NVIDIA / pyxis

Container plugin for Slurm Workload Manager
Apache License 2.0
281 stars 31 forks source link

Pyxis Installation Error #128

Closed leela-uppuluri closed 8 months ago

leela-uppuluri commented 11 months ago

Hi, I am trying to install Pyxis on my BCM cluster and running into the below error. I was following this documentation from Bright to install it.

root@bcm10:~/pyxis-0.16.1# make rpm
rm -rf common.o args.o pyxis_slurmstepd.o pyxis_slurmd.o pyxis_srun.o pyxis_alloc.o pyxis_dispatch.o config.o enroot.o common.d args.d pyxis_slurmstepd.d pyxis_slurmd.d pyxis_srun.d pyxis_alloc.d pyxis_dispatch.d config.d enroot.d spank_pyxis.so
test -e x86_64 || ln -s . x86_64
rpmbuild --target=x86_64 --clean -ba -D"_topdir /root/pyxis-0.16.1/rpm" -D"VERSION 0.16.1" pyxis.spec
Building target platforms: x86_64
Building for target x86_64
error: Failed build dependencies:
        gcc is needed by nvslurm-plugin-pyxis-0.16.1-1.x86_64
        make is needed by nvslurm-plugin-pyxis-0.16.1-1.x86_64
        slurm-devel is needed by nvslurm-plugin-pyxis-0.16.1-1.x86_64
make: *** [Makefile:56: rpm] Error 11

I already tried loading slurm and gcc modules on BCM before running make rpm.

root@bcm10:~# module load slurm
root@bcm10:~# sinfo --version
slurm 23.02.4
root@bcm10:~#
root@bcm10:~# module load gcc
root@bcm10:~# gcc --version
gcc (GCC) 13.1.0
Copyright (C) 2023 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

root@bcm10:~#
root@bcm10:~# apt list --installed | grep make

WARNING: apt does not have a stable CLI interface. Use with caution in scripts.

automake/jammy,now 1:1.16.5-1.3 all [installed]
make/jammy,now 4.3-4.1build1 amd64 [installed]
flx42 commented 11 months ago

Please contact the BCM support then. It might be simpler for you to build the RPM packages on a rpm-based system (that could be inside a Docker container too).

leela-uppuluri commented 11 months ago

@flx42 already reached out. Thanks for the suggestion. The issue is fixed now. Please feel free to close this thread.