zookeeper 1.2.6-java → 1.2.7-java

data/CHANGELOG

@@ -1,3 +1,22 @@
+v1.2.7 further lock adjustments, deadlock risk reduction
+
+  * Refactor ZookeeperBase to not hold onto the mutex while waiting
+    for the dispatch thread to exit and the CZookeeper instance to close.
+    Instead, lock, nil out @czk, and unlock (which will cause all calls to
+    raise NotConnected), and then carry on with the shutdown procedure, greatly
+    reducing the chances of a deadlock. Also add a hardcoded 30 second timeout
+    to the join of the shutdown thread, that way we won't hang indefinitely in
+    the case of an unforseen condition.
+
+  * Improve the CZookeeper#wait_until_connected to use a deadline approach
+    to waiting for both running and connected states. Also, handle the 
+    'nil' (wait forever) timeout properly.
+
+  * Wake all waiting threads on all ConditionVariables when CZookeeper#shut_down!
+    is called
+
+v1.2.6 fix build on fedora
+
 v1.2.5 cleanup locking in ZookeeperBase
 
   * There were several situations where we would hold the lock before calling

data/ext/c_zookeeper.rb

@@ -161,7 +161,7 @@ class CZookeeper
 
   def state
     return ZOO_CLOSED_STATE if closed?
-    submit_and_block(:state)
+    @mutex.synchronize { @state }
   end
 
   # this implementation is gross, but i don't really see another way of doing it
@@ -172,12 +172,22 @@ class CZookeeper
   # if timeout is nil, we never time out, and wait forever for CONNECTED state
   #
   def wait_until_connected(timeout=10)
+    time_to_stop = timeout ? Time.now + timeout : nil
 
     return false unless wait_until_running(timeout)
 
     @mutex.synchronize do
-      # TODO: use deadline here
-      @state_cond.wait(timeout) unless (@state == ZOO_CONNECTED_STATE)
+      while true 
+        if timeout
+          now = Time.now
+          break if (@state == ZOO_CONNECTED_STATE) || @_shutting_down || @_closed || (now > time_to_stop)
+          delay = time_to_stop.to_f - now.to_f
+          @state_cond.wait(delay)
+        else
+          break if (@state == ZOO_CONNECTED_STATE) || @_shutting_down || @_closed
+          @state_cond.wait
+        end
+      end
     end
 
     connected?
@@ -343,7 +353,12 @@ class CZookeeper
     def shut_down!
       logger.debug { "##{__method__}" }
 
-      @mutex.synchronize { @_shutting_down = true }
+      @mutex.synchronize do 
+        @_shutting_down = true
+        # ollie ollie oxen all home free!
+        @state_cond.broadcast
+        @running_cond.broadcast
+      end
     end
 
     # called by underlying C code to signal we're running
@@ -355,11 +370,5 @@ class CZookeeper
         @running_cond.broadcast
       end
     end
-
-#     def notify_state_change!
-#       @mutex.synchronize do
-#         @state_cond.broadcast
-#       end
-#     end
 end
 end

data/ext/zookeeper_base.rb

@@ -130,20 +130,34 @@ class ZookeeperBase
   # potentially dangerous and should only be called after a fork() to close
   # this instance
   def close!
-    @czk && @czk.close
+    inst, @czk = @czk, nil
+    inst && inst.close
   end
 
   # close the connection normally, stops the dispatch thread and closes the
   # underlying connection cleanly
   def close
-    shutdown_thread = Thread.new do
-      @mutex.synchronize do
+    sd_thread = nil
+
+    @mutex.synchronize do
+      return unless @czk
+      inst, @czk = @czk, nil
+    
+      sd_thread = Thread.new(inst) do |_inst|
         stop_dispatch_thread!
-        close!
+        _inst.close
+      end
+    end
+
+    # if we're on the event dispatch thread for some stupid reason, then don't join
+    unless event_dispatch_thread?
+      # hard-coded 30 second delay, don't hang forever
+      if sd_thread.join(30) != sd_thread 
+        logger.error { "timed out waiting for shutdown thread to exit" }
       end
     end
 
