On Fri, 3 Mar 2023 12:07:43 GMT, Daniel Fuchs <[email protected]> wrote:
>> Please find here a fix that address a robustness issue in our test HTTP/2
>> server.
>>
>> Somewhat unrelated to the issue - but making the debug log easier to read -
>> I changed the toString() of OutgoingHeaders to be more meaningful.
>> OutgoingHeaders is a synthetic frame - the toString of the parent class was
>> making it appear as an UNKNOWN frame - which was confusing.
>>
>> The CancelRequestTest has been observed failing intermittently with this
>> exception:
>> Got expected exception: java.util.concurrent.ExecutionException:
>> java.io.EOFException: EOF reached while reading
>> cancelled
>> [jdk.internal.net.http.common.MinimalFuture@3564a02c]jdk.internal.net.http.common.MinimalFuture@3564a02c)[Completed
>> exceptionally: java.util.concurrent.CompletionException:
>> java.io.EOFException: EOF reached while reading] (id=6669)
>> test
>> CancelRequestTest.testPostSendAsync("http://localhost:49317/http2/x/same/interrupt",
>> true, true): failure
>> java.lang.AssertionError: expected [true] but found [false]
>> at org.testng.Assert.fail(Assert.java:99)
>> at org.testng.Assert.failNotEquals(Assert.java:1037)
>> at org.testng.Assert.assertTrue(Assert.java:45)
>> at org.testng.Assert.assertTrue(Assert.java:55)
>> at CancelRequestTest.testPostSendAsync(CancelRequestTest.java:453)
>>
>> Analysis of the logs shows that this is due to the Http2TestServer closing
>> the associated connection after failing to queue up a frame for the
>> previously cancelled stream in its connection readLoop. The frame could not
>> be queued because the previous stream was already closed.
>>
>> DEBUG: [pool-1-thread-4] [10s 773ms] Http2Connection(SocketTube(19)) Closed
>> stream 27
>> Sent response headers 200
>> DEBUG: [readLoop] [10s 773ms] FramesDecoder decodes: 9
>> DEBUG: [readLoop] [10s 773ms] FramesDecoder Tail size is now: 0, current=
>> DEBUG: [readLoop] [10s 773ms] FramesDecoder Got frame: DATA: length=0,
>> streamid=27, flags=END_STREAM
>> DEBUG: [readLoop] [10s 773ms] FramesDecoder decodes: 0
>> DEBUG: [readLoop] [10s 773ms] FramesDecoder Tail size is now: 0, current=
>> java.io.IOException: closed
>> at
>> jdk.httpclient.test.lib.http2.BodyOutputStream.write(BodyOutputStream.java:84)
>> at java.base/java.io.OutputStream.write(OutputStream.java:124)
>> at CancelRequestTest$HTTPSlowHandler.handle(CancelRequestTest.java:676)
>> at
>> jdk.httpclient.test.lib.common.HttpServerAdapters$HttpChain$Http2Chain.doFilter(HttpServerAdapters.java:476)
>> at
>> jdk.httpclient.test.lib.common.HttpServerAdapters$HttpTestServer$Http2TestContext.handle(HttpServerAdapters.java:746)
>> at
>> jdk.httpclient.test.lib.common.HttpServerAdapters$HttpTestHandler.doHandle(HttpServerAdapters.java:395)
>> at
>> jdk.httpclient.test.lib.common.HttpServerAdapters$HttpTestHandler.lambda$toHttp2Handler$1(HttpServerAdapters.java:382)
>> at
>> jdk.httpclient.test.lib.http2.Http2TestServerConnection.handleRequest(Http2TestServerConnection.java:721)
>> at
>> jdk.httpclient.test.lib.http2.Http2TestServerConnection.lambda$createStream$4(Http2TestServerConnection.java:671)
>> at
>> java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:577)
>> at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:317)
>> at
>> java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1144)
>> at
>> java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:642)
>> at java.base/java.lang.Thread.run(Thread.java:1623)
>> Http server reader thread shutdown
>> java.io.IOException: stream closed
>> at jdk.httpclient.test.lib.http2.Queue.put(Queue.java:64)
>> at
>> jdk.httpclient.test.lib.http2.Http2TestServerConnection.readLoop(Http2TestServerConnection.java:833)
>> at
>> jdk.httpclient.test.lib.http2.Http2TestServerConnection$ConnectionThread.run(Http2TestServerConnection.java:466)
>>
>> The readLoop should be more robust to asynchronous closing of HTTP/2 stream,
>> and just drop the frame if the stream is already closed.
>
> Daniel Fuchs has updated the pull request incrementally with one additional
> commit since the last revision:
>
> Fix trailing whitespaces
Marked as reviewed by jpai (Reviewer).
-------------
PR: https://git.openjdk.org/jdk/pull/12850