Issue with get_NamedList logging

[michelpons]

I am using PRO/II 9.1, together with a TEA Property Package and the COUSCOUS Mixer. The TEA Thermo System and the COUSCOUS Mixer are both logged with COLTT. I start from a case file where the Material Streams have been defined with the CEA C1_C2 (EOS) as Property Package. I am adding the COUSCOUS Mixer, connecting the two inlets and the outlet ports to streams and then I launch the Calculation which ends up in error.

The log I get in the Viewer shows that Calculate is not as the same level as Validate which is odd. Going to the raw file I see that there is no return for the get_NamedValueList call.

When logging is not enabled on either components, the calculation does not end up in error.

Attached are the log file obtained and the screenshot of the Viewer as obtained.

[michelpons]

Log file obtained with COLTT v2.0 Build Sep 11, 2011

[michelpons]

View obtained in Viewer with COLTT v2.0 Build Sep 11, 2011

[michaelhalloran]

The COUSCOUS mixer is trying to get the value of \"DefaultThermoVersion\" based on what it does when running in COFE. Presumably PRO/11 does support the simulation context methods? I couldn't find any logs that would prove that one way or the other but I expect it is supported. Then the log looks like there are two threads running one executing Calculate and one executing the interaction with the SimulationContext. Does PRO/II execute Calculate in its own thread? I suspect some log output is getting lost due to a lack of thread safety on the logging code. One way to find out is to change the format of log output to include thread id against each line. That's done by changing CAPE-OPENLogs.ini to modify the line:
log4j.appender.A1.layout.ConversionPattern=%c %m%n
to be
log4j.appender.A1.layout.ConversionPattern=[%t] %c %m%n
COLTT will then add thread id to the beginning of each line in subsequent logs. Note that this file will not then load into the viewer but if you could repeat the run with this setting and upload the file I will take a look.

[michelpons]

I have rerun the case after modifying the CAPE-OPENLogs.ini file and I have obtained the attached log file. It deos not seem that two threads are used. I am using COLTT 2.0 with commit 212.

[michelpons]

Log file with thread indicated

[michelpons]

Log file obtained with COLTT 2.0 build Oct 1, 2011 at 17:27:02

[michelpons]

I have retried the scenario of this ticket using the Loggers as of build made on Oct 1, 2011 at 17:27:02. I am still viewing a non-convergence reported in PRO/II when the TEA Thermo System and the COUSCOUS Mixer have COLTT enabled on them. It gives the attached log file that shows a strange arrangement of the calls in Viewer. I have checked once more that without COLTT enabled on TEA Thermo System and on COUSCOUS Mixer, I have a calculation going to its end without any error.

[michaelhalloran]

Fixed at commit 214 for the 2.1 release: COLTT no longer assumes that a COSE will implement the COSE interfaces.

[michelpons]

I have installed COLTT 2.1 Build June 16, 2012 and I have rerun the scenario where a TEA Thermo System and a COUSCOUS Mixer are used and logged in PRO/II 9/1. The attached log file shows the following:

Unit <CO1> : Call to put_simulationContext

put_simulationContext takes input argument: 000E0344

SimulationContext <CO1> : Call to QueryInterface

Requested Interface: ICapeCOSEUtilities

SimulationContext <CO1> : Error returned from QueryInterface - 0x80004002
Unit <CO1> : Return from put_simulationContext - Succeeded

Hence PRO/II does not support ICapeCOSEUtilities and there is no call upon any of the methods of this interface made by COLTT. Hence the calculation with TEA Thermo System and COUSCOUS Mixer logged ends up fine.
The issue is indeed fixed and can be closed. For the sake of completeness I attach the log file obtained.

[michelpons]

Log file obtained with COLTT 2.1 Build June 16, 2012