OpenStack

Edit Jan. 26, 2015: After discussing this with the Oslo team some more, I discovered that almost everything I said here was already covered quite well in the drop namespaces spec (that I reviewed... ;-). You're probably just as well off going and reading that, but I'm going to leave this here just in case it happens to help get the word out.

The Why

During the Kilo cycle the Oslo team has been busy deprecating the oslo. namespace in all of our projects. The first question this probably raises for most people is: why? Unfortunately the answer to that is not simple, and I'm not sure I understand some of the deeper details of it myself. However, after having talked to some of the Python packaging folks that hang around the OpenStack project we came to the conclusion that namespace packages in Python 2 are just too broken for us to continue using them.

So it's been a couple of weeks since I got back from the Paris Summit, but it's been a busy time and I haven't had time to write up all of my thoughts about it. Here's my belated attempt to do that.

Edit: Updated 2015/3/19 with more current diffs of my changes.

At the Atlanta OpenStack Summit we had a session on something called QuintupleO, otherwise known as "TripleO wasn't confusing enough, let's add another layer" :-) Barring a few specific concerns from other teams, which I believe have now been addressed to their satisfaction, everyone seemed to be on board with the idea. But what exactly is QuintupleO, and where does it stand today? Read on to find out.

As I mentioned in a previous post, about a year ago I picked up a 1U server from EBay to use as a local single-node OpenStack environment. In general I was quite happy with it, but at some point I got tired of paying for electricity to run a fairly power-hungry server that sits idle, or close to it, about 95% of the time. The fans also picked up an annoying whine somewhere along the line, so once I discovered how much more efficient a modern desktop processor would be, while actually performing better than the old dual server ones, I decided it was time for a new box. This post will be the story of my journey to get Tempest running in an acceptable fashion on it, and what I learned along the way.

Most OpenStack projects use a combination of tox, testr, and subunit to run their unit tests. This works fine as long as you don't have any simple errors in your code, such as a typo in an import or a missing closing paren. In those cases testr or subunit may swallow the error and you won't know exactly where it is (I hear this may be fixed in some version of the tools, but the one I'm currently using still has this limitation). To fix this, we need another way to run unit tests that will expose the actual error to us.

One of the unintended consequences of Feature Freeze in OpenStack is that a lot of new features get merged in a hurry at the end of the cycle. That's obviously not ideal, but it's a reality we all have to live with for now. TripleO has been bitten by this a few times recently, and the long gate queue that also goes along with Feature Freeze means sometimes fixes for problems take a long time (days) to merge in the other projects. In the gate we have tempreverts and cherry-picks to mitigate that problem, but for local devtest those features are not available. Fortunately there's another solution.

I attended my first OpenStack Summit in Atlanta last week, and now that I'm (mostly) recovered I wanted to write down some of my thoughts about the experience.

I said in a previous post that I would write something up about my devtest workflow once I had it nailed down a bit more, and although it's an always-evolving thing, I think I've got a pretty good setup at this point so I figured I'd go ahead and write it up. Instead of including scripts and such here, I'm just going to link to my Github repo where all of this stuff is stored and reference those scripts in this post.

Just ran into an issue running tox in diskimage-builder on a CentOS 6.5 instance I booted to test a change that failed the python26 gate test, so I thought I'd post the fix in case anyone else hits the same thing.

A full run of TripleO's devtest takes a long time - around an hour or more on my i7/16GB box even with a hot squid cache. Quite a bit of the time is spent building images, and there are a few ways to speed that up, some of which are easier than others.