diff options
Diffstat (limited to 'lib')
| -rw-r--r-- | lib/gitlab/background_migration/delete_diff_files.rb | 83 | ||||
| -rw-r--r-- | lib/gitlab/background_migration/schedule_diff_files_deletion.rb | 61 |
2 files changed, 58 insertions, 86 deletions
diff --git a/lib/gitlab/background_migration/delete_diff_files.rb b/lib/gitlab/background_migration/delete_diff_files.rb index e2c90fce6b1..36161285aac 100644 --- a/lib/gitlab/background_migration/delete_diff_files.rb +++ b/lib/gitlab/background_migration/delete_diff_files.rb @@ -8,6 +8,7 @@ module Gitlab self.table_name = 'merge_request_diffs' belongs_to :merge_request + has_many :merge_request_diff_files include EachBatch end @@ -18,41 +19,73 @@ module Gitlab include EachBatch end - def perform(merge_request_diff_id) - merge_request_diff = MergeRequestDiff.find_by(id: merge_request_diff_id) + BATCH = 5_000 + DEAD_TUPLES_THRESHOLD = 50_000 + VACUUM_WAIT_TIME = 5.minutes - return unless merge_request_diff - return unless should_delete_diff_files?(merge_request_diff) + def perform + diffs_with_files = MergeRequestDiff + .joins(:merge_request) + .where("merge_requests.state = 'merged'") + .where('merge_requests.latest_merge_request_diff_id IS NOT NULL') + .where('merge_requests.latest_merge_request_diff_id != merge_request_diffs.id') + .where("merge_request_diffs.state NOT IN ('without_files', 'empty')") - MergeRequestDiff.transaction do - merge_request_diff.update_column(:state, 'without_files') - - # explain (analyze, buffers) when deleting 453 diff files: - # - # Delete on merge_request_diff_files (cost=0.57..8487.35 rows=4846 width=6) (actual time=43.265..43.265 rows=0 loops=1) - # Buffers: shared hit=2043 read=259 dirtied=254 - # -> Index Scan using index_merge_request_diff_files_on_mr_diff_id_and_order on merge_request_diff_files (cost=0.57..8487.35 rows=4846 width=6) (actu - # al time=0.466..26.317 rows=453 loops=1) - # Index Cond: (merge_request_diff_id = 463448) - # Buffers: shared hit=17 read=84 - # Planning time: 0.107 ms - # Execution time: 43.287 ms - # - MergeRequestDiffFile.where(merge_request_diff_id: merge_request_diff.id).delete_all + diffs_with_files.each_batch(of: BATCH) do |batch, index| + wait_deadtuple_vacuum(index) + prune_diff_files(batch, index) + end + end + + def wait_deadtuple_vacuum(index) + db_klass = Gitlab::Database + + if defined?(db_klass) && db_klass.respond_to?(:postgresql?) && db_klass.postgresql? + while diff_files_dead_tuples_count >= DEAD_TUPLES_THRESHOLD + log_info("Dead tuple threshold hit on merge_request_diff_files (#{index}th batch): " \ + "#{diff_files_dead_tuples_count}, waiting 5 minutes") + sleep VACUUM_WAIT_TIME + end end end private - def should_delete_diff_files?(merge_request_diff) - return false if merge_request_diff.state == 'without_files' + def diff_files_dead_tuples_count + dead_tuple = + execute_statement("SELECT n_dead_tup FROM pg_stat_all_tables "\ + "WHERE relname = 'merge_request_diff_files'")[0] + + if dead_tuple.present? + dead_tuple['n_dead_tup'].to_i + else + 0 + end + end - merge_request = merge_request_diff.merge_request + def prune_diff_files(batch, index) + diff_ids = batch.pluck(:id) - return false unless merge_request.state == 'merged' - return false if merge_request_diff.id == merge_request.latest_merge_request_diff_id + removed = 0 + updated = 0 + + MergeRequestDiff.transaction do + updated = MergeRequestDiff.where(id: diff_ids) + .update_all(state: 'without_files') + removed = MergeRequestDiffFile.where(merge_request_diff_id: diff_ids) + .delete_all + end + + log_info("#{index}th batch - Removed #{removed} merge_request_diff_files rows, "\ + "updated #{updated} merge_request_diffs rows") + end + + def execute_statement(sql) + ActiveRecord::Base.connection.execute(sql) + end - true + def log_info(message) + Rails.logger.info("BackgroundMigration::DeleteDiffFiles - #{message}") end end end diff --git a/lib/gitlab/background_migration/schedule_diff_files_deletion.rb b/lib/gitlab/background_migration/schedule_diff_files_deletion.rb deleted file mode 100644 index 6442468836b..00000000000 --- a/lib/gitlab/background_migration/schedule_diff_files_deletion.rb +++ /dev/null @@ -1,61 +0,0 @@ -# frozen_string_literal: true -# rubocop:disable Metrics/AbcSize -# rubocop:disable Style/Documentation - -module Gitlab - module BackgroundMigration - class ScheduleDiffFilesDeletion - class MergeRequestDiff < ActiveRecord::Base - self.table_name = 'merge_request_diffs' - - has_many :merge_request_diff_files - - include EachBatch - end - - ITERATION_BATCH = 1000 - DELETION_BATCH = 1000 # per minute - MIGRATION = 'DeleteDiffFiles' - - # Considering query times and Redis writings, this should take around 2 - # hours to complete. - def perform - diffs_with_files = MergeRequestDiff.where.not(state: %w(without_files empty)) - - # This will be increased for each scheduled job - process_job_in = 1.second - - # explain (analyze, buffers) example for the iteration: - # - # Index Only Scan using tmp_index_20013 on merge_request_diffs (cost=0.43..1630.19 rows=60567 width=4) (actual time=0.047..9.572 rows=56976 loops=1) - # Index Cond: ((id >= 764586) AND (id < 835298)) - # Heap Fetches: 8 - # Buffers: shared hit=18188 - # Planning time: 0.752 ms - # Execution time: 12.430 ms - # - diffs_with_files.reorder(nil).each_batch(of: ITERATION_BATCH) do |relation, scheduler_index| - relation.each do |diff| - BackgroundMigrationWorker.perform_in(process_job_in, MIGRATION, [diff.id]) - - diff_files_count = diff.merge_request_diff_files.reorder(nil).count - - # We should limit on 1000 diff files deletion per minute to avoid - # replication lag issues. - # - interval = (diff_files_count.to_f / DELETION_BATCH).minutes - process_job_in += interval - end - end - - log_days_to_process_all_jobs(process_job_in) - end - - def log_days_to_process_all_jobs(seconds_to_process) - days_to_process_all_jobs = (seconds_to_process / 60 / 60 / 24).to_i - Rails.logger.info("Gitlab::BackgroundMigration::DeleteDiffFiles will take " \ - "#{days_to_process_all_jobs} days to be processed") - end - end - end -end |