Holy crap that was a great conference. It took a couple of days to recover, but I think it went really well.
I taught Practitioner Sat-Mon, had a workshop Tuesday, then I did session operator duty on Wednesday. In between, I attended a couple of talks and sat at the Walk-In Clinic dispensing advice. I had much more fun than I should have, but there it is.
Here, have some links:
My workshop Github repo
I attended Jeff McCune’s talk about using Puppet with containers. Here’s his repo.
Day 1 wrap-up with great tweets
Day 2 wrap-up with tweets
My talk was not accepted for PuppetConf 2014, but I’ll be delivering a workshop about the basics of organizing your Puppet code with git. You can find the schedule item here.
Sign up as soon as possible for PuppetConf 2014 because the price keeps going up on tickets. Use this link to get an earlybird discount.
I just played around with installing a Puppet agent using PE 3.2, using the new simplified installer. It’s nifty once you figure out how it works. Here’s what I did.
First of all, I’m using two nodes, a master and agent, using CentOS 6.5, running virtualized on VMWare Fusion. First, I installed the PE master using the usual bundled install script.
Then, I used PE console to add the proper package class to the available classes. I then classified the Master to use the class. In this case, I used pe_repo::platform::el_6_x86_64. I then ran puppet agent -t on the master to get the packages and scripts in the proper place. Just for shits n’ giggles, I did a service pe-httpd restart to be sure everything would work.
On the node, I then ran:
curl -k https://master.puppetlabs.vm:8140/packages/current/install.bash | bash
That was it. Everything installed and Puppet is running. Of course, we’re just curling a bash file, and we’re still using a bash script to install Puppet once more…BUT this beats the heck out of an answers file when you think about massive deployments that need fast provisioning.