Controller¶
Full build-out of the controller (master) node — server02. This one machine also hosts NFS and LDAP for the rest of the cluster.
-
NFS
Hostname/FQDN, NFS server export, and disk quota setup.
-
LDAP
slapd install, TLS/CA certificates, and the LDAP Account Manager web UI.
-
MUNGE
Authentication service for SLURM — key generation happens here.
-
SLURM
Database, building from source, configuration, and debugging.
-
GPU / CUDA
CUDA toolkit installation —
server02has its own GPUs. -
SSH Access
NFS-server SFTP-only policy for
slurmUsers. -
clush
Install, key generation, lockdown, and node groups.