I've been doing business with Godaddy for many years. Yes, I know it's easy to Google up pages of complaints about them, but they were the devil I know, and I'd had no problems with them, until last week. True, it was only one serious problem, but it was a doozy, and they choose to avoid communication until after the technical issues had been resolved.

After nearly three years of worry-free operation, one VPS server suddenly went totally nuts. Web and DB services continued to run, but DNS (domain name services) repeatedly kept failing (shutting down) and so too did mail and other services, though not as consistently as DNS, which at times stayed up for only as little as three minutes after restarting.

It also happened to be the server that hosts this site as well as a number of clients. Nothing anywhere near this bad had ever happened before, and the timing couldn't have been worse, as I and a number of associates were all at DrupalCamp Nashville, a Drupal conference event.

The problems had begun earlier in the week, but only very sporadically (about 4 times in 3 days). Then the day before the Saturday conference, the day of the conference and for two more days it became a nightmare.

I submitted a support ticket on Friday morning. The auto-reply email said I could expect a response in 24 hours. Huh? I called their support number, but they said they really couldn't deal with server issues, and suggested I do a chat with another department. The person I chatted with hinted at possible problems with memory on "some linux servers" but could not say whether or not ours was affected, nor would I get any notification once the problem was believed to be fixed. In the meantime, she suggested I start investigating our logs and check configs to find the problem.

I simply couldn't do that at the time, but also did not understand how/why a configuration that had been stable for three years would suddenly be the root cause of our trouble.

A day later, about at the 24 hour mark or longer, I received an email from a support tech. He pointed out that the log file for the DNS service had lots of messages in it. Really? No freeking kidding? A log file for the DNS service (the busiest service on an internet connected server) full of messages is hardly a surprise. He suggested I put together my plan of action.

Meanwhile, I was at the Drupal Camp, handing out business cards with my broken email address and business URL. The most I could do was to notify my clients that I was aware of their sites being down, and doing what I could to get the situation resolved. I fully admitted to them that my attention was focused elsewhere, had been for some time, but that that was also why I strongly suggested this sudden nightmare was ultimately beyond our control.

Sunday and Monday were spent double-checking logs and config files, making sure there was not some fatal condition that we could address ourselves. Found nothing. I also reached out to one of the guys at 7Sudos.com. Overall he reinforced my belief that the situation was not in Raindear Media's control. The last thing we wanted to do was start making willy-nilly changes and risk breaking what was not already broken. We didn't have much choice other than to sit tight and wait for some "magic."

Magic arrived at about 6:20 pm that Monday night. That was the last time the server was restarted, and, miraculously, it has been up since.

I forgot to mention that I had also received an invitation from Godaddy to complete a customer satisfaction survey and tell them what I thought of their "first class service." That email came within minutes of my first phone call on Friday, long before I actually got any service. The only thing Godaddy was in a hurry for was my praise. Well, I hung on to it and let them have it after I was convinced the troubles were over, at least temporarily.

The next day I received an email from a "concierge" telling me they had observed the server for 12 hours and found no problems. He didn't mention which 12 hour period he was referring to. The whole tone of the email seemed to imply that there never had been a problem or that it must have been something we'd taken care of ourselves, or maybe just a figment of my (and my client's) imagination.

So I blasted back again and let them know how unhappy I was with the whole ordeal. After that, finally, I received yet another email and it said, in so many words, we already told you we found a config problem on the parent server (the overlord master of our server's universe), fixed it and observed it to make sure it was fixed. It took long enough, but that was the first time they ever specifically indicated that there had ever been a problem, and that it was indeed beyond my control, and that they fixed it.

They didn't seem to get how important communication would have been during this server event. Had they told me on Friday there was a problem that would take four days to fix, yes, I would have been very annoyed, but not nearly as upset as I was from the lacking communication and the finger-pointing in what little there had been. Keeping me informed would have at least assured me that they gave a damn.

So, that does it. Before the present agreement expires, Raindear Media and clients will be looking for a new hosting service to call home.