Slurm has some optional scripts to ease the transition from a PBS-based system (like torque). The package is called slurm-torque in the EPEL repos (or come from contribs/torque if compiling from source) and they simply provide wrapper scripts (e.g. qsub, qalter, pbsnodes) for the equivalent slurm command. I'm not sure how its packaged in Rocks but if you can exclude the slurm-torque package, you should be able to run both at the same time. We had both running on our cluster while we were transitioning from torque to slurm.
From: Mahmood Naderan <[email protected]> Reply-To: slurm-dev <[email protected]> Date: Thursday, May 18, 2017 at 10:02 AM To: slurm-dev <[email protected]> Subject: [slurm-dev] Slurm and Torque Hi, I tried to install slurm-roll for a Rocks based cluster however it failed! I know this is not generally related to this mailing list since the roll has not been published by slurm guys, however I have a question and any feedback is greatly appreciated. It seems that on a system where both Torque and Slurm are running, Torque commands are overwritten by Slurm. Please see the following output: mahmood@cluster:~$ which qsub /usr/bin/qsub mahmood@cluster:~$ qsub -h Usage: qsub [-a start_time] [-A account] [-b y|n] [-e err_path] [-I] [-l resource_list] [-m mail_options] [-M user_list] [-N job_name] [-o out_path] [-p priority] [-pe shm task_cnt] [-P wckey] [-q destination] [-r y|n] [-v variable_list] [-V] [-wd workdir] [-W additional_attributes] [-h] [script] mahmood@cluster:~$ cat tor.sh #!/bin/bash #PBS -q default #PBS -l nodes=cluster:ppn=1 #PBS -N test #PBS -V #PBS -o test.out #PBS -j oe cd $PBS_O_WORKDIR ls mahmood@cluster:~$ qsub tor.sh There was an error running the SLURM sbatch command. The command was: '/usr/bin/sbatch -e tor.sh.e%A -o tor.sh.o%A tor.sh 2>&1' and the output was: 'sbatch: error: Parse error in file /etc/slurm/headnode.conf line 1: "ClusterName=" sbatch: error: Parse error in file /etc/slurm/headnode.conf line 2: "ControlMachine=" sbatch: error: Parse error in file /etc/slurm/headnode.conf line 3: "ControlAddr=" sbatch: error: Parse error in file /etc/slurm/headnode.conf line 5: "DefaultStorageHost=" sbatch: error: "Include" failed in file /etc/slurm/slurm.conf line 95 sbatch: fatal: Unable to process configuration file ' mahmood@cluster:~$ cat /etc/slurm/headnode.conf ClusterName= ControlMachine= ControlAddr= DefaultStorageType=slurmdbd DefaultStorageHost= mahmood@cluster:~$ Any idea? Regards, Mahmood
