mikal – Page 2 – Made by Mikal

Why is Azure so slow to start instances?

I’ve been playing with terraform recently, and decided to see how different the terraform for launching a simple Ubuntu instance in various clouds is. There are two big questions there for me — how big is the variation between OpenStack derived clouds; and how painful is it to move between the proprietary clouds? Part of this is because terraform doesn’t present a standardised layer of cloud functionality, it has a provider per cloud.

(Although, I suspect there’s nothing stopping someone from writing a libcloud provider or something like that. It is an interesting idea which requires some additional thought.)

My terraform implementations for each cloud are on github if you’re interested. I don’t want to spend a lot of analysis on the actual terraform, because I think the really interesting thing I found isn’t where I expected it to be (there’s a hint in the title for this post). That said, the OpenStack clouds vary mostly by capabilities. vexxhost for example seems to only offer flavors that require boot-from-volume. The proprietary clouds are complete re-writes, but are generally relatively simple and well documented.

However, that interesting accidental thing — as best as I can tell, Microsoft Azure is really really slow to launch instances. The graph below presents five instance launches on each cloud I tested:

As you can see, Vault, Vexxhost, and AWS are basically all in the same ballpark. Google and Azure are outliers, with Google being crazy fast (but also very slow to delete instances, a metric not presented here), and Azure being more than three times slower than everyone else.

Instance launch time isn’t a great metric to be honest, but it does matter. For example if you were trying to autoscale a web tier or a kubernetes cluster, then waiting over two minutes just for the instance to boot before it can be configured and added to the cluster is probably not ok.

I wonder why Azure is so slow?

I did some further exploring after writing this post and was able to improve performance by changing how I handled resource groups in the terraform. The performance still isn’t great though. You can read more about that in a separate post if you’d like.

Coming to grips with Kubernetes in 2020: podcasts

It has become clear to me that it is time to care about Kubernetes more. I’m sure many people have cared for ages, but the things I want to build at the moment are starting to be more container based now that I am thinking more at the application layer than the cloud infrastructure layer. So how to do that? I thought I’d write down some notes on what has worked (or not) for me, in the hope it will help others. In this post, podcasts.

I thought podcasts would be an interesting way to get started with some nice overviews. This is especially true because I’m already a pretty heavy podcast user, so it was easy to slot into my existing routine. Unfortunately this hasn’t really worked out. I started with the podctl podcast, but they only ever talk about Red Hat stuff. It is very rare for a guest to not be a Red Hat employee for example. The presenters of this podcast seem to also really dislike OpenStack for reasons they never explain, which is annoying.

Then I figured maybe the Google Kubernetes podcast would be better, but it often lacks the depth I am interested in.

I am yet to find a good podcast which deep dives into technology instead of just talking about what is in the latest release. So maybe these podcasts are useful if you’re interested in what things dropped in the most recent release, but they’re not a good nor systematic way to get introduced to Kubernetes.

That said, I only just discovered the TGI Kubernetes youtube channel yesterday. It is not really what I wanted in a podcast given its a video blog, but I think it has prospects to be interesting. I will update this post when I’ve had a chance to check it out in more depth.

Have you found a good Kubernetes podcast? Am I being wildly unfair?

If I Understood You, Would I Have This Look on My Face?

This book discusses science and technical communication from the perspective of someone who comes from professional theatre and acting. Alan explains how his accidental discovery of the application of theatre sports to communication created an opportunity to teach technical communicators how to be more effective. Essentially, the argument is that empathy is essential to communication — you need to be able to understand where your audience is starting and and where they’re likely to get stuck before you can take them on the journey.

Unsurprisingly given the topic of the book, this is a well written and engaging read. The book is nicely structured and uses regular anecdotes (some of them humorous) to get its message across.

A detailed and fun read.

