Fix E2E Intel MPI integ tests #676
Conversation
GonzaloSaez
force-pushed
the
gonzalo/run_ci
branch
from
acc2970 to
a1636d7
Compare
| @@ -18,7 +18,7 @@ RUN apt update \ | |||
| && apt install -y --no-install-recommends \ | |||
| libstdc++-12-dev binutils procps clang \ | |||
| intel-oneapi-compiler-dpcpp-cpp \ | |||
| intel-oneapi-mpi-devel \ | |||
| intel-oneapi-mpi-devel-2021.13 \ | |||
There was a problem hiding this comment.
I'm ok with this as a temporary solution, but eventually this version will be out of support or have CVEs, so we still need to figure what's wrong with newer versions.
WDYT @tenzen-y?
There was a problem hiding this comment.
I did not find the exhaustive changelog for release 2021.14, but it seems they mention that they changed the default IPC exchange mechanism to be pidfd instead of drmfd which was the default. It's not clear to me if that could be the issue.
There was a problem hiding this comment.
Thank you for this great investigation!
I'm ok with this solution as a short term. Could we open an issue to upgrade to 2021-.14?
There was a problem hiding this comment.
I'm curious about the reason why we face this issue only in the github actions, though.
GonzaloSaez
force-pushed
the
gonzalo/run_ci
branch
from
2a5339d to
f3737bb
Compare
|
Could you squash all commits please? |
GonzaloSaez
force-pushed
the
gonzalo/run_ci
branch
from
f3737bb to
57eb324
Compare
Signed-off-by: GonzaloSaez <[email protected]>
GonzaloSaez
force-pushed
the
gonzalo/run_ci
branch
from
57eb324 to
9dfc55b
Compare
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: tenzen-y The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
This should fix #675 by pinning the intel mpi packages to 2021.13 instead of the 2021.14 version that was being installed in the github actions. It is not clear to me what changed in the 2021.14 to make these tests fail but I don't think it should be a big deal. See https://github.com/GonzaloSaez/mpi-operator/actions/runs/12709381529/job/35428272102 for a build e2e run that shows the intel mpi tests passing.