-    shutdown_thread.join unless event_dispatch_thread?
+    nil
   end
 
   # the C lib doesn't strip the chroot path off of returned path values, which

data/lib/zookeeper/common.rb

@@ -10,15 +10,7 @@ module Common
     @dispatcher && (@dispatcher == Thread.current)
   end
 
-protected
-  def get_next_event(blocking=true)
-    @event_queue.pop(!blocking).tap do |event|
-      logger.debug { "#{self.class}##{__method__} delivering event #{event.inspect}" }
-    end
-  rescue ThreadError
-    nil
-  end
-
+private
   def setup_call(meth_name, opts)
     req_id = nil
     @mutex.synchronize {
@@ -71,21 +63,9 @@ protected
         return
       end
 
-      logger.debug {  "starting dispatch thread" }
-
-      @dispatcher = Thread.new do
-        while true
-          begin
-            dispatch_next_callback(get_next_event(true))
-          rescue QueueWithPipe::ShutdownException
-            logger.info { "dispatch thread exiting, got shutdown exception" }
-            break
-          rescue Exception => e
-            $stderr.puts ["#{e.class}: #{e.message}", e.backtrace.map { |n| "\t#{n}" }.join("\n")].join("\n")
-          end
-        end
-        signal_dispatch_thread_exit!
-      end
+      logger.debug { "starting dispatch thread" }
+
+      @dispatcher = Thread.new(&method(:dispatch_thread_body))
     end
   end
   
@@ -120,11 +100,12 @@ protected
     end
   end
 
-  def signal_dispatch_thread_exit!
-    @mutex.synchronize do
-      logger.debug { "dispatch thread exiting!" }
-      @dispatch_shutdown_cond.broadcast
+  def get_next_event(blocking=true)
+    @event_queue.pop(!blocking).tap do |event|
+      logger.debug { "#{self.class}##{__method__} delivering event #{event.inspect}" }
     end
+  rescue ThreadError
+    nil
   end
 
   def dispatch_next_callback(hash)
@@ -177,7 +158,28 @@ protected
     end
   end
 
-private
+  def dispatch_thread_body
+    while true
+      begin
+        dispatch_next_callback(get_next_event(true))
+      rescue QueueWithPipe::ShutdownException
+        logger.info { "dispatch thread exiting, got shutdown exception" }
+        return
+      rescue Exception => e
+        $stderr.puts ["#{e.class}: #{e.message}", e.backtrace.map { |n| "\t#{n}" }.join("\n")].join("\n")
+      end
+    end
+  ensure
+    signal_dispatch_thread_exit!
+  end
+
+  def signal_dispatch_thread_exit!
+    @mutex.synchronize do
+      logger.debug { "dispatch thread exiting!" }
+      @dispatch_shutdown_cond.broadcast
+    end
+  end
+
   def prettify_event(hash)
     hash.dup.tap do |h|
       # pretty up the event display

data/lib/zookeeper/version.rb

@@ -1,4 +1,4 @@
 module Zookeeper
-  VERSION = '1.2.6'
+  VERSION = '1.2.7'
   DRIVER_VERSION = '3.3.5'
 end

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification 
 name: zookeeper
 version: !ruby/object:Gem::Version 
-  hash: 19
+  hash: 17
   prerelease: 
   segments: 
   - 1
   - 2
-  - 6
-  version: 1.2.6
+  - 7
+  version: 1.2.7
 platform: java
 authors: 
 - Phillip Pearson
@@ -20,7 +20,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2012-06-04 00:00:00 Z
+date: 2012-06-05 00:00:00 Z
 dependencies: 
 - !ruby/object:Gem::Dependency 
   name: backports