1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
|
# frozen_string_literal: true
module Ci
module PipelineProcessing
class AtomicProcessingService
include Gitlab::Utils::StrongMemoize
include ExclusiveLeaseGuard
attr_reader :pipeline
DEFAULT_LEASE_TIMEOUT = 1.minute
BATCH_SIZE = 20
def initialize(pipeline)
@pipeline = pipeline
@collection = AtomicProcessingService::StatusCollection.new(pipeline)
end
def execute
return unless pipeline.needs_processing?
# Run the process only if we can obtain an exclusive lease; returns nil if lease is unavailable
success = try_obtain_lease { process! }
if success
if ::Feature.enabled?(:ci_reset_skipped_jobs_in_atomic_processing, project)
# If any jobs changed from stopped to alive status during pipeline processing, we must
# re-reset their dependent jobs; see https://gitlab.com/gitlab-org/gitlab/-/issues/388539.
new_alive_jobs.group_by(&:user).each do |user, jobs|
log_running_reset_skipped_jobs_service(jobs)
ResetSkippedJobsService.new(project, user).execute(jobs)
end
end
# Re-schedule if we need further processing
PipelineProcessWorker.perform_async(pipeline.id) if pipeline.needs_processing?
end
success
end
private
def process!
update_stages!
update_pipeline!
update_jobs_processed!
Ci::ExpirePipelineCacheService.new.execute(pipeline)
true
end
def update_stages!
pipeline.stages.ordered.each { |stage| update_stage!(stage) }
end
def update_stage!(stage)
# Update jobs for a given stage in bulk/slices
@collection
.created_job_ids_in_stage(stage.position)
.in_groups_of(BATCH_SIZE, false) { |ids| update_jobs!(ids) }
status = @collection.status_of_stage(stage.position)
stage.set_status(status)
end
def update_jobs!(ids)
created_jobs = pipeline
.current_processable_jobs
.id_in(ids)
.with_project_preload
.created
.ordered_by_stage
.select_with_aggregated_needs(project)
created_jobs.each { |job| update_job!(job) }
end
def update_pipeline!
pipeline.set_status(@collection.status_of_all)
end
def update_jobs_processed!
processing = @collection.processing_jobs
processing.each_slice(BATCH_SIZE) do |slice|
pipeline.all_jobs.match_id_and_lock_version(slice)
.update_as_processed!
end
end
def update_job!(job)
previous_status = status_of_previous_jobs(job)
# We do not continue to process the job if the previous status is not completed
return unless Ci::HasStatus::COMPLETED_STATUSES.include?(previous_status)
Gitlab::OptimisticLocking.retry_lock(job, name: 'atomic_processing_update_job') do |subject|
Ci::ProcessBuildService.new(project, subject.user)
.execute(subject, previous_status)
# update internal representation of job
# to make the status change of job to be taken into account during further processing
@collection.set_job_status(job.id, job.status, job.lock_version)
end
end
def status_of_previous_jobs(job)
if job.scheduling_type_dag?
# job uses DAG, get status of all dependent needs
@collection.status_of_jobs(job.aggregated_needs_names.to_a)
else
# job uses Stages, get status of prior stage
@collection.status_of_jobs_prior_to_stage(job.stage_idx.to_i)
end
end
# Gets the jobs that changed from stopped to alive status since the initial status collection
# was evaluated. We determine this by checking if their current status is no longer stopped.
def new_alive_jobs
initial_stopped_job_names = @collection.stopped_job_names
return [] if initial_stopped_job_names.empty?
new_collection = AtomicProcessingService::StatusCollection.new(pipeline)
new_alive_job_names = initial_stopped_job_names - new_collection.stopped_job_names
return [] if new_alive_job_names.empty?
pipeline
.current_jobs
.by_name(new_alive_job_names)
.preload(:user) # rubocop: disable CodeReuse/ActiveRecord
.to_a
end
def project
pipeline.project
end
def lease_key
"#{super}::pipeline_id:#{pipeline.id}"
end
def lease_timeout
DEFAULT_LEASE_TIMEOUT
end
def log_running_reset_skipped_jobs_service(jobs)
Gitlab::AppJsonLogger.info(
class: self.class.name.to_s,
message: 'Running ResetSkippedJobsService on new alive jobs',
project_id: project.id,
pipeline_id: pipeline.id,
user_id: jobs.first.user.id,
jobs_count: jobs.count
)
end
end
end
end
|