New HttpClient PUT request fails oddly - is it size is it number of invocations!
Jaikiran Pai
jai.forums2013 at gmail.com
Tue Jun 26 14:41:49 UTC 2018
Hi Daniel,
You are right- I just added:
diff -r 955a66f0f04a test/jdk/java/net/httpclient/PUTRequestSizeTest.java
--- a/test/jdk/java/net/httpclient/PUTRequestSizeTest.java Tue Jun
26 18:33:18 2018 +0530
+++ b/test/jdk/java/net/httpclient/PUTRequestSizeTest.java Tue Jun
26 20:05:20 2018 +0530
@@ -92,6 +92,7 @@
public void handle(final HttpExchange httpExchange) throws
IOException {
final URI requestURI = httpExchange.getRequestURI();
System.out.println("Handling " +
httpExchange.getRequestMethod() + " request " + requestURI);
+ httpExchange.getRequestBody().readAllBytes();
httpExchange.sendResponseHeaders(200, -1);
}
}
(and no more Thread.sleep())
and it started passing.
So is this a case of the server implementation misbehaving? I haven't
checked what the expectations are from a spec point of view for handling
such requests on the server. Or is it a combination of a misbehaving
server plus the HttpClient implementation having some issue? I think the
fact that it fails after what appear to be a couple of successful
invocations and the odd nature of the exception message, makes it hard
to narrow this down from a client usage point of view.Thank you
narrowing this down quickly and for the -ea hints.
-Jaikiran
On 26/06/18 8:02 PM, Daniel Fuchs wrote:
> Hi Jaikiran,
>
> I believe you need to drain the request body in the server
> before sending the response.
>
> If you pass -ea -esa options to jtreg, you should see the
> following assertion printed by the server:
>
> java.lang.AssertionError: State is not RESPONSE (REQUEST)
> at
> jdk.httpserver/sun.net.httpserver.ServerImpl.responseCompleted(ServerImpl.java:814)
> at
> jdk.httpserver/sun.net.httpserver.ServerImpl$Dispatcher.handleEvent(ServerImpl.java:297)
> at
> jdk.httpserver/sun.net.httpserver.ServerImpl$Dispatcher.run(ServerImpl.java:356)
> at java.base/java.lang.Thread.run(Thread.java:834)
>
> I added this line in the server handler (before calling sendResponse()):
>
> httpExchange.getRequestBody().readAllBytes();
>
> and with that the assertion disappears and the test passes for me.
>
> Could you try adding this line to your test, and run it again
> in your environment to confirm whether that might be the actual
> issue ? (I'm obviously running on a different machine than you, so
> if it's really a race condition I am concerned that I might not
> be seeing it)
>
> best regards,
>
> -- daniel
>
> On 26/06/2018 14:52, Jaikiran Pai wrote:
>> So it looks like there some race condition somewhere in the
>> HttpClient implementation. I just addeda smalldelay between the PUT
>> requests, in the test that I sent in the patch and it's now started
>> passing:
>>
>> diff -r 955a66f0f04a
>> test/jdk/java/net/httpclient/PUTRequestSizeTest.java
>> --- a/test/jdk/java/net/httpclient/PUTRequestSizeTest.java Tue Jun
>> 26 18:33:18 2018 +0530
>> +++ b/test/jdk/java/net/httpclient/PUTRequestSizeTest.java Tue Jun
>> 26 19:19:54 2018 +0530
>> @@ -59,6 +59,9 @@
>> for (int i = 0; i < 10; i++) {
>> System.out.println("Sending PUT request " + (i +1));
>> issuePUT(httpClient, requestURL);
>> + // TODO: This shouldn't be needed but is here to
>> demonstrate a race
>> + // condition somewhere in the HttpClient implementation
>> + Thread.sleep(500);
>> }
>> } finally {
>> server.stop(0);
>>
>>
>>
>> -Jaikiran
>> On 26/06/18 7:12 PM, Jaikiran Pai wrote:
>>> In my random experimentation with the new HttpClient API usage, I
>>> have ended up running into an odd and hard to decipher exception
>>> when dealing with PUT requests. I am noticing that if I issue
>>> multiple PUT requests using the same HttpClient instance, the first
>>> 2 invocations succeed while the 3rd one fails with an exception[1].
>>> Initially when I ran into this, I saw this happening depending on
>>> the size of the data being uploaded via PUT request. As soon as it
>>> hit 16385 bytes (16 KB + 1), it would end up throwing the odd
>>> exception. However, when I decided to narrow it down to a testcase,
>>> I was able to reproduce this without the size of the data playing a
>>> role. And weirdly, it now keeps failing for the third invocation.
