hello, I want to config the job_submit.so plugin in my cluster, and I add JobSubmitPlugins=job_submit/lua in slurm.conf,but when I restart slurmctld, Getting the error below:
Nov 05 02:24:03 mgt systemd[1]: Starting Slurm controller daemon... Nov 05 02:24:03 mgt systemd[1]: PID file /var/run/slurmctld.pid not readable (yet?) after start. Nov 05 02:24:03 mgt systemd[1]: Started Slurm controller daemon. Nov 05 02:24:03 mgt slurmctld[24793]: fatal: failed to initialize job_submit plugin Nov 05 02:24:03 mgt systemd[1]: slurmctld.service: main process exited, code=exited, status=1...LURE Nov 05 02:24:03 mgt.hpc.pku.edu.cn systemd[1]: Unit slurmctld.service entered failed state. Nov 05 02:24:03 mgt.hpc.pku.edu.cn systemd[1]: slurmctld.service failed. I google a lot, but I couldn't find a clearly post about how to config this. In my cluster, I didn't have job_submit.so plugin. And I see someone said in this case, I should rebuild slurm. So, I have two questions about this, firstly, how to rebuild slurm with job_submit.so plugin? secondly, If I rebuild slurm , how can I save the former accounts,users and database informations? Thanks for any suggestions. All the best. Ashlee
