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

Re: HTTPAPI calls slower in our production environment



   Charles,
   If I understand Venter's message correctly, he's saying that he's using
   host names that are only defined in the "host table", and that on the
   "slow" system, it was configured to look at DNS first and look at the
   host table only if not found in DNS.
   If that's the case, the DNS lookup has to fail before it can look in
   the host table.  Since DNS is a large database distributed across the
   entire internet, looking up a host that's not defined in the DNS server
   may take a fairly long time -- and that would certainly explain the
   delays.
   I could be misinterpreting the message, of course.

   On 1/8/2014 8:34 AM, Charles Wilt wrote:

   Not sure that I'd call this resolved quite yet...
   DNS queries, especially to an internal server for an internal server
   name should be pretty quick.  I'd even say you shouldn't notice a
   difference.
   It sounds like the i might have an invalid DNS server setup.  My guess
   is that at least the first DNS server listed in CHGTCPDMN is no longer
   valid.  The i will attempt to query that and has to wait till the
   request times out before trying the second one listed.  If the second
   is invalid, it waits again before trying the third.  Actually it's
   worse than that as there's a DMNSVRRTY parm that controls how many
   times the i trys and how long it waits before trying another.
   While you are in there, make sure DMNNAME and DMNSCHLIST are correct.
   Charles

   On Wed, Jan 8, 2014 at 4:10 AM, [1]<[1]Venter.Derick@xxxxxxxxxxxxxx>
   wrote:

        Hi Mike / Charles
        We found the problem.
        We can see that the AS400_TEST server is configured to first look
     up a
        host name in the local host table.  AS400_PRODUCTION is set up to
     use
        the DNS server first for a name look up.
        When we ping the WINDOWS_TEST server from AS400_TEST server  we
     get a
        quick response with the server name and no domain name at the end
     of
        the server name.
        When we ping the WINDOWS_PRODUCTION server from AS400_PRODUCTION
     server
         we get a slow response with the server name and the domain name
     added
        at the end of the server name.
        We re-configured the AS400_PRODUCTION server and it is now as
     fast as
        we expect it to be.
        Thanks for then help.

      Regards
      Derick Venter
      Applications Developer IV

        [[2]cid:_4_07E90E0007E90A000032667442257C5A]

      Systems Integration
      Tel: [2]+27 (13) 247 2816 Fax: [3]+27 (0) 86 573 2274
      Cell: [4]+27 (0) 83 458 6599
      Email: [1][[3]5]derick.venter@xxxxxxxxxx
      [2][6]www.gijima.com

        From:        Mike Krebs [4]<[7]mkrebs@xxxxxxxxxxxxxxxxxx>
        To:        HTTPAPI and FTPAPI Projects
     [5]<[8]ftpapi@xxxxxxxxxxxxxxxxxxxxxx>
        Date:        2014/01/07 07:32 PM
        Subject:        RE: HTTPAPI calls slower in our production
     environment
        Sent by:        [[6]9]ftpapi-bounces@xxxxxxxxxxxxxxxxxxxxxx

     __________________________________________________________________

      +1 on adding timestamps. Running in all three environments and
   checking
      the differences will help to narrow down the issue. The archives
   have a
      simple change to one procedure that will be a good spot to start:


     [3][10][7]http://www.scottklement.com/archives/ftpapi/201310/msg00080.h
     tml

      The network should not be a problem no matter what, but just to make
      sure the basics are covered, trace route in each environment to make
      sure the routing to the production box is not messed up. TRACEROUTE
      RMTSYS([11]WINDOWSERVER.COMPANY.COM)
      Another thought, in production, is work management "normal" or do
   you
      have special setup? The program making the httpapi call set up with
      defined activation group or *caller or *new? Any changes between
   that
      and development?
      -----Original Message-----
      From: [[8]12]ftpapi-bounces@xxxxxxxxxxxxxxxxxxxxxx

