Learn Chef, Puppet, configuration management, and automation | Hacker News Reader

Learn Chef, Puppet, configuration management, and automation

makerops.com

22 pointsmakerops12 comments

Hi, I have a 9-5 in "devops" writing puppet/chef code, and do freelancing on the side, I am putting together a site to teach entry level, or even experienced admins who don't mess with development much, a way to get through the learning curve. If you write ruby, you have probably heard of railscasts, my goal is the be the railscasts for system admins, teaching openstack/aws's APIs, how to write chef/puppet providers and recipes/manifests respectively, and how to manage things on the server platform (open source chef/puppet servers) as well.

On the flip side, I aim to teach devs how to roll your own systems, in a safe, scalable manner.

What do you guys think, any suggestions?

Open articleView on HN

> Makerops.com will help you or your team, decide whether puppet or chef is the right path for you.

The deployment for an app I have involves building EC2 machines from scratch by means of a Fabric script, which is doing nothing more than connect to an EC2 instance through SSH and then invoke shell commands, like installing packages through "apt-get install", cloning the Git repo, compiling the project and copying configuration files.

This Fabric script is invoked by a Ruby script I wrote, that's using the AWS command-line tools under the hood, first creating an EC2 machine, then running the Fabric script on it, then creating an AMI (image) out of that machine.

Another Ruby script I have, also using the command line AWS tools, is creating auto-scaling groups out of prepared AMI images, properly configured, with auto-scaling policies based on latency, attaching those auto-scaling groups to our ELB load balancer.

This workflow of first creating an AMI image seems like overkill, however I like immutable snapshots of deployments, such that (a) we can run different versions in parallel to test new versions on only a small percentage of traffic and (b) in case shit happens, we can instantly revert to an older snapshot that is known to work (and problems can happen not only when you're breaking things in your code-base, but also when you install different versions of software packages or with faulty system-wide configuration files, so having a snapshot of the whole system is useful).

For me the question of Puppet or Chef has been neither. Both are way too complicated and problematic for what I wanted to do. I'm a developer and I solved my deployment needs by writing a bunch of scripts.

This is an awesome point, and I need to change the copy; the site won't teach just puppet/chef, even though that is my specialty, there is a lot of room between going full fledged with a config stack, and writing some thing in between. It will be down the road, but these topics are something, I'll also be pursuing. Especially how to take a vagrant instance to the cloud, that is a workflow I really like.

As a recent Puppet/Chef advocate I think this is a great idea. I've recently gone to a lot of DevOps conferences/meetups in the London area and they are becoming quite popular.

I am sure a similar thing is happening where you are based and it might be worth you going along to give talks and sharing your experience. It will help your site get exposure to the right audience if you respectfully mention it at the beginning or end of your talk.

Some constructive points: The copy on your site could use a tweak. There are a lot of unnecessary commas and strange grammar usage.

The mouse-over nav on your blog makes my eyes bleed, especially the colour of the tag line.

Consider putting your blog under your main subdomain (http://www.makerops.com/blog) rather than a separate sub-domain (http://blog.makerops.com). We did this recently and it dramatically improved our SEO as Google started using the content on our blog to rank our whole site.

Edit: Fixing my own grammar!

Thanks for the input, first thing on the agenda after mastering producing screen casts, is to find a good copywriter, and designer for the launch. If anyone is looking for some freelance work in these areas, feel free to get in touch anthony(at)makerops.com

What about other tools like: Open Source PaaSs, Open Virtual Machine Formats, OpenShift, OpenStack, Cloudify, CloudFoundry, Hyper V, KVM, Xen (XCP), Virtual Box, AMI, LXC, OpenVZ, GlusterFS, Ceph, OpenStack Object Storage (SWIFT), Sheepdog, OpenFiler, Provisioning and Monitoring tools, like you said Orchestration/Automation & Configuration Management tools, Cobbler, Spacewalk, Kickstart, others like Cfengine, Cacti/RRDTool, Nagios, Zabbix, Zenoss, Capistrano, RunDeck, Func, MCollective, and others. Then probably choices between choosing the OS: Debian, Fedora, maybe Arch Linux for those that like that.

These, and probably others that I don't know, are the choices that someone, who believes in not pursuing in anything proprietary, that also wants to learn to build an OpenSource IaaS, or PaaS stack for their product- are researching into but cannot find a straightforward guide. Advocating working with OSS, and teaching people how to use OSS tools will lead to more jobs, so keep at it. I subscribed to your email list...

It looks like you are double-decoding: your signup form turned a plus sign in my email into a space and then complained that it was invalid. And that has to be the blandest "thank you" page I've ever seen.

I think that's a niche that needs to be filled. We don't have half as many devops as we need right now, partly because many companies haven't even realized they need this position in the team, but also because it's so hard to train people.

Thanks for the tips, I'm not the best marketer/designer etc. I forgot to dig in to the + in the emails, Ill check that out when I get a chance.

I agree there is a niche that needs to be filled, especially from a devs perspective, after reading about the startup that lost almost everything due to no backup, it re-affirmed my belief that even more so from a devs perspective, a good railscasts-like learning site will hopefully meet a need.

This seems to be an extremely needed tutorial. makerops, do you have an ETA on when the lessons will begin? I am very excited for the premiere.

Minor thing: would you mind having the logos for Chef, Puppet, etc. actually link to those sites? Seems kind of mean not to.

Let me play devil's advocate: This is a landing page. The very last thing you want in a landing page is to send your visitor to another site. They have enough traffic as it is; they don't need the OP's.

This site offers (if I read it correctly, which I may not have) training on these devops tools and platforms, right?

Being able to go and look at those sites to learn what the tools are and why I might need training would be useful--seeing the Chef docs might make me appreciate this service even more.

I agree with you, I should have linked to them; when I get a chance I'll do so; it was one of those things that just didn't occur to me.

Reply on news.ycombinator.com