1a952a4e7a
option for docker host access to cluster
2021-06-28 17:51:45 +02:00
7e767c3716
memory / cpu ressources fix
2021-06-25 12:19:00 +02:00
9237d736d8
tags
2021-06-24 14:17:16 +02:00
e979ea4d6e
fix hostname of cobald slurm node
...
made cobald be able to run slurm jobs, previously failing with
permission denied.
2021-06-24 14:07:35 +02:00
1f4dfe1821
build cobald image from slurm role, separated tags
2021-06-22 16:48:56 +02:00
78850d4636
merged slurm_dockerimage back into slurm role
2021-06-22 00:26:00 +02:00
e78e184375
WIP: cobald container containing and using slurm
2021-06-21 19:19:19 +02:00
02e87d7c40
cleanup, requisite files instead of startupscripts
2021-06-18 12:03:14 +02:00
4450c9bb65
WIP: separate slurm base and docker images
2021-06-17 22:50:40 +02:00
6eb6984d6a
new startup for cobald containers
2021-06-17 14:55:34 +02:00
f7dd3bcf02
run slurmctld as user
...
Notice: also trying to run slurmd on execute nodes as user makes no
sense because it breaks sbatch. Furthermore there is another necessary
to run mpi jobs (just tried MpiDefault=none). I don't consider running
slurmd as root a good idea, but there seems to be no other choice at the
moment.
2021-04-30 17:15:57 +02:00
bc27dec00e
shared volume
2021-04-27 22:36:06 +02:00
aa0fe4d039
variable username
2021-04-27 15:00:12 +02:00
89423edf25
fully variable execute nodes
2021-04-27 14:18:14 +02:00
077988e03d
privileged container switch
2021-04-26 18:00:40 +02:00
4586fa7092
slurm startup
2021-04-26 17:20:54 +02:00
53502213bc
WIP: slurm
2021-04-24 00:13:57 +02:00
