Resolve "Hashed Storage: Make possible to migrate single project"

app/workers/storage_migrator_worker.rb

 class StorageMigratorWorker
   include ApplicationWorker
 
-  BATCH_SIZE = 100
-
   def perform(start, finish)
-    projects = build_relation(start, finish)
-
-    projects.with_route.find_each(batch_size: BATCH_SIZE) do |project|
-      Rails.logger.info "Starting storage migration of #{project.full_path} (ID=#{project.id})..."
-
-      begin
-        project.migrate_to_hashed_storage!
-      rescue => err
-        Rails.logger.error("#{err.message} migrating storage of #{project.full_path} (ID=#{project.id}), trace - #{err.backtrace}")
-      end
-    end
-  end
-
-  def build_relation(start, finish)
-    relation = Project
-    table = Project.arel_table
-
-    relation = relation.where(table[:id].gteq(start)) if start
-    relation = relation.where(table[:id].lteq(finish)) if finish
-
-    relation
+    migrator = Gitlab::HashedStorage::Migrator.new
+    migrator.bulk_migrate(start, finish)
   end
 end

changelogs/unreleased/46922-hashed-storage-single-project.yml

+---
+title: 'Hashed Storage: migration rake task now can be executed to specific project'
+merge_request: 19268
+author:
+type: changed

doc/administration/raketasks/storage.md

@@ -17,13 +17,21 @@ This task will schedule all your existing projects and attachments associated wi
 **Omnibus Installation**
 
 ```bash
-gitlab-rake gitlab:storage:migrate_to_hashed
+sudo gitlab-rake gitlab:storage:migrate_to_hashed
 ```
 
 **Source Installation**
 
 ```bash
-rake gitlab:storage:migrate_to_hashed
+sudo -u git -H bundle exec rake gitlab:storage:migrate_to_hashed RAILS_ENV=production
+```
+
+They both also accept a range as environment variable:
+
+```bash
+# to migrate any non migrated project from ID 20 to 50.
+export ID_FROM=20 
+export ID_TO=50
 ```
 
 You can monitor the progress in the _Admin > Monitoring > Background jobs_ screen.
@@ -44,13 +52,13 @@ To have a simple summary of projects using **Legacy** storage:
 **Omnibus Installation**
 
 ```bash
-gitlab-rake gitlab:storage:legacy_projects
+sudo gitlab-rake gitlab:storage:legacy_projects
 ```
 
 **Source Installation**
 
 ```bash
-rake gitlab:storage:legacy_projects
+sudo -u git -H bundle exec rake gitlab:storage:legacy_projects RAILS_ENV=production
 ```
 
 ------
@@ -60,13 +68,13 @@ To list projects using **Legacy** storage:
 **Omnibus Installation**
 
 ```bash
-gitlab-rake gitlab:storage:list_legacy_projects
+sudo gitlab-rake gitlab:storage:list_legacy_projects
 ```
 
 **Source Installation**
 
 ```bash
-rake gitlab:storage:list_legacy_projects
+sudo -u git -H bundle exec rake gitlab:storage:list_legacy_projects RAILS_ENV=production
 
 ```
 
@@ -77,13 +85,13 @@ To have a simple summary of projects using **Hashed** storage:
 **Omnibus Installation**
 
 ```bash
-gitlab-rake gitlab:storage:hashed_projects
+sudo gitlab-rake gitlab:storage:hashed_projects
 ```
 
 **Source Installation**
 
 ```bash
-rake gitlab:storage:hashed_projects
+sudo -u git -H bundle exec rake gitlab:storage:hashed_projects RAILS_ENV=production
 ```
 
 ------
@@ -93,14 +101,13 @@ To list projects using **Hashed** storage:
 **Omnibus Installation**
 
 ```bash
-gitlab-rake gitlab:storage:list_hashed_projects
+sudo gitlab-rake gitlab:storage:list_hashed_projects
 ```
 
 **Source Installation**
 
 ```bash
-rake gitlab:storage:list_hashed_projects
-
+sudo -u git -H bundle exec rake gitlab:storage:list_hashed_projects RAILS_ENV=production
 ```
 
 ## List attachments on Legacy storage
@@ -110,13 +117,13 @@ To have a simple summary of project attachments using **Legacy** storage:
 **Omnibus Installation**
 
 ```bash
-gitlab-rake gitlab:storage:legacy_attachments
+sudo gitlab-rake gitlab:storage:legacy_attachments
 ```
 
 **Source Installation**
 
 ```bash
-rake gitlab:storage:legacy_attachments
+sudo -u git -H bundle exec rake gitlab:storage:legacy_attachments RAILS_ENV=production
 ```
 
 ------
@@ -126,13 +133,13 @@ To list project attachments using **Legacy** storage:
 **Omnibus Installation**
 
 ```bash
-gitlab-rake gitlab:storage:list_legacy_attachments
+sudo gitlab-rake gitlab:storage:list_legacy_attachments
 ```
 
