Merge pull request resque#693 from 37signals/1-x-stable

Backported fixes from 37s changes in master

Author: steveklabnik

lib/resque/worker.rb

@@ -149,7 +149,7 @@ def work(interval = 5.0, &block)
           else
             unregister_signal_handlers if will_fork? && term_child
             procline "Processing #{job.queue} since #{Time.now.to_i}"
-            redis.client.reconnect # Don't share connection with parent
+            reconnect
             perform(job, &block)
             exit! if will_fork?
           end
@@ -164,8 +164,9 @@ def work(interval = 5.0, &block)
         end
       end
 
-    ensure
       unregister_worker
+    rescue Exception => exception
+      unregister_worker(exception)
     end
 
     # DEPRECATED. Processes a single job. If none is given, it will
@@ -218,6 +219,24 @@ def reserve
       raise e
     end
 
+    # Reconnect to Redis to avoid sharing a connection with the parent,
+    # retry up to 3 times with increasing delay before giving up.
+    def reconnect
+      tries = 0
+      begin
+        redis.client.reconnect
+      rescue Redis::BaseConnectionError
+        if (tries += 1) <= 3
+          log "Error reconnecting to Redis; retrying"
+          sleep(tries)
+          retry
+        else
+          log "Error reconnecting to Redis; quitting"
+          raise
+        end
+      end
+    end
+
     # Returns a list of queues to use when searching for a job.
     # A splat ("*") means you want every queue (in alpha order) - this
     # can be useful for dynamically adding new queues.
@@ -425,15 +444,15 @@ def run_hook(name, *args)
     end
 
     # Unregisters ourself as a worker. Useful when shutting down.
-    def unregister_worker
+    def unregister_worker(exception = nil)
       # If we're still processing a job, make sure it gets logged as a
       # failure.
       if (hash = processing) && !hash.empty?
         job = Job.new(hash['queue'], hash['payload'])
         # Ensure the proper worker is attached to this job, even if
         # it's not the precise instance that died.
         job.worker = self
-        job.fail(DirtyExit.new)
+        job.fail(exception || DirtyExit.new)
       end
 
       redis.srem(:workers, self)

test/test_helper.rb

@@ -161,14 +161,18 @@ def now
   self.fake_time = nil
 end
 
-def capture_stderr
-  # The output stream must be an IO-like object. In this case we capture it in
-  # an in-memory IO object so we can return the string value. You can assign any
-  # IO object here.
-  previous_stderr, $stderr = $stderr, StringIO.new
+# From minitest/unit
+def capture_io
+  require 'stringio'
+
+  orig_stdout, orig_stderr         = $stdout, $stderr
+  captured_stdout, captured_stderr = StringIO.new, StringIO.new
+  $stdout, $stderr                 = captured_stdout, captured_stderr
+
   yield
-  $stderr.string
+
+  return captured_stdout.string, captured_stderr.string
 ensure
-  # Restore the previous value of stderr (typically equal to STDERR).
-  $stderr = previous_stderr
+  $stdout = orig_stdout
+  $stderr = orig_stderr
 end

test/worker_test.rb

@@ -32,11 +32,20 @@
     end
   end
 
-  test "fails uncompleted jobs on exit" do
+  test "fails uncompleted jobs with DirtyExit by default on exit" do
     job = Resque::Job.new(:jobs, {'class' => 'GoodJob', 'args' => "blah"})
     @worker.working_on(job)
     @worker.unregister_worker
     assert_equal 1, Resque::Failure.count
+    assert_equal('Resque::DirtyExit', Resque::Failure.all['exception'])
+  end
+
+  test "fails uncompleted jobs with worker exception on exit" do
+    job = Resque::Job.new(:jobs, {'class' => 'GoodJob', 'args' => "blah"})
+    @worker.working_on(job)
+    @worker.unregister_worker(StandardError.new)
+    assert_equal 1, Resque::Failure.count
+    assert_equal('StandardError', Resque::Failure.all['exception'])
   end
 
   class ::SimpleJobWithFailureHandling
@@ -464,6 +473,40 @@ def self.perform
     assert_not_equal original_connection, Resque.redis.client.connection.instance_variable_get("@sock")
   end
 
+  test "tries to reconnect three times before giving up" do
+    begin
+      class Redis::Client
+        alias_method :original_reconnect, :reconnect
+
+        def reconnect
+          raise Redis::BaseConnectionError
+        end
+      end
+
+      class Resque::Worker
+        alias_method :original_sleep, :sleep
+
+        def sleep(duration = nil)
+          # noop
+        end
+      end
+
+      @worker.very_verbose = true
+      stdout, stderr = capture_io { @worker.work(0) }
+
+      assert_equal 3, stdout.scan(/retrying/).count
+      assert_equal 1, stdout.scan(/quitting/).count
+    ensure
+      class Redis::Client
+        alias_method :reconnect, :original_reconnect
+      end
+
+      class Resque::Worker
+        alias_method :sleep, :original_sleep
+      end
+    end
+  end
+
   if !defined?(RUBY_ENGINE) || defined?(RUBY_ENGINE) && RUBY_ENGINE != "jruby"
     test "old signal handling is the default" do
       rescue_time = nil
@@ -599,14 +642,14 @@ def self.perform( run_time, rescue_time=nil )
     end
 
     test "displays warning when not using term_child" do
-      stderr = capture_stderr { @worker.work(0) }
+      stdout, stderr = capture_io { @worker.work(0) }
 
       assert stderr.match(/^WARNING:/)
     end
 
     test "it does not display warning when using term_child" do
       @worker.term_child = "1"
-      stderr = capture_stderr { @worker.work(0) }
+      stdout, stderr = capture_io { @worker.work(0) }
 
       assert !stderr.match(/^WARNING:/)
     end