IBM webMethods Hybrid Integration

IBM webMethods Hybrid Integration

Join this online group to communicate across IBM product users and experts by sharing advice and best practices with peers and staying up to date regarding product enhancements.


#TechXchangePresenter
 View Only
Expand all | Collapse all

Integration Server | Crash or stalled or stackOverflow upon

  • 1.  Integration Server | Crash or stalled or stackOverflow upon

    Posted Mon January 12, 2004 10:55 AM

    Hi all (and happy new year).

    I’m currently facing some difficulties in Integration Server (technicals reference later in this post).
    In only one session, i realise the following :

    begin loop over 3 Xmls (let’s name it “A”)


    #webMethods-General
    #Integration-Server-and-ESB
    #webMethods


  • 2.  RE: Integration Server | Crash or stalled or stackOverflow upon

    Posted Tue January 13, 2004 07:11 AM

    Here are some add-on upon the error.
    In fact, the TransformSerialXml service seems to take a look at all the parameters in the pipeline.
    The problem was coming from a quite complex document that was in the pipeline (it’s the return value of the second WS call, not dropped for some personal purpose.)

    If this document (not parameter in or out, nor with a name that could be “reserved” → “traiterPTEReturn”) is present in the input pipeline, the Integration Server become unstable, and this often leads to the crash of the IS :frowning:

    We are able to reproduce this problem, and will soon drop a Case on this particular point.

    But it seems quite awful that a simple parameter (even an “alien” one) in pipeline leads to a crash/stalling of the entire application.

    Does anybody have similar experiences, or comments to share?

    Regards

    Jean-Marc LE TOUX
    PS: excuse my poor english skill


    #webMethods-General
    #Integration-Server-and-ESB
    #webMethods


  • 3.  RE: Integration Server | Crash or stalled or stackOverflow upon

    Posted Tue January 13, 2004 12:12 PM

    You are correct that a StackOverFlow should not be happening.

    It sounds like the problem is in the serialization of the pipeline when creating the SOAP message for the second call.

    Is it possible to send a re-create of this problem to webMethods Support?

    If you add a pub.flow:savePipelineToFile after the first call, then all that should be needed to recreate the problem is a pub.flow:restorePipelineFromFile, the call to the second service and the Web Service Connector service and input and output documents that the Web Service Connector uses.


    #webMethods-General
    #webMethods
    #Integration-Server-and-ESB


  • 4.  RE: Integration Server | Crash or stalled or stackOverflow upon

    Posted Wed January 14, 2004 04:36 AM

    I perhaps missed something in my explanation, but first loop is OK : the beginning of the second one leads to a crash, not the first one.

    I have saved (with quite some IS reboots :frowning: ) the
    But it seem to be quite… BIG :
    -rw-rw-r-- 1 user group 149315584 Jan 14 11:03 coreProducerPipeline

    As the previous one (first loop) was only 37873, i think “Houston, we’ve got a problem” :wink:

    This will be send to support as soon as i’ll be authorized to.

    Thank for your help Fred, and if anybody got comments or help, feel free to respond my post.

    JeanMarc LE TOUX
    PS: i’m currently waiting for being added to my company’s “support request authorized” list.
    As usual, please excuse my poor english.


    #webMethods-General
    #webMethods
    #Integration-Server-and-ESB