Docker Service Discovery and Load Balancing with Haystack

Haystack provides service discovery and automatic HTTP load balancing in a Docker environment. It uses its own DNS to provide service discovery and dynamically updates its embedded load balancer as services are started or stopped.

Haystack solves two problems:

  • Service Discovery – Haystack automatically manages service registration within DNS by monitoring the Docker container lifecycle. Haystack introspects container metadata exposing available services within its own DNS. Haystack automatically manages the registration of services into a common DNS for discovery by other services.
  • Load balancing of a service as it is scaled up or down to meet demand – Haystack registers DNS service names that are load balanced over the available instances as the service is scaled up or down to meet demand. Each service name is registered in Haystack’s own DNS and Haystack manages the load balancing of the containers providing that service, from one to thousands and back again – dynamically as demand rises or falls.

This post walks through a demo of some of capabilities that Haystack provides using a simple HTTP based micro service. The service accepts HTTP GET requests and responds with the hostname of the container and the HTTP path used in the request.

To start a number of services to demonstrate load balancing in Haystack:

for i in {1..5}; do
     docker run \
     --name demo-$(printf %03d $i) \
     -d shortishly/haystack_demo;
done

Start Haystack – replace 172.16.1.218 with the location of your Docker Engine that is already listening on a tcp port.

docker run \
    -p 8080:80 \
    -e DOCKER_HOST=tcp://172.16.1.218:2375 \
    -d \
    --name haystack \
    shortishly/haystack

You should now have Haystack and 5 demo micro services running within Docker:

# docker ps --format="{{.Names}}"

haystack
demo-001
demo-002
demo-003
demo-004
demo-005

Now start a busybox that uses Haystack for DNS resolution as follows:

  docker run \
  --dns=$(docker inspect --format='{{.NetworkSettings.IPAddress}}' haystack) \
  --tty \
  --interactive \
  --rm busybox /bin/sh

In the busybox shell:

wget \
   -q \
   -O /dev/stdout \
    http://demo.haystack_demo.services.haystack/this/is/a/demo

Haystack has registered demo.haystack_demo.services.haystack in
its own DNS service, and is load balancing any
request randomly to one of the demo-001, demo-002,
demo-003, demo-003, demo-004 or demo-005
containers.

If you make a number of wget requests to the same URL you will
get responses from the different containers at random:

# wget -q -O /dev/stdout http://demo.haystack_demo.services.haystack/load/balancing
d617596e70da: /load/balancing
# wget -q -O /dev/stdout http://demo.haystack_demo.services.haystack/load/balancing
96c6c6f27f03: /load/balancing
# wget -q -O /dev/stdout http://demo.haystack_demo.services.haystack/load/balancing
296b5208edf9: /load/balancing
# wget -q -O /dev/stdout http://demo.haystack_demo.services.haystack/load/balancing
1166e110e70d: /load/balancing
# wget -q -O /dev/stdout http://demo.haystack_demo.services.haystack/load/balancing
9909343b937a: /load/balancing

You can verify which services are available in Haystack by curling /api/info:

curl -s http://$(docker inspect --format='{{.NetworkSettings.IPAddress}}' haystack)/api/info|python -m json.tool

Use you bowser to visit the Haystack UI running on your Docker host:

http://YOUR_DOCKER_HOST:8080/ui

Building Erlang/OTP docker releases with gitlab CI

Gitlab has awesome integrated support for CI. Building your next Erlang/OTP release into a Docker image could be as simple as committing a .gitlab-ci.yml to your project.

To run Gitlab in Docker locally follow these instructions. The command that I use is:

docker run --detach \
    --hostname YOUR_GITLAB_HOSTNAME \
    --publish 443:443 --publish 80:80 --publish 22:22 \
    --name gitlab \
    --restart always \
    --volume /data/gitlab/config:/etc/gitlab:Z \
    --volume /data/gitlab/logs:/var/log/gitlab:Z \
    --volume /data/gitlab/data:/var/opt/gitlab:Z \
    --volume /data/gitlab/logs/reconfigure:/var/log/gitlab/reconfigure:Z \
    gitlab/gitlab-ce:latest

My docker host is running Fedora with SE linux enabled, the “Z” on the volume mounts does some labelling magic making docker and SE happy together.

Our next step is to start a gitlab-runner to orchestrate our builds, which also can run as a Docker container:

docker run -d --name gitlab-runner --restart always \
  -v /var/run/docker.sock:/var/run/docker.sock:Z \
  -v /data/gitlab-runner/config:/etc/gitlab-runner:Z \
  gitlab/gitlab-runner:latest

Note that I use /data as the host volume mount – replace with your preferred local mount point. We will be editing the configuration created in /data/gitlab-runner/config in a moment. We need to register this runner with gitlab, as follows:

docker exec -it gitlab-runner gitlab-runner register \
  --url "http://YOUR_GITLAB_HOSTNAME/ci" \
  --registration-token "YOUR_GITLAB_TOKEN" \
  --description "docker-ruby-2.1" \
  --executor "docker" \
  --docker-image ruby:2.1

