Merge branch 'ab/partitioning-detect-range' into 'master'

Detect ranges for time-space partitioning

See merge request gitlab-org/gitlab!34652

lib/gitlab/database/partitioning_migration_helpers/table_management_helpers.rb

@@ -24,14 +24,25 @@ module Gitlab
         #
         #   partition_table_by_date :audit_events, :created_at, min_date: Date.new(2020, 1), max_date: Date.new(2020, 6)
         #
-        # Required options are:
+        # Options are:
         #   :min_date - a date specifying the lower bounds of the partition range
-        #   :max_date - a date specifying the upper bounds of the partitioning range
+        #   :max_date - a date specifying the upper bounds of the partitioning range, defaults to today + 1 month
         #
-        def partition_table_by_date(table_name, column_name, min_date:, max_date:)
+        # Unless min_date is specified explicitly, we default to
+        # 1. The minimum value for the partitioning column in the table
+        # 2. If no data is present yet, the current month
+        def partition_table_by_date(table_name, column_name, min_date: nil, max_date: nil)
           assert_table_is_allowed(table_name)
+
           assert_not_in_transaction_block(scope: ERROR_SCOPE)
 
+          max_date ||= Date.today + 1.month
+
+          min_date ||= connection.select_one(<<~SQL)['minimum'] || max_date - 1.month
+            SELECT date_trunc('MONTH', MIN(#{column_name})) AS minimum
+            FROM #{table_name}
+          SQL
+
           raise "max_date #{max_date} must be greater than min_date #{min_date}" if min_date >= max_date
 
           primary_key = connection.primary_key(table_name)

spec/lib/gitlab/database/partitioning_migration_helpers/table_management_helpers_spec.rb

@@ -173,6 +173,77 @@ RSpec.describe Gitlab::Database::PartitioningMigrationHelpers::TableManagementHe
           '202003' => ["'2020-03-01 00:00:00'", "'2020-04-01 00:00:00'"]
         })
       end
+
+      context 'when min_date is not given' do
+        let(:template_table) { :todos }
+
+        context 'with records present already' do
+          before do
+            create(:todo, created_at: Date.parse('2019-11-05'))
+          end
+
+          it 'creates a partition spanning over each month from the first record' do
+            migration.partition_table_by_date template_table, partition_column, max_date: max_date
+
+            expect_range_partitions_for(partitioned_table, {
+              '000000' => ['MINVALUE', "'2019-11-01 00:00:00'"],
+              '201911' => ["'2019-11-01 00:00:00'", "'2019-12-01 00:00:00'"],
+              '201912' => ["'2019-12-01 00:00:00'", "'2020-01-01 00:00:00'"],
+              '202001' => ["'2020-01-01 00:00:00'", "'2020-02-01 00:00:00'"],
+              '202002' => ["'2020-02-01 00:00:00'", "'2020-03-01 00:00:00'"],
+              '202003' => ["'2020-03-01 00:00:00'", "'2020-04-01 00:00:00'"]
+            })
+          end
+        end
+
+        context 'without data' do
+          it 'creates the catchall partition plus two actual partition' do
+            migration.partition_table_by_date template_table, partition_column, max_date: max_date
+
+            expect_range_partitions_for(partitioned_table, {
+              '000000' => ['MINVALUE', "'2020-02-01 00:00:00'"],
+              '202002' => ["'2020-02-01 00:00:00'", "'2020-03-01 00:00:00'"],
+              '202003' => ["'2020-03-01 00:00:00'", "'2020-04-01 00:00:00'"]
+            })
+          end
+        end
+      end
+
+      context 'when max_date is not given' do
+        it 'creates partitions including the next month from today' do
+          today = Date.new(2020, 5, 8)
+
+          Timecop.freeze(today) do
+            migration.partition_table_by_date template_table, partition_column, min_date: min_date
+
+            expect_range_partitions_for(partitioned_table, {
+              '000000' => ['MINVALUE', "'2019-12-01 00:00:00'"],
+              '201912' => ["'2019-12-01 00:00:00'", "'2020-01-01 00:00:00'"],
+              '202001' => ["'2020-01-01 00:00:00'", "'2020-02-01 00:00:00'"],
+              '202002' => ["'2020-02-01 00:00:00'", "'2020-03-01 00:00:00'"],
+              '202003' => ["'2020-03-01 00:00:00'", "'2020-04-01 00:00:00'"],
+              '202004' => ["'2020-04-01 00:00:00'", "'2020-05-01 00:00:00'"],
+              '202005' => ["'2020-05-01 00:00:00'", "'2020-06-01 00:00:00'"],
+              '202006' => ["'2020-06-01 00:00:00'", "'2020-07-01 00:00:00'"]
+            })
+          end
+        end
+      end
+
+      context 'without min_date, max_date' do
+        it 'creates partitions for the current and next month' do
+          current_date = Date.new(2020, 05, 22)
+          Timecop.freeze(current_date.to_time) do
+            migration.partition_table_by_date template_table, partition_column
+
+            expect_range_partitions_for(partitioned_table, {
+              '000000' => ['MINVALUE', "'2020-05-01 00:00:00'"],
+              '202005' => ["'2020-05-01 00:00:00'", "'2020-06-01 00:00:00'"],
+              '202006' => ["'2020-06-01 00:00:00'", "'2020-07-01 00:00:00'"]
+            })
+          end
+        end
+      end
     end
 
     describe 'keeping data in sync with the partitioned table' do