      [[4]mailto:[[9]13]ftpapi-bounces@xxxxxxxxxxxxxxxxxxxxxx] On Behalf Of
   Charles
      Wilt
      Sent: Tuesday, January 07, 2014 7:45 AM
      To: HTTPAPI and FTPAPI Projects
      Subject: Re: HTTPAPI calls slower in our production environment
      Do you have debug on in production by chance?
      That's the only HTTPAPI related slowdown I can think of.
      Other than that, I'd look to make sure it's not the Windows side.
   Can
      you
      call the windows service with SoapUI or maybe call the DEV/TEST from
      your
      production i?
      Assuming you can confirm it is the i, one thing to check would be
   the
      configuration of IP6 on production.  Users on the midrange list have
      reported poor network performance (granted mostly Java apps) when
   IP6
      is
      incompletely configured.
      Can you add some dsply op-codes to display time stamps (or use the
      debug
      message in HTTPAPI) to see if the slowdown is in the network code or
      your
      RPG code?  If you have performance tools, you could use those also.
      You might consider making sure the times on the machines are
   reasonably
      in
      sync (hopefully both are using the same NTP service) and then run a
      comm
      trace on both sides.  That would let you see if the network itself
   is
      the
      issue.
      Lastly, look at the environment where the HTTPAPI using app is
   running,
      QINTER or QBATCH?  Do you have enough activity levels?  You may want
   to
      try
      moving it to it's own dedicated subsystem with dedicated memory at
      least
      temporarily to ensure that you have total control of it's
   environment.
      HTH,
      CHarles
      On Tue, Jan 7, 2014 at 4:51 AM, [10]<[14]Venter.Derick@xxxxxxxxxxxxxx>
   wrote:
      >    Hello  everyone.
      >    We are in the process of re-writing all our web service calls
      using RPG
      >    functions and HTTPAPI. The new web service calls are much
   faster
      than
      >    the previous technology used but have a major performance
   problem
      in
      >    our production environment.
      >    Our development an test environments (RPG web service calls)
   are
      faster
      >    than our production environment. Doesn't make sense because the
      >    production server is bigger, better and faster.
      >    Dev environment
      >    RPGLE function on AS400_DEV server calling a web service on
      WINDOWS_DEV
      >    server.
      >    Test environmenrt
      >    RPGLE function on AS400_TEST server calling a web service on
      >    WINDOWS_TEST server.
      >    Production environment
      >    RPGLE function on AS400_PRODUCTION server calling a web service
   on
      >    WINDOWS_PRODUCTION server.
      >    I am using the same test function in all environments.
      >    Any idea where to look..start...
      >    Regards
      >    Derick Venter
      >    Applications Developer IV
      >

        >    [[5][11]cid:_4_09AE216409AE1D64003623EB42257C59]

      >
      >    Systems Integration
      >    Tel: [15]+27 (13) 247 2816 Fax: [16]+27 (0) 86 573 2274
      >    Cell: [17]+27 (0) 83 458 6599
      >    Email: [1][[12]18]derick.venter@xxxxxxxxxx

        >    [2][6][19]www.gijima.com

      >
      __________________________________________________________________
      >
      >    This e-mail is subject to the Columbus Stainless [Pty] Ltd
   Email
      Legal
      >    Notices available at:

        >    [3][7][20][13]http://www.columbus.co.za/EmailLegalNotice.htm.

      >
      __________________________________________________________________
      >
      >    This e-mail message has been scanned for Viruses and Content
   and
      >    cleared by MailMarshal
      >
      __________________________________________________________________
      >
      > References
      >

        >    1. [8]mailto:[[14]21]derick.venter@xxxxxxxxxx
        >    2. [9][22][15]http://www.gijima.com/
        >    3. [10][23][16]http://www.columbus.co.za/EmailLegalNotice.htm

      >
      >

   -----------------------------------------------------------------------
      > This is the FTPAPI mailing list.  To unsubscribe, please go to:

        > [11][24][17]http://www.scottklement.com/mailman/listinfo/ftpapi

      >

   -----------------------------------------------------------------------
      >
      >

   -----------------------------------------------------------------------
      This is the FTPAPI mailing list.  To unsubscribe, please go to:

        [12][25][18]http://www.scottklement.com/mailman/listinfo/ftpapi

     --------------------------------------------------------------------
     ---

        __________________________________________________________________
      This e-mail is subject to the Columbus Stainless [Pty] Ltd Email
   Legal
      Notices available at:

        [13][26][19]http://www.columbus.co.za/EmailLegalNotice.htm.

