From cb8f648c320c758a207c3d72ae6e7680e2ce498e Mon Sep 17 00:00:00 2001
From: Grzegorz Bizon <grzesiek.bizon@gmail.com>
Date: Wed, 9 May 2018 12:47:26 +0200
Subject: [PATCH] Add tmp index to ci_builds to optimize stages migration

---
 ...420010616_cleanup_build_stage_migration.rb | 35 ++++++++++++++++++-
 1 file changed, 34 insertions(+), 1 deletion(-)

diff --git a/db/migrate/20180420010616_cleanup_build_stage_migration.rb b/db/migrate/20180420010616_cleanup_build_stage_migration.rb
index 0342695ec3d..24777294101 100644
--- a/db/migrate/20180420010616_cleanup_build_stage_migration.rb
+++ b/db/migrate/20180420010616_cleanup_build_stage_migration.rb
@@ -2,6 +2,7 @@ class CleanupBuildStageMigration < ActiveRecord::Migration
   include Gitlab::Database::MigrationHelpers
 
   DOWNTIME = false
+  TMP_INDEX = 'tmp_id_stage_partial_null_index'.freeze
 
   disable_ddl_transaction!
 
@@ -13,16 +14,48 @@ class CleanupBuildStageMigration < ActiveRecord::Migration
   end
 
   def up
+    disable_statement_timeout
+
+    ##
+    # We steal from the background migrations queue to catch up with the
+    # scheduled migrations set.
+    #
     Gitlab::BackgroundMigration.steal('MigrateBuildStage')
 
+    ##
+    # We add temporary index, to make iteration over batches more performant.
+    # Conditional here is to avoid the need of doing that in a separate
+    # migration file to make this operation idempotent.
+    #
+    unless index_exists_by_name?(:ci_builds, TMP_INDEX)
+      add_concurrent_index(:ci_builds, :id, where: 'stage_id IS NULL', name: TMP_INDEX)
+    end
+
+    ##
+    # We check if there are remaining rows that should be migrated (for example
+    # if Sidekiq / Redis fails / is restarted, what could result in not all
+    # background migrations being executed correctly.
+    #
+    # We migrate remaining rows synchronously in a blocking way, to make sure
+    # that when this migration is done we are confident that all rows are
+    # already migrated.
+    #
     Build.where('stage_id IS NULL').each_batch(of: 50) do |batch|
       range = batch.pluck('MIN(id)', 'MAX(id)').first
 
       Gitlab::BackgroundMigration::MigrateBuildStage.new.perform(*range)
     end
+
+    ##
+    # We remove temporary index, because it is not required during standard
+    # operations and runtime.
+    #
+    remove_concurrent_index_by_name(:ci_builds, TMP_INDEX)
   end
 
   def down
-    # noop
+    if index_exists_by_name?(:ci_builds, TMP_INDEX)
+      remove_concurrent_index_by_name(:ci_builds, TMP_INDEX)
+    end
   end
 end