Hi, Have you tried putting the NFS mount unit as an option in After=. For example, say the mount point is /soft/slurm then there should be a mount unit called soft-slurm.mount.
Also, I'm not aware of a command to resume all nodes in a partition but you could try … for NODE in $( sinfo -hNe --partition=<PARTITION> --format=%n | sort | uniq ) do scontrol update nodename=$NODE state=resume done --- Samuel Gallop From: Daniel Ruiz Molina [mailto:[email protected]] Sent: 30 January 2017 13:37 To: slurm-dev <[email protected]> Subject: [slurm-dev] Daemon doesn't start at boot Hello, I have noticed that in all my cluster (running CentOS-7.x x886_64), slurmd daemon doesn't start automatically, even though it is configured as "enabled" at boot time. My slurmd systemd file is: # /usr/lib/systemd/system/slurmd.service [Unit] Description=Slurm node daemon After=ypbind.service nfs-mountd.service rpcbind.service nfs-lock.service nfslock.service nfs.target remote-fs.target ConditionPathExists=/soft/slurm-16.05/etc/slurm.conf [Service] Type=forking EnvironmentFile=-/etc/sysconfig/slurmd ExecStart=/soft/slurm-16.05/sbin/slurmd $SLURMD_OPTIONS ExecReload=/bin/kill -HUP $MAINPID PIDFile=/var/run/slurmd.pid KillMode=process LimitNOFILE=51200 LimitMEMLOCK=infinity LimitSTACK=infinity [Install] WantedBy=multi-user.target I need this daemon starts after ypbind and nfs-mount because slurm path is located in a NFS shared folder. Because of this, I added some days the "ConditionPathExits" attribute in Unit section, but daemon doesn't start. I think this is the reason my partition always are in "down" state... Is there a way for executing a "make node idle" for all the computes in a partition? Could you help me? Thanks.
