Make Installation Error - v11.1.0

[hootmet]

Met installation goes well right down to what literally appears to be the last step.

cli comand:
./compile_MET_all.sh install_met_env.jupyter-env

End of compile script:
configure_cmd="${configure_cmd} ${OPT_ARGS}"

echo "cd pwd"
run_cmd "${configure_cmd} > configure.log 2>&1"
run_cmd "make ${MAKE_ARGS} > met.make.log 2>&1"
run_cmd "make install > met.make_install.log 2>&1"
run_cmd "make test > met.make_test.log 2>&1"

echo "Finished compiling at date"

Error message:
make test > met.make_test.log 2>&1
ERROR: Command returned with non-zero (2) status: make test > met.make_test.log 2>&1

Log output:
cat ./external_libs/netcdf/netcdf-cxx4-4.3.1/met.make_test.log
make: *** No rule to make target 'test'. Stop.

[jprestop]

@hootmet Could you please provide the following files for us to review?

1. The entire screen output from the run of the script
2. Your install_met_env.jupyter-env file
3. The config.log file (not the configure.log file) from the MET directory
4. The met.make.log file from the MET directory
5. The met.make_test.log file from the MET directory

Please feel free to either attach them via GitHub Discussions here in this post or provide them via FTP server as described here.

[hootmet]

Attached find the requested information.
requested_info.zip

Thank you for the help

[jprestop]

@hootmet Thank you for providing these files. It was helpful to debug the problem. In your met_screen.txt file, I see the following:

Compiling MET at Tue Jun 18 22:56:15 UTC 2024
tar (child): /efs/share/jhthomps/tar_files/v11.1.0.tar.gz: Cannot open: No such file or directory
tar (child): Error is not recoverable: exiting now
tar: Child returned status 2
tar: Error is not recoverable: exiting now
./compile_MET_all.sh: line 671: cd: /efs/share/jhthomps/MET*: No such file or directory

It looks like you perhaps you have not downloaded the v11.1.0.tar.gz file and placed it in the expected tar_files subdirectory. Please download the file and place it in a tar_files directory that is at the same level as the compile_MET_all.sh script (The /efs/share/jhthomps/tar_files directory from what I can tell from your configuration file). Then, try running the script again. Please let us know how it goes.

[hootmet]

Well that certainly helped but now errors out per the attached file....

met.make.log

[jprestop]

Thank you @hootmet. Could you also please pass along the config.log file located in the same directory as your met.make.log file?

[hootmet]

here you go....

config.log

[jprestop]

Thank you @hootmet. In your install_met_env file, please try changing the value of MET_PYTHON_CC to
-I/srv/conda/envs/notebook/include/python3.1. Then, try recompiling and let us know how it goes.

[hootmet]

Failed in the Make Met section with ....

python_line.h:23:10: fatal error: Python.h: No such file or directory

[jprestop]

@hootmet Thank you. I see I had a typo in what I told you to set MET_PYTHON_CC to. Instead of setting it to
-I/srv/conda/envs/notebook/include/python3.1
it should have been
-I/srv/conda/envs/notebook/include/python3.11.

Please give that a try.

[hootmet]

Still issues, see attached....

met.make.log

[jprestop]

Thank you @hootmet. I'm sorry to see you're still having trouble. Could you please also pass along your config.log file from the same directory as your met.make.log file?

[hootmet]

config.log
configure.log

[hootmet]

Any feedback on logs previously sent?

[jprestop]

We haven't encountered this problem before, but we have had users encounter similar problems. I believe the problem you are experiencing is likely due to a mismatch in compilers in /usr/bin and the conda environment. I see the following in your config.log file:

CC='/usr/bin/gcc'
CPP='/srv/conda/envs/notebook/bin/x86_64-conda-linux-gnu-cpp'
CXX='/usr/bin/c++'

In your install_met_env.jupyter-mvu file, please try setting the following:

export CPP=/usr/bin/c++

as '/srv/conda/envs/notebook/bin/x86_64-conda-linux-gnu-cpp' is currently being used, but we don't want it to be used.

I also see that the following is being set in your install_met_env.jupyter-mvu file:

export COMPILER=gnu_12.3.0

but in your config.log file it shows that the following compiler is actually being used:

gcc version 11.4.0

so please update your install_met_env.jupyter-mvu file to:

export COMPILER=gnu_11.4.0

and try recompiling. Please let us know how it goes.

[hootmet]

No joy I'm afraid, see log.

Thanks for all the help. It's much appreciated.
gsl.configure.log

[jprestop]

Hi @hootmet. I'm sorry to hear that. If GSL did not compile, could you please pass along the output from the compilation script along with the config.log file in the GSL directory? Unfortunately, the configure.log files don't contain any useful information. They simply say at the end See 'config.log' for more details.

[hootmet]

Here you go

compile_06262024.txt
config.log

[jprestop]

Thanks @hootmet. Can you please add the following line to the end of your install_met_env.jupyter-mvu file and send me the output?

printenv

[hootmet]

here you go...

printenv.txt

[jprestop]

@hootmet Thank you. Can you please tell me - does cpp exist in /usr/bin?

[hootmet]

(notebook) hootmet@jupyter-hootmet:/usr/bin$ ls /usr/bin | grep cpp
cpp
cpp-11
x86_64-linux-gnu-cpp
x86_64-linux-gnu-cpp-11

[jprestop]

Great, @hootmet!
Please update your install_met_env.jupyter-mvu file to be:

export CPP=/usr/bin/cpp

and rerun.

