documentcloud-cloud-crowd 0.1.1 → 0.2.0

data/lib/cloud_crowd/worker.rb

@@ -1,13 +1,14 @@
 module CloudCrowd
   
-  # The Worker, run at intervals by the Daemon, fetches WorkUnits from the
-  # central server and dispatches Actions to process them. Workers only fetch
-  # units that they are able to handle (for which they have an action in their
-  # actions directory). If communication with the central server is interrupted, 
-  # the WorkUnit will repeatedly attempt to complete its unit -- every 
-  # Worker::RETRY_WAIT seconds. Any exceptions that take place during 
+  # The Worker, forked off from the Node when a new WorkUnit is received, 
+  # launches an Action for processing. Workers will only ever receive WorkUnits
+  # that they are able to handle (for which they have a corresponding action in 
+  # their actions directory). If communication with the central server is 
+  # interrupted, the Worker will repeatedly attempt to complete its unit -- 
+  # every Worker::RETRY_WAIT seconds. Any exceptions that take place during 
   # the course of the Action will cause the Worker to mark the WorkUnit as 
-  # having failed.
+  # having failed. When finished, the Worker's process exits, minimizing the
+  # potential for memory leaks.
   class Worker
     
     # Wait five seconds to retry, after internal communcication errors.
@@ -15,40 +16,28 @@ module CloudCrowd
             
     attr_reader :action
     
-    # Spinning up a worker will create a new AssetStore with a persistent
-    # connection to S3. This AssetStore gets passed into each action, for use
-    # as it is run.
+    # A new Worker begins processing its WorkUnit straight off.
     def initialize(node, work_unit)
-      Signal.trap('INT') { shut_down }
-      Signal.trap('KILL') { shut_down }
-      Signal.trap('TERM') { shut_down }
       @pid  = $$
       @node = node
+      trap_signals
       setup_work_unit(work_unit)
       run
     end
     
-    # # Ask the central server for the first WorkUnit in line.
-    # def fetch_work_unit
-    #   keep_trying_to "fetch a new work unit" do
-    #     unit_json = @server['/work'].post(base_params)
-    #     setup_work_unit(unit_json)
-    #   end
-    # end
-    
-    # Return output to the central server, marking the current work unit as done.
+    # Return output to the central server, marking the WorkUnit done.
     def complete_work_unit(result)
       keep_trying_to "complete work unit" do
-        data = completion_params.merge({:status => 'succeeded', :output => result})
+        data = base_params.merge({:status => 'succeeded', :output => result})
         @node.server["/work/#{data[:id]}"].put(data)
         log "finished #{display_work_unit} in #{data[:time]} seconds"
       end
     end
     
-    # Mark the current work unit as failed, returning the exception to central.
+    # Mark the WorkUnit failed, returning the exception to central.
     def fail_work_unit(exception)
       keep_trying_to "mark work unit as failed" do
-        data = completion_params.merge({:status => 'failed', :output => {'output' => exception.message}.to_json})
+        data = base_params.merge({:status => 'failed', :output => {'output' => exception.message}.to_json})
         @node.server["/work/#{data[:id]}"].put(data)
         log "failed #{display_work_unit} in #{data[:time]} seconds\n#{exception.message}\n#{exception.backtrace}"
       end
@@ -69,12 +58,14 @@ module CloudCrowd
       end
     end
     
-    # Loggable string of the current work unit.
+    # Loggable details describing what the Worker is up to.
     def display_work_unit
       "unit ##{@options['work_unit_id']} (#{@action_name}/#{CloudCrowd.display_status(@status)})"
     end
     
-    # Executes the current work unit, catching all exceptions as failures.
+    # Executes the WorkUnit by running the Action, catching all exceptions as 
+    # failures. We capture the thread so that we can kill it from the outside,
+    # when exiting.
     def run_work_unit
       @worker_thread = Thread.new do
         begin
@@ -91,12 +82,14 @@ module CloudCrowd
           complete_work_unit({'output' => result}.to_json)
         rescue Exception => e
           fail_work_unit(e)
+        ensure
+          @action.cleanup_work_directory
         end
       end
       @worker_thread.join
     end
     
-    # Wraps <tt>run_work_unit</tt> to benchmark the execution time, if requested.
+    # Wraps run_work_unit to benchmark the execution time, if requested.
     def run
       return run_work_unit unless @options['benchmark']
       status = CloudCrowd.display_status(@status)
@@ -106,23 +99,15 @@ module CloudCrowd
     
     private
     
-    # Common parameters to send back to central.
-    def base_params
-      @base_params ||= {
-        :pid => @pid 
-      }
-    end
-    
     # Common parameters to send back to central upon unit completion, 
     # regardless of success or failure.
-    def completion_params
-      base_params.merge({
-        :id       => @options['work_unit_id'], 
-        :time     => Time.now - @start_time 
-      })
+    def base_params
+      { :pid  => @pid,
+        :id   => @options['work_unit_id'], 
+        :time => Time.now - @start_time }
     end
     
-    # Extract our instance variables from a WorkUnit's JSON.
+    # Extract the Worker's instance variables from a WorkUnit's JSON.
     def setup_work_unit(unit)
       return false unless unit
       @start_time = Time.now
@@ -139,15 +124,15 @@ module CloudCrowd
       puts "Worker ##{@pid}: #{message}" unless ENV['RACK_ENV'] == 'test'
     end
     
-    # When we're done with a unit, clear out our instance variables to make way 
-    # for the next one. Also, remove all of the unit's temporary storage.
-    def clear_work_unit
-      @action.cleanup_work_directory
-      @action, @action_name, @input, @options, @start_time = nil, nil, nil, nil, nil
+    # When signaled to exit, make sure that the Worker shuts down cleanly.
+    def trap_signals
+      Signal.trap('INT') { shut_down }
+      Signal.trap('KILL') { shut_down }
+      Signal.trap('TERM') { shut_down }
     end
     
