Filter legacy attachments correctly

An attachment from an unmigrated project will only be using the legacy
storage when stored locally. Any file on object storage already use
hashed storage format.

app/models/project.rb

@@ -374,9 +374,17 @@ class Project < ApplicationRecord
   scope :pending_delete, -> { where(pending_delete: true) }
   scope :without_deleted, -> { where(pending_delete: false) }
 
-  scope :with_storage_feature, ->(feature) { where('storage_version >= :version', version: HASHED_STORAGE_FEATURES[feature]) }
-  scope :without_storage_feature, ->(feature) { where('storage_version < :version OR storage_version IS NULL', version: HASHED_STORAGE_FEATURES[feature]) }
-  scope :with_unmigrated_storage, -> { where('storage_version < :version OR storage_version IS NULL', version: LATEST_STORAGE_VERSION) }
+  scope :with_storage_feature, ->(feature) do
+    where(arel_table[:storage_version].gteq(HASHED_STORAGE_FEATURES[feature]))
+  end
+  scope :without_storage_feature, ->(feature) do
+    where(arel_table[:storage_version].lt(HASHED_STORAGE_FEATURES[feature])
+        .or(arel_table[:storage_version].eq(nil)))
+  end
+  scope :with_unmigrated_storage, -> do
+    where(arel_table[:storage_version].lt(LATEST_STORAGE_VERSION)
+        .or(arel_table[:storage_version].eq(nil)))
+  end
 
   # last_activity_at is throttled every minute, but last_repository_updated_at is updated with every push
   scope :sorted_by_activity, -> { reorder(Arel.sql("GREATEST(COALESCE(last_activity_at, '1970-01-01'), COALESCE(last_repository_updated_at, '1970-01-01')) DESC")) }

app/models/upload.rb

@@ -23,6 +23,21 @@ class Upload < ApplicationRecord
   after_destroy :delete_file!, if: -> { uploader_class <= FileUploader }
 
   class << self
+    def inner_join_local_uploads_projects
+      upload_table = Upload.arel_table
+      project_table = Project.arel_table
+
+      join_statement = upload_table.project(upload_table[Arel.star])
+                         .join(project_table)
+                         .on(
+                           upload_table[:model_type].eq('Project')
+                             .and(upload_table[:model_id].eq(project_table[:id]))
+                             .and(upload_table[:store].eq(ObjectStorage::Store::LOCAL))
+                         )
+
+      joins(join_statement.join_sources)
+    end
+
     ##
     # FastDestroyAll concerns
     def begin_fast_destroy

app/workers/hashed_storage/project_migrate_worker.rb

@@ -14,7 +14,7 @@ module HashedStorage
 
       try_obtain_lease do
         project = Project.without_deleted.find_by(id: project_id)
-        break unless project
+        break unless project && project.storage_upgradable?
 
         old_disk_path ||= Storage::LegacyProject.new(project).disk_path
 

changelogs/unreleased/35789-hashed-storage-attachments-migration-object-storage.yml

+---
+title: 'Hashed Storage attachments migration: exclude files in object storage as they
+  are all hashed already'
+merge_request: 20338
+author:
+type: changed

lib/gitlab/hashed_storage/rake_helper.rb

@@ -47,23 +47,13 @@ module Gitlab
       end
       # rubocop: enable CodeReuse/ActiveRecord
 
-      # rubocop: disable CodeReuse/ActiveRecord
       def self.legacy_attachments_relation
-        Upload.joins(<<~SQL).where('projects.storage_version < :version OR projects.storage_version IS NULL', version: Project::HASHED_STORAGE_FEATURES[:attachments])
-          JOIN projects
-            ON (uploads.model_type='Project' AND uploads.model_id=projects.id)
-        SQL
+        Upload.inner_join_local_uploads_projects.merge(Project.without_storage_feature(:attachments))
       end
-      # rubocop: enable CodeReuse/ActiveRecord
 