If I Understood You, Would I Have This Look on My Face? Book Cover

Title: If I Understood You, Would I Have This Look on My Face?
Author: Alan Alda
Genre: Self-Help
Publisher: Random House
Release Date: June 6, 2017
Pages: 240

NEW YORK TIMES BESTSELLER • Award-winning actor Alan Alda tells the fascinating story of his quest to learn how to communicate better, and to teach others to do the same. With his trademark humor and candor, he explores how to develop empathy as the key factor. “Invaluable.”—Deborah Tannen, #1 New York Times bestselling author of You’re the Only One I Can Tell and You Just Don’t Understand Alan Alda has been on a decades-long journey to discover new ways to help people communicate and relate to one another more effectively. If I Understood You, Would I Have This Look on My Face? is the warm, witty, and informative chronicle of how Alda found inspiration in everything from cutting-edge science to classic acting methods. His search began when he was host of PBS’s Scientific American Frontiers, where he interviewed thousands of scientists and developed a knack for helping them communicate complex ideas in ways a wide audience could understand—and Alda wondered if those techniques held a clue to better communication for the rest of us. In his wry and wise voice, Alda reflects on moments of miscommunication in his own life, when an absence of understanding resulted in problems both big and small. He guides us through his discoveries, showing how communication can be improved through learning to relate to the other person: listening with our eyes, looking for clues in another’s face, using the power of a compelling story, avoiding jargon, and reading another person so well that you become “in sync” with them, and know what they are thinking and feeling—especially when you’re talking about the hard stuff. Drawing on improvisation training, theater, and storytelling techniques from a life of acting, and with insights from recent scientific studies, Alda describes ways we can build empathy, nurture our innate mind-reading abilities, and improve the way we relate and talk with others. Exploring empathy-boosting games and exercises, If I Understood You is a funny, thought-provoking guide that can be used by all of us, in every aspect of our lives—with our friends, lovers, and families, with our doctors, in business settings, and beyond. “Alda uses his trademark humor and a well-honed ability to get to the point, to help us all learn how to leverage the better communicator inside each of us.”—Forbes “Alda, with his laudable curiosity, has learned something you and I can use right now.”—Charlie Rose

Prometheus 2.12, query logging, and startup failures on macos

Prometheus v2.12 added active query logging. The basic idea is that there is a mmaped JSON file that contains all of the queries currently running. If prometheus was to crash, that file would therefore be a list of the queries running at the time of the crash. Overall, not a bad idea.

Some friends had recently added prometheus to their development environments. This is wired up to grafana dashboards for their microservices, and prometheus is configured to store 14 days worth of time series data via a persistent volume from the developer desktops. We did this because it is valuable for the developers to be able to see the history of metrics before and after their changes.

Now we have a developer using macos as their primary development platform, and since prometheus 2.12 it hasn’t worked. Specifically this developer is using parallels to provide the docker virtual machine on his mac. You can summarise the startup for prometheus in the dev environment like this:

$ docker run ...stuff...
...snip...
level=error ts=2019-09-15T02:20:23.520Z caller=query_logger.go:94 component=activeQueryTracker msg="Failed to mmap" file=/prometheus-data/data/queries.active Attemptedsize=20001 err="invalid argument"
panic: Unable to create mmap-ed active query log

goroutine 1 [running]:
github.com/prometheus/prometheus/promql.NewActiveQueryTracker(0x7fff9917af38, 0x15, 0x14, 0x2a6b7c0, 0xc00003c7e0, 0x2a6b7c0)
	/app/promql/query_logger.go:112 +0x4d2
main.main()
	/app/cmd/prometheus/main.go:361 +0x52bd

And here’s the underlying problem — because of the way the persistent data is mapped into this container (via parallels sharing in this case), the mmap of the active queries file fails and prometheus fails to start.

In other words, since prometheus 2.12 your prometheus data files have to be stored on a filesystem which supports mmap. Additionally, there is no flag to just disable the active query logger.

So how do we work around this? Well, here’s a horrible workaround — in the data directory that is volume mapped into the container, create a symlink that is to a path that is mmapable inside the docker container, even if that path doesn’t exist outside the container. For example, given that we store the prometheus time series at $CONFIG/prometheus-data:

$ ln -s /tmp/queries.active "$CONFIG/prometheus-data/queries.active"

Note that /tmp/queries.active does not exist on the developer’s mac. Prometheus now starts and its puppies and kittens the whole way down.

