Various tweaks to improve server stability

4 files changed, 73 insertions, 39 deletions

diff --git a/src/ruby/.rubocop_todo.yml b/src/ruby/.rubocop_todo.yml
index ed4a4438b3..c35e970df6 100644
--- a/src/ruby/.rubocop_todo.yml
+++ b/src/ruby/.rubocop_todo.yml
@@ -1,5 +1,5 @@
 # This configuration was generated by `rubocop --auto-gen-config`
-# on 2015-04-17 14:43:27 -0700 using RuboCop version 0.30.0.
+# on 2015-05-22 13:23:34 -0700 using RuboCop version 0.30.1.
 # The point is for the user to remove these configuration records
 # one by one as the offenses are removed from the code base.
 # Note that changes in the inspected code, or installation of new
@@ -7,12 +7,12 @@
 
 # Offense count: 30
 Metrics/AbcSize:
-  Max: 40
+  Max: 38
 
 # Offense count: 3
 # Configuration parameters: CountComments.
 Metrics/ClassLength:
-  Max: 184
+  Max: 192
 
 # Offense count: 35
 # Configuration parameters: CountComments.
diff --git a/src/ruby/lib/grpc/generic/rpc_server.rb b/src/ruby/lib/grpc/generic/rpc_server.rb
index 665c144432..dcb11bfbef 100644
--- a/src/ruby/lib/grpc/generic/rpc_server.rb
+++ b/src/ruby/lib/grpc/generic/rpc_server.rb
@@ -76,7 +76,7 @@ module GRPC
       @jobs = Queue.new
       @size = size
       @stopped = false
-      @stop_mutex = Mutex.new
+      @stop_mutex = Mutex.new # needs to be held when accessing @stopped
       @stop_cond = ConditionVariable.new
       @workers = []
       @keep_alive = keep_alive
@@ -92,10 +92,15 @@ module GRPC
     # @param args the args passed blk when it is called
     # @param blk the block to call
     def schedule(*args, &blk)
-      fail 'already stopped' if @stopped
       return if blk.nil?
-      GRPC.logger.info('schedule another job')
-      @jobs << [blk, args]
+      @stop_mutex.synchronize do
+        if @stopped
+          GRPC.logger.warn('did not schedule job, already stopped')
+          return
+        end
+        GRPC.logger.info('schedule another job')
+        @jobs << [blk, args]
+      end
     end
 
     # Starts running the jobs in the thread pool.
@@ -116,8 +121,8 @@ module GRPC
     def stop
       GRPC.logger.info('stopping, will wait for all the workers to exit')
       @workers.size.times { schedule { throw :exit } }
-      @stopped = true
       @stop_mutex.synchronize do  # wait @keep_alive for works to stop
+        @stopped = true
         @stop_cond.wait(@stop_mutex, @keep_alive) if @workers.size > 0
       end
       forcibly_stop_workers
@@ -249,15 +254,18 @@ module GRPC
                    server_override:nil,
                    connect_md_proc:nil,
                    **kw)
-      @cq = RpcServer.setup_cq(completion_queue_override)
-      @server = RpcServer.setup_srv(server_override, @cq, **kw)
       @connect_md_proc = RpcServer.setup_connect_md_proc(connect_md_proc)
-      @pool_size = pool_size
+      @cq = RpcServer.setup_cq(completion_queue_override)
       @max_waiting_requests = max_waiting_requests
       @poll_period = poll_period
-      @run_mutex = Mutex.new
-      @run_cond = ConditionVariable.new
+      @pool_size = pool_size
       @pool = Pool.new(@pool_size)
+      @run_cond = ConditionVariable.new
+      @run_mutex = Mutex.new
+      @running = false
+      @server = RpcServer.setup_srv(server_override, @cq, **kw)
+      @stopped = false
+      @stop_mutex = Mutex.new
     end
 
     # stops a running server
@@ -266,20 +274,23 @@ module GRPC
     # server's current call loop is it's last.
     def stop
       return unless @running
-      @stopped = true
+      @stop_mutex.synchronize do
+        @stopped = true
+      end
       @pool.stop
+      @server.close
+    end
 
-      # TODO: uncomment this:
-      #
-      # This segfaults in the c layer, so its commented out for now.  Shutdown
-      # still occurs, but the c layer has to do the cleanup.
-      #
-      # @server.close
+    # determines if the server has been stopped
+    def stopped?
+      @stop_mutex.synchronize do
+        return @stopped
+      end
     end
 
     # determines if the server is currently running
     def running?
-      @running ||= false
+      @running
     end
 
     # Is called from other threads to wait for #run to start up the server.
@@ -311,11 +322,6 @@ module GRPC
       t.join
     end
 
-    # Determines if the server is currently stopped
-    def stopped?
-      @stopped ||= false
-    end
-
     # handle registration of classes
     #
     # service is either a class that includes GRPC::GenericService and whose
@@ -407,7 +413,13 @@ module GRPC
       request_call_tag = Object.new
       until stopped?
         deadline = from_relative_time(@poll_period)
-        an_rpc = @server.request_call(@cq, request_call_tag, deadline)
+        begin
+          an_rpc = @server.request_call(@cq, request_call_tag, deadline)
+        rescue Core::CallError, RuntimeError => e
+          # can happen during server shutdown
+          GRPC.logger.warn("server call failed: #{e}")
+          next
+        end
         c = new_active_server_call(an_rpc)
         unless c.nil?
           mth = an_rpc.method.to_sym
diff --git a/src/ruby/spec/generic/rpc_server_pool_spec.rb b/src/ruby/spec/generic/rpc_server_pool_spec.rb
index aae3a7d7cb..b67008de48 100644
--- a/src/ruby/spec/generic/rpc_server_pool_spec.rb
+++ b/src/ruby/spec/generic/rpc_server_pool_spec.rb
@@ -74,11 +74,11 @@ describe GRPC::Pool do
   end
 
