Adding a counter of pushes since last garbage collection, incremented when pushed to and reset when gc-ed

The algorithm for finding gc candidates is unchanged, will run this in production and collect metrics for changing this the next few days.

- Also tracking disk usage for repositories, calculated when pushed to and stored in the database

5 files changed, 75 insertions, 8 deletions

diff --git a/app/models/repository.rb b/app/models/repository.rb
index 4c61148..ee4f1d4 100644
--- a/app/models/repository.rb
+++ b/app/models/repository.rb
@@ -662,11 +662,26 @@ class Repository < ActiveRecord::Base
   def gc!
     Grit::Git.with_timeout(nil) do
       if self.git.gc_auto
-        return update_attribute(:last_gc_at, Time.now.utc)
+        self.last_gc_at = Time.now
+        self.push_count_since_gc = 0
+        return save
       end
     end
   end
 
+  def register_push
+    self.last_pushed_at = Time.now.utc
+    self.push_count_since_gc = push_count_since_gc.to_i + 1
+  end
+  
+  def update_disk_usage
+    self.disk_usage = calculate_disk_usage
+  end
+
+  def calculate_disk_usage
+    @calculated_disk_usage ||= `du -sb #{full_repository_path} 2>/dev/null`.chomp.to_i
+  end
+
   def matches_regexp?(term)
     return user.login =~ term ||
       name =~ term ||
diff --git a/app/processors/push_event_processor.rb b/app/processors/push_event_processor.rb
index 5ff15f5..a7d49ef 100644
--- a/app/processors/push_event_processor.rb
+++ b/app/processors/push_event_processor.rb
@@ -79,7 +79,9 @@ class PushEventProcessor < ApplicationProcessor
     r, name, @identifier = @revname.split("/", 3)
     @target = {'tags' => :tag, 'heads' => :head, 'merge-requests' => :review}[name]
     if @target != :review && @repository
-      @repository.update_attribute(:last_pushed_at, Time.now.utc)
+      @repository.update_disk_usage
+      @repository.register_push
+      @repository.save
     end
     process_push
   end
diff --git a/db/migrate/20100203132425_adding_disk_size_and_push_count_to_repositories.rb b/db/migrate/20100203132425_adding_disk_size_and_push_count_to_repositories.rb
new file mode 100644
index 0000000..2381f64
--- /dev/null
+++ b/db/migrate/20100203132425_adding_disk_size_and_push_count_to_repositories.rb
@@ -0,0 +1,11 @@
+class AddingDiskSizeAndPushCountToRepositories < ActiveRecord::Migration
+  def self.up
+    add_column :repositories, :disk_usage, :integer
+    add_column :repositories, :push_count_since_gc, :integer
+  end
+
+  def self.down
+    remove_column :repositories, :disk_usage
+    remove_column :repositories, :push_count_since_gc
+  end
+end
diff --git a/db/schema.rb b/db/schema.rb
index d0b5c2d..188607c 100644
--- a/db/schema.rb
+++ b/db/schema.rb
@@ -9,7 +9,7 @@
 #
 # It's strongly recommended to check this file into your version control system.
 
-ActiveRecord::Schema.define(:version => 20100126094829) do
+ActiveRecord::Schema.define(:version => 20100203132425) do
 
   create_table "cloners", :force => true do |t|
     t.string   "ip"
@@ -279,6 +279,8 @@ ActiveRecord::Schema.define(:version => 20100126094829) do
     t.boolean  "notify_committers_on_new_merge_request", :default => true
     t.datetime "last_gc_at"
     t.boolean  "merge_requests_enabled",                 :default => true
+    t.integer  "disk_usage"
+    t.integer  "push_count_since_gc"
   end
 
   add_index "repositories", ["hashed_path"], :name => "index_repositories_on_hashed_path", :unique => true
diff --git a/test/unit/repository_test.rb b/test/unit/repository_test.rb
index da1d938..50d7794 100644
--- a/test/unit/repository_test.rb
+++ b/test/unit/repository_test.rb
@@ -960,17 +960,23 @@ class RepositoryTest < ActiveSupport::TestCase
   context "garbage collection" do
     setup do
       @repository = repositories(:johans)
+      @now = Time.now
+      Time.stubs(:now).returns(@now)
+      @repository.stubs(:git).returns(stub())
+      @repository.git.expects(:gc_auto).returns(true)
     end
 
     should "have a gc! method that updates last_gc_at" do
-      now = Time.now
-      Time.stubs(:now).returns(now)
-      @repository.stubs(:git).returns(stub())
-      @repository.git.expects(:gc_auto).returns(true)
       assert_nil @repository.last_gc_at
       assert @repository.gc!
       assert_not_nil @repository.last_gc_at
-      assert_equal now, @repository.last_gc_at
+      assert_equal @now, @repository.last_gc_at
+    end
+
+    should "set push_count_since_gc to 0 when doing gc" do
+      @repository.push_count_since_gc = 10
+      @repository.gc!
+      assert_equal 0, @repository.push_count_since_gc
     end
   end
 
@@ -1118,6 +1124,37 @@ class RepositoryTest < ActiveSupport::TestCase
     end
   end
 
+  context "Calculation of disk usage" do
+    setup do
+      @repository = repositories(:johans)
+      @bytes = 90129
+    end
+
+    should "save the bytes used" do
+      @repository.expects(:calculate_disk_usage).returns(@bytes)
+      @repository.update_disk_usage
+      assert_equal @bytes, @repository.disk_usage
+    end
+  end
+
+  context "Pushing" do
+    setup do
+      @repository = repositories(:johans)
+    end
+
+    should "update last_pushed_at" do
+      @repository.last_pushed_at = 1.hour.ago.utc
+      @repository.register_push
+      assert @repository.last_pushed_at > 1.hour.ago.utc
+    end
+
+    should "increment the number of pushes" do
+      @repository.push_count_since_gc = 2
+      @repository.register_push
+      assert_equal 3, @repository.push_count_since_gc
+    end
+  end
+
 end