This will not be typical blog post, but rather some kind of a diary. Of course, I’ll try not to write complete nonsense, but as it will be rather exploratory effort it may happen. Sorry. My interest in the cloud is personally-professional. I may use it in my line of work, but I wanted to “touch” the cloud myself to get the feeling of it. Reading is cool, but it’s just not enough. Everybody is talking about it, I know I’m way behind – but lo, here I am going to experiment with the cloud as a curious developer.
My personal needs
My goals? I want to run some Docker containers in the cloud, maybe connect them somehow – so it will be about learning Docker better as well. Last year was very fast in that field (it was only the second year for Docker after all!) and there’s a lot I missed since I covered some basics of the technology in the summer 2014.
And I also wanted to know what it is to manage my cloud – even if it is small, preferably free. I didn’t check many players. It is difficult to orient among them and I’m not going to enter my credit card details to some company unknown to me. Talking about Docker, I needed either some direct container hosting (not sure if/how it is provided), or IaaS – that is some Linux machine where I run the Docker.
Finding my provider
Originally I wanted to test Google cloud, but that one is not available for personal use – which surprised me a bit. I knew Rackspace, but wasn’t able to find reasonable offer – something under 10 bucks a month. So I tried Amazon’s AWS – I like Amazon after all. 🙂 You may also check this article – 10 IaaS providers who provide free cloud resources.
What attracted me to AWS the most is their Free Tier offering that can be used for the whole one year! I don’t need much power, I don’t need excessive connectivity, I just needed some quiet developer’s machine. Also, one year is an incredible option there, many times I start some trial and don’t get enough of “real CPU/brain time” to play with the trial because of other obligations. But a year?! I’ll definitely be able to use AWS at least a bit during that time. Longer span also gives you better perspective.
Finally, AWS has a lot of documentation, instructions, videos, help pages, etc. I also quickly got a feeling that they are trying hard to give you all the tools to know how much you’d pay (which is probably feature of any good cloud) – even if you are using Free Tier. We’ll get to that later.
First steps with AWS
So I registered at AWS – it took couple of straightforward steps and one call they made to let me enter some PIN. I chose Basic support (no change in my zero price), confirmed here, confirmed there – and suddenly I was introduced to my AWS Console with tons of icons.
What next? Let’s stick to the plan and Google something about Docker and AWS. Actually I did this before I even started with AWS, of course. This page reads all the necessary steps. So we need some EC2 Linux instance – t2.micro with Amazon Linux AMI is available for Free Tier. Let’s launch it! Or, because I’m the manual kind of guy (kinda rare, reportedly), let’s check how to do it or see the video. It doesn’t seem that difficult after all.
I chose my closes region in AWS Console (Frankfurt, or eu-central) and created my instance. I clicked through all the Next buttons just to walk and read through the setup, but I didn’t change anything. In the course of this I created my key pair, downloaded the private key and – my instance was ready! Even if you don’t name it right away, it’s easy to do it later in the table of instances when click into the cell in Name column:
In general – I have to say I’m very satisfied with overall look-and-feel of AWS web. It provides tons of options and information, but so far I found what I wanted. You either look around the screen and see what you want, or click on the top menu (e.g. Billing & Cost Management is under your login name on the right) – or Google it really quickly.
Let’s connect to my Linux box!
If you know how to use private key with SSH, this will be extra easy for you. You just need to know what user to use for login. Again – I googled – and you can find either this article from their Getting Started section, or another article from Instance Lifecycle section.
Whatever method or client you used to SSH into your instance (I used command line ssh that comes with git bash), you’ll find yourself in an ever familiar CLI. And because we wanted to play with Docker, we – of course – follow another help page. This time called Docker Basics.
Well, docker info command has run, everything seems OK, we will continue some other time when the night is younger.
Next morning first impressions…
I mentioned already, that you have full control over the spending and AWS guys really offer you all the information needed. For example, when I woke up today after setting and launching my EC2 t2.micro instance late last night, I saw this report in my Billing & Cost Management Dashboard:
Sure, I don’t know exactly how to read it yet – or how it’s calculated exactly, because it must have been running for more than 5 hours. But at least I know where I am with my “spending”.
…and evening observations
I definitely should read how those instances are billed – and when. When I logged out from the instance I saw no change on the report (not even for hours), but when I logged in later, it suddenly jumped by couple of hours. This is probably about when the billing update is triggered. Talking about couple of hours, this is no problem at all.
Another thing to remember is that when you restart the instance it will take an instance-hour right away. That is, every started hour is billed. We’re still talking about prices like 0.0something for an t2.micro hour, so again, not a real deal-breaker. But with 750 hour limit for Free Tier, one should not try to shutdown/restart the instance 750 times in a day. 🙂
The pricing information is detailed but clear and it also plainly states: “Pricing is per instance-hour consumed for each instance, from the time an instance is launched until it is terminated or stopped. Each partial instance-hour consumed will be billed as a full hour.”
One day is not enough time to draw any conclusions, but so far I’m extremely satisfied with AWS. Next we will try to deploy some Docker images to it.