[hootmet]

Failed in the make MET section again.....

/srv/conda/envs/notebook/bin/ld: /srv/conda/envs/notebook/bin/../x86_64-conda-linux-gnu/sysroot/lib64/libdl.so.2: undefined reference to _dl_addr@GLIBC_PRIVATE' /srv/conda/envs/notebook/bin/ld: /srv/conda/envs/notebook/bin/../x86_64-conda-linux-gnu/sysroot/lib64/[libdl.so](http://libdl.so/).2: undefined reference to _dl_vsym@GLIBC_PRIVATE'
/srv/conda/envs/notebook/bin/ld: /srv/conda/envs/notebook/bin/../x86_64-conda-linux-gnu/sysroot/lib64/libpthread.so.0: undefined reference to __libc_pthread_init@GLIBC_PRIVATE' /srv/conda/envs/notebook/bin/ld: /srv/conda/envs/notebook/bin/../x86_64-conda-linux-gnu/sysroot/lib64/[libpthread.so](http://libpthread.so/).0: undefined reference to _dl_make_stack_executable@GLIBC_PRIVATE'
collect2: error: ld returned 1 exit status
make[5]: *** [Makefile:436: ensemble_stat] Error 1
make[5]: Leaving directory '/efs/share/jhthomps/MET-11.1.0/src/tools/core/ensemble_stat'
make[4]: *** [Makefile:381: all-recursive] Error 1
make[4]: Leaving directory '/efs/share/jhthomps/MET-11.1.0/src/tools/core'
make[3]: *** [Makefile:375: all-recursive] Error 1
make[3]: Leaving directory '/efs/share/jhthomps/MET-11.1.0/src/tools'
make[2]: *** [Makefile:372: all-recursive] Error 1
make[2]: Leaving directory '/efs/share/jhthomps/MET-11.1.0/src'
make[1]: *** [Makefile:429: all-recursive] Error 1
make[1]: Leaving directory '/efs/share/jhthomps/MET-11.1.0'
make: *** [Makefile:370: all] Error 2

[jprestop]

@hootmet Can you please send the config.log file and the met.make.log file?

[hootmet]

config.log
met.make.log

[jprestop]

@hootmet it looks like that is the config.log file for GSL. Could you please send the config.log file for MET? It will be in the same directory as the met.make.log file.

[jprestop]

@hootmet Actually, my apologies. It was the correct file. I was looking at an older file. Sorry about that!

[jprestop]

Thanks @hootmet.  I see the following in your printenv.txt file:

PATH=/srv/conda/envs/notebook/bin:/srv/conda/condabin:/srv/conda/envs/notebook/bin:/srv/conda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/opt/goofys/bin

Please add

export PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/opt/goofys/bin:/srv/conda/envs/notebook/bin:/srv/conda/condabin:/srv/conda/envs/notebook/bin:/srv/conda/bin:

above

export GCC=/usr/bin/gcc

to prioritize the system level executables in your install_met_env.jupyter-mvu.

Before rerunning, there's no need for you to recompile all of the external libraries again, so please change

#export EXTERNAL_LIBS=${TEST_BASE}/external_libs
#export MET_GSL=${EXTERNAL_LIBS}
#export MET_BUFRLIB=${EXTERNAL_LIBS}/lib
#export BUFRLIB_NAME=-lbufr
#export LIB_JASPER=${EXTERNAL_LIBS}/lib
#export LIB_LIBPNG=${EXTERNAL_LIBS}/lib
#export LIB_Z=${EXTERNAL_LIBS}/lib
#export MET_GRIB2CLIB=${EXTERNAL_LIBS}/lib
#export MET_GRIB2CINC=${EXTERNAL_LIBS}/include
#export GRIB2CLIB_NAME=-lgrib2c
#export MET_HDF5=${EXTERNAL_LIBS}
#export MET_NETCDF=${EXTERNAL_LIBS}

in your install_met_env.jupyter-mvu to be

export EXTERNAL_LIBS=${TEST_BASE}/external_libs
export MET_GSL=${EXTERNAL_LIBS}
export MET_BUFRLIB=${EXTERNAL_LIBS}/lib
export BUFRLIB_NAME=-lbufr
export LIB_JASPER=${EXTERNAL_LIBS}/lib
export LIB_LIBPNG=${EXTERNAL_LIBS}/lib
export LIB_Z=${EXTERNAL_LIBS}/lib
export MET_GRIB2CLIB=${EXTERNAL_LIBS}/lib
export MET_GRIB2CINC=${EXTERNAL_LIBS}/include
export GRIB2CLIB_NAME=-lgrib2c
export MET_HDF5=${EXTERNAL_LIBS}
export MET_NETCDF=${EXTERNAL_LIBS}

That way, the script will only recompile MET. Please let us know how it goes.

[hootmet]

And there was joy throughout the land!!!

It seemed to compile cleanly. Is there a quick test that can be run?

[hootmet]

Actually, we found the test scripts. All looks good!

Thank you very much for your incredible help, support and patience.

[jprestop]

That's great news @hootmet! Thank you for letting us know. I'll go ahead and select an answer and mark this conversation as locked to prevent future posts here. Please feel free to start a new discussion with any new questions.