Friday, March 02, 2007


Another Thursday night at PLUG

Feeling guilty over missing so many get-togethers last month, I had to go to last night's Portland Linux/UNIX Group meeting in order to keep my techie credentials somewhat believable. The presentation was by Brian Martin, who talked about the challenge of implementing his recovery plan after his employer was struck by the Loma Linda earthquake.

What I'm going to provide here are the contents of the Powerpoint slides of his presentation -- well, what they might have read if he had Powerpoint slides for his presentation.

First was the three steps of disaster management, in the order of importance Brian assigned them:

  1. Take regular backups;
  2. Move these backups off-site; and
  3. Develop a disaster recovery plan.

His reasons for this order was simple: You can recover better from a disaster if you have backups and no disaster recovery plan than the other way around. And without regular backups, there is no reason (except an obsession with useless procedure) to move tapes off-site unless they contain backups.

Next would come filler material: photos of the earthquake damage, airport interiors, Brian after wearing the same set of clothes for three days. (poor guy never had a chance between the earthquake and arriving at the hot backup site in Philadelphia to go home and pack). While this eye candy flashed on the screen, Brian could tell the story of what happened, and since he's a better story teller than I am, I'll just let you know that he's planning on repeating this presentation to other groups.

And at the end, would come a set of slides emphasizing the message of his presentation. He didn't bother to talk much about the technical points; that part went without a hitch, especially because his team had been practicing for this this on a regular basis and worked out any bugs with that. His emphasis was on the non-technical problems that he and his fellow workers spent five to six days coping with. I'd summarize them as follows:

  1. When the manager of the data center declares a disaster, starting the recovery process, he will start spending lots and lots of money -- and much of it will be wasted because the goal is to get the business back online.
  2. There will be a number of unforeseen events. For example, vendors will fail to do what you need them to do, a disaster will effect a lot more than just the computer center, and so on.
  3. Create your own personal recovery plan. Each member of the team needs to take care of himself and his family first. (Remember, Brian spent three days wearing the same three-piece suit because he never had a chance to get home.)
  4. The disaster recovery plan should not start with the words "In the case of a disaster" because that supposes the plan implementer is comfortable making that determination. It should also include guidelines so that person does not lose time worrying whether he should declare an emergency.

After that we all went out for beer, which some consider is the most important part of the meeting.


Technorati tags:


Comments: Post a Comment

Links to this post:

Create a Link

<< Home

This page is powered by Blogger. Isn't yours? Site Meter