class method
self.parallelize
Ruby on Rails edge
Since v6.0.6.1Signature
self.parallelize(workers: :number_of_processors, with: :processes, threshold: ActiveSupport.test_parallelization_threshold, parallelize_databases: ActiveSupport.parallelize_test_databases, work_stealing: false)Parallelizes the test suite.
Takes a workers argument that controls how many times the process is forked. For each process a new database will be created suffixed with the worker number.
test-database_0
test-database_1
If ENV["PARALLEL_WORKERS"] is set the workers argument will be ignored and the environment variable will be used instead. This is useful for CI environments, or other environments where you may need more workers than you do for local testing. Note that setting PARALLEL_WORKERS will also bypass the threshold check, enabling parallelization regardless of test count.
If the number of workers is set to 1 or fewer, the tests will not be parallelized.
If workers is set to :number_of_processors, the number of workers will be set to the actual core count on the machine you are on.
The default parallelization method is to fork processes. If you’d like to use threads instead you can pass with: :threads to the parallelize method. Note the threaded parallelization does not create multiple databases and will not work with system tests.
parallelize(workers: :number_of_processors, with: :threads)
The threaded parallelization uses minitest’s parallel executor directly. The processes parallelization uses a Ruby DRb server.
Because parallelization presents an overhead, it is only enabled when the number of tests to run is above the threshold param. The default value is 50, and it’s configurable via config.active_support.test_parallelization_threshold.
Parallelization shuffles tests and distributes them to workers in round-robin order, so given the same seed value and worker count, the same sequence of tests will run on each worker. This makes flaky tests caused by parallel test interdependence somewhat easier to reproduce, though it is not a guarantee of deterministic test execution.
This can make test runtime slower and spikier when one worker gets most of the slow tests. Enable work_stealing to allow idle workers to steal tests from busy workers, smoothing out runtime at the cost of less reproducible flaky-test failures:
parallelize(workers: :number_of_processors, work_stealing: true)
If you want to skip Rails default creation of one database per process in favor of writing your own implementation, you can set parallelize_databases, or configure it via config.active_support.parallelize_test_databases.
parallelize(workers: :number_of_processors, parallelize_databases: false)
Note that your test suite may deadlock if you attempt to use only one database with multiple processes.
Parameters
-
workerskey = :number_of_processors -
withkey = :processes -
thresholdkey = ActiveSupport.test_parallelization_threshold -
parallelize_databaseskey = ActiveSupport.parallelize_test_databases -
work_stealingkey = false
Source
# File activesupport/lib/active_support/test_case.rb, line 128
def parallelize(workers: :number_of_processors, with: :processes, threshold: ActiveSupport.test_parallelization_threshold, parallelize_databases: ActiveSupport.parallelize_test_databases, work_stealing: false)
case
when ENV["PARALLEL_WORKERS"]
workers = ENV["PARALLEL_WORKERS"].to_i
when workers == :number_of_processors
workers = (Concurrent.available_processor_count || Concurrent.processor_count).floor
end
if with == :processes
ActiveSupport.parallelize_test_databases = parallelize_databases
end
Minitest.parallel_executor = ActiveSupport::Testing::ParallelizeExecutor.new(size: workers, with: with, threshold: threshold, work_stealing: work_stealing)
end
Defined in activesupport/lib/active_support/test_case.rb line 128
· View on GitHub
· Improve this page
· Find usages on GitHub
Defined in ActiveSupport::TestCase