FW: HTTPAPI question

["Fasnacht, Mike J (TOG)" <mjfasnacht@xxxxxxxxxxxxxxxxxxxxx>]

   An FYI on some further findings.

   Per Scott's response, we changed the hosts table and were able to
   successfully resolve the host name.  Here is the bad news, after X
   amount of times successfully sending data, the socket() function fails
   100% of the time.  This is very similar to the failures we were having
   on the gethostbyname() function.  We ran the program on a different
   machine with the same results.  Some more news, when we tested on our
   production machine it seemd to have an adverse affect to other
   applications that use sockets.

   We also noticed that the more we run tests the slower the machine
   gets.  We are examining the architecture of how we are using the
   HTTPAPI to see if something needs adjusted.

   If anyone has seen this issues, please share your stories.

   Thanks

   ______________________________________________
   From:   Fasnacht, Mike J (TOG)
   Sent:   Friday, April 06, 2007 7:53 AM
   To:     'ftpapi@xxxxxxxxxxxxxxxxxxxxxx'
   Subject:        FW: HTTPAPI question

   Hey all,
   We have been using the HTTPAPI service program for some time now.
   Recently, we created a program to do many posts to an outside domain.
   When we start up a new job and post these entries, the program gets to
   a specific spot then quits.  When I say "quits" I mean the
   gethostbyname function returns a null for a host.  What is funny about
   it is the previous attempt at a gethostbyname for the same host was
   successful.  In fact the host does not change for all the posts we are
   attempting.  Once this null is returned each and every subsequent
   gethostbyname request in that job returns a null.  Which eventually
   ends up with a host not found message.  The subprocedure we are
   finding this issue in is http_build_sockaddr.

   We have been exploring problems with data with no success.
   Those explorations included getting rid of suspected bad data,
   repeated attempts with the same data from one successful attempt and
   many attempts with smaller batches of data.  There is only one
   consistent piece to it.  The process has never been able to post more
   that 360 successful times.  There is no consistency to when the
   gethostbyname fails.  It could fail after 150, 231, 360 attempts.

   My question is has any one seen this issue?  If so, what solutions did
   you use?  Should we install a more current version?  Any help would be
   appreciated.

   Thanks,

   Michael J. Fasnacht
   mjfasnacht@xxxxxxxxxxxxxxxxxxxxx
   PH. 507 386-5827
   FAX 507 625-8937

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