An animated GIF resume

The graphic designer at work and I were talking, and I challenged him to come up with a resume as an animated GIF. This is where he landed…

I think its quite clever. Need a graphic designer or video team? Consider onefishsea.

The wonderful world of machine learning automated lego sorting

Inspired by Alastair D’Silva‘s cunning plans for world domination, I’ve been googling around for automated lego sorting systems recently. This seems like a nice tractable machine learning problem with some robotics thrown in for fun.

Some cool projects if you’re that way inclined:

- Sorting 2 Metric Tons of Lego
- A lego sorter using tensorflow

This sounds like a great way to misspend some evenings to me…

What is the Spotify model for Agile?

The other day someone said to me that “they use the Spotify development model”, and I said “you who the what now?”. It was a super productive conversation that I am quite proud of.

So… in order to look like less of a n00b in the next conversation, what is the “Spotify development model”? Well, it turns out the Spotify came up with a series of tweaks to the standard Agile process in order to scale their engineering teams. If you google for “spotify development model” or “spotify agile” you’ll get lots and lots of third party blog posts about what Spotify did (I guess a bit like this one), but its surprisingly hard to find primary sources. The best I’ve found so far is this Quora answer from a former VP of Engineering at Spotify, although some of the resources he links to no longer exist.

Continue reading “What is the Spotify model for Agile?”

Quick hack: extracting the contents of a Docker image to disk

For various reasons, I wanted to inspect the contents of a Docker image without starting a container. Docker makes it easy to get an image as a tar file, like this:

docker save -o foo.tar image

But if you extract that tar file you’ll find a configuration file and manifest as JSON files, and then a series of tar files, one per image layer. You use the manifest to determine in what order you extract the tar files to build the container filesystem.

That’s fiddly and annoying. So I wrote this quick python hack to extract an image tarball into a directory on disk that I could inspect:

#!/usr/bin/python3

# Call me like this:
#  docker-image-extract tarfile.tar extracted

import tarfile
import json
import os
import sys

image_path = sys.argv[1]
extracted_path = sys.argv[2]

image = tarfile.open(image_path)
manifest = json.loads(image.extractfile('manifest.json').read())

for layer in manifest[0]['Layers']:
    print('Found layer: %s' % layer)
    layer_tar = tarfile.open(fileobj=image.extractfile(layer))

    for tarinfo in layer_tar:
        print('  ... %s' % tarinfo.name)
        if tarinfo.isdev():
            print('  --> skip device files')
            continue

        dest = os.path.join(extracted_path, tarinfo.name)
        if not tarinfo.isdir() and os.path.exists(dest):
            print('  --> remove old version of file')
            os.unlink(dest)

        layer_tar.extract(tarinfo, path=extracted_path)

Hopefully that’s useful to someone else (or future me).

Mastermind in JavaScript

I’ve been learning JavaScript for the last few days, and I figured I’d implement Jacqui’s favourite board game as a learning exercise. Jacqui loves a simple colour guessing game called Mastermind. In the game someone picks four coloured pins and then the player has to progressively guess what those colours are.

In my JavaScript version the computer picks four colours, and you need to work out what they are. Click on the white squares to cycle through colours and then hit the “guess” button when you’re ready to see how many you got right. The gray boxes in the top row will progressively reveal their colours as you guess them.

The code is here, and the game can be played here.

A nerd snipe, in which I reverse engineer the Aussie Broadband usage API

I was curious about the newly available FTTN NBN service in my area, so I signed up to see what’s what. Of course, I need a usage API so that I can graph my usage in prometheus and grafana as everyone does these days. So I asked Aussie. The response I got was that I was welcome to reverse engineer the REST API that the customer portal uses.

So I did.

I give you my super simple implementation of an Aussie Broadband usage client in Python. Patches of course are welcome.

I’ve now released the library on pypi under the rather innovative name of “aussiebb”, so installing it is as simple as:

$ pip install aussiebb