Speedup Test::Unit + RSpec + Cucumber by running parallel on multiple CPUs (or cores).
ParallelTests splits tests into even groups(by number of tests or runtime) and runs each group in a single process with its own database.
If you use RSpec: ensure you got >= 2.4
As gem
# add to Gemfile
gem "parallel_tests", :group => :development
OR as plugin
rails plugin install git://github.com/grosser/parallel_tests.git
# add to Gemfile
gem "parallel", :group => :development
As gem
gem install parallel_tests
# add to config/environments/development.rb
config.gem "parallel_tests"
# add to Rakefile
begin; require 'parallel_tests/tasks'; rescue LoadError; end
OR as plugin
gem install parallel
# add to config/environments/development.rb
config.gem "parallel"
./script/plugin install git://github.com/grosser/parallel_tests.git
ParallelTests uses 1 database per test-process, 2 processes will use *_test
and *_test2
.
test:
database: xxx_test<%= ENV['TEST_ENV_NUMBER'] %>
rake parallel:create
rake parallel:prepare
rake parallel:test # Test::Unit
rake parallel:spec # RSpec
rake parallel:features # Cucumber
rake parallel:test[1] --> force 1 CPU --> 86 seconds
rake parallel:test --> got 2 CPUs? --> 47 seconds
rake parallel:test --> got 4 CPUs? --> 26 seconds
...
Test by pattern (e.g. use one integration server per subfolder / see if you broke any 'user'-related tests)
rake parallel:test[^unit] # everything in test/unit folder (every test file matching /^unit/)
rake parallel:test[user] # run users_controller + user_helper + user tests
rake parallel:test['user|product'] # run user and product related tests
2 processes for 210 specs, ~ 105 specs per process
... test output ...
843 examples, 0 failures, 1 pending
Took 29.925333 seconds
Log test runtime to give each process the same runtime.
Add to your spec/parallel_spec.opts
(or spec/spec.opts
) :
RSpec 1.x:
--format progress
--format ParallelSpecs::SpecRuntimeLogger:tmp/parallel_profile.log
RSpec >= 2.4:
If installed as plugin: -I vendor/plugins/parallel_tests/lib
--format progress
--format ParallelSpecs::SpecRuntimeLogger --out tmp/parallel_profile.log
This logger logs the test output without the different processes overwriting each other.
Add the following to your spec/parallel_spec.opts
(or spec/spec.opts
) :
RSpec 1.x:
--format progress
--format ParallelSpecs::SpecSummaryLogger:tmp/spec_summary.log
RSpec >= 2.2:
If installed as plugin: -I vendor/plugins/parallel_tests/lib
--format progress
--format ParallelSpecs::SpecSummaryLogger --out tmp/spec_summary.log
This logger produces pasteable command-line snippets for each failed example.
E.g.
spec /path/to/my_spec.rb -e "should do something"
Add the following to your spec/parallel_spec.opts
(or spec/spec.opts
) :
RSpec 1.x:
--format progress
--format ParallelSpecs::SpecFailuresLogger:tmp/failing_specs.log
RSpec >= 2.4:
If installed as plugin: -I vendor/plugins/parallel_tests/lib
--format progress
--format ParallelSpecs::SpecFailuresLogger --out tmp/failing_specs.log
sudo gem install parallel_tests
# go to your project dir
parallel_test OR parallel_spec OR parallel_cucumber
# [Optional] use ENV['TEST_ENV_NUMBER'] inside your tests to select separate db/memcache/etc.
[optional] Only run selected files & folders:
parallel_test test/bar test/baz/xxx_text.rb
Options are:
-n [PROCESSES] How many processes to use, default: available CPUs
-p, --path [PATH] run tests inside this path only
--no-sort do not sort files before running them
-m, --multiply-processes [FLOAT] use given number as a multiplier of processes to run
-r, --root [PATH] execute test commands from this path
-e, --exec [COMMAND] execute this code parallel and with ENV['TEST_ENV_NUM']
-o, --test-options '[OPTIONS]' execute test commands with those options
-t, --type [TYPE] which type of tests to run? test, spec or features
--non-parallel execute same commands but do not in parallel, needs --exec
-v, --version Show Version
-h, --help Show this.
You can run any kind of code with -e / --execute
parallel_test -n 5 -e 'ruby -e "puts %[hello from process #{ENV[:TEST_ENV_NUMBER.to_s].inspect}]"'
hello from process "2"
hello from process ""
hello from process "3"
hello from process "5"
hello from process "4"
1 Process | 2 Processes | 4 Processes | |
RSpec spec-suite | 18s | 14s | 10s |
Rails-ActionPack | 88s | 53s | 44s |
- [Capybara + Selenium] add to env.rb:
Capybara.server_port = 8888 + ENV['TEST_ENV_NUMBER'].to_i
- [RSpec] add a
spec/parallel_spec.opts
to use different options, e.g. no --drb (default:spec/spec.opts
) - [RSpec] if something looks fishy try to delete
script/spec
- [RSpec] if
script/spec
is missing parallel:spec uses justspec
(which solves some issues with double-loaded environment.rb) - [RSpec] 'script/spec_server' or spork do not work in parallel
- [RSpec]
./script/generate rspec
if you are running rspec from gems (this plugin uses script/spec which may fail if rspec files are outdated) - [RSpec] remove --loadby from you spec/*.opts
- [Bundler] if you have a
Gemfile
thenbundle exec
will be used to run tests - Capybara setup
- Sphinx setup
- [SQL schema format] use :ruby schema format to get faster parallel:prepare`
export PARALLEL_TEST_PROCESSORS=X
in your environment and parallel_tests will use this number of processors by default- with zsh this would be
rake "parallel:prepare[3]"
- make jRuby compatible basics
- make windows compatible
inspired by pivotal labs
- Charles Finkel
- Indrek Juhkam
- Jason Morrison
- jinzhu
- Joakim Kolsjö
- Kevin Scaldeferri
- Kpumuk
- Maksim Horbul
- Pivotal Labs
- Rohan Deshpande
- Tchandy
- Terence Lee
- Will Bryant
- Fred Wu
- xxx
- Levent Ali
- Michael Kintzer
- nathansobo
- Joe Yates
- asmega
- Doug Barth
- Geoffrey Hichborn
Michael Grosser
[email protected]
Hereby placed under public domain, do what you want, just do not hold me accountable...