make deb undefined symbol: slurm_spank_log

[microbioticajon]

Hi Guys,

I am trying to build and install pyxis on my ubuntu 20.04 based machine and am running into issues.

With the following apt packages slurm-wlm libslurm-dev devscripts debhelper

I build pyix v0.9.1 as a deb package and install as per the installation instructions. However, once I do this slurmd refuses to start:

sudo systemctl restart slurmd
Job for slurmd.service failed because the control process exited with error code.
See "systemctl status slurmd.service" and "journalctl -xe" for details.

/var/log/slurm-llnl/slurmd.log

...
[2021-02-18T00:15:28.112] error: plugin_load_from_file: dlopen(/usr/lib/x86_64-linux-gnu/slurm/spank_pyxis.so): /usr/lib/x86_64-linux-gnu/slurm/spank_pyxis.so: undefined symbol: slurm_spank_log
[2021-02-18T00:15:28.112] error: spank: /usr/lib/x86_64-linux-gnu/slurm/spank_pyxis.so: Dlopen of plugin file failed
[2021-02-18T00:15:28.112] error: spank: /etc/slurm-llnl/plugstack.conf.d/pyxis.conf:1: Failed to load plugin /usr/lib/x86_64-linux-gnu/slurm/spank_pyxis.so. Aborting.
[2021-02-18T00:15:28.112] error: slurmd initialization failed

I have tried building from source and get the following warning:

sudo make  clean install
rm -rf common.o args.o pyxis_slurmstepd.o pyxis_slurmd.o pyxis_srun.o pyxis_dispatch.o config.o enroot.o common.d args.d pyxis_slurmstepd.d pyxis_slurmd.d pyxis_srun.d pyxis_dispatch.d config.d enroot.d spank_pyxis.so
cc -std=gnu11 -O2 -g -Wall -Wunused-variable -fstack-protector-strong -fpic  -D_GNU_SOURCE -D_FORTIFY_SOURCE=2  -MMD -MF common.d -c common.c
cc -std=gnu11 -O2 -g -Wall -Wunused-variable -fstack-protector-strong -fpic  -D_GNU_SOURCE -D_FORTIFY_SOURCE=2  -MMD -MF args.d -c args.c
cc -std=gnu11 -O2 -g -Wall -Wunused-variable -fstack-protector-strong -fpic  -D_GNU_SOURCE -D_FORTIFY_SOURCE=2  -MMD -MF pyxis_slurmstepd.d -c pyxis_slurmstepd.c
pyxis_slurmstepd.c: In function ‘enroot_container_create’:
pyxis_slurmstepd.c:528:3: warning: implicit declaration of function ‘slurm_spank_log’; did you mean ‘slurm_spank_exit’? [-Wimplicit-function-declaration]
  528 |   slurm_spank_log("pyxis: importing docker image ...");
      |   ^~~~~~~~~~~~~~~
      |   slurm_spank_exit
cc -std=gnu11 -O2 -g -Wall -Wunused-variable -fstack-protector-strong -fpic  -D_GNU_SOURCE -D_FORTIFY_SOURCE=2  -MMD -MF pyxis_slurmd.d -c pyxis_slurmd.c
cc -std=gnu11 -O2 -g -Wall -Wunused-variable -fstack-protector-strong -fpic  -D_GNU_SOURCE -D_FORTIFY_SOURCE=2  -MMD -MF pyxis_srun.d -c pyxis_srun.c
cc -std=gnu11 -O2 -g -Wall -Wunused-variable -fstack-protector-strong -fpic  -D_GNU_SOURCE -D_FORTIFY_SOURCE=2  -MMD -MF pyxis_dispatch.d -c pyxis_dispatch.c
cc -std=gnu11 -O2 -g -Wall -Wunused-variable -fstack-protector-strong -fpic  -D_GNU_SOURCE -D_FORTIFY_SOURCE=2  -MMD -MF config.d -c config.c
cc -std=gnu11 -O2 -g -Wall -Wunused-variable -fstack-protector-strong -fpic  -D_GNU_SOURCE -D_FORTIFY_SOURCE=2  -MMD -MF enroot.d -c enroot.c
cc -shared -Wl,-znoexecstack -Wl,-zrelro -Wl,-znow  -o spank_pyxis.so spank_pyxis.lds common.o args.o pyxis_slurmstepd.o pyxis_slurmd.o pyxis_srun.o pyxis_dispatch.o config.o enroot.o
strip --strip-unneeded -R .comment spank_pyxis.so
install -d -m 755 /usr/local/lib/slurm
install -m 644 spank_pyxis.so /usr/local/lib/slurm
install -d -m 755 /usr/local/share/pyxis
echo 'required /usr/local/lib/slurm/spank_pyxis.so' | install -m 644 /dev/stdin /usr/local/share/pyxis/pyxis.conf

Im not sure where to go from here :-/ Best, Jon

[lukeyeager]

You need Slurm >= 20.02. From the wiki:

Requirements: Slurm version 20.02 or later. https://github.com/NVIDIA/pyxis/wiki/Installation

[flx42]

The latest releases require Slurm 20.02 indeed, but if you want to quickly test pyxis you can compile tag version 0.7.0 and it should work with older versions of Slurm.

[microbioticajon]

...... that was the one thing I didn't check.....

Many thanks @lukeyeager, @flx42 for the fast reply - apologies for generating noise.