Fail to run due to seccomp_filter error

basvandervlies commented 4 years ago

I have installed this plugin on a Debian host stretch (9.12) with slurm 19.05.5. . SECCOMP is enabled:

oot# grep SECCOMP /boot/config-$(uname -r)
CONFIG_HAVE_ARCH_SECCOMP_FILTER=y
CONFIG_SECCOMP_FILTER=y
CONFIG_SECCOMP=y

I have patched the source to print the error code and removed the exit/ So the error messages are printed and the container is still started:

bas@r33n4:~/enroot$ srun --container-name=testje --container-image=$HOME/enroot/ubuntu.sqsh --pty cat /etc/os-release
slurmstepd: task_p_pre_launch: Using sched_affinity for tasks
slurmstepd: error: pyxis: seccomp filter failed: Invalid argument:22
NAME="Ubuntu"
VERSION="18.04.4 LTS (Bionic Beaver)"
ID=ubuntu
ID_LIKE=debian
PRETTY_NAME="Ubuntu 18.04.4 LTS"
VERSION_ID="18.04"
HOME_URL="https://www.ubuntu.com/"
SUPPORT_URL="https://help.ubuntu.com/"
BUG_REPORT_URL="https://bugs.launchpad.net/ubuntu/"
PRIVACY_POLICY_URL="https://www.ubuntu.com/legal/terms-and-policies/privacy-policy"
VERSION_CODENAME=bionic
UBUNTU_CODENAME=bionic

flx42 commented 4 years ago

@basvandervlies thanks for reporting! Could you also try by replacing SECCOMP_FILTER_FLAG_SPEC_ALLOW with 0 here? https://github.com/NVIDIA/pyxis/blob/07cf4f37877fb12932e0d3f72fac93a28df62f8c/seccomp_filter.h#L63

Thank you

basvandervlies commented 4 years ago

@flx42 Thanks that solves the problem.

flx42 commented 4 years ago

Thanks for checking! I fixed it in https://github.com/NVIDIA/pyxis/commit/6bd15a0ba36e48c21bf22c494dec0ac8c6e895b6

NVIDIA / pyxis

Fail to run due to seccomp_filter error #17