In my current setup we have TC corporate server with DispatcherClient installed, and Scheduler and Module running on a separate machine doing NX to JT translations. And that is working very well.
However, my problem is that there are more translation requests than the current setup can handle. So at the end of the working day there might be 5000 requests queued, and next morning there might be 1-2000 requests still in queue for processing. So we are always behind.
From the documentation it seems like I can add one or more machines with Module installed and let the scheduler distribute the tasks to the two (or more) module machines. But I struggle to find detailed information about how this is done. Has somebody here got some guidelines on how to do this in practical steps?
Solved! Go to Solution.
Thanks alot for the input. That seems to work kind of..
The new module machine is now able to connect to scheduler machine, and both modules are processing requests now.
I have also copied the tessUG.config file from the existing machine to make sure translation settings are identical.
However, on the new module machine, all translations fail with this error message:
PVTRANS: *** Error encountered during Translation-> 940607: An error occurred while trying to initialize the Teamcenter SOA client
Any idea how to solve this?
There is a 4 tier TC client install on that machine (if that is relevant)
The error message appeared because the 2-tier client install was missing (I ran it on a machine with 4-tier install).
Eventually I ended up just copying the whole TC folder from the main dispatcher machine because I still got some error messages on the new install. So that solved the problems.
So now both the module machines are translating simultaneously.
However, I notice that I have plenty of cpu and ram available on the module machines. And I read something about enabling parallell translation jobs via async services? But I struggle to understand how it is set up. Have you guys set up that successfully?
I realized that 2 module machines is not enough to process all the translation requests.
So I added a 3rd one. It successfully connects to scheduler and starts translating immediately.
BUT when it connects, the 'oldest' module session finishes the current translation and starts no more translation jobs. Now, if I close the module that has paused and restart it, it will start translating again, but then one of the other modules will stop running new translations.
If I look at the Admin client there is never more than 2 modules connected. A 3rd one will always replace on of the current ones. Is there a setting somewhere that limits the scheduler not to have more than two current modules connected?
(or can this be caused by the fact that I have just copied the module folder from the main dispatcher machine to the new module machines so that there is a setting somewhere that is confusing the scheduler?)
Just to give you an update:
The problem with connecting more than 2 modules was caused by the ModuleSession.properties file in the module\cache folder. I deleted it on all modules, and after it was automatically recreated by the module it started working fine.
In order to run several parallell translations in one module machine I added maxlimit attribute in the translator.xml file
NxToPvDirect provider="SIEMENS" service="nxtopvdirect" maxlimit="6" isactive="true" OutputNeeded="false"
And also to the transmodule.properties file
So now I'm running 10 parallell translations processing approx 1700 parts per hour.