gitlab-org--gitlab-foss/spec/services/ci/retry_pipeline_service_spec.rb

464 lines
16 KiB
Ruby

# frozen_string_literal: true
require 'spec_helper'
RSpec.describe Ci::RetryPipelineService, '#execute' do
include ProjectForksHelper
let(:user) { create(:user) }
let(:project) { create(:project) }
let(:pipeline) { create(:ci_pipeline, project: project) }
let(:service) { described_class.new(project, user) }
let(:build_stage) { create(:ci_stage, name: 'build', position: 0, pipeline: pipeline) }
let(:test_stage) { create(:ci_stage, name: 'test', position: 1, pipeline: pipeline) }
let(:deploy_stage) { create(:ci_stage, name: 'deploy', position: 2, pipeline: pipeline) }
context 'when user has full ability to modify pipeline' do
before do
project.add_developer(user)
create(:protected_branch, :developers_can_merge,
name: pipeline.ref, project: project)
end
context 'when there are already retried jobs present' do
before do
create_build('rspec', :canceled, build_stage, retried: true)
create_build('rspec', :failed, build_stage)
end
it 'does not retry jobs that has already been retried' do
expect(statuses.first).to be_retried
expect { service.execute(pipeline) }
.to change { CommitStatus.count }.by(1)
end
end
context 'when there are failed builds in the last stage' do
before do
create_build('rspec 1', :success, build_stage)
create_build('rspec 2', :failed, test_stage)
create_build('rspec 3', :canceled, test_stage)
end
it 'enqueues all builds in the last stage' do
service.execute(pipeline)
expect(build('rspec 2')).to be_pending
expect(build('rspec 3')).to be_pending
expect(pipeline.reload).to be_running
end
end
context 'when there are failed or canceled builds in the first stage' do
before do
create_build('rspec 1', :failed, build_stage)
create_build('rspec 2', :canceled, build_stage)
create_build('rspec 3', :canceled, test_stage)
create_build('spinach 1', :canceled, deploy_stage)
end
it 'retries builds failed builds and marks subsequent for processing' do
service.execute(pipeline)
expect(build('rspec 1')).to be_pending
expect(build('rspec 2')).to be_pending
expect(build('rspec 3')).to be_created
expect(build('spinach 1')).to be_created
expect(pipeline.reload).to be_running
end
it 'changes ownership of subsequent builds' do
expect(build('rspec 2').user).not_to eq(user)
expect(build('rspec 3').user).not_to eq(user)
expect(build('spinach 1').user).not_to eq(user)
service.execute(pipeline)
expect(build('rspec 2').user).to eq(user)
expect(build('rspec 3').user).to eq(user)
expect(build('spinach 1').user).to eq(user)
end
end
context 'when there is failed build present which was run on failure' do
before do
create_build('rspec 1', :failed, build_stage)
create_build('rspec 2', :canceled, build_stage)
create_build('rspec 3', :canceled, test_stage)
create_build('report 1', :failed, deploy_stage)
end
it 'retries builds only in the first stage' do
service.execute(pipeline)
expect(build('rspec 1')).to be_pending
expect(build('rspec 2')).to be_pending
expect(build('rspec 3')).to be_created
expect(build('report 1')).to be_created
expect(pipeline.reload).to be_running
end
it 'creates a new job for report job in this case' do
service.execute(pipeline)
expect(statuses.find_by(name: 'report 1', status: 'failed')).to be_retried
end
end
context 'when there is a failed test in a DAG' do
before do
create_build('build', :success, build_stage)
create_build('build2', :success, build_stage)
test_build = create_build('test', :failed, test_stage, scheduling_type: :dag)
create(:ci_build_need, build: test_build, name: 'build')
create(:ci_build_need, build: test_build, name: 'build2')
end
it 'retries the test' do
service.execute(pipeline)
expect(build('build')).to be_success
expect(build('build2')).to be_success
expect(build('test')).to be_pending
expect(build('test').needs.map(&:name)).to match_array(%w(build build2))
end
context 'when there is a failed DAG test without needs' do
before do
create_build('deploy', :failed, deploy_stage, scheduling_type: :dag)
end
it 'retries the test' do
service.execute(pipeline)
expect(build('build')).to be_success
expect(build('build2')).to be_success
expect(build('test')).to be_pending
expect(build('deploy')).to be_pending
end
end
end
context 'when the last stage was skipped' do
before do
create_build('build 1', :success, build_stage)
create_build('test 2', :failed, test_stage)
create_build('report 3', :skipped, deploy_stage)
