search

pmodels / casper

Process-based Asynchronous Progress Model for MPI Communication
https://pmodels.github.io/casper-www/
Other
9 stars 4 forks source link

Travis CI #20

Closed jeffhammond closed 7 years ago

jeffhammond commented 7 years ago

Do not merge this yet. I rebased and need to verify that it is not broken.

minsii commented 7 years ago

It looks all tests passed except ompi + osx. Search all passed for other logs. About the error in ompi + osx, have you seem this before ? I did not see such error on Jenkins ompi + osx run. "ORTE_ERROR_LOG: Bad parameter in file orted/pmix/pmix_server.c at line 262"

jeffhammond commented 7 years ago

I don't see anything failing in https://travis-ci.org/pmodels/casper/builds/211946540. Where are you seeing the error?

However, the error you describe is https://github.com/open-mpi/ompi/issues/2956#issuecomment-285343579 and I am very familiar, having had to push the same fix to all of MPI-based projects recently.

jeffhammond commented 7 years ago

I pushed the fix for this in https://github.com/pmodels/casper/pull/20/commits/dbde7d2029f15b369b5cf4e5ef8d20e6dd093d14.

minsii commented 7 years ago

Not sure why travis does not correctly cache such failing. But you can find the error message in raw log. I guess junit xml output might be easier for travis to cache error.

jeffhammond commented 7 years ago

I don't know what is wrong with OMPI yet but I just set to XFAIL in Travis. This PR is ready for merge.

minsii commented 7 years ago

What do you think about the comments in travis/build-run.sh ? The lines after # Run real tests seem unnecessary.

jeffhammond commented 7 years ago

Hmm, those lines might be leftovers from MPICH Travis. I removed them.