Revert "Revert migration dropping columns in one release"

This reverts commit a2bcb0bec1de4a854f77656be9ca023e4b8b1a0e.

db/schema.rb

@@ -1063,7 +1063,6 @@ ActiveRecord::Schema.define(version: 20180608201435) do
   create_table "geo_hashed_storage_migrated_events", id: :bigserial, force: :cascade do |t|
     t.integer "project_id", null: false
     t.text "repository_storage_name", null: false
-    t.text "repository_storage_path"
     t.text "old_disk_path", null: false
     t.text "new_disk_path", null: false
     t.text "old_wiki_disk_path", null: false
@@ -1175,7 +1174,6 @@ ActiveRecord::Schema.define(version: 20180608201435) do
   create_table "geo_repository_created_events", id: :bigserial, force: :cascade do |t|
     t.integer "project_id", null: false
     t.text "repository_storage_name", null: false
-    t.text "repository_storage_path"
     t.text "repo_path", null: false
     t.text "wiki_path"
     t.text "project_name", null: false
@@ -1186,7 +1184,6 @@ ActiveRecord::Schema.define(version: 20180608201435) do
   create_table "geo_repository_deleted_events", id: :bigserial, force: :cascade do |t|
     t.integer "project_id", null: false
     t.text "repository_storage_name", null: false
-    t.text "repository_storage_path"
     t.text "deleted_path", null: false
     t.text "deleted_wiki_path"
     t.text "deleted_project_name", null: false
@@ -1197,7 +1194,6 @@ ActiveRecord::Schema.define(version: 20180608201435) do
   create_table "geo_repository_renamed_events", id: :bigserial, force: :cascade do |t|
     t.integer "project_id", null: false
     t.text "repository_storage_name", null: false
-    t.text "repository_storage_path"
     t.text "old_path_with_namespace", null: false
     t.text "new_path_with_namespace", null: false
     t.text "old_wiki_path_with_namespace", null: false

ee/db/post_migrate/20180417102933_drop_repository_storage_events_for_geo_events.rb

+# See http://doc.gitlab.com/ce/development/migration_style_guide.html
+# for more information on how to write migrations for GitLab.
+
+class DropRepositoryStorageEventsForGeoEvents < ActiveRecord::Migration
+  include Gitlab::Database::MigrationHelpers
+
+  DOWNTIME = false
+  BATCH_SIZE = 5_000
+  TABLES = %i(geo_hashed_storage_migrated_events geo_repository_created_events
+              geo_repository_deleted_events geo_repository_renamed_events)
+
+  def up
+    TABLES.each { |t| remove_column(t, :repository_storage_path) }
+  end
+
+  def down
+    TABLES.each { |t| add_column(t, :repository_storage_path, :text) }
+
+    add_shard_path
+  end
+
+  private
+
+  def add_shard_path
+    TABLES.each do |t|
+      min_id = 0
+
+      loop do
+        newest = newest_entry(t)
+        break unless newest
+        break if newest['repository_storage_path'].present?
+
+        new_batch = batch(t, min_id)
+        update_batch(t, new_batch)
+
+        min_id = new_batch.last.to_i
+      end
+
+      change_column_null(t, :repository_storage_path, true)
+    end
+  end
+
+  def newest_entry(table)
+    execute(
+      <<~SQL
+      SELECT id, repository_storage_path
+      FROM #{table}
+      ORDER BY id DESC
+      LIMIT 1;
+      SQL
+    ).first
+  end
+
+  def batch(table, min_id)
+    execute(
+      <<~SQL
+        SELECT id
+        FROM #{table}
+        WHERE id > #{min_id}
+        ORDER BY id ASC
+        LIMIT #{BATCH_SIZE};
+      SQL
+    ).map { |row| row['id'] }
+  end
+
+  def update_batch(table, ids)
+    execute(
+      <<~SQL
+      UPDATE #{table}
+      SET repository_storage_path =
+        CASE repository_storage_name
+          #{case_statements}
+        END
+      WHERE id IN (#{ids.join(',')})
+      SQL
+    )
+  end
+
+  def case_statements
+    statement = ""
+    Gitlab.config.repositories.storages.each do |shard, data|
+      statement << "WHEN '#{shard}' THEN '#{data.legacy_disk_path}'\n"
+    end
+
+    statement
+  end
+end

ee/spec/migrations/drop_repository_storage_events_for_geo_events_spec.rb

+# encoding: utf-8
+
+require 'spec_helper'
+require Rails.root.join('ee', 'db', 'post_migrate', '20180417102933_drop_repository_storage_events_for_geo_events.rb')
+
+describe DropRepositoryStorageEventsForGeoEvents, :migration do
+  let(:migration) { described_class.new }
+
+  describe '#up' do
+    before do
+      migration.up
+    end
+
+    it 'dropped the repository_storage_path column' do
+      described_class::TABLES.each do |table_name|
+        columns = table(table_name).columns.map(&:name)
+
+        expect(columns).not_to include("repository_storage_path")
+      end
+    end
+  end
+
+  describe '#down' do
+    let(:event_name) { :geo_repository_created_event }
+
+    before do
+      described_class.const_set(:BATCH_SIZE, 1)
+
+      Gitlab.config.repositories.storages.each do |name, _|
+        described_class.execute(<<~SQL
+                                INSERT INTO #{event_name}s (project_id, project_name, repository_storage_name, repo_path)
+                                VALUES (1, 'mepmep', '#{name}', 'path/to/gitlab-org')
+                                SQL
+                               )
+      end
+
+      migration.down
+      reset_column_in_all_models
+    end
+
+    it 'created the repository_storage_path column' do
+      described_class::TABLES.each do |table_name|
+        columns = table(table_name).columns.map(&:name)
+
+        expect(columns).to include("repository_storage_path")
+      end
+
+      null_columns = described_class
+        .execute("SELECT COUNT(*) FROM #{event_name}s WHERE repository_storage_path IS NULL;")
+        .first['count']
+
+      expect(null_columns.to_i).to be(0)
+    end
+  end
+end