axis-c-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Carsten Blecken" <cblec...@macrovision.com>
Subject RE: timeout exception because of no Content-Length
Date Mon, 11 Jul 2005 15:05:28 GMT
+1 for adding an exception in case the transport returns TRANSPORT_FINISHED. This case needs
to be
   handled by the axis c engine and not the transport.
 
Carsten

-----Original Message-----
From: Fred Preston [mailto:PRESTONF@uk.ibm.com]
Sent: Monday, July 11, 2005 5:33 AM
To: Apache AXIS C Developers List
Subject: RE: timeout exception because of no Content-Length



Hi Bernd, 
        Good idea, but the parser works on a 'pull' model where information is only 'pulled'
from the socket when Xerces internally runs out of information to parse and needs to draw
more information from the socket.  Thus, it does not matter what proportion of the message
has been received because if the parser needs more information to complete, it will attempt
to read more information from the socket.  What could go into readBytes is some exception
testing that would throw an exception if a readBytes was called once the Transport had read
all of the expected bytes and returned with a TRANSPORT_FINISHED.  The iMaxToRead is a maximum
size of pcToFill array and not an 'expecting to read' value.  BUFFER_SIZE reflects the size
of a TCP packet, i.e. the maximum amount of data that can be transmitted per packet or the
maximum granularity of the message data size. 

As a footnote to all this, if extra bytes are transmitted after the end of the response message
(that is not part of the XML message), who will drain these additional bytes? 

Regards,

Fred Preston.




	bernd.weber04@polizei.bayern.de 


07/11/2005 12:29 PM 
Please respond to "Apache AXIS C Developers List" 

        
        To:        axis-c-dev@ws.apache.org 
        cc:         
        Subject:        RE: timeout exception because of no Content-Length 

       


Hi All,

I am using axis-c-final-1.5 and encountered problems of that kind, too,
but it seemed not to be a time-out problem. I took a look at
src/xml/xerces/SoapBinInputStream.cpp:41 - method readBytes

unsigned int SoapBinInputStream::readBytes(XMLByte* const pcToFill,
                                const unsigned int iMaxToRead)
{
                int iRead = iMaxToRead;
                m_pInputStream->getBytes((char*)pcToFill, &iRead);
                m_nByteCount += iRead;
                return iRead;
}

This method doesn't recognize if transport is still in progress after
m_pInputStream->getBytes was called. If iRead is less than iMaxToRead
and there are some more bytes left to receive, they get lost. If you
think of the buffer size of HTTPChannel (BUF_SIZE == 8096) and the value
of iMaxToRead (49152), you see that this must lead to truncated data
if the amount to receive is greater that BUF_SIZE.
This in mind, I rewrote that method:

unsigned int SoapBinInputStream::readBytes(XMLByte* const pcToFill,
                                          const unsigned int iMaxToRead)
{
                // Bernd Weber: call m_pInputStream->getBytes in a loop checking
                // AXIS_TRANSPORT_STATUS and the remaining space

                unsigned iLoopRead = 0;
                AXIS_TRANSPORT_STATUS status = TRANSPORT_IN_PROGRESS;
                do {
                                 int iRead = iMaxToRead - iLoopRead;
                                 char* pcBuffer = (char*)(pcToFill + iLoopRead);
                                 status = m_pInputStream->getBytes(pcBuffer, &iRead);
                                 m_nByteCount += iRead;                                  //
this instance's byte
amount
                                 iLoopRead += iRead;                                  // this
loop's byte amount
                                 if (iLoopRead < iMaxToRead) {
                                                  pcToFill[iLoopRead] = 0;
                                 }
                } while (status == TRANSPORT_IN_PROGRESS && iMaxToRead > iLoopRead);
                return iLoopRead;
}


Please check if these changes fix your problems, too, and if they could
get part of the next Axis release.


Best regards and thanks for Axis-c!

Bernd Weber
bernd.weber@t-systems.com
(bernd.weber04@polizei.bayern.de)




