ODE Solver error causes job to fail

[rwest]

In the middle of solving reaction system 3 of 12 I get this:

Error running the ODESolver: ***MESSAGE FROM ROUTINE DDASSL IN LIBRARY SLATEC.
Error running the ODESolver: ***POTENTIALLY RECOVERABLE ERROR, PROGRAM CONTINUES.
Error running the ODESolver: *  AT T =    1.543983D-01  AND STEPSIZE H =    6.087239D-09  THE CORRECTOR
Error running the ODESolver: *  FAILED TO CONVERGE REPEATEDLY OR WITH ABS(H)=HMIN
Error running the ODESolver: *  ERROR NUMBER = -7
Error running the ODESolver: *
Error running the ODESolver: ***END OF MESSAGE
Error running the ODESolver:
Error running the ODESolver: ******ODESOLVER FAILED : IDID= -7
ODE solver integrated to 0.15439833930556138 sec.
After ODE: from 0.0 SEC to 0.15439833930556138SEC

Then at the end of the log file: Reaction System 3 has not reached its termination criterion although it seems to be valid (complete), so it was not interrupted for being invalid. This probably means there was an error with the ODE solver, and we risk entering an endless loop. Stopping. Exception in thread "main" java.lang.Error at jing.rxnSys.ReactionModelGenerator.modelGeneration(ReactionModelGenerator.java:1638) at RMG.main(RMG.java:57)

It would be nice if RMG could recover from an ODE error in one reaction system. It may even be possible to recover that reaction system, depending on what DASSL's IDID=-7 represents.

[shamelmerchant]

My Jobs also failed due to this error how did you work your way around this ?

[rwest]

It turns out RMG only stops in this way if all of the non-converged systems fail. I.e. every system either converged or failed. So my request that RMG recovers from an ODE error in one reaction system has already been met.

Carefully check the last few hundred lines of your log and you will probably find that all the other systems had met their convergence criterion.

[rwest]

Documented in commit 17013ece7390771ee90a100e476f38214b0f1c8d