create_build('report 4', :skipped, deploy_stage)
end
it 'retries builds only in the first stage' do
service.execute(pipeline)
expect(build('build 1')).to be_success
expect(build('test 2')).to be_pending
expect(build('report 3')).to be_created
expect(build('report 4')).to be_created
expect(pipeline.reload).to be_running
end
end
context 'when pipeline contains manual actions' do
context 'when there are optional manual actions only' do
context 'when there is a canceled manual action in first stage' do
before do
create_build('rspec 1', :failed, build_stage)
create_build('staging', :canceled, build_stage, when: :manual, allow_failure: true)
create_build('rspec 2', :canceled, test_stage)
end
it 'retries failed builds and marks subsequent for processing' do
service.execute(pipeline)
expect(build('rspec 1')).to be_pending
expect(build('staging')).to be_manual
expect(build('rspec 2')).to be_created
expect(pipeline.reload).to be_running
end
it 'changes ownership of subsequent builds' do
expect(build('staging').user).not_to eq(user)
expect(build('rspec 2').user).not_to eq(user)
service.execute(pipeline)
expect(build('staging').user).to eq(user)
expect(build('rspec 2').user).to eq(user)
end
end
end
context 'when pipeline has blocking manual actions defined' do
context 'when pipeline retry should enqueue builds' do
before do
create_build('test', :failed, build_stage)
create_build('deploy', :canceled, build_stage, when: :manual, allow_failure: false)
create_build('verify', :canceled, test_stage)
end
it 'retries failed builds' do
service.execute(pipeline)
expect(build('test')).to be_pending
expect(build('deploy')).to be_manual
expect(build('verify')).to be_created
expect(pipeline.reload).to be_running
end
end
context 'when pipeline retry should block pipeline immediately' do
before do
create_build('test', :success, build_stage)
create_build('deploy:1', :success, test_stage, when: :manual, allow_failure: false)
create_build('deploy:2', :failed, test_stage, when: :manual, allow_failure: false)
create_build('verify', :canceled, deploy_stage)
end
it 'reprocesses blocking manual action and blocks pipeline' do
service.execute(pipeline)
expect(build('deploy:1')).to be_success
expect(build('deploy:2')).to be_manual
expect(build('verify')).to be_created
expect(pipeline.reload).to be_blocked
end
end
end
context 'when there is a skipped manual action in last stage' do
before do
create_build('rspec 1', :canceled, build_stage)
create_build('rspec 2', :skipped, build_stage, when: :manual, allow_failure: true)
create_build('staging', :skipped, test_stage, when: :manual, allow_failure: true)
end
it 'retries canceled job and reprocesses manual actions' do
service.execute(pipeline)
expect(build('rspec 1')).to be_pending
expect(build('rspec 2')).to be_manual
expect(build('staging')).to be_created
expect(pipeline.reload).to be_running
end
end
context 'when there is a created manual action in the last stage' do
before do
create_build('rspec 1', :canceled, build_stage)
create_build('staging', :created, test_stage, when: :manual, allow_failure: true)
end
it 'retries canceled job and does not update the manual action' do
service.execute(pipeline)
expect(build('rspec 1')).to be_pending
expect(build('staging')).to be_created
expect(pipeline.reload).to be_running
end
end
context 'when there is a created manual action in the first stage' do
before do
create_build('rspec 1', :canceled, build_stage)
create_build('staging', :created, build_stage, when: :manual, allow_failure: true)
end
it 'retries canceled job and processes the manual action' do
service.execute(pipeline)
expect(build('rspec 1')).to be_pending
expect(build('staging')).to be_manual
expect(pipeline.reload).to be_running
end
end
end
it 'closes all todos about failed jobs for pipeline' do
expect(::MergeRequests::AddTodoWhenBuildFailsService)
.to receive_message_chain(:new, :close_all)
service.execute(pipeline)
end
it 'reprocesses the pipeline' do
expect_any_instance_of(Ci::ProcessPipelineService).to receive(:execute)
service.execute(pipeline)
end
context 'when pipeline has processables with nil scheduling_type' do
let!(:build1) { create_build('build1', :success, build_stage) }
let!(:build2) { create_build('build2', :failed, build_stage) }
let!(:build3) { create_build('build3', :failed, test_stage) }
let!(:build3_needs_build1) { create(:ci_build_need, build: build3, name: build1.name) }
before do
statuses.update_all(scheduling_type: nil)
end
