Pulling the Strings

In the latest Puppet Podcast, Puppet Community Lead Dawn Foster is joined by Eric Johnson, program manager of the Google Compute Engine team and Aaron Armstrong, an engineering manager at Puppet Labs.

Last Tuesday, Google announced that its Google Compute Engine is now generally available to the public. Listen in to this podcast to hear Eric, Aaron and Dawn discuss and explore the integrations between Google Cloud Platform and Puppet.

GCP is a collection of services created using Google's experience scaling and supporting their own services: platform-as-a-service and infrastructure-as-a-service options, storage solutions and a range of other services and developer tools.

In Puppet Enterprise 3.1, we updated our cloud provisioner to work natively with Google Compute Engine, and there's now a node type for GCE. Watch the video below to see how you can spin up and interact with Google Compute instances directly from the command line.

Google is also committed to helping make sure that its Puppet modules are kept up to date to take advantage of the full power of GCE. Eric says,

"Google is committed to providing support for Puppet users. So we'll be contributing support for both GCE_Compute and Node_GCE to ensure that it's updated to take advantage of the latest features that GCE is releasing."

Eric helped to update the GCE compute module, which offers Puppet users the opportunity to integrate GCE and Puppet. You can create additional persistent disks, networks and firewall rules, and set up a load balancer. Aaron notes that employees within Puppet Labs are particularly excited at how fast it is to provision new nodes on GCE, as well as the more granular, per-minute billing that's available after the first 10 minutes.

See the rest of the podcasts.

• Watch Eric spin up 1000 nodes in 4 minutes in this PuppetConf talk on GCE

We caught up with Google's Brad Fitzpatrick at OSCON this year to talk about what has developers and operations folks so excited about the Go Language (aka golang). Go came out of Google in 2009, and has been slowly been adding different services to the language.

Fitzpatrick says, "I was disappointed in the different compromises that you had to make between the different languages. You either got this safe little playground that was slow and generated lots of garbage. Or have this like scripting language world where you would waste a lot of CPU, and you could do only one thing at a time or you were in event-callback hell. Or in you're in C++ land where you have to do everything where you have total control of everything but it's also a pain." Go seems to provide a nice middle ground where it is fun like a scripting language, but it also allows you go do low-level tasks and provide concurrency that's lacking in other scripting languages.

Fitzpatrick says that people in operations tend to like it because Go compiles down to a self-contained binary, which makes it easier to deploy. Take a listen to the interview above for how Google is using Go, more reasons why he finds it so refreshing, and how it's made him feel like a productive programmer again.

• Subscribe to the Puppet Labs podcast.

Tim O'Reilly—founder of O'Reilly Media—started the Open Source Conference (OSCON) fifteen years ago. In this interview, he reflects on the how the cloud has changed the nature of open source software. Because the Internet is the primary operating system, he says, there's a mindset about licensing and development models that is fundamentally different from that found in open source culture 15 years ago.

According to O'Reilly, the identity of the open source community centered around various licenses. Now, a significant portion of developers are posting their code on GitHub without an explicitly declared open source license. It's a big enough issue that GitHub started a campaign to help people choose a license at ChooseALicense.com.

He also talks about the rise of data in applications, the importance of DevOps, and how the next wave of open source hardware and operating systems will take us in unexpected directions.

• Subscribe to the Puppet Labs podcast.

This week on the Puppet Podcast, authors William Van Hevelingen and Spencer Krum join Puppet Community team Dawn Foster and Kara Sowles to give us some insight into the upcoming new edition of the book Pro Puppet. The third author, Mozilla sysadmin Ben Kero, wasn't able to join us.

The first three chapters of the new Pro Puppet will be focused on anyone setting up Puppet for the first time. The rest of the book goes deeper, including how to run Puppet without a master, how to get your workflow down, an updated section on the MCollective orchestration engine, and an entire chapter dedicated to Hiera.

The original Pro Puppet was written by James Turnbull and Jeff McCune and published by Apress in 2011, so it was overdue for an update. Listen to how Spencer, William, and Ben ended up writing the book - involving a piece of misplaced spam mail - and what book is coming up next.

The new Pro Puppet will be available Nov. 20th on Amazon.com, and you can pre-order your copy now!

• Preorder the second edition Pro Puppet or add it to your wishlist

We recently sat down with Devon Peters, a tools engineer at Jive Software to talk about how they use MCollective for orchestration along with Puppet to manage their infrastructure. You might remember Devon from his recent PuppetConf talk, "Intro to Systems Orchestration with MCollective." We were joined by Pieter Loubser and Richard Clamp from the MCollective engineering team here at Puppet Labs to talk about a wide variety of orchestration topics.

This was a fun podcast because we talked about some of the performance problems Jive Software has faced in the past, along with how the team here at Puppet Labs is working closely with Jive Software to improve MCollective performance, which will benefit a lot of other companies as well. We also talked about the future of MCollective and other things we are doing to improve the product.

