When things break, you need to fix them.

When things break and you are not near them, you need stop what you are doing to get to them… and fix them.

When things break, you’re on vacation, you absolutely have to fix the problem, and you don’t have a lot of money to pay others to fix it for you, you are screwed. We did not want to be screwed.

For our hosting company, we were operating in Worcester, MA and the servers were north of Boston in Lynn, MA. Every now and then, one would crash, and we would get on the phone with technical support for them to run into the server room and reboot it. Oh, and they would bill us a nice, hefty fee for their trouble. If we were fortunate to have the problem at 10am on a Wednesday, we could be back up and running 10 minutes after the call was made; 2am on a Friday, however, was a different beast. That’s when we would get in the car, and drive the dreaded drive. From Worcester to north of Boston.

Uphill.

In the snow.

Both ways.

It only took a couple of times in a month and a couple of irate customer calls for us to know that something had to be done. We first tried a common off-the-shelf solution of using a controllable power strip, essentially a series of power outlets that you could remotely control either over the phone or over the Web.

Oh, young and foolish Cory… how I wish I could go back and smack you on the head. 

We used the controllable power strips quite successfully to remotely reboot crashed servers for a couple of weeks, but we started noticing a trend. A computer would stop responding, we would remotely reboot it by pulling the AC power out, but it would not come back up.

Uh oh.

Next step… get in the car, drive to the datacenter, and find your power supply dead. A couple of times, it was a hard drive that failed to work. In retrospect, all of our wonderful power conditioning and UPS equipment didn’t mean squat when we were literally yanking the power cord out of the server while under heavy load and throwing it back in. We went from meager software freeze-ups and crashes to full blown hardware failures that required us to make the trip to the DC. Oh yea, and buy replacement parts.

We stopped using the remote reboot equipment, consequently eliminating our all too common hardware failures, and went back to dealing with getting tech support on the phone or getting in the car.

Necessity is the Mother of Invention

Whenever the datacenter tech support folks were called upon to reboot our servers, all they did was open the front of the cabinet and press the reset switch on the front of the chassis. We never had a power supply failure or hard drive failure whenever they did a reboot for us. Hmmm…

It hit me like Terry Tate defending the office supply closet. I could whip together a quick little box that will do that for me! I just need to press that reset or power button from the web. And so the baby was born.

And this baby was named WebReboot 1.0. It was crude, it was ugly, it was hand-made, and it lived in a Radio Shack project box that I cut out in my father’s workshop and spray-painted. But it worked.

It worked well enough that nearly every other customer in our little datacenter wanted one. We got a patent. We setup shop. And we grew…

As our team grew, we were able to listen to more and more customers, pimp more datacenters (as covered by InfoWorld), and so we made more and better products. 

In 2008, however, we ran into some insurmountable brick walls, and it was time for me and the rest of the founders to move on. As the sun set on my Servprise venture, I began to look where the sun would rise again for my next one. We had one heck of a journey, and I would love to share the rest of the story with you, the lessons learned, and where I’m going next, on this site.