How to adding a new "state" variable in PeleC?

[Soobin120596]

Hello everyone,
Following Q&A #487, I want to define a new "state" variable to store the maximum pressure along time.
I added a new macro in IndexDefines.H and modifying the NVARS：
#define USOOT 8 //"state" variable I want to add
#define UFS (USOOT + NUM_ADV)
//there I add a variable
#define NVAR (NTHERM + NUM_ADV + NUM_SPECIES + NUM_AUX + NUM_LIN + 1)

In the prob.H/pc_initdata, the new state is initialized: state(i, j, k, USOOT) = prob_parm.P_amb;, and something is done in pc_poststep

However, after the modification, the case that previously ran successfully now encounters CVODE errors:
From CVODE: At t = 0 and h = 8.77244e-20, the corrector convergence test failed repeatedly or with |h| = hmin. amrex::Abort::30::Aborting from CVODE !!!

Is there something wrong with my method of adding a new 'state' variable?

[baperry2]

To add a new state variable, the easiest way is to use the existing ADV (advected) and AUX (not advected) capabilities. In your GNUmakefile, set PELE_NUM_ADV or PELE_NUM_AUX to the number of new variables you would like. You can provide names for these variables by defining a custom ProblemSpecificFunctions::set_adv_names function in your prob.H. Then it's up to you to provide appropriate initial conditions for these components of the state vector in pc_initdata, user-defined boundary conditions if needed in bcnormal, and user-defined source terms if desired in ProblemSpecificFunctions::problem_modify_ext_sources.

See Exec/RegTests/AuxQuantities for an example.

With what you are doing, you are likely misaligning the indices expected in the state vector, which will cause lots of issues.

[Soobin120596]

Thanks for the answers here. I finally got what I wanted working!