it 'populates scheduling_type of processables' do
service.execute(pipeline)
expect(build1.reload.scheduling_type).to eq('stage')
expect(build2.reload.scheduling_type).to eq('stage')
expect(build3.reload.scheduling_type).to eq('dag')
end
end
context 'when the pipeline is a downstream pipeline and the bridge is depended' do
let!(:bridge) { create(:ci_bridge, :strategy_depend, status: 'success') }
before do
create(:ci_sources_pipeline, pipeline: pipeline, source_job: bridge)
end
it 'marks source bridge as pending' do
service.execute(pipeline)
expect(bridge.reload).to be_pending
end
end
context 'when there are skipped jobs in later stages' do
before do
create_build('build 1', :success, build_stage)
create_build('test 2', :failed, test_stage)
create_build('report 3', :skipped, deploy_stage)
create_bridge('deploy 4', :skipped, deploy_stage)
end
it 'retries failed jobs and processes skipped jobs' do
service.execute(pipeline)
expect(build('build 1')).to be_success
expect(build('test 2')).to be_pending
expect(build('report 3')).to be_created
expect(build('deploy 4')).to be_created
expect(pipeline.reload).to be_running
end
end
context 'when user is not allowed to retry build' do
before do
build = create(:ci_build, pipeline: pipeline, status: :failed)
allow_next_instance_of(Ci::RetryJobService) do |service|
allow(service).to receive(:can?).with(user, :update_build, build).and_return(false)
end
end
it 'returns an error' do
response = service.execute(pipeline)
expect(response.http_status).to eq(:forbidden)
expect(response.errors).to include('403 Forbidden')
expect(pipeline.reload).not_to be_running
end
end
end
context 'when user is not allowed to retry pipeline' do
it 'returns an error' do
response = service.execute(pipeline)
expect(response.http_status).to eq(:forbidden)
expect(response.errors).to include('403 Forbidden')
expect(pipeline.reload).not_to be_running
end
end
context 'when user is not allowed to trigger manual action' do
before do
project.add_developer(user)
create(:protected_branch, :maintainers_can_push,
name: pipeline.ref, project: project)
end
context 'when there is a failed manual action present' do
before do
create_build('test', :failed, build_stage)
create_build('deploy', :failed, build_stage, when: :manual)
create_build('verify', :canceled, test_stage)
end
it 'returns an error' do
response = service.execute(pipeline)
expect(response.http_status).to eq(:forbidden)
expect(response.errors).to include('403 Forbidden')
expect(pipeline.reload).not_to be_running
end
end
context 'when there is a failed manual action in later stage' do
before do
create_build('test', :failed, build_stage)
create_build('deploy', :failed, test_stage, when: :manual)
create_build('verify', :canceled, deploy_stage)
end
it 'returns an error' do
response = service.execute(pipeline)
expect(response.http_status).to eq(:forbidden)
expect(response.errors).to include('403 Forbidden')
expect(pipeline.reload).not_to be_running
end
end
end
context 'when maintainer is allowed to push to forked project' do
let(:user) { create(:user) }
let(:project) { create(:project, :public) }
let(:forked_project) { fork_project(project) }
let(:pipeline) { create(:ci_pipeline, project: forked_project, ref: 'fixes') }
before do
project.add_maintainer(user)
create(:merge_request,
source_project: forked_project,
target_project: project,
source_branch: 'fixes',
allow_collaboration: true)
create_build('rspec 1', :failed, test_stage)
end
it 'allows to retry failed pipeline' do
allow_any_instance_of(Project).to receive(:branch_allows_collaboration?).and_return(true)
allow_any_instance_of(Project).to receive(:empty_repo?).and_return(false)
service.execute(pipeline)
expect(build('rspec 1')).to be_pending
expect(pipeline.reload).to be_running
end
end
def statuses
pipeline.reload.statuses
end
# The method name can be confusing because this can actually return both Ci::Build and Ci::Bridge
def build(name)
statuses.latest.find_by(name: name)
end
def create_build(name, status, stage, **opts)
create_processable(:ci_build, name, status, stage, **opts)
end
def create_bridge(name, status, stage, **opts)
create_processable(:ci_bridge, name, status, stage, **opts)
end
def create_processable(type, name, status, stage, **opts)
create(type, name: name,
status: status,
ci_stage: stage,
stage_idx: stage.position,
pipeline: pipeline, **opts) do |_job|
::Ci::ProcessPipelineService.new(pipeline).execute
end
end
end