[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Help - HTTP_URL_POST_XML recvresp hanging



   I probably did not explain it correctly,  but the symptoms were the
   same : application locked.. staying in wait status... we had to cancel
   the job...
   I observed that after 2 hours, the connection was dropped, and an error
   was sent to the program...  In my situation, changing the TCP keep
   alive value did solve the issue... (error sent to the application after
   1 minute)

   My understanding/guess was that the target application was no longer
   connected but that the disconnect packet was lost/missing. So the
   connection was still existing/locked in the TCP stack level and the
   local application was still in wait status...
   From what I remember, after the time specified in TCPkeep alive, the
   TCP stack sends a RR(?) packet and if there is no answer the
   communication is dropped..
   Of course, I could be confusing different things ,my knowledge of
   TCP/IP is very superficial and very old. More over, I have only used
   HTTPAPI (great tool! thank you!) and never dig into it...  so sorry if
   I was misleading.
   Paul
   From:        Scott Klement <sk@xxxxxxxxxxxxxxxx>
   To:        HTTPAPI and FTPAPI Projects <ftpapi@xxxxxxxxxxxxxxxxxxxxxx>
   Date:        09/01/2016 08:27
   Subject:        Re: Help - HTTP_URL_POST_XML recvresp hanging
   Sent by:        ftpapi-bounces@xxxxxxxxxxxxxxxxxxxxxx
     __________________________________________________________________

   Paul Roy,
   I can't see what this has to do with the error he's describing? I
   wonder
   if you're confusing two different things?
   With Java applications (and only with Java... I suspect this is a bug
   somewhere deep in the JVM) I've seen issues where TCP connections will
   sporadically disconnect, and setting the keep-alive value (i.e. the
   CHGTCPA option you mention) to a small value works around this bug.
   But that's not at all related to what Paul Kenosky is describing. He's
   saying that his session never ends or times out -- so pretty much the
   opposite.  It seems unlikely that this keep-alive value would solve
   that, unless there's something that I'm not thinking of?
   -SK
   On 1/8/2016 5:32 AM, Paul Roy wrote:
   >     Hi,
   >     this looks to me like a communication error...   either the
   remote did
   >     not receive your data or you did not receive the response...
   >     I had the same kind of issues (not with HTTPAPI but a Java pgm)
   that
   >     was solved by shortening the tcp keep alive value (if i remember
   >     correctly, the time that the TCP stack waits before closing the
   socket
   >     . if there is no traffic)
   >     the default value is 2 hours and  can be changed with CHGTCPA
   >     TCPKEEPALV(1) to set 1 minute...  you can also change the option
   of the
   >     socket wit set_sock_options()... but I do not know if te option
   is
   >     available in HTTPAPI
   >     then you should add some logic in your program to try again...
   >
   -----------------------------------------------------------------------
   This is the FTPAPI mailing list.  To unsubscribe, please go to:
   [1]http://www.scottklement.com/mailman/listinfo/ftpapi
   -----------------------------------------------------------------------

References

   1. http://www.scottklement.com/mailman/listinfo/ftpapi
-----------------------------------------------------------------------
This is the FTPAPI mailing list.  To unsubscribe, please go to:
http://www.scottklement.com/mailman/listinfo/ftpapi
-----------------------------------------------------------------------