Amazon EC2 Personal Experiment Terminated

Thanks also to my ignorance of the tool, my experiment in the Amazon EC2 environment terminated abruptly this night with a “termination” of my server instance: I terminated it, thinking that “terminated” meant an equivalent to a hardware “halt”.

No, it does not only mean “halt”, it means also the complete loss of all data inside the instance, as clearly stated here (and in another zillion places like here, if you are lucky and stumble upon them before it’s too late):

Instance – The running system based on an AMI is referred to as an instance. All instances based on the same AMI begin executing identically. Any information on them is lost when the instances are terminated or if they fail.

What led me to the termination was the loss of network connectivity that happened during a reboot (yes, I did a reboot before destroying my server, and rebooted a couple of times before without any problems.) I did not have the time to fix it , but probably it was related to some bug in network detection startup scripts.

So tonight I understood the real nature of EC2: it’s a “volatile” service and it’s not good for my simple web hosting needs and other similar tasks; indeed I must say that it’s very good and stable (in 45 days there wasn’t a glitch), all in all it was a good experience to do.

Now, after some hours rebuilding lost services on my home server from memory (the one embedded in my body!), I’ll search for a piece of hardware and a stable IP address or two: I could “halt” my server (kindly asking someone to switch it on again) and, if the operating system or the hardware fails, it’s not said that I’ll lose all my data.

One thought on “Amazon EC2 Personal Experiment Terminated”

Comments are closed.