none
Calling WIP from HIP RRS feed

  • Question

  • Hi. 

    We have used HIP and WIP (HIS 2009) for quite some time but now I've come across a new situation that's causing me headaches.

    I have a HIP module that's being called from host through Cics. This module invokes a webservice which somewhere along the line calls one of our WIP methods before returning to the HIP module. My problem is that the WIP method won't start until the HIP module times out. When the HIP module has timed out the WIP method runs like it should and performs the task it should on host. Clearly there is some kind of collision between the HIP and WIP but I can't find any error-messages on eventlogs or anywhere. I've tried to activate tracing in the Transaction Integrator (TI) on the WIP but nothing is being logged. I made a small test-program in my HIP-module and all it does is to call the WIP method directly. I'm then able to catch an errormessage after the HIP-module has timed out and it only reads "Unable to connect to the remote server". The WIP method itself has been tested and is working when initiated from the server and not from host through HIP.

    Is it even possible to invoke a WIP method from a HIP module?

    Any tips/suggestions would be greatly appreciated :)

     

    Regards

    John

     

    Wednesday, October 5, 2011 2:13 PM

Answers

  • Hi John,

        as to the host side, no idea...someone else might be able to help.

        TI Manager and lots of objects...ah...there's the rub. I could go into a long long explanation of why it ended up like that - but I won't :-) - suffice it to say: the NEXT time you open TI Manager, right click on the WIP Node in the left hand pane, and select "Consistency Checking..." (or whatever the exact wording is), and then in the resulting dialog, select "Don't do it never, no more" ? :-). With this setting, we have successfully cut down those 30 minute coffee breaks to about 5 seconds (for 800 objects).

    Caveat: if you are naughty, and e.g. rename the physical directory of a Virtual Directory containing a Web or WCF Service hosted TI Object, well, tough, we will not notice that for you (and automatically correct our administrative entries)...the consistency checking handles "stop the user shooting himself in the foot too easily while doing things outside the manager".

    HTH, Rob

    Friday, October 7, 2011 3:15 PM
    Moderator

All replies

  • Hi John,

        is the WebService that you are calling hosted on another machine, or on the machine where the HIP Service is running? When you run the WebService/WIP without calling from inside HIP, do you get tracing? If not, make sure that the WebService is running under an acccount that is in the HIS Runtime Users group.

    You are assuming that it is HIP/WIP that is causing the problems - what if the WIP call to the host is blocking (on the host) due to the not yet finished call to HIP from the host? Is the WIP call via TCP/IP, and if so, have you tried a NetMon trace to see if the WIP call has started going to the host?

    Rob

    Wednesday, October 5, 2011 3:19 PM
    Moderator
  • Hi John,

        is the WebService that you are calling hosted on another machine, or on the machine where the HIP Service is running? When you run the WebService/WIP without calling from inside HIP, do you get tracing? If not, make sure that the WebService is running under an acccount that is in the HIS Runtime Users group.

    You are assuming that it is HIP/WIP that is causing the problems - what if the WIP call to the host is blocking (on the host) due to the not yet finished call to HIP from the host? Is the WIP call via TCP/IP, and if so, have you tried a NetMon trace to see if the WIP call has started going to the host?

    Rob

    Hi Rob

    Thank you for quick reply :)

    We've done some tracing on TCP/IP and CICS and you were right in suspecting that the WIP call had actually reached the host and that the problem is located there somewhere. The HIP method has started a CICS task and the WIP method has dispatched another task but this task is not executed until the first task has ended. 

    Do you know if there are any special requirements in CICS for running tasks the way we're trying to do here? (Probably not the right forum for this question but...)

    I also have a slightly different "problem" that you might know the answer to. Our TI Managers are running unbelievably slow. To open TI Manager, delete an object and then install a new object took me 30 minutes today, most of which is spent waiting. I assume that it's not supposed to be like this? When TI Manager was empty (newly installed) everything worked great but once we started adding objects it slowed down dramatically. We've currently defined around 100 objects. We have no such issues when running other software like SQL Server Management Studio or Visual Studio on the same machine.

    John

    • Edited by J.Haland Friday, October 7, 2011 1:04 PM
    Friday, October 7, 2011 1:02 PM
  • Hi John,

        as to the host side, no idea...someone else might be able to help.

        TI Manager and lots of objects...ah...there's the rub. I could go into a long long explanation of why it ended up like that - but I won't :-) - suffice it to say: the NEXT time you open TI Manager, right click on the WIP Node in the left hand pane, and select "Consistency Checking..." (or whatever the exact wording is), and then in the resulting dialog, select "Don't do it never, no more" ? :-). With this setting, we have successfully cut down those 30 minute coffee breaks to about 5 seconds (for 800 objects).

    Caveat: if you are naughty, and e.g. rename the physical directory of a Virtual Directory containing a Web or WCF Service hosted TI Object, well, tough, we will not notice that for you (and automatically correct our administrative entries)...the consistency checking handles "stop the user shooting himself in the foot too easily while doing things outside the manager".

    HTH, Rob

    Friday, October 7, 2011 3:15 PM
    Moderator
  • Thanks Rob, that tip on Consistency Checking helped a lot and will save us countless hours not to mention frustration :)

    Regards John

    Friday, October 7, 2011 6:22 PM