-      # rubocop: disable CodeReuse/ActiveRecord
       def self.hashed_attachments_relation
-        Upload.joins(<<~SQL).where('projects.storage_version >= :version', version: Project::HASHED_STORAGE_FEATURES[:attachments])
-          JOIN projects
-          ON (uploads.model_type='Project' AND uploads.model_id=projects.id)
-        SQL
+        Upload.inner_join_local_uploads_projects.merge(Project.with_storage_feature(:attachments))
       end
-      # rubocop: enable CodeReuse/ActiveRecord
 
       def self.relation_summary(relation_name, relation)
         relation_count = relation.count

spec/workers/hashed_storage/project_migrate_worker_spec.rb

@@ -5,13 +5,13 @@ require 'spec_helper'
 describe HashedStorage::ProjectMigrateWorker, :clean_gitlab_redis_shared_state do
   include ExclusiveLeaseHelpers
 
-  describe '#perform' do
-    let(:project) { create(:project, :empty_repo, :legacy_storage) }
-    let(:lease_key) { "project_migrate_hashed_storage_worker:#{project.id}" }
-    let(:lease_timeout) { described_class::LEASE_TIMEOUT }
   let(:migration_service) { ::Projects::HashedStorage::MigrationService }
+  let(:lease_timeout) { described_class::LEASE_TIMEOUT }
 
+  describe '#perform' do
     it 'skips when project no longer exists' do
+      stub_exclusive_lease(lease_key(-1), 'uuid', timeout: lease_timeout)
+
       expect(migration_service).not_to receive(:new)
 
       subject.perform(-1)
@@ -19,14 +19,41 @@ describe HashedStorage::ProjectMigrateWorker, :clean_gitlab_redis_shared_state d
 
     it 'skips when project is pending delete' do
       pending_delete_project = create(:project, :empty_repo, pending_delete: true)
+      stub_exclusive_lease(lease_key(pending_delete_project.id), 'uuid', timeout: lease_timeout)
 
       expect(migration_service).not_to receive(:new)
 
       subject.perform(pending_delete_project.id)
     end
 
-    it 'delegates migration to service class when we have exclusive lease' do
-      stub_exclusive_lease(lease_key, 'uuid', timeout: lease_timeout)
+    it 'skips when project is already migrated' do
+      migrated_project = create(:project, :empty_repo)
+      stub_exclusive_lease(lease_key(migrated_project.id), 'uuid', timeout: lease_timeout)
+
+      expect(migration_service).not_to receive(:new)
+
+      subject.perform(migrated_project.id)
+    end
+
+    context 'with exclusive lease available' do
+      it 'delegates migration to service class' do
+        project = create(:project, :empty_repo, :legacy_storage)
+        stub_exclusive_lease(lease_key(project.id), 'uuid', timeout: lease_timeout)
+
+        service_spy = spy
+
+        allow(migration_service)
+          .to receive(:new).with(project, project.full_path, logger: subject.logger)
+                .and_return(service_spy)
+
+        subject.perform(project.id)
+
+        expect(service_spy).to have_received(:execute)
+      end
+
+      it 'delegates migration to service class with correct path in a partially migrated project' do
+        project = create(:project, :empty_repo, storage_version: 1)
+        stub_exclusive_lease(lease_key(project.id), 'uuid', timeout: lease_timeout)
 
         service_spy = spy
 
@@ -38,13 +65,21 @@ describe HashedStorage::ProjectMigrateWorker, :clean_gitlab_redis_shared_state d
 
         expect(service_spy).to have_received(:execute)
       end
+    end
 
+    context 'with exclusive lease taken' do
       it 'skips when it cant acquire the exclusive lease' do
-      stub_exclusive_lease_taken(lease_key, timeout: lease_timeout)
+        project = create(:project, :empty_repo, :legacy_storage)
+        stub_exclusive_lease_taken(lease_key(project.id), timeout: lease_timeout)
 
         expect(migration_service).not_to receive(:new)
 
         subject.perform(project.id)
       end
     end
+  end
+
+  def lease_key(key)
+    "project_migrate_hashed_storage_worker:#{key}"
+  end
 end