   describe '#schedule' do
-    it 'throws if the pool is already stopped' do
+    it 'return if the pool is already stopped' do
       p = Pool.new(1)
       p.stop
       job = proc {}
-      expect { p.schedule(&job) }.to raise_error
+      expect { p.schedule(&job) }.to_not raise_error
     end
 
     it 'adds jobs that get run by the pool' do
diff --git a/src/ruby/spec/generic/rpc_server_spec.rb b/src/ruby/spec/generic/rpc_server_spec.rb
index 640b0f656c..e60a8b27c3 100644
--- a/src/ruby/spec/generic/rpc_server_spec.rb
+++ b/src/ruby/spec/generic/rpc_server_spec.rb
@@ -212,10 +212,14 @@ describe GRPC::RpcServer do
 
   describe '#stopped?' do
     before(:each) do
-      opts = { a_channel_arg: 'an_arg', poll_period: 1 }
+      opts = { a_channel_arg: 'an_arg', poll_period: 1.5 }
       @srv = RpcServer.new(**opts)
     end
 
+    after(:each) do
+      @srv.stop
+    end
+
     it 'starts out false' do
       expect(@srv.stopped?).to be(false)
     end
@@ -225,7 +229,7 @@ describe GRPC::RpcServer do
       expect(@srv.stopped?).to be(false)
     end
 
-    it 'stays false after the server starts running' do
+    it 'stays false after the server starts running', server: true do
       @srv.handle(EchoService)
       t = Thread.new { @srv.run }
       @srv.wait_till_running
@@ -234,7 +238,7 @@ describe GRPC::RpcServer do
       t.join
     end
 
-    it 'is true after a running server is stopped' do
+    it 'is true after a running server is stopped', server: true do
       @srv.handle(EchoService)
       t = Thread.new { @srv.run }
       @srv.wait_till_running
@@ -251,21 +255,22 @@ describe GRPC::RpcServer do
       expect(r.running?).to be(false)
     end
 
-    it 'is false after run is called with no services registered' do
+    it 'is false if run is called with no services registered', server: true do
       opts = {
         a_channel_arg: 'an_arg',
-        poll_period: 1,
+        poll_period: 2,
         server_override: @server
       }
       r = RpcServer.new(**opts)
       r.run
       expect(r.running?).to be(false)
+      r.stop
     end
 
     it 'is true after run is called with a registered service' do
       opts = {
         a_channel_arg: 'an_arg',
-        poll_period: 1,
+        poll_period: 2.5,
         server_override: @server
       }
       r = RpcServer.new(**opts)
@@ -284,6 +289,10 @@ describe GRPC::RpcServer do
       @srv = RpcServer.new(**@opts)
     end
 
+    after(:each) do
+      @srv.stop
+    end
+
     it 'raises if #run has already been called' do
       @srv.handle(EchoService)
       t = Thread.new { @srv.run }
@@ -335,6 +344,10 @@ describe GRPC::RpcServer do
         @srv = RpcServer.new(**server_opts)
       end
 
+      after(:each) do
+        @srv.stop
+      end
+
       it 'should return NOT_FOUND status on unknown methods', server: true do
         @srv.handle(EchoService)
         t = Thread.new { @srv.run }
@@ -376,7 +389,7 @@ describe GRPC::RpcServer do
         t.join
       end
 
-      it 'should receive metadata when a deadline is specified', server: true do
+      it 'should receive metadata if a deadline is specified', server: true do
         service = SlowService.new
         @srv.handle(service)
         t = Thread.new { @srv.run }
@@ -445,11 +458,11 @@ describe GRPC::RpcServer do
 
       it 'should handle multiple parallel requests', server: true do
         @srv.handle(EchoService)
-        Thread.new { @srv.run }
+        t = Thread.new { @srv.run }
         @srv.wait_till_running
         req, q = EchoMsg.new, Queue.new
         n = 5  # arbitrary
-        threads = []
+        threads = [t]
         n.times do
           threads << Thread.new do
             stub = EchoStub.new(@host, **client_opts)
@@ -472,7 +485,7 @@ describe GRPC::RpcServer do
         }
         alt_srv = RpcServer.new(**opts)
         alt_srv.handle(SlowService)
-        Thread.new { alt_srv.run }
+        t = Thread.new { alt_srv.run }
         alt_srv.wait_till_running
         req = EchoMsg.new
         n = 5  # arbitrary, use as many to ensure the server pool is exceeded
@@ -490,6 +503,7 @@ describe GRPC::RpcServer do
         end
         threads.each(&:join)
         alt_srv.stop
+        t.join
         expect(one_failed_as_unavailable).to be(true)
       end
     end
@@ -513,6 +527,10 @@ describe GRPC::RpcServer do
         @srv = RpcServer.new(**server_opts)
       end
 
+      after(:each) do
+        @srv.stop
+      end
+
       it 'should send connect metadata to the client', server: true do
         service = EchoService.new
         @srv.handle(service)
@@ -545,6 +563,10 @@ describe GRPC::RpcServer do
         @srv = RpcServer.new(**server_opts)
       end
 
+      after(:each) do
+        @srv.stop
+      end
+
       it 'should be added to BadStatus when requests fail', server: true do
         service = FailingService.new
         @srv.handle(service)