Finally we need to volume mount /var/run/docker.sock into the runner so that we can run docker within the runner. Modify /data/gitlab-runner/config/config.toml so that the volumes include:

  volumes = ["/cache", "/var/run/docker.sock:/var/run/docker.sock"]

Restart the runner with:

docker stop gitlab-runner && docker start gitlab-runner

Phew! You now have a gitlab runner setup ready to build some software! Gitlab will build a project providing it has a .gitlab-ci.yml present in the root directory of that project:

image: shortishly/docker-erlang

stages:
  - build

app:
  stage: build
  script:
    - make
    - docker login -e $DOCKER_EMAIL -u $DOCKER_USERNAME -p $DOCKER_PASSWORD
    - docker tag $(bin/release_name):$(bin/version) $DOCKER_USERNAME/$(bin/app):$(bin/version)
    - docker push $DOCKER_USERNAME/$(bin/app):$(bin/version)
  only:
    - master

The image shortishly/docker-erlang is a container (built from this Dockerfile) that is from the official erlang docker image, but also has Docker installed so that we can build, tag and push.

The bin/release_name, bin/version are bin/app helper scripts from this project. The above is an erlang.mk based build using a fork that automatically creates a Docker container from Erlang/OTP release. To use in your own project replace the existing erlang.mk by using:

wget -q https://github.com/shortishly/erlang.mk/raw/master/erlang.mk

Commit the new erlang.mk and .gitlab-ci.yml and check Gitlab to see whether the build is progressing. Assuming everything is OK a new docker image will be pushed with your release.

docker from scratch in erlang.mk

I have been using erlang.mk for a couple of years now on various projects – just a Makefile, but what a Makefile! I’ve recently added a fork that includes support for creating docker containers from scratch.

On MacOS you might want to run shortishly/docker-erlang which will give you erlang and docker packaged together in a shell (brew install docker-machine if you don’t already have it):

docker run \
       -v /var/run/docker.sock:/var/run/docker.sock \
       -t \
       -i \
       --rm \
       shortishly/docker-erlang \
       /bin/bash

This is a short tutorial that assumes you are running on Linux, with both erlang and docker installed:

First lets create and move into a directory for this tutorial:

mkdir demo && cd demo

We need a copy of erlang.mk from the fork that I have created:

wget -q https://github.com/shortishly/erlang.mk/raw/master/erlang.mk

Bootstrap a new erlang.mk project as you would do normally, including a release template:

make -f erlang.mk bootstrap
make bootstrap-rel

Also bootstrap this project to create a Docker container:

make bootstrap-docker

The “bootstrap-docker” target has created a Dockerfile for us:

FROM scratch

ARG REL_NAME
ARG REL_VSN=1
ARG ERTS_VSN

ENV BINDIR /erts-7.2.1/bin
ENV BOOT /releases/1/demo_release
ENV CONFIG /releases/1/sys.config
ENV ARGS_FILE /releases/1/vm.args

ENV TZ=GMT

ENTRYPOINT exec ${BINDIR}/erlexec -boot_var /lib -boot ${BOOT} -noinput -config ${CONFIG} -args_file ${ARGS_FILE}

ADD _rel/demo_release/ /

Finally! Lets build the demo:

make

There should be lots of output at this point. At the end you should see:

===> Starting relx build process ...
===> Resolving OTP Applications from directories:
          /home/pmorgan/src/git/demo/ebin
          /home/pmorgan/opt/erlang/18.2.1/lib
          /home/pmorgan/src/git/demo/apps
          /home/pmorgan/src/git/demo/deps
===> Resolved demo_release-1
===> Including Erts from /home/pmorgan/opt/erlang/18.2.1
===> release successfully created!
 GEN    docker-scratch-cp-dynamic-libs
 GEN    docker-scratch-cp-link-loader
 GEN    docker-scratch-cp-sh
 GEN    docker-strip-erts-binaries
 GEN    docker-build
sha256:6440d984e0efb5a7f10168e805cdfcf47f670b03ce953427ba2613b177072d8c

A release has been successfully created “demo_release-1”, together with a bunch of new docker based targets and a new docker image (you will have a different sh256 output). You can check the output of docker images:

docker images

REPOSITORY                 TAG                 IMAGE ID            CREATED             SIZE
demo_release               0.0.1               6440d984e0ef        3 minutes ago       22.02 MB

You can run the image with make docker-run:

make docker-run

 GEN    docker-rm
 GEN    docker-run
b9deb1f7ef337540a6d17731ef309768c22cc0fa9703ea68be819895e4d78732

Again, you will have a different hash shown for the container that is now running – you can check that it is running with docker ps -a:

docker ps -a

CONTAINER ID        IMAGE                COMMAND                  CREATED              STATUS              PORTS               NAMES
b9deb1f7ef33        demo_release:0.0.1   "/bin/sh -c 'exec ${B"   About a minute ago   Up About a minute                       demo_release

You can use the shortcut make docker-logs to display the logs from your container:

make docker-logs

 GEN    docker-logs
heart_beat_kill_pid = 1

So with the help of erlang.mk we have been able to bootstrap an erlang project, building and running a docker container in about 10 shell commands.

erlang in docker from scratch

When packaging an application as a Docker container it is too easy to just be lazy and put FROM debian (other distributions are available, replace debian with your distribution of choice). For sure it is going to work, but you have just included dozens of libraries and binaries that your application just does not need. An image that could be tens of megabytes is now at least several hundred – we are building containers not virtual machines here!

