Slurm changed its db schema again in v19.
This came up in:
Bockjoo shared the sql he is using to work around the problem; we can work this into the slurm probe like we did for the v18 breakage.
Workaround code from Bockjoo:
GROUP BY j.id_job,j.exit_code,j.id_group,j.id_user,j.job_name,j.tres_alloc,j.partition,j.state,a.acct,a.user,j.job_db_inx
GROUP BY id_job