July 1999

When Good 3000s Go Bad

By Scott Hirsh

Welcome to Worst Practices in HP 3000 system management. Why worst practices and not best practices? For one thing, opinion regarding this management consulting buzzword has now diminished to the point where some consider it a fad; too subjective a concept to be broadly applied. On the other hand, there is much more likely to be consensus on worst practices, those (in this case) system management death marches we have all participated in, or even led.

By focusing on the big no-nos, we can narrow down all practices to a subset that can help you compare the way you currently manage your systems with other ways that may be better now — or perhaps in the future when the political planets of your shop are in perfect alignment. And last, it’s much more fun to point out the stinkers, as evidenced by the popularity of “worst-dressed” lists, “Web sites that suck” and such.

If you recognize a worst practice as one of your favorite techniques for system management, feel free to write and offer your dissenting opinion (e-mail me at scott@precisionsg.com). In fact, I welcome any and all nominations for worst HP 3000 system management practices.

After 12 years in captivity at a shop in San Francisco I have had the opportunity over the last four years to visit many IT shops across the US in a systems management consulting role. While the names will be changed to disguise the innocent, in my travels I have seen the sometimes good, the often bad and the occasionally downright ugly sides of how to manage a system. (Of course, my shop was a systems management Shangri La, with 100 percent uptime, maximum efficiency, and a blissfully happy staff.)

The 3000 NewsWire has already established a high standard of presenting great tips and techniques for getting more from your HP 3000. However, when good systems go bad, generally speaking the least of the problems are technical. Rather, the systems become victims of less glamorous management maladies, such as (I’m not making these up):

1. The “not invented here” syndrome; ignorance and inbreeding as virtue. We don’t belong to Interex, don’t attend conferences, don’t read HP3000-L, don’t even read The NewsWire (gasp!)… and we’re proud of it. Even calling the HP Response Center is often viewed as a sign of weakness. The motto of this type of shop: “Wheels reinvented here.”

2. The revolving door. The vicious cycle of “shop from hell creates employee burnout that perpetuates shop from hell.” Usually accompanied by “we document nothing!” either because the management jerks don’t deserve it, or everyone’s too paranoid that they’ll be replaced as soon as the ink on the procedures is dry.

3. Only chumps buy system management software tools. We would rather spend hundreds of (unpaid) staff hours to develop our own tools that will eventually be unsupportable and either fail miserably and cause an interruption (at the worst possible moment) or fall into disuse to be replaced by good old time-consuming manual processes.

4. Us vs. them. Us (the virtuous, the misunderstood) is operations/production control. Them is the programmers or developers. They couldn’t be bothered (supposedly) to test properly, and we pay the price when the job blows up at night. We can only do wrong, like forgetting to reply at the console for over two hours the other night. Where is Ward Cleaver when you need him?

Worst Practices will continue to explore both technical and non-technical issues that plague our datacenters. I hope you enjoy the trip through management hell.