> -----Original Message-----
> From: krishna [mailto:krishna@lac.uic.edu]
> Sent: Friday, July 08, 2005 3:59 PM
> To: Apache AXIS C Developers List
> Subject: Re: timeout exception because of no Content-Length
> 
> 
> Hi Samisa/All
> 
> I varied the timeout with the following values - 0,10,100,1000 -
> and received the same amount of data in the SOAP-RESPONSE each time.
> 
> Cheers
> Krishna
> 
> Samisa Abeysinghe wrote:
> 
> >Hi,
> >                 Even though the variable name used in the implementation is
> >m_iContentLength, it has nothing to do with the HTTP Content-Length
> >header. As Carsten has pointed out, there is no need to have the
> >Content-Length HTTP header, if the message is chunked.
> >                 Our transport has been fixed to work with chunked data long time
> >back and it has been working OK since then.
> >                 Please vary your time out and see if you get lager message with
> >larger timeout, or if you get the same amount of message 
> with different
> >timeouts.
> >                 If you get larger message with larger timeout, then you can
> >solve the problem by increasing the timeout. However, if 
> that is not the
> >case, then there may be a bug which we would have to dig into.
> >
> >Thanks,
> >Samisa...
> >
> >-----Original Message-----
> >From: Carsten Blecken [mailto:cblecken@macrovision.com] 
> >Sent: Friday, July 08, 2005 4:37 AM
> >To: Apache AXIS C Developers List
> >Subject: RE: timeout exception because of no Content-Length
> >
> >Hi,
> >
> >if you are saying that the chunked code path depends on
> >Content-Length to be set, then this is likely a bug in
> >the code. The http spec does NOT require content-length
> >to be set if transfer encoding is chunked. There a several
> >app servers out there who use chunking (I know Jetty for
> >sure) and no content length. The logic should be to look
> >for 'Transfer-Encoding : chunked' first and only after 
> >none was found to look for Content-length. 
> >
> >Carsten
> >
> >P.S> The relevant HTTP spec part -
> >
> >--------------------------------------------
> >
> >When a message-body is included with a message, the 
> transfer-length of
> >that body is determined by one of the following (in order of
> >precedence):
> >
> >1.Any response message which "MUST NOT" include a 
> message-body (such as
> >the 1xx, 204, and 304 responses and any response to a HEAD 
> request) is
> >always terminated by the first empty line after the header fields,
> >regardless of the entity-header fields present in the message.
> >
> >2.If a Transfer-Encoding header field (section 14.41) is 
> present and has
> >any value other than "identity", then the transfer-length is 
> defined by
> >use of the "chunked" transfer-coding (section 3.6), unless 
> the message
> >is terminated by closing the connection.
> >
> >3.If a Content-Length header field (section 14.13) is present, its
> >decimal value in OCTETs represents both the entity-length and the
> >transfer-length. The Content-Length header field MUST NOT be sent if
> >these two lengths are different (i.e., if a Transfer-Encoding
> >
> >     header field is present). If a message is received with both a
> >     Transfer-Encoding header field and a Content-Length 
> header field,
> >     the latter MUST be ignored.
> >
> >4.If the message uses the media type "multipart/byteranges", and the
> >ransfer-length is not otherwise specified, then this self- elimiting
> >media type defines the transfer-length. This media type UST 
> NOT be used
> >unless the sender knows that the recipient can arse it; the 
> presence in
> >a request of a Range header with ultiple byte- range 
> specifiers from a
> >1.1 client implies that the lient can parse multipart/byteranges
> >responses.
> >
> >       A range header might be forwarded by a 1.0 proxy that does not
> >       understand multipart/byteranges; in this case the server MUST
> >       delimit the message using methods defined in items 1,3 or 5 of
> >       this section.
> >
> >5.By the server closing the connection. (Closing the 
> connection cannot
> >be used to indicate the end of a request body, since that 
> would leave no
> >possibility for the server to send back a response.) 
> >
> >
> >-----Original Message-----
> >From: krishna [mailto:krishna@lac.uic.edu]
> >Sent: Thursday, July 07, 2005 3:23 PM
> >To: Apache AXIS C Developers List
> >Subject: timeout exception because of no Content-Length
> >
> >
> >Hi Samisa/Fred/All,
> >
> >Both the methods
> >1. Using the stubObj to setTransportTimeout
> >2. Calling setTimeout from the stub code
> >
> >Gave me the following error.
> >
> >Error
> >HTTPTransportException:Input streaming error while getting 
> data Timed 
> >out waiting for SOAP message (1).
> >
> >My analysis showed that the code with for http chunking has 2 parts.
> >1. For the initial chunk
> >2. For the subsequent chunks
> >
> >It relies on the content length field to decide whether the current 
> >chunk is the first chunk or not
> >Please find the lines below from HTTPTransport.cpp