>>>
>>> I have now isolated this into a jtreg testcase and created a patch
>>> against the latest upstream jdk to reproduce this issue. I've
>>> attached the patch in this mail. In the test, I create a local
>>> server and keep sending PUT requests to the server using the same
>>> instance of HttpClient. As soon as it hits the 3rd invocation, I end
>>> up seeing the exception[1].
>>>
>>> It's really odd, since, initially I thought it could be a size based
>>> issue, which would have been more understandable. However with the
>>> way it's failing now in this jtreg test, I am starting to wonder if
>>> I have got some basics wrong.
>>>
>>>
>>> [1]
>>>
>>> java.io.IOException: HTTP/1.1 header parser received no bytes
>>> at
>>> java.net.http/jdk.internal.net.http.HttpClientImpl.send(HttpClientImpl.java:546)
>>>
>>> at
>>> java.net.http/jdk.internal.net.http.HttpClientFacade.send(HttpClientFacade.java:113)
>>>
>>> at PUTRequestSizeTest.issuePUT(PUTRequestSizeTest.java:76)
>>> at PUTRequestSizeTest.main(PUTRequestSizeTest.java:61)
>>> at
>>> java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native
>>> Method)
>>> at
>>> java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
>>>
>>> at
>>> java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
>>>
>>> at java.base/java.lang.reflect.Method.invoke(Method.java:566)
>>> at
>>> com.sun.javatest.regtest.agent.MainWrapper$MainThread.run(MainWrapper.java:115)
>>>
>>> at java.base/java.lang.Thread.run(Thread.java:834)
>>> Caused by: java.io.IOException: HTTP/1.1 header parser received no
>>> bytes
>>> at
>>> java.net.http/jdk.internal.net.http.common.Utils.wrapWithExtraDetail(Utils.java:293)
>>>
>>> at
>>> java.net.http/jdk.internal.net.http.Http1Response$HeadersReader.onReadError(Http1Response.java:646)
>>>
>>> at
>>> java.net.http/jdk.internal.net.http.Http1AsyncReceiver.checkForErrors(Http1AsyncReceiver.java:297)
>>>
>>> at
>>> java.net.http/jdk.internal.net.http.Http1AsyncReceiver.flush(Http1AsyncReceiver.java:263)
>>>
>>> at
>>> java.net.http/jdk.internal.net.http.common.SequentialScheduler$SynchronizedRestartableTask.run(SequentialScheduler.java:175)
>>>
>>> at
>>> java.net.http/jdk.internal.net.http.common.SequentialScheduler$CompleteRestartableTask.run(SequentialScheduler.java:147)
>>>
>>> at
>>> java.net.http/jdk.internal.net.http.common.SequentialScheduler$SchedulableTask.run(SequentialScheduler.java:198)
>>>
>>> at
>>> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
>>>
>>> at
>>> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
>>>
>>> ... 1 more
>>> Caused by: java.io.EOFException: EOF reached while reading
>>> at
>>> java.net.http/jdk.internal.net.http.Http1AsyncReceiver$Http1TubeSubscriber.onComplete(Http1AsyncReceiver.java:587)
>>>
>>> at
>>> java.net.http/jdk.internal.net.http.SocketTube$InternalReadPublisher$ReadSubscription.signalCompletion(SocketTube.java:629)
>>>
>>> at
>>> java.net.http/jdk.internal.net.http.SocketTube$InternalReadPublisher$InternalReadSubscription.read(SocketTube.java:830)
>>>
>>> at
>>> java.net.http/jdk.internal.net.http.SocketTube$SocketFlowTask.run(SocketTube.java:175)
>>>
>>> at
>>> java.net.http/jdk.internal.net.http.common.SequentialScheduler$SchedulableTask.run(SequentialScheduler.java:198)
>>>
>>> at
>>> java.net.http/jdk.internal.net.http.common.SequentialScheduler.runOrSchedule(SequentialScheduler.java:271)
>>>
>>> at
>>> java.net.http/jdk.internal.net.http.common.SequentialScheduler.runOrSchedule(SequentialScheduler.java:224)
>>>
>>> at
>>> java.net.http/jdk.internal.net.http.SocketTube$InternalReadPublisher$InternalReadSubscription.handleSubscribeEvent(SocketTube.java:687)
>>>
>>> at
>>> java.net.http/jdk.internal.net.http.AsyncTriggerEvent.handle(AsyncTriggerEvent.java:54)
>>>
>>> at
>>> java.net.http/jdk.internal.net.http.HttpClientImpl$SelectorManager.run(HttpClientImpl.java:796)
>>>
>>>
>>
>
More information about the net-dev
mailing list