-    # Force the worker to quit, even if it's in the middle of processing.
-    # If it had checked out a work unit, the node should have released it on
+    # Force the Worker to quit, even if it's in the middle of processing.
+    # If it had a checked-out WorkUnit, the Node should have released it on
     # the central server already.
     def shut_down
       if @worker_thread

data/public/css/admin_console.css

@@ -158,23 +158,33 @@ body {
       border: 1px solid transparent;
       margin: 1px 7px;
       padding-left: 18px;
+      background-position: left center;
+      background-repeat: no-repeat;
     }
       #nodes .node {
         font-size: 11px;
         line-height: 22px;
-        background: url(/images/server.png) no-repeat left center;
+        background-image: url(/images/server.png);
       }
+        #nodes .node.busy {
+          background-image: url(/images/server_busy.png);
+        }
+          #nodes .node.busy span.busy {
+            font-size: 9px;
+            color: #7f7f7f;
+            text-transform: uppercase;
+          }
       #nodes .worker {
         font-size: 10px;
         line-height: 18px;
         cursor: pointer;
-        background: url(/images/bullet_green.png) no-repeat left center;
-      }
-      #nodes .worker:hover {
-        border: 1px solid #aaa;
-        border-radius: 4px; -moz-border-radius: 4px; -webkit-border-radius: 4px;
-        background-color: #ccc;
+        background-image: url(/images/bullet_green.png);
       }
+        #nodes .worker:hover {
+          border: 1px solid #aaa;
+          border-radius: 4px; -moz-border-radius: 4px; -webkit-border-radius: 4px;
+          background-color: #ccc;
+        }
       
 #worker_info {
   position: absolute;

data/public/js/admin_console.js

@@ -117,6 +117,7 @@ window.Console = {
   },
   
   // Re-render all workers from scratch each time.
+  // This method is desperately in need of Javascript templates...
   renderNodes : function() {
     var header = $('#sidebar_header');
     var nc = this._nodes.length, wc = this._workerCount;
@@ -124,7 +125,8 @@ window.Console = {
     header.toggleClass('no_nodes', this._nodes.length <= 0);
     $('#nodes').html($.map(this._nodes, function(node) { 
       var html = "";
-      html += '<div class="node">' + node.host + '</div>';
+      var extra = node.status == 'busy' ? ' <span class="busy">[busy]</span>' : '';
+      html += '<div class="node ' + node.status + '">' + node.host + extra + '</div>';
       html += $.map(node.workers, function(pid) {
         var name = pid + '@' + node.host;
         return '<div class="worker" rel="' + name + '">' + name + '</div>';

data/test/config/config.yml

@@ -2,4 +2,4 @@
 :work_unit_retries:       3
 
 :central_server:          http://localhost:9173
-:storage:                 filesystem
+:storage:                 filesystem

data/test/unit/test_action.rb

@@ -27,7 +27,7 @@ class ActionTest < Test::Unit::TestCase
     end
     
     should "be able to save (to the filesystem while testing)" do
-      assert @action.save(@action.input_path) == "file://#{CloudCrowd::AssetStore::LOCAL_STORAGE_PATH}/word_count/job_1/unit_1/test_action.rb"
+      assert @action.save(@action.input_path) == "file://#{@store.local_storage_path}/word_count/job_1/unit_1/test_action.rb"
     end
     
     should "be able to clean up after itself" do

data/test/unit/test_job.rb

@@ -61,6 +61,8 @@ class JobTest < Test::Unit::TestCase
     end
     
     should "fire a callback when a job has finished, successfully or not" do
+      @job.update_attribute(:callback_url, 'http://example.com/callback')
+      CloudCrowd::Job.any_instance.stubs(:fire_callback).returns(true)
       CloudCrowd::Job.any_instance.expects(:fire_callback)
       @job.work_units.first.finish('{"output":"output"}', 10)
       assert @job.all_work_units_complete?

data/views/{index.erb → operations_center.erb}

@@ -55,11 +55,11 @@
       </div>
       <div class="graph_container">
         <div class="graph_title">
-					Active Nodes
-					(<span id="nodes_legend" class="legend_box"></span>)
-					and Workers
-					(<span id="workers_legend" class="legend_box"></span>)
-				</div>
+          Active Nodes
+          (<span id="nodes_legend" class="legend_box"></span>)
+          and Workers
+          (<span id="workers_legend" class="legend_box"></span>)
+        </div>
         <div id="workers_graph" class="graph"></div>
       </div>
     </div>

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification 
 name: documentcloud-cloud-crowd
 version: !ruby/object:Gem::Version 
-  version: 0.1.1
+  version: 0.2.0
 platform: ruby
 authors: 
 - Jeremy Ashkenas
@@ -9,7 +9,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2009-09-15 00:00:00 -07:00
+date: 2009-09-17 00:00:00 -07:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -171,6 +171,7 @@ files:
 - public/images/logo.png
 - public/images/queue_fill.png
 - public/images/server.png
+- public/images/server_busy.png
 - public/images/server_error.png
 - public/images/sidebar_bottom.png
 - public/images/sidebar_top.png
@@ -194,7 +195,7 @@ files:
 - test/unit/test_configuration.rb
 - test/unit/test_job.rb
 - test/unit/test_work_unit.rb
-- views/index.erb
+- views/operations_center.erb
 has_rdoc: true
 homepage: http://wiki.github.com/documentcloud/cloud-crowd
 licenses: