[sampaio]

Dear all
When it comes the time to write data, my simulation crashes, issueing the following output (using a SGE queue system):


n-1<4053> ssi:boot:base:linear: booting n55 (tpol59)
n-1<4053> ssi:boot:base:linear: booting n56 (tpol61)
n-1<4053> ssi:boot:base:linear: booting n57 (tpol62)
n-1<4053> ssi:boot:base:linear: finished


--> FOAM FATAL ERROR : NO_READ specified for read-constructor of object U_0 of class IOobject

Function: regIOobject::readStream(const word&)
in file: db/regIOobject/regIOobjectRead.C at line: 53.

FOAM parallel run aborting

-----------------------------------------------------------------------------
One of the processes started by mpirun has exited with a nonzero exit
code. This typically indicates that the process finished in error.
If your process did not finish in error, be sure to include a "return
0" or "exit(0)" in your C code before exiting the application.

PID 23695 failed on node n0 (10.1.1.64) with exit status 1.
-----------------------------------------------------------------------------
[anieckele@lcad10 oodles]$


It seems to be a read problem (NO_READ on U_0), but it happens at the exact instant of writing the fields. p, k, etc.. are writen, but when it comes to U_0, it crashes...

Do you have any clue?

[mattijs]

Have a search on the site for NO_READ or nfs. This problem has been discussed before quite a few times.

[valentine]

Hi,

am trying to incomperate stress analysis in the icoFoam solver but after copying the solver into my user application bin and modifying, its giving me this error "getApplicationClass::Invalid application class name" followed by the name of my application. Can somebody please help

[valentine]

Hi,

am trying to inco-operate stress analysis in the icoFoam solver but after copying the solver into my user application bin and modifying, its giving me this error "getApplicationClass::Invalid application class name" followed by the name of my application. Can somebody please help

[mattijs]

This message is from FoamX?

The application class is whatever is in system/controlDict in the 'application' setting.

If you have created a new solver and want to use it from FoamX you will also need to create the corresponding configuration files (try to find e.g. icoFoam.cfg and have a look at that)

[sampaio]

Thanks Mattijs,
I found the posts about the NO_READ problems, and now OpenFOAM is able to write the time directories.

The only thing is that now if I change writeInterval in controlDict (after stopping the simulation), it doesnot use the new values in the next run...

I also observed that it actually updates to the new values on the next days or two... which is weird.

I checked the date on the machines and the time on the files and seems normal.

Is there anything else I can check?

Is this weird behaviour caused by a change in fileModificationSkew? (I changed it to 1000000 and later to 100)

Thanks a lot

[mattijs]

Check your nfs update interval. This is usually a few seconds. Just e.g. 'touch' a file and see when it appears on the other machines.

[mattijs]

Just saw this one:

For people whose cluster is not time-synchronized: check out NFStimesync.

http://nfstimesync.sourceforge.net/

It runs in user space and makes all computers appear to run the same time.

[valentine]

Cac someone help me please. I have this problem when running Stressed Foam. The only changes I have made is on E and nu, which are now time dependant. the problem seems to be with exp(t/T) function but don't know how to clear it.

dimensionedScalar rho(mechanicalProperties.lookup("rho"));
dimensionedScalar rhoE(mechanicalProperties.lookup("E0"));

dimensionedScalar t = runTime++;

Info<< "Normalising E : E/rho\n" << endl;
dimensionedScalar E = rhoE/rho*
(
1.0 - 0.015*(1.0 - exp(-(t/T1))) - 0.08*(1.0 - exp(-(t/T2))) - 0.0266*(1.0 - exp(-(t/T3)))
);

dimensionedScalar nu = ((74.01 - E)/148.02);

Info<< "Calculating Lame's coefficients\n" << endl;

dimensionedScalar mu = E/(2.0*(1.0 + nu));
dimensionedScalar lambda = nu*E/((1.0 + nu)*(1.0 - 2.0*nu));
dimensionedScalar threeK = E/(1.0 - 2.0*nu);

the constants T1, T2, T3 are declared as below

Info<< "Reading relaxation time constants\n" << endl;

IOdictionary relaxationTimeConstants
(
IOobject
(
"relaxationTimeConstants",
runTime.constant(),
runTime,
IOobject::MUST_READ,
IOobject::NO_WRITE
)
);
dimensionedScalar T1(relaxationTimeConstants.lookup("T1"));
dimensionedScalar T2(relaxationTimeConstants.lookup("T2"));
dimensionedScalar T3(relaxationTimeConstants.lookup("T3"));

but when I run the solver I get the following error

FOAM FATAL ERROR : ds not dimensionless

Function: exp(const dimensionedScalar& ds)
in file: dimensionedTypes/dimensionedScalar/dimensionedScalar.C at line 157.

FOAM aborting

I thought t = runTime++ is in seconds and hence T are dimensioned in seconds.