AberdeenCloud is gone! We have 36hs to recover your backups!

What's important to learn about the recent AberdeenCloud meltdown? You can not trust your backups and site to a single provider!


Well it is all over the blogosphere, AberdeenCloud breakdown. I already read a couple of articles on the subject and I couldn't agree more with the one from annertech, only way to avoid this kind of issues is offsite backup. I would normally write how to recover from something like this, but sadly as codeenigma explained there isn't much we could do, their servers just collapsed and were gone. But now for 36hs we'll be able to recover our data.

Sadly we didn't notice problems with AberdeenCloud until it was too late. Due to other reasons we had a plan to pull away, but we didn't get to apply it for the only site that remained there. We did have backups, but not automatic. So they weren't as frequent as we would have needed given the total meltdown on AberdeenCloud. Since we had some backups and it was only one site it didn't take that long to rebuild what we lost. It wasn't really an option to wait for them to come back as it didn't seem it would happen. I was really surprised and pleased to see they finally replied, but before that we used AberdeenCloud cache (wasn't realiable but with patience you could get something) and also use the Wayback Machine.

So what's important about this? We need to learn we should ALWAYS have an alternate site backup on a different provider. You'll need it for disaster recovery, that's the only way to guard your client's data. You can not trust a single provider with both your site and backups. We shouldn't be naive about this, it is also important that you track back to learn which company your provider trust on. If for example your sites are hosted by Pantheon and you want to trust your backups to NodeSquirrel, think again as they are both the same company.

Bottom line, be serious about backups even if you are only trying the service. In the company I work for we do it for the other services we have, and from now one we won't try a service without implementing offsite backups first. That needs to be part of the initial implementation, not a post launch burn down list task.