Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Dockerized ED2 fails with MPI error messages #2146

Closed
ashiklom opened this issue Oct 11, 2018 · 0 comments · Fixed by #2183
Closed

Dockerized ED2 fails with MPI error messages #2146

ashiklom opened this issue Oct 11, 2018 · 0 comments · Fixed by #2183

Comments

@ashiklom
Copy link
Member

ashiklom commented Oct 11, 2018

Describe the bug
Dockerized ED2 crashes immediately with a bunch of cryptic OpenMPI-related error messages.

To Reproduce
Run Dockerized ED2 (git).

Expected behavior
Model should at least initialize and fail with ED-related errors. These errors come before ED logs any output.

Machine (please complete the following information):
Docker.

Temporary Workaround
Edit the models/ed/inst/template.job file so that the @BINARY is called as @BINARY -s. AFAICT, @robkooper added the -s flag to delay MPI initialization in ED, which seems to fix the problem. Note that you will then have to rebuild the pecan/base and pecan/executor images to make these changes take effect.

A nicer, more-permanent fix might be to allow users to pass additional flags to the model binary via pecan.xml. We already have a <binary> tag -- it should be easy to add an optional <binary-flags> tag and then change our template.job to call @BINARY as @BINARY @BINARYFLAGS.

ashiklom added a commit to ashiklom/pecan that referenced this issue Nov 6, 2018
Allows passing arbitrary flags to ED executable.

Closes PecanProject#2146.
ashiklom added a commit to ashiklom/pecan that referenced this issue Nov 6, 2018
Allows passing arbitrary flags to ED executable.

Closes PecanProject#2146.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging a pull request may close this issue.

1 participant