        __________________________________________________________________
      This e-mail message has been scanned for Viruses and Content and
      cleared by MailMarshal
        __________________________________________________________________
   References
      1. mailto:[[20]27]derick.venter@xxxxxxxxxx
      2. [28][21]http://www.gijima.com/

        3.
     [29][22]http://www.scottklement.com/archives/ftpapi/201310/msg00080.html
        4. mailto:[[23]30]ftpapi-bounces@xxxxxxxxxxxxxxxxxxxxxx
        5. [24]cid:_4_09AE216409AE1D64003623EB42257C59
        6. [25]file://localhost/tmp/[31]www.gijima.com
        7. [32][26]http://www.columbus.co.za/EmailLegalNotice.htm
        8. mailto:[[27]33]derick.venter@xxxxxxxxxx
        9. [34][28]http://www.gijima.com/
       10. [35][29]http://www.columbus.co.za/EmailLegalNotice.htm
       11. [36][30]http://www.scottklement.com/mailman/listinfo/ftpapi
       12. [37][31]http://www.scottklement.com/mailman/listinfo/ftpapi
       13. [38][32]http://www.columbus.co.za/EmailLegalNotice.htm
     --------------------------------------------------------------------
     ---
     This is the FTPAPI mailing list.  To unsubscribe, please go to:
     [39][33]http://www.scottklement.com/mailman/listinfo/ftpapi
     --------------------------------------------------------------------
     ---

References

   1. [34]mailto:Venter.Derick@xxxxxxxxxxxxxx
   2. [35]tel:%2B27%20%2813%29%20247%202816
   3. [36]tel:%2B27%20%280%29%2086%20573%202274
   4. [37]tel:%2B27%20%280%29%2083%20458%206599
   5. [38]mailto:derick.venter@xxxxxxxxxx
   6. [39]http://www.gijima.com/
   7. [40]mailto:mkrebs@xxxxxxxxxxxxxxxxxx
   8. [41]mailto:ftpapi@xxxxxxxxxxxxxxxxxxxxxx
   9. [42]mailto:ftpapi-bounces@xxxxxxxxxxxxxxxxxxxxxx
  10. [43]http://www.scottklement.com/archives/ftpapi/201310/msg00080.html
  11. [44]http://WINDOWSERVER.COMPANY.COM/
  12. [45]mailto:ftpapi-bounces@xxxxxxxxxxxxxxxxxxxxxx
  13. [46]mailto:ftpapi-bounces@xxxxxxxxxxxxxxxxxxxxxx
  14. [47]mailto:Venter.Derick@xxxxxxxxxxxxxx
  15. [48]tel:%2B27%20%2813%29%20247%202816
  16. [49]tel:%2B27%20%280%29%2086%20573%202274
  17. [50]tel:%2B27%20%280%29%2083%20458%206599
  18. [51]mailto:derick.venter@xxxxxxxxxx
  19. [52]http://www.gijima.com/
  20. [53]http://www.columbus.co.za/EmailLegalNotice.htm
  21. [54]mailto:derick.venter@xxxxxxxxxx
  22. [55]http://www.gijima.com/
  23. [56]http://www.columbus.co.za/EmailLegalNotice.htm
  24. [57]http://www.scottklement.com/mailman/listinfo/ftpapi
  25. [58]http://www.scottklement.com/mailman/listinfo/ftpapi
  26. [59]http://www.columbus.co.za/EmailLegalNotice.htm
  27. [60]mailto:derick.venter@xxxxxxxxxx
  28. [61]http://www.gijima.com/
  29. [62]http://www.scottklement.com/archives/ftpapi/201310/msg00080.html
  30. [63]mailto:ftpapi-bounces@xxxxxxxxxxxxxxxxxxxxxx
  31. [64]http://www.gijima.com/
  32. [65]http://www.columbus.co.za/EmailLegalNotice.htm
  33. [66]mailto:derick.venter@xxxxxxxxxx
  34. [67]http://www.gijima.com/
  35. [68]http://www.columbus.co.za/EmailLegalNotice.htm
  36. [69]http://www.scottklement.com/mailman/listinfo/ftpapi
  37. [70]http://www.scottklement.com/mailman/listinfo/ftpapi
  38. [71]http://www.columbus.co.za/EmailLegalNotice.htm
  39. [72]http://www.scottklement.com/mailman/listinfo/ftpapi