> >
> >if( m_strReceived.length () > 0)
> >{
> >627
> >628 if( m_bChunked && m_iContentLength < 1) // Read first chunk
> >629 {
> >
> >
> >.......
> >722 }
> >723 else if( m_bChunked) // read continued portions of a chunk
> >724 {
> >
> >As the SOAP - RESPONSE does not contain a Content-Length 
> field. It keeps
> >
> >repeating the algorithm for the first chunk
> >and lands in an exception.
> >
> >If my analysis is right then my question is why does the 
> server not send
> >
> >a Content-Length field for chunked messages and how can this be 
> >incorporated?
> >
> >Cheers
> >Krishna
> >
> >Samisa Abeysinghe wrote:
> >
> >  
> >
> >>Well, without editing the generated code, you can change the user
> >>    
> >>
> >code:
> >  
> >
> >>stubObj->setTransportTimeout(30);//assuming the stub object is a
> >>    
> >>
> >pointer
> >  
> >
> >>Samisa...
> >>
> >>-----Original Message-----
> >>*From:* Fred Preston [mailto:PRESTONF@uk.ibm.com]
> >>*Sent:* Thursday, July 07, 2005 3:34 PM
> >>*To:* Apache AXIS C Developers List
> >>*Subject:* Re: how to set a timeout - update
> >>
> >>
> >>Hi Krishna,
> >>The IterationCountdown in the HTTPTransport::getBytes() 
> method is not 
> >>a timeout, but is used to catch a potential deadlock 
> situation where 
> >>the client is waiting for a response message from the 
> server and the 
> >>server either never or only partially responds. The timeout for the 
> >>socket is set by HTTPTransport::setTimeout( long lSeconds). The 
> >>default channel timeout is 10 seconds. To change the 
> timeout you have 
> >>to be able to call the setTimeout method. You cannot do 
> this directly 
> >>from the client application because it does not have access to the 
> >>transport layer. You can however change the timeout at the 
> generated 
> >>stub level. The best time to change the timeout is before 
> the invoke, 
> >>i.e.
> >>
> >>*In the stub code for the web service method, change:*
> >>:
> >>if (AXIS_SUCCESS == m_pCall->invoke())
> >>{
> >>:
> >>
> >>*To:*
> >>:
> >>m_pCall->getTransport()->setTimeout( 30); // Change the 
> socket timeout
> >>    
> >>
> >
> >  
> >
> >>to 30 seconds
> >>if (AXIS_SUCCESS == m_pCall->invoke())
> >>{
> >>:
> >>
> >>Regards,
> >>
> >>Fred Preston.
> >>
> >>
> >>                 
> >>
> >>*krishna <krishna@lac.uic.edu>*
> >>
> >>07/07/2005 01:24 AM
> >>Please respond to "Apache AXIS C Developers List"
> >>
> >>                 
> >>
> >>
> >>To: Apache AXIS C Developers List <axis-c-dev@ws.apache.org>
> >>cc:
> >>Subject: Re: how to set a timeout - update
> >>
> >>
> >>
> >>Hi
> >>
> >>Sorry for the traffic in the mailing list.
> >>
> >>I traced the exception back to its where it was thrown. It 
> looks like
> >>the timeout is hardcoded to 100 by setting 
> iIterationCountdown=100 in
> >>the HTTPTransport.cpp:655 file.
> >>
> >>Is there a way of resetting this timeout without recompiling axis or
> >>Is there a way of recompiling the transport alone??
> >>
> >>Cheers
> >>Krishna
> >>
> >>krishna wrote:
> >>
> >>    
> >>
> >>>Hi All
> >>>
> >>>The problem why the Array example was failing was due to 
> timeouts. It
> >>>is "HTTPTransportException:Input streaming error while getting data
> >>>Timed out waiting for SOAP message (1)." I saw a discussion in the
> >>>lists about connection timeouts and response timeouts. Would this
> >>>      
> >>>
> >fall
> >  
> >
> >>>under a response timeout ?
> >>>
> >>>I saw that Axis Java has a setTimeout functionality as part of the
> >>>stubs. I could not find it in Stub.hpp. Does this 
> functionality exist
> >>>in Axis C++??
> >>>
> >>>Cheers
> >>>Krishna
> >>>
> >>>
> >>>krishna wrote:
> >>>
> >>>      
> >>>
> >>>>Hi All
> >>>>
> >>>>I was implementing the Array example on Axis 1.6 alpha. (
> >>>>echoIntArray - The client sends an array of numbers and the server
> >>>>echoes the array back to the client ). This is a RPC based
> >>>>        
> >>>>
> >webservice
> >  
> >
> >>>>The example seemed to work fine for < 200 numbers. For 
> numbers > 200
> >>>>the I had the following problems
> >>>>I have attached the source code and stubs along with this email.
> >>>>
> >>>>Please find parts of the SOAP_RESPONSE below.
> >>>>
> >>>>1. The server did not echo back all the numbers
> >>>>
> >>>><item>1159</item>
> >>>><item>1160</item>
> >>>><item>1161</item>
> >>>><item>1162</item>
> >>>><item
> >>>>
> >>>>The message stopped here it was supposed to return 10000 items.
> >>>>
> >>>>2. Some text - "ffb" came into the message
> >>>>
> >>>><item>208</item>
> >>>><item>209</item>
> >>>><item
> >>>>ffb
> >>>>        
> >>>>
> >>>>>210</item>
> >>>>>          
> >>>>>
> >>>><item>211</item>
> >>>><item>212</item>
> >>>>
> >>>><item>449</item>
> >>>><item>450</item>
> >>>>ffb
> >>>>
> >>>><item>451</item>
> >>>><item>452</item>
> >>>>
> >>>>Is this a known issue? The output above was captured 
> using ethereal.
> >>>>
> >>>>Cheers
> >>>>Krishna
> >>>>
> >>>>        
> >>>>
> >>>
> >>>      
> >>>
> >
> >
> >
> >  
> >
> 





Mime
View raw message