If you are curious and want to learn more, you can check out our new and improved MCollective documentation or watch Devon's PuppetConf talk, "Intro to Systems Orchestration with MCollective."  You can also join us in the community on the MCollective-Users Group or in the #mcollective IRC channel on Freenode to ask questions or just learn more about how people are using MCollective.

See the rest of the podcasts.

Learn More

• Find a local Puppet User Group or start one in your city

• Visit our community Ask site to get answers and advice for using Puppet

• PuppetConf talk:  Intro to Systems Orchestration with MCollective

• Talk to us on the MCollective-Users Group or in the #mcollective IRC channel on Freenode

• Read the new and improved MCollective documentation

We've been dusting off some audio from OSCON in July, and we found this gem for your listening pleasure.

Kent Bye, videographer & media maven at Puppet Labs, took some time out of his busy schedule at OSCON to interview Ben Kero, senior sysadmin at Mozilla. They spent some time talking about how Ben and the rest of the ops team there use Puppet.

Ben also talked about his session at OSCON, "How to Use Puppet Like an Adult," which he co-presented with fellow Mozillian, Daniel Maher. If you missed their talk, you can still have a look through their slides.

A huge thanks to Ben and Kent for doing this, and we hope you enjoy the podcast!

See the rest of the podcasts.

• Find a local Puppet User Group or start one in your city

• Visit our community Ask site to get answers and advice for using Puppet

• How to use Puppet like an adult

• Read about featured community member Dustin Mitchell from Mozilla

In this week's Puppet Podcast, I'm joined by Ryan Coleman, product manager for the Puppet Forge, and module engineering team members Ashley Penney and Hunter Haugen. They dive into an exploration of the module testing frameworks the Puppet Forge team uses.

What happens when you submit a patch to a module maintained by Puppet Labs? Why bother testing your Puppet modules at all?

Learn the answers to these questions, as well as the value of unit testing a module with rspec-puppet and integration testing with rspec-system, the tools available to make sure your manifests pass a syntax test, and how to take advantage of real-world data from Facter during testing.

Listen to the Puppet Forge team pretend for your benefit that they are bad at their jobs, and don't believe a word of it!

See the rest of the podcasts.

• Find a local Puppet User Group or start one in your city

Florian Haas, CEO, Principal Consultant from Hastexo joined Chris Hoge, OpenStack Developer for Puppet Labs, and Dawn Foster, Community Lead at Puppet Labs for a discussion during OSCON on the love affair between OpenStack and Puppet.

"The reason why we find Puppet particularly interesting and appealing," says Florian, "is that when we look at all these tools, Puppet has a massive mindshare ... there's some very interesting work that has been done there."

At Puppet Labs we've been maintaining a set of modules that allow you to deploy a whole OpenStack system with Puppet, but it's part of a much larger body of contributors. The modules are a community effort, developed under the OpenStack StackForge community.

The OpenStack Tour de Force tutorial during OSCON used these modules as part of KickStack. KickStack is a custom deployment wrapper built on the Puppet Labs OpenStack modules, which enables installation of an Openstack cloud in as little as 20 minutes

Looking to get some more OpenStack and Puppet goodness?  Florian Haas from Hastexo is holding a class on deploying OpenStack with Puppet in Portland. Register now to attend Puppet Cloud Jumpstart for OpenStack.

At Developer Day, the day before PuppetConf, we'll devote one of the many hacker tables to OpenStack. During PuppetConf  there are more OpenStack talks than I can list here.  See you then!

Attend PuppetConf, our annual operations conference at the Fairmont in San Francisco on August 22 and 23, and come a day early for Developer Day!

Find a local Solutions: Puppet and OpenStack

Blog Post: Puppet OpenStack modules 2.0 released to ForgePuppet User Group

Find a local Puppet User Group or start one in your city

Visit our community Ask site to get answers and advice for using Puppet.

We would like to use our podcasts to showcase more cool, new projects from the community, so last week Hunter Haugen and Dawn Foster talked to Daniele Sluijters (a.k.a. "daenney") about his newly released Puppetboard project.

Just last week, Daniele announced Puppetboard on the puppet-users list, humbly beginning,  "it’s a lovely grey and rainy day here in the Dutch summer, as good a day as any to release a new little project."

Puppetboard is a web interface to PuppetDB aiming to replace the reporting functionality of Puppet Dashboard. Currently, Puppetboard can show you all the info that PuppetDB can show you: a list of nodes and all known facts for each node in your infrastructure. You can drill down into them by fact, node, and run reports.

This is so new that it comes with some significant caveats and some of the code is experimental, so contributions and bug reports are welcome!

Daniele will be at PuppetConf starting on Wednesday at Developer Day. Track him down at the conference to learn more about Puppetboard.

See the rest of the podcasts.

Attend PuppetConf, our annual operations conference at the Fairmont in San Francisco on August 22 and 23, and come a day early for Developer Day!

Find a local Puppet User Group or start one in your city

Visit our community Ask site to get answers and advice for using Puppet.