 **Source Installation**
 
 ```bash
-rake gitlab:storage:list_legacy_attachments
+sudo -u git -H bundle exec rake gitlab:storage:list_legacy_attachments RAILS_ENV=production
 ```
 
 ## List attachments on Hashed storage
@@ -142,13 +149,13 @@ To have a simple summary of project attachments using **Hashed** storage:
 **Omnibus Installation**
 
 ```bash
-gitlab-rake gitlab:storage:hashed_attachments
+sudo gitlab-rake gitlab:storage:hashed_attachments
 ```
 
 **Source Installation**
 
 ```bash
-rake gitlab:storage:hashed_attachments
+sudo -u git -H bundle exec rake gitlab:storage:hashed_attachments RAILS_ENV=production
 ```
 
 ------
@@ -158,13 +165,13 @@ To list project attachments using **Hashed** storage:
 **Omnibus Installation**
 
 ```bash
-gitlab-rake gitlab:storage:list_hashed_attachments
+sudo gitlab-rake gitlab:storage:list_hashed_attachments
 ```
 
 **Source Installation**
 
 ```bash
-rake gitlab:storage:list_hashed_attachments
+sudo -u git -H bundle exec rake gitlab:storage:list_hashed_attachments RAILS_ENV=production
 ```
 
 [storage-types]: ../repository_storage_types.md

lib/gitlab/hashed_storage/migrator.rb

+module Gitlab
+  module HashedStorage
+    # Hashed Storage Migrator
+    #
+    # This is responsible for scheduling and flagging projects
+    # to be migrated from Legacy to Hashed storage, either one by one or in bulk.
+    class Migrator
+      BATCH_SIZE = 100
+
+      # Schedule a range of projects to be bulk migrated with #bulk_migrate asynchronously
+      #
+      # @param [Object] start first project id for the range
+      # @param [Object] finish last project id for the range
+      def bulk_schedule(start, finish)
+        StorageMigratorWorker.perform_async(start, finish)
+      end
+
+      # Start migration of projects from specified range
+      #
+      # Flagging a project to be migrated is a synchronous action,
+      # but the migration runs through async jobs
+      #
+      # @param [Object] start first project id for the range
+      # @param [Object] finish last project id for the range
+      def bulk_migrate(start, finish)
+        projects = build_relation(start, finish)
+
+        projects.with_route.find_each(batch_size: BATCH_SIZE) do |project|
+          migrate(project)
+        end
+      end
+
+      # Flag a project to me migrated
+      #
+      # @param [Object] project that will be migrated
+      def migrate(project)
+        Rails.logger.info "Starting storage migration of #{project.full_path} (ID=#{project.id})..."
+
+        project.migrate_to_hashed_storage!
+      rescue => err
+        Rails.logger.error("#{err.message} migrating storage of #{project.full_path} (ID=#{project.id}), trace - #{err.backtrace}")
+      end
+
+      private
+
+      def build_relation(start, finish)
+        relation = Project
+        table = Project.arel_table
+
+        relation = relation.where(table[:id].gteq(start)) if start
+        relation = relation.where(table[:id].lteq(finish)) if finish
+
+        relation
+      end
+    end
+  end
+end

lib/gitlab/hashed_storage/rake_helper.rb

@@ -9,8 +9,20 @@ module Gitlab
         ENV.fetch('LIMIT', 500).to_i
       end
 
+      def self.range_from
+        ENV['ID_FROM']
+      end
+
+      def self.range_to
+        ENV['ID_TO']
+      end
+
+      def self.range_single_item?
+        !range_from.nil? && range_from == range_to
+      end
+
       def self.project_id_batches(&block)
-        Project.with_unmigrated_storage.in_batches(of: batch_size, start: ENV['ID_FROM'], finish: ENV['ID_TO']) do |relation| # rubocop: disable Cop/InBatches
+        Project.with_unmigrated_storage.in_batches(of: batch_size, start: range_from, finish: range_to) do |relation| # rubocop: disable Cop/InBatches
           ids = relation.pluck(:id)
 
           yield ids.min, ids.max

lib/tasks/gitlab/storage.rake

@@ -2,9 +2,26 @@ namespace :gitlab do
   namespace :storage do
     desc 'GitLab | Storage | Migrate existing projects to Hashed Storage'
     task migrate_to_hashed: :environment do
-      legacy_projects_count = Project.with_unmigrated_storage.count
+      storage_migrator = Gitlab::HashedStorage::Migrator.new
       helper = Gitlab::HashedStorage::RakeHelper
 
+      if helper.range_single_item?
+        project = Project.with_unmigrated_storage.find_by(id: helper.range_from)
+
+        unless project
+          puts "There are no projects requiring storage migration with ID=#{helper.range_from}"
+
+          next
+        end
+
+        puts "Enqueueing storage migration of #{project.full_path} (ID=#{project.id})..."
+        storage_migrator.migrate(project)
+
+        next
+      end
+
+      legacy_projects_count = Project.with_unmigrated_storage.count
+
       if legacy_projects_count == 0
         puts 'There are no projects requiring storage migration. Nothing to do!'
 