One of the things I like about Go is that typical application binaries are small with no runtime dependencies. Fewer dependencies mean less patching and security headaches. The less friction in the CI build cycle, the better. Go achieves this by having statically linked applications meaning that just one binary is necessary in ADD, and they are typically built from scratch (etcd as a good example).

Erlang was designed to be embedded in telecoms equipment, so we must be able to package applications in Docker with a small footprint too?

An example repository containing a regular erlang application that is packaged in a docker container from scratch. You will need both erlang installed and docker service running preferably on a Linux environment. The release needs to be built on Linux to be able to run on Linux because we are going include the ERTS.

On MacOS you might want to run shortishly/docker-erlang which will give you erlang and docker packaged together in a shell (brew install docker-machine if you don’t already have it):

docker run \
       -v /var/run/docker.sock:/var/run/docker.sock \
       -t \
       -i \
       --rm \
       shortishly/docker-erlang \
       /bin/bash

Clone and build the erlang-in-docker-from-scratch repository, which contains a minimal erlang application that builds a release into the _rel directory:

git clone https://github.com/shortishly/erlang-in-docker-from-scratch.git eidfs
cd eidfs
make

At the end of make a standard erlang release for the eidfs application is now present in the _rel directory. To make it run inside a scratch container we need to include any runtime dependencies too. This is where mkimage comes in:

./bin/mkimage

The ./bin/mkimage script copies in any dynamic libraries that ERTS needs to run the erlang release:

'/lib/x86_64-linux-gnu/libc.so.6' -> '_rel/eidfs/lib/x86_64-linux-gnu/libc.so.6'
'/lib/x86_64-linux-gnu/libdl.so.2' -> '_rel/eidfs/lib/x86_64-linux-gnu/libdl.so.2'
'/lib/x86_64-linux-gnu/libm.so.6' -> '_rel/eidfs/lib/x86_64-linux-gnu/libm.so.6'
'/lib/x86_64-linux-gnu/libpthread.so.0' -> '_rel/eidfs/lib/x86_64-linux-gnu/libpthread.so.0'
'/lib/x86_64-linux-gnu/librt.so.1' -> '_rel/eidfs/lib/x86_64-linux-gnu/librt.so.1'
'/lib/x86_64-linux-gnu/libtinfo.so.5' -> '_rel/eidfs/lib/x86_64-linux-gnu/libtinfo.so.5'
'/lib/x86_64-linux-gnu/libutil.so.1' -> '_rel/eidfs/lib/x86_64-linux-gnu/libutil.so.1'
'/lib/x86_64-linux-gnu/libz.so.1' -> '_rel/eidfs/lib/x86_64-linux-gnu/libz.so.1'

It also copies /bin/sh so that we can run the release too. We can build a docker image for the release using the following command:

docker build \
       --build-arg REL_NAME=$(bin/release_name) \
       --build-arg ERTS_VSN=$(bin/system_version) \
       --pull=true \
       --no-cache=true \
       --force-rm=true \
       -t $(bin/release_name):$(bin/version) .

The $(bin/release_name), $(bin/system_version) and $(bin/version) are short escripts that respond with the release name, system ERTS version and the application version respectively.

Quite a lot of effort, what is the reward? Try docker images and look at the size of the resultant container:

REPOSITORY                 TAG                 IMAGE ID            CREATED             SIZE
eidfs                      0.0.1               6748931f94e4        4 seconds ago       16.74 MB

We have a docker packaged erlang release in ~17MB. Lets run it!

docker run \
       --name $(bin/release_name) \
       -d \
       $(bin/release_name):$(bin/version)

Check the logs using docker logs $(bin/release_name) and you will see lots of application startup messages from SASL.

You might notice that the ENTRYPOINT used in the Dockerfile directly invokes erlexec. I have done this to reduce dependencies further so that the release, ERTS dynamic libraries, and /bin/bash only are present in the container.

FROM scratch
MAINTAINER Peter Morgan <peter.james.morgan@gmail.com>

ARG REL_NAME
ARG REL_VSN=1
ARG ERTS_VSN

ENV BINDIR /erts-${ERTS_VSN}/bin
ENV BOOT /releases/1/${REL_NAME}
ENV CONFIG /releases/${REL_VSN}/sys.config
ENV ARGS_FILE /releases/${REL_VSN}/vm.args

ENV TZ=GMT

ENTRYPOINT exec ${BINDIR}/erlexec \
           -boot_var /lib \
           -boot ${BOOT} \
           -noinput \
           -config ${CONFIG} \
           -args_file ${ARGS_FILE}

ADD _rel/${REL_NAME}/ /