to
serving our nation's veterans," said Tom Iannotti, senior vice president
and general manager, Consulting & Integration, HP Services. "The VA's
teamwork, combined with HP's Adaptive Enterprise strategy, has resulted
in VistA being recognized as one of the premier hospital information
systems in the world."

The VA's VistA solution is implemented at all VA medical centers. VistA
provides automation and record keeping for almost every clinical and
administrative office and function in the VA through the many custom
integrated software modules running from a single integrated database at
individual medical or regional computing centers. The VA continues to
demand more access, speed, manageability, scalability, and high
availability from the systems on which it implements VistA.

A strong record of collaboration

HP has been an infrastructure, consulting and services provider to the
VA since 1983 through its work on the Decentralized Hospital Computer
Program (DHCP) and Enhanced Decentralized Hospital Computer Program
(EDHCP) contracts. Over time, HP and the VA have collaborated closely to
continually evolve the VA's IT environment and have successfully
deployed HP's OpenVMS clusters on AlphaServer systems to build an
adaptive environment that has increased performance, utilizes 64-bit
architecture and has enhanced reliability and up-time. As part of this
latest agreement, HP takes responsibility for maintenance and support
for all hardware and software products that comprise the VistA solution.


"The HP team has worked closely with VA over the last 20 years as VA has
grown VistA from three core applications to a single integrated system
which services over 100 clinical and administrative functions," said a
VA representative. "We look forward to teaming with the HP architects,
engineers, and service teams as we continue to ensure VistA provides
world class automation for veterans healthcare needs."

HP has also worked with the VA in other areas related to the VistA
solution, using technology to improve patient care. This includes the
implementation of BCMA (Bar-Code Medication Administration), MAF (Mumps
Audio-Fax), automated patient information and interaction, and VistA
Imaging.

Information about VistA is available at www.va.gov/vista_monograph/.

About HP

HP is a technology solutions provider to consumers, businesses and
institutions globally. The company's offerings span IT infrastructure,
personal computing and access devices, global services and imaging and
printing. For the fiscal year ending on Oct. 31, 2003, HP revenue
totaled $73.1 billion. More information about HP is available at
www.hp.com.

The information contained herein is subject to change without notice.
The only warranties for HP products and services are set forth in the
express warranty statements accompanying such products and services.
Nothing herein should be construed as constituting an additional
warranty. HP shall not be liable for technical or editorial errors or
omissions contained herein.

SOURCE: HP

HP
Brad Bass, 240-744-8119
brad.bass@hp.com


Customize your Business Wire news & multimedia to match your needs. Get
breaking news from companies and organizations worldwide. Logon for FREE
today at www.BusinessWire.com.


Copyright (C) 2004 Business Wire. All rights reserved.


Mark

VA Team - Solutions Architect
OpenVMS Ambassador
Consulting & Integration
Federal Government Region
hp
(301) 918-5577 (office)
(240) 472-1385 (cell)
Mark.Zimmermann@hp.com


-----Original Message-----
From: Zimmermann, Mark
Sent: Tuesday, March 23, 2004 1:50 PM
To: Skonetski, Susan
Subject: RE: Mark do you still think the press release will go out this
week?


I haven't heard anymore about it. And my boss (Donna) is out today, so
I can't check. Is it important?

Mark

VA Team - Solutions Architect
OpenVMS Ambassador
Consulting & Integration
Federal Government Region
hp
(301) 918-5577 (office)
(240) 472-1385 (cell)
Mark.Zimmermann@hp.com


-----Original Message-----
From: Skonetski, Susan
Sent: Tuesday, March 23, 2004 12:15 PM
To: Zimmermann, Mark
Subject: Mark do you still think the press release will go out this
week?

[Brad McCusker]

Paul,

Have you contacted any of your former Ambassador collegues for the latest and greatest VMS stories?

If you aren't sure just whom that may be in NZ these days, send me mail back channel and I'll try to hook you up.

Brad McCusker
OpenVMS Engineering

[Antoniov.]

Paul,
my little success story ...

My biggest customer, in 1995 had one VAX 3300, one VAX 3100 and one VAX 2000, mixed V5.2 and V4.6.
After DEC presented alpha processor my customer asked me what had to do and how make. I made porting of my software from VAX to AXP in only three months and he migrated quickly to DS400. I had to made only minimal updated on source.
VMS has granted investiment of my little customers; it's very very important bacause it's easy grant big business such as banks, assurance or others but when entry level customers(now they use DS10) can evolve without problems, money is in money-box!
Don't forget, HP has a big focus on its customers: everytime I contatted HP I've always received an answer.

Regards
@Antoniov