Docker Community Forums

Share and learn in the Docker community.

Mounting Rails app through volume slows app precipitously


(Mitcholiver) #1

Expected behavior

A request through rails should respond in reasonable amount of time. Here, “reasonable amount of time” is measured by making a single request to a known URL in an environment built in Vagrant and Virtualbox. Each request made as the first request in a cold-started environment.

time curl -o - > /dev/null
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current  Dload  Upload   Total   Spent    Left  Speed
100  582k    0  582k    0     0  58401      0 --:--:--  0:00:10 --:--:--  148k

real    0m10.226s
user    0m0.004s
sys    0m0.010s

When placing my working directory in the container using ADD in my Dockerfile, I was able to achieve response speed close to that provided by the Vagrant/VM environment, averaging closer to 12s than 10s.

From my Dockerfile:


RUN mkdir /vagrant && mkdir /bundle
WORKDIR /vagrant

ADD . /vagrant

RUN bundle install --path /bundle --deployment

Actual behavior

When mounting my working directory using a volume, response times for the sample request reach an average of 56s.

From my Dockerfile:


RUN mkdir /vagrant && mkdir /bundle
WORKDIR /vagrant

ADD Gemfile /vagrant/Gemfile
ADD Gemfile.lock /vagrant/Gemfile.lock

RUN bundle install --path /bundle --deployment

VOLUME /vagrant

Using docker-compose, I point the PWD to /vagrant:

    build: .
    command: bundle exec rails s -p 3000 -b ''
      - ".:/vagrant"
      - "3000:3000"


Docker for Mac: version: mac-v1.12.0-beta20
OS X: version 10.11.5 (build: 15F34)
logs: /tmp/20160721-113714.tar.gz
[OK]     docker-cli
[OK]     app
[OK]     moby-syslog
[OK]     virtualization
[OK]     menubar
[OK]     system
[OK]     osxfs
[OK]     db
[OK]     slirp
[OK]     moby-console
[OK]     logs
[OK]     vmnetd
[OK]     env
[OK]     moby
[OK]     driver.amd64-linux