[ 
https://issues.apache.org/jira/browse/CXF-9271?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18141446#comment-18141446
 ] 

Freeman Yue Fang edited comment on CXF-9271 at 11/9/26, 5:01 PM:
-----------------------------------------------------------------

Hi [~reta],

I just reproduced with the test case and with the fix for CXF-9268
{code}
--- 
a/rt/transports/http-hc5/src/test/java/org/apache/cxf/transport/http/asyncclient/hc5/SharedInputBufferTest.java
+++ 
b/rt/transports/http-hc5/src/test/java/org/apache/cxf/transport/http/asyncclient/hc5/SharedInputBufferTest.java
@@ -27,10 +27,13 @@ import java.util.concurrent.ExecutorService;
 import java.util.concurrent.Executors;
 import java.util.concurrent.Future;
 import java.util.concurrent.TimeUnit;
+import java.util.function.Consumer;
 
 import org.junit.Test;
 
 import static org.junit.Assert.assertArrayEquals;
+import static org.junit.Assert.assertEquals;
+import static org.junit.Assert.assertFalse;
 
 public class SharedInputBufferTest {
 
@@ -81,4 +84,38 @@ public class SharedInputBufferTest {
             executor.shutdownNow();
         }
     }
+
+    @Test
+    public void testShutdownWakesUpConsumerWaitingForBufferToBeDrained() 
throws Exception {
+        assertConsumerWokenUp(SharedInputBuffer::shutdown);
+    }
+
+    @Test
+    public void testCloseWakesUpConsumerWaitingForBufferToBeDrained() throws 
Exception {
+        assertConsumerWokenUp(SharedInputBuffer::close);
+    }
+
+    private static void assertConsumerWokenUp(Consumer<SharedInputBuffer> 
abort) throws Exception {
+        final SharedInputBuffer inbuf = new SharedInputBuffer(1024);
+        // A block larger than the buffer fills it, and with no reader 
draining it the I/O thread
+        // waits until the reader has drained it or the response is aborted
+        final Thread io = new Thread(() -> {
+            try {
+                inbuf.consumeContent(ByteBuffer.allocate(4096), false);
+            } catch (IOException ex) {
+                // aborted
+            }
+        });
+        io.setDaemon(true);
+        io.start();
+        final long deadline = System.nanoTime() + TimeUnit.SECONDS.toNanos(10);
+        while (io.getState() != Thread.State.WAITING && System.nanoTime() < 
deadline) {
+            Thread.sleep(10);
+        }
+        assertEquals(Thread.State.WAITING, io.getState());
+
+        abort.accept(inbuf);
+        io.join(TimeUnit.SECONDS.toMillis(10));
+        assertFalse("The I/O thread is still waiting for the buffer to be 
drained", io.isAlive());
+    }

{code}

So the fix for CXF-9268 didn't cover this. And the suggested fix here seems 
work.

Best Regards
Freeman


was (Author: ffang):
Hi [~reta],

I just verified with the test case
{code}
--- 
a/rt/transports/http-hc5/src/test/java/org/apache/cxf/transport/http/asyncclient/hc5/SharedInputBufferTest.java
+++ 
b/rt/transports/http-hc5/src/test/java/org/apache/cxf/transport/http/asyncclient/hc5/SharedInputBufferTest.java
@@ -27,10 +27,13 @@ import java.util.concurrent.ExecutorService;
 import java.util.concurrent.Executors;
 import java.util.concurrent.Future;
 import java.util.concurrent.TimeUnit;
+import java.util.function.Consumer;
 
 import org.junit.Test;
 
 import static org.junit.Assert.assertArrayEquals;
+import static org.junit.Assert.assertEquals;
+import static org.junit.Assert.assertFalse;
 
 public class SharedInputBufferTest {
 
@@ -81,4 +84,38 @@ public class SharedInputBufferTest {
             executor.shutdownNow();
         }
     }
+
+    @Test
+    public void testShutdownWakesUpConsumerWaitingForBufferToBeDrained() 
throws Exception {
+        assertConsumerWokenUp(SharedInputBuffer::shutdown);
+    }
+
+    @Test
+    public void testCloseWakesUpConsumerWaitingForBufferToBeDrained() throws 
Exception {
+        assertConsumerWokenUp(SharedInputBuffer::close);
+    }
+
+    private static void assertConsumerWokenUp(Consumer<SharedInputBuffer> 
abort) throws Exception {
+        final SharedInputBuffer inbuf = new SharedInputBuffer(1024);
+        // A block larger than the buffer fills it, and with no reader 
draining it the I/O thread
+        // waits until the reader has drained it or the response is aborted
+        final Thread io = new Thread(() -> {
+            try {
+                inbuf.consumeContent(ByteBuffer.allocate(4096), false);
+            } catch (IOException ex) {
+                // aborted
+            }
+        });
+        io.setDaemon(true);
+        io.start();
+        final long deadline = System.nanoTime() + TimeUnit.SECONDS.toNanos(10);
+        while (io.getState() != Thread.State.WAITING && System.nanoTime() < 
deadline) {
+            Thread.sleep(10);
+        }
+        assertEquals(Thread.State.WAITING, io.getState());
+
+        abort.accept(inbuf);
+        io.join(TimeUnit.SECONDS.toMillis(10));
+        assertFalse("The I/O thread is still waiting for the buffer to be 
drained", io.isAlive());
+    }

{code}

So the fix for CXF-9268 didn't cover this. And the suggested fix here seems 
work.

Best Regards
Freeman

> http-hc5: SharedInputBuffer.consumeContent() can park the I/O reactor thread 
> forever — close()/shutdown() never signal suspendInput
> -----------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: CXF-9271
>                 URL: https://issues.apache.org/jira/browse/CXF-9271
>             Project: CXF
>          Issue Type: Bug
>          Components: Transports
>    Affects Versions: 4.2.4
>         Environment: Apache CXF 4.2.4 (also reproduces on 4.2.x)
> Module: rt/transports/http-hc5
> httpclient5 / httpcore5 5.4.x
>            Reporter: Alexander Glazovskiy
>            Assignee: Andriy Redko
>            Priority: Critical
>
> SharedInputBuffer.consumeContent() waits on the suspendInput condition when 
> the internal buffer is full, but close() and shutdown() only signal the other 
> condition. If either method is called while a dispatch thread is parked in 
> consumeContent(), that thread is never woken up and the I/O reactor 
> permanently loses one of its threads.
> Under load this compounds into complete connection-pool starvation and 
> OutOfMemoryError.
> It can be reproduced with any SOAP call whose response body is larger than 
> chunkLength, where the subscription is cancelled while the body is still 
> being received.



--
This message was sent by Atlassian Jira
(v10.3.26#10030026)

Reply via email to