Crowbar modularized: latest changes that make clouds even easier to create, update, and maintain

In the last week, my team at Dell completed a major refactoring of Crowbar that significantly improves our ability to bring in community contributions and field customizations.  Today, we merged it into Crowbar’s public repo(s).

From the very first versions, our objective for Crowbar was to create the fastest and most reliable cloud deployments. Along the way, we realized Crowbar’s true potential lay in embracing DevOps as an operational model for maintaining clouds. That meant building up cloud deployments in layers from pieces that we call barclamps (extensions of Chef cookbooks). Our first version, centered on OpenStack Cactus, leveraged barclamps but was still created as a single system. This unified system was a huge step forward in cloud deployments, but did not live up to our CloudOps vision of continuous delivery.

In this version, each Crowbar barclamp is an independent delivery unit that can be integrated before, while or after installing Crowbar.

The core of the change is each barclamp, including the most core ones, are stored in independent code repositories. Putting the code into distinct repos means that each barclamp can have its own life cycle, its own maintainer site and its own dependency tree. This modularization allows customers to manage their Crowbar deployments with a very fine brush: they may choose to customize parts of the system, they could lock components to specific tag and they can bring in barclamps from other vendors.

While the core barclamps are automatically integrated into the Crowbar build using git submodules; other barclamps are installed into the system as needed. This allows you to pull in the suite of OpenStack barclamps at build time or to wait until your Crowbar system is running before installing. Once you install a barclamp, you are able to retrieve an updated barclamp and reapply it to the system.

This feature gives you the ability to 1) choose exactly what you want to include and 2) perform field updates to a live Crowbar system.

Let’s look at some examples:

  1. The Cloud Foundry barclamp can be sourced Cloud Foundry instead of bundled into the Crowbar repository. This allows the team working on the cloud application to take ownership for their own deployment. As a continuous delivery proponent, I believe strongly that the development team should be responsible for ensuring that their code is deployable (refer to my OpenStack “Deployer API” blue print attempting to codify this).
  2. DreamHost, maintainers of Ceph Storage, can maintain their own local barclamp repos for OpenStack that are cloned from our community Swift barclamp. This allows them to innovate and customize OpenStack deployments for their business and choose which updates to merge back to the community.
  3. Rackspace Cloud Builders can work on the most leading edge OpenStack features and maintaining workable deployments on branches. As the code stabilizes, they simply merge in their changes.
  4. Dell BIOS and RAID barclamps only support the PowerEdge C line today. When we offer PowerEdge R support, you will be able to install or update the barclamps to add that capability. If another hardware vendor creates a barclamp for their hardware then you can install that into your existing system.

I believe that these changes to Crowbar are a huge step forwards on our journey of creating a community supportable Open Operations framework. I hope that you are as excited as I am about these changes.

I encourage you to take the first step by trying out Crowbar and, ultimately, writing your own barclamps.

Post Scripts:

  • In addition to the modularization, the updated code includes RHEL as a deployment platform. At present, you must choose to be either RHEL or Ubuntu at build time.
  • We have enhanced the network barclamp to describe connections as more abstract connections, called conduits, between nodes. This is a powerful change, but requires some understanding before you start making changes.
  • We have only begun testing the change as of 9/12, we expect the system to be fully stabilized by 10/3. If you are not willing to deal with bugs then I recommend building the Crowbar “v1.0” tag (or using the ISOs from our July launch).

Crowbar near-term features: increasing DevOps mojo and brewing Diablo

We’ve been so busy working on getting RHEL support ready to drop into the Crowbar repos that I have not had time to post about what’s coming next for Crowbar. The RHEL addition has required a substantial amount of work to accommodate different packaging models and capabilities. This change moves Crowbar closer to being able allow nodes’ operating systems (the allocated TFTP Boot Image) to be unique per node.

I will post more forward looking details soon but wanted to prime the pump and invite suggestions from our community.

We are tracking two major features for delivery by the OpenStack October Design Conference

  1. OpenStack Diablo Barclamps. Expect to see individual barclamps for various components like Keystone, Dashboard, Glace, Nova, Swift, etc)
  2. Barclamp versioning / connected imports. This feature will enable Crowbar to pull in the latest components for barclamps from remote repositories. I consider this a critical feature for Crowbar’s core DevOps/CloudOps capabilities and to support more community development for barclamps.

We are also working on some UI enhancements

  • Merging together the barclamps/proposals/active views into a single view
  • Enabling bulk actions for nodes (description, BIOS types, and allocate)
  • Allowing users to set node names and showing the names throughout the UI
  • More clarity on state of proposal application process (stretch goal)

I am planning to post more about our design ideas as work begins.

If you want to help with Diablo barclamps, these will be worked in the open and we’d be happy to collaborate. We’re also open to suggestion for what’s next.

Rob’s rules for good APIs (with explanations!):

or “being a provider in an age of mass consumption.”

While blogging to abstraction, I kept thinking of ways to I should have avoided those awkward and painful APIs that I’ve written in the past.  Short of a mouthful of bad tasting SOAP, here are my top API considerations.

Before you get started:

  1. Take are to design the semantics of you API because APIs are abstraction that hide complexity from, simplify work of and encourage good behaviors in its consumers.
  2. Think security and auditing first because it’s difficult to “see” all the options that an API will expose on a quick review.  You don’t want deal with the mess when Venkman crosses the streams.
  3. Write tests for your API because you want to find out that you broke it before your consumers do.  Even better, also add simulators that pump data into your API so that you can run system tests.
  4. Use versioned APIs even at 1.0 because you and your consumer must expect APIs to evolve.
  5. Separate verbs (taking actions) from nouns (editing things) in your API because they have fundamentally different use, validation, and notification models.   This also helps eliminate side-effects.
  6. Do not anticipate consumers’ appetite for data because there is a cost for returning information and maintaining compatibility.

Now that you’re committed:

  1. Offer the minimum options to service the consumer’s use-case because it’s easier to add than remove options.
  2. Avoid exposing details to your consumer because it will lock-in your implementation and limit your flexibility.
  3. Use natural keys over system assigned keys because users will have to make extra calls to figure out your assignments and it will lock your implementation.
  4. Never break your API’s existing calls when adding options because you never really know who is using which API calls.
  5. Object versioning (timestamping is enough!) reduces confusion because you can’t control how long a user will hang onto an object before they decide to send it back to you.
  6. Fail loudly with obvious delight because you want to make it hard for bad data to get into your system and you also consumers to know what went wrong. 
  7. And, when failing loudly use HTTP error codes correctly because they really do mean something.
  8. Finally, be text tolerant because strong typing makes your API harder to use and text is very robust and flexible.

Whew.  Now for a REST.