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

References

   1. mailto:[1]Venter.Derick@xxxxxxxxxxxxxx
   2. cid:_4_07E90E0007E90A000032667442257C5A
   3. mailto:5]derick.venter@xxxxxxxxxx
   4. mailto:[7]mkrebs@xxxxxxxxxxxxxxxxxx
   5. mailto:[8]ftpapi@xxxxxxxxxxxxxxxxxxxxxx
   6. mailto:9]ftpapi-bounces@xxxxxxxxxxxxxxxxxxxxxx
   7. http://www.scottklement.com/archives/ftpapi/201310/msg00080.h
   8. mailto:12]ftpapi-bounces@xxxxxxxxxxxxxxxxxxxxxx
   9. mailto:13]ftpapi-bounces@xxxxxxxxxxxxxxxxxxxxxx
  10. mailto:[14]Venter.Derick@xxxxxxxxxxxxxx
  11. cid:_4_09AE216409AE1D64003623EB42257C59
  12. mailto:18]derick.venter@xxxxxxxxxx
  13. http://www.columbus.co.za/EmailLegalNotice.htm
  14. mailto:21]derick.venter@xxxxxxxxxx
  15. http://www.gijima.com/
  16. http://www.columbus.co.za/EmailLegalNotice.htm
  17. http://www.scottklement.com/mailman/listinfo/ftpapi
  18. http://www.scottklement.com/mailman/listinfo/ftpapi
  19. http://www.columbus.co.za/EmailLegalNotice.htm
  20. mailto:27]derick.venter@xxxxxxxxxx
  21. http://www.gijima.com/
  22. http://www.scottklement.com/archives/ftpapi/201310/msg00080.html
  23. mailto:30]ftpapi-bounces@xxxxxxxxxxxxxxxxxxxxxx
  24. cid:_4_09AE216409AE1D64003623EB42257C59
  25. file://localhost/tmp/
  26. http://www.columbus.co.za/EmailLegalNotice.htm
  27. mailto:33]derick.venter@xxxxxxxxxx
  28. http://www.gijima.com/
  29. http://www.columbus.co.za/EmailLegalNotice.htm
  30. http://www.scottklement.com/mailman/listinfo/ftpapi
  31. http://www.scottklement.com/mailman/listinfo/ftpapi
  32. http://www.columbus.co.za/EmailLegalNotice.htm
  33. http://www.scottklement.com/mailman/listinfo/ftpapi
  34. mailto:Venter.Derick@xxxxxxxxxxxxxx
  35. tel:%2B27%20%2813%29%20247%202816
  36. tel:%2B27%20%280%29%2086%20573%202274
  37. tel:%2B27%20%280%29%2083%20458%206599
  38. mailto:derick.venter@xxxxxxxxxx
  39. http://www.gijima.com/
  40. mailto:mkrebs@xxxxxxxxxxxxxxxxxx
  41. mailto:ftpapi@xxxxxxxxxxxxxxxxxxxxxx
  42. mailto:ftpapi-bounces@xxxxxxxxxxxxxxxxxxxxxx
  43. http://www.scottklement.com/archives/ftpapi/201310/msg00080.html
  44. http://WINDOWSERVER.COMPANY.COM/
  45. mailto:ftpapi-bounces@xxxxxxxxxxxxxxxxxxxxxx
  46. mailto:ftpapi-bounces@xxxxxxxxxxxxxxxxxxxxxx
  47. mailto:Venter.Derick@xxxxxxxxxxxxxx
  48. tel:%2B27%20%2813%29%20247%202816
  49. tel:%2B27%20%280%29%2086%20573%202274
  50. tel:%2B27%20%280%29%2083%20458%206599
  51. mailto:derick.venter@xxxxxxxxxx
  52. http://www.gijima.com/
  53. http://www.columbus.co.za/EmailLegalNotice.htm
  54. mailto:derick.venter@xxxxxxxxxx
  55. http://www.gijima.com/
  56. http://www.columbus.co.za/EmailLegalNotice.htm
  57. http://www.scottklement.com/mailman/listinfo/ftpapi
  58. http://www.scottklement.com/mailman/listinfo/ftpapi
  59. http://www.columbus.co.za/EmailLegalNotice.htm
  60. mailto:derick.venter@xxxxxxxxxx
  61. http://www.gijima.com/
  62. http://www.scottklement.com/archives/ftpapi/201310/msg00080.html
  63. mailto:ftpapi-bounces@xxxxxxxxxxxxxxxxxxxxxx
  64. http://www.gijima.com/
  65. http://www.columbus.co.za/EmailLegalNotice.htm
  66. mailto:derick.venter@xxxxxxxxxx
  67. http://www.gijima.com/
  68. http://www.columbus.co.za/EmailLegalNotice.htm
  69. http://www.scottklement.com/mailman/listinfo/ftpapi
  70. http://www.scottklement.com/mailman/listinfo/ftpapi
  71. http://www.columbus.co.za/EmailLegalNotice.htm
  72. http://www.scottklement.com/mailman/listinfo/ftpapi
  73. 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
-----------------------------------------------------------------------