In case you seem to experience a déjà-vu right now, it is not your imagination, this is the second instantiation of this post. More about this below. But first, here we go, I’ve done it again: For the fifth time in its almost 20 year history, this blog has moved again to another physical location.
As you might be aware from previous posts, I’ve been slowly running out of disk space on my bare metal server in Finland due to other projects besides this blog running on it, and I’ve experienced network slowdowns over recent months that do not get any better. So I’m jumping ship and the new home for the Docker containers of this blog is a virtual machine running on a bare metal server in a Scaleway data center in Paris, France!
While transferring all my services from one bare metal server to another is a bit of work, I’m nevertheless quite happy to do it. And that is because just for such an occasion I’ve always been careful not to use any features that would tie me to a particular cloud infrastructure provider. So moving from one data center to another basically meant just copying over virtual machines and docker directories and update DNS entries to new IP addresses.
To reduce technical-debt, I didn’t copy all of my virtual machines, but rather decided to install those that just run containers from scratch, so I would have a clean operating system and reverse proxy installation on them going forward. Which brings me to the déjà-vu of this post: After transferring the blog to the virtual machine on the server in Paris, I made a mistake and accidentally overwrote the disk image of the VM on the bare metal server. A silly mistake that has cost me a few hours of work and the original version of this blog post. All other data was nicely backed-up and during re-installation of the VM, I could just switch all services running on that VM back to the old server for a while. Outage time created by the silly mistake: 5 minutes, so I still count it as a win!