@@ -14,7 +31,7 @@ namespace :gitlab do
       print "Enqueuing migration of #{legacy_projects_count} projects in batches of #{helper.batch_size}"
 
       helper.project_id_batches do |start, finish|
-        StorageMigratorWorker.perform_async(start, finish)
+        storage_migrator.bulk_schedule(start, finish)
 
         print '.'
       end

spec/lib/gitlab/hashed_storage/migrator_spec.rb

+require 'spec_helper'
+
+describe Gitlab::HashedStorage::Migrator do
+  describe '#bulk_schedule' do
+    it 'schedules job to StorageMigratorWorker' do
+      Sidekiq::Testing.fake! do
+        expect { subject.bulk_schedule(1, 5) }.to change(StorageMigratorWorker.jobs, :size).by(1)
+      end
+    end
+  end
+
+  describe '#bulk_migrate' do
+    let(:projects) { create_list(:project, 2, :legacy_storage) }
+    let(:ids) { projects.map(&:id) }
+
+    it 'enqueue jobs to ProjectMigrateHashedStorageWorker' do
+      Sidekiq::Testing.fake! do
+        expect { subject.bulk_migrate(ids.min, ids.max) }.to change(ProjectMigrateHashedStorageWorker.jobs, :size).by(2)
+      end
+    end
+
+    it 'sets projects as read only' do
+      allow(ProjectMigrateHashedStorageWorker).to receive(:perform_async).twice
+      subject.bulk_migrate(ids.min, ids.max)
+
+      projects.each do |project|
+        expect(project.reload.repository_read_only?).to be_truthy
+      end
+    end
+
+    it 'rescues and log exceptions' do
+      allow_any_instance_of(Project).to receive(:migrate_to_hashed_storage!).and_raise(StandardError)
+      expect { subject.bulk_migrate(ids.min, ids.max) }.not_to raise_error
+    end
+
+    it 'delegates each project in specified range to #migrate' do
+      projects.each do |project|
+        expect(subject).to receive(:migrate).with(project)
+      end
+
+      subject.bulk_migrate(ids.min, ids.max)
+    end
+  end
+
+  describe '#migrate' do
+    let(:project) { create(:project, :legacy_storage, :empty_repo) }
+
+    it 'enqueues job to ProjectMigrateHashedStorageWorker' do
+      Sidekiq::Testing.fake! do
+        expect { subject.migrate(project) }.to change(ProjectMigrateHashedStorageWorker.jobs, :size).by(1)
+      end
+    end
+
+    it 'rescues and log exceptions' do
+      allow(project).to receive(:migrate_to_hashed_storage!).and_raise(StandardError)
+
+      expect { subject.migrate(project) }.not_to raise_error
+    end
+
+    it 'sets project as read only' do
+      allow(ProjectMigrateHashedStorageWorker).to receive(:perform_async)
+      subject.migrate(project)
+
+      expect(project.reload.repository_read_only?).to be_truthy
+    end
+
+    it 'migrate project' do
+      Sidekiq::Testing.inline! do
+        subject.migrate(project)
+      end
+
+      expect(project.reload.hashed_storage?(:attachments)).to be_truthy
+    end
+  end
+end

spec/tasks/gitlab/storage_rake_spec.rb

 require 'rake_helper'
 
-describe 'gitlab:storage:*' do
+describe 'rake gitlab:storage:*' do
   before do
     Rake.application.rake_require 'tasks/gitlab/storage'
 
@@ -44,16 +44,18 @@ describe 'gitlab:storage:*' do
   end
 
   describe 'gitlab:storage:migrate_to_hashed' do
+    let(:task) { 'gitlab:storage:migrate_to_hashed' }
+
     context '0 legacy projects' do
       it 'does nothing' do
         expect(StorageMigratorWorker).not_to receive(:perform_async)
 
-        run_rake_task('gitlab:storage:migrate_to_hashed')
+        run_rake_task(task)
       end
     end
 
     context '3 legacy projects' do
-      let(:projects) { create_list(:project, 3, storage_version: 0) }
+      let(:projects) { create_list(:project, 3, :legacy_storage) }
 
       context 'in batches of 1' do
         before do
@@ -65,7 +67,7 @@ describe 'gitlab:storage:*' do
             expect(StorageMigratorWorker).to receive(:perform_async).with(project.id, project.id)
           end
 
