Thursday, January 22, 2009
GoGrid Sucks! If You Value Your Sanity - Do Not Use Them!
Yesterday was not a good day. We received a message from GoGrid that our server was about to be rebooted. When it came back up, the filesystem was corrupted. What's this mean? It means our server was a useless pile of transistors. But that really wasn't the beginning or the end of it.
A month ago when our server was rebooted, our filesystem had a few files missing. We filed a ticket and went back and forth with GoGrid support. They kept pushing the blame back at us. This was the final correspondence we had when they decided to close the case, even though, as you can see, we weren't prepared to close it. My comments are in red, GoGrid's in blue. I was pretty sure this wasn't our problem:
I was quite surprised when they admitted to me on the phone yesterday that they have had sporadic read only issues with their servers - thus, the problem we experienced back in November was their fault. Well that's nice. They took responsibility for their inaction. Would have been nice a month ago.
Today they again took fault for the corruption issue. But this is where it gets really out of hand. They knew that there were potential issues with their hardware. These issues had the potential to cause server corruption. Yet they decided not to notify their customers or to implement a backup strategy for their customers. Instead, as Raj (one of their VP's) explained to me on the phone today, they didn't reach their threshold of the number of customers who may have been impacted, so they didn't notify us. I chided him for that. What an utterly absurd notion. If there is ANY potential impact to ANY of your client's data, you have a responsibility to notify your customers. This isn't rocket science. It's damn common sense. Even their support staff told me they couldn't believe that management decided to withhold this information. Who the hell is running the show there?
What is the result of all of this? Well, I can't code right now because our SVN server was lost, so we have no code repo. I can't build our server because they're in the middle of upgrading hardware on our node that's "supposed" to fix this issue. And I don't quite have our strategy figured out because Fajar is probably sleeping somewhere in Jakarta. And I want to run everything through him - since he is our CTO.
It's putting us back a few days to say the least. As you can imagine, we are looking to alternative hosting services. They gave us a $400 credit. I asked for the next 6 months, at least, free. I'm not sure what it would take to feel confident in their service. But I do know, if I were you, I wouldn't put anything up in GoGrid's cloud. Beta my ass... this isn't close to a beta.