Possible bug regarding XXXXXX.clean_build and compiler files using beta1.3.17

[zarzycki]

Not 100% sure if this is a bug, but I'll leave it for you guys to decide.Using beta1.3.17.XXXXXXXXX.clean_build does not delete the compiler directory (within $SCRATCH/$CASE/bld).This is problematic on machines like Yellowstone where data is purged at particular intervals.Ex: using intel on YS....Attempting to rebuild the model 92 days after the initial build throws this error.Tue Jun 30 14:32:51 MDT 2015 /glade/scratch/zarzycki/nhemi_30_x4_slab/bld/mct.bldlog.150630-143241
ERROR: buildlib.mct failed, see /glade/scratch/zarzycki/nhemi_30_x4_slab/bld/mct.bldlog.150630-143241
ERROR: cat /glade/scratch/zarzycki/nhemi_30_x4_slab/bld/mct.bldlog.150630-143241------------------- (in mct.bldlog.******) -------------gmake[1]: Entering directory '/glade/scratch/zarzycki/nhemi_30_x4_slab/bld/intel/mpich2/nodebug/threads/mct/mpeu'
mpicc   -c -DLINUX  -DNDEBUG -DMCT_INTERFACE -DHAVE_MPI -DTHREADED_OMP -DFORTRANUNDERSCORE -DNO_R16 -DLINUX -DCPRINTEL  -DHAVE_SLASHPROC -DSYSLINUX -DCPR  -O2  -no-opt-dynamic-align -fp-model precise   -xHost  -openmp   -DLINUX  -DNDEBUG -DMCT_INTERFACE -DHAVE_MPI -DTHREADED_OMP -DFORTRANUNDERSCORE -DNO_R16 -DLINUX -DCPRINTEL  -DHAVE_SLASHPROC -I.. -I.  -I/glade/scratch/zarzycki/nhemi_30_x4_slab/bld/intel/mpich2/nodebug/threads/include -I/glade/scratch/zarzycki/nhemi_30_x4_slab/bld/intel/mpich2/nodebug/threads/MCT/noesmf/a1l1r1i1o1g1w1/csm_share -I/glade/apps/opt/netcdf-mpi/4.3.2/intel/default/include -I/glade/apps/opt/pnetcdf/1.4.1/intel/default/include -I/glade/scratch/zarzycki/nhemi_30_x4_slab/bld/intel/mpich2/nodebug/threads/include -I/glade/p/work/zarzycki/cesm1_3_beta17/models/csm_share/shr -I/glade/p/work/zarzycki/cesm1_3_beta17/models/csm_share/include   -I. -I../ /glade/p/work/zarzycki/cesm1_3_beta17/models/utils/mct/mpeu/get_zeits.c
/glade/p/work/zarzycki/cesm1_3_beta17/models/utils/mct/mpeu/get_zeits.c(25): catastrophic error: cannot open source file "config.h"  #include "config.h"                     ^ compilation aborted for /glade/p/work/zarzycki/cesm1_3_beta17/models/utils/mct/mpeu/get_zeits.c (code 4)----------------------Running .clean_build and then .build does not remedy this problem -- however, purging ${SCRATCH}/${CASE}/${BUILD}/intel and then .build allowed the model to build successfully. It appears that some components of /intel were purged by the scratch auto-delete (90 days) routines (ex: config.h), but since the folder itself isn't cleaned during ./clean_build, the compiler didn't see the need to rebuild config.h

 

[jedwards]

It's a bug or maybe a feature or unintended consequence.   In any case there is an option 'all' to clean_build that will resolve the problem.> $CASE.clean_build all

 

[jedwards]

It's a bug or maybe a feature or unintended consequence.   In any case there is an option 'all' to clean_build that will resolve the problem.> $CASE.clean_build all

 

[jedwards]

It's a bug or maybe a feature or unintended consequence.   In any case there is an option 'all' to clean_build that will resolve the problem.> $CASE.clean_build all