-          run_rake_task('gitlab:storage:migrate_to_hashed')
+          run_rake_task(task)
         end
       end
 
@@ -80,8 +82,33 @@ describe 'gitlab:storage:*' do
             expect(StorageMigratorWorker).to receive(:perform_async).with(first, last)
           end
 
-          run_rake_task('gitlab:storage:migrate_to_hashed')
+          run_rake_task(task)
+        end
+      end
+    end
+
+    context 'with same id in range' do
+      it 'displays message when project cant be found' do
+        stub_env('ID_FROM', 99999)
+        stub_env('ID_TO', 99999)
+
+        expect { run_rake_task(task) }.to output(/There are no projects requiring storage migration with ID=99999/).to_stdout
       end
+
+      it 'displays a message when project exists but its already migrated' do
+        project = create(:project)
+        stub_env('ID_FROM', project.id)
+        stub_env('ID_TO', project.id)
+
+        expect { run_rake_task(task) }.to output(/There are no projects requiring storage migration with ID=#{project.id}/).to_stdout
+      end
+
+      it 'enqueues migration when project can be found' do
+        project = create(:project, :legacy_storage)
+        stub_env('ID_FROM', project.id)
+        stub_env('ID_TO', project.id)
+
+        expect { run_rake_task(task) }.to output(/Enqueueing storage migration .* \(ID=#{project.id}\)/).to_stdout
       end
     end
   end
@@ -89,14 +116,14 @@ describe 'gitlab:storage:*' do
   describe 'gitlab:storage:legacy_projects' do
     it_behaves_like 'rake entities summary', 'projects', 'Legacy' do
       let(:task) { 'gitlab:storage:legacy_projects' }
-      let(:create_collection) { create_list(:project, 3, storage_version: 0) }
+      let(:create_collection) { create_list(:project, 3, :legacy_storage) }
     end
   end
 
   describe 'gitlab:storage:list_legacy_projects' do
     it_behaves_like 'rake listing entities', 'projects', 'Legacy' do
       let(:task) { 'gitlab:storage:list_legacy_projects' }
-      let(:create_collection) { create_list(:project, 3, storage_version: 0) }
+      let(:create_collection) { create_list(:project, 3, :legacy_storage) }
     end
   end
 
@@ -133,7 +160,7 @@ describe 'gitlab:storage:*' do
   describe 'gitlab:storage:hashed_attachments' do
     it_behaves_like 'rake entities summary', 'attachments', 'Hashed' do
       let(:task) { 'gitlab:storage:hashed_attachments' }
-      let(:project) { create(:project, storage_version: 2) }
+      let(:project) { create(:project) }
       let(:create_collection) { create_list(:upload, 3, model: project) }
     end
   end
@@ -141,7 +168,7 @@ describe 'gitlab:storage:*' do
   describe 'gitlab:storage:list_hashed_attachments' do
     it_behaves_like 'rake listing entities', 'attachments', 'Hashed' do
       let(:task) { 'gitlab:storage:list_hashed_attachments' }
-      let(:project) { create(:project, storage_version: 2) }
+      let(:project) { create(:project) }
       let(:create_collection) { create_list(:upload, 3, model: project) }
     end
   end

spec/workers/storage_migrator_worker_spec.rb

@@ -2,29 +2,24 @@ require 'spec_helper'
 
 describe StorageMigratorWorker do
   subject(:worker) { described_class.new }
-  let(:projects) { create_list(:project, 2, :legacy_storage) }
-
-  describe '#perform' do
+  let(:projects) { create_list(:project, 2, :legacy_storage, :empty_repo) }
   let(:ids) { projects.map(&:id) }
 
-    it 'enqueue jobs to ProjectMigrateHashedStorageWorker' do
-      expect(ProjectMigrateHashedStorageWorker).to receive(:perform_async).twice
+  describe '#perform' do
+    it 'delegates to MigratorService' do
+      expect_any_instance_of(Gitlab::HashedStorage::Migrator).to receive(:bulk_migrate).with(5, 10)
 
-      worker.perform(ids.min, ids.max)
+      worker.perform(5, 10)
     end
 
-    it 'sets projects as read only' do
-      allow(ProjectMigrateHashedStorageWorker).to receive(:perform_async).twice
+    it 'migrates projects in the specified range' do
+      Sidekiq::Testing.inline! do
         worker.perform(ids.min, ids.max)
+      end
 
       projects.each do |project|
-        expect(project.reload.repository_read_only?).to be_truthy
-      end
+        expect(project.reload.hashed_storage?(:attachments)).to be_truthy
       end
-
-    it 'rescues and log exceptions' do
-      allow_any_instance_of(Project).to receive(:migrate_to_hashed_storage!).and_raise(StandardError)
-      expect { worker.perform(ids.min, ids.max) }.not_to raise_error
     end
   end
 end