On 17/03/2014 09:52, Martin Grigorov wrote: > On Mon, Mar 17, 2014 at 11:29 AM, Mark Thomas <ma...@apache.org> wrote: > >> On 17/03/2014 09:06, Martin Grigorov wrote: >>> Hi, >>> >>> A user of Apache Wicket has reported a problem where Tomcat hangs when >>> writing the response to the browser - >>> https://issues.apache.org/jira/browse/WICKET-5530 >>> >>> The application tries to upload a file (with Ajax) with bigger size than >>> the allowed one. Then in the Ajax response it tries to write 16K+ bytes >> and >>> Tomcat hangs with tracktrace like: >>> >>> Thread [qtp263823398-19] (Suspended) >>> owns: ByteArrayOutputStream2 (id=69) >>> owns: AbstractHttpConnection$OutputWriter (id=70) >>> SocketOutputStream.socketWrite0(FileDescriptor, byte[], int, int) >>> line: not available [native method] >>> SocketOutputStream.socketWrite(byte[], int, int) line: 113 >>> SocketOutputStream.write(byte[], int, int) line: 159 >>> ByteArrayBuffer.writeTo(OutputStream) line: 375 >>> SocketConnector$ConnectorEndPoint(StreamEndPoint).flush(Buffer) >> line: 164 >>> HttpGenerator.flushBuffer() line: 850 >>> HttpGenerator(AbstractGenerator).blockForOutput(long) line: 507 >>> AbstractHttpConnection$Output(HttpOutput).write(Buffer) line: 170 >>> AbstractHttpConnection$Output(HttpOutput).write(byte[], int, int) >> line: 107 >>> >> ByteArrayOutputStream2(ByteArrayOutputStream).writeTo(OutputStream) line: >> 154 >>> AbstractHttpConnection$OutputWriter(HttpWriter).write(char[], int, >>> int) line: 288 >>> AbstractHttpConnection$OutputWriter(HttpWriter).write(String, int, >>> int) line: 112 >>> AbstractHttpConnection$OutputWriter(HttpWriter).write(String, int, >>> int) line: 101 >>> AbstractHttpConnection$1(PrintWriter).write(String, int, int) >> line: 456 >>> >>> .... >>> >>> >>> >>> There is a demo app attached to the ticket and an attachment with the >>> full stacktrace. >>> >>> I personally wasn't able to reproduce it but two users have been able >>> to do it with Tomcat 7.0.47 on Windows. >>> >>> >>> Any help/hints are welcome! >> >> Best guess, the client isn't reading the response because it hasn't >> fully written the request. Tomcat isn't going to close the connection as >> that would prevent it writing a response. Tomcat writes the response but >> because it is bigger than the various network buffers, Tomcat hangs >> writing the response. >> >> I assume something times out eventually and closes the connection. >> >> Mark >> > > Thanks for the answer! > > What is the current default behavior of Tomcat 7/8 about swallowing the > file upload input in case its size is bigger than the configured max size? > I've found discussion [1] but it doesn't say whether the behavior has > changed in later versions of Tomcat. > > 1. http://markmail.org/message/c4qo4g2wc4sm7g53
http://tomcat.apache.org/tomcat-8.0-doc/config/context.html http://tomcat.apache.org/tomcat-7.0-doc/config/context.html Look for swallowAbortedUploads However, all of this is moot if it is the application reading the request body via getInputStream() rather than Tomcat. In that case it is up to the application to handle this case. Mark --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
