-
I get the following error message when launching a simple elasticluster ( the instances are there and I can SSH into them, but SLURM isn't working for example)
```
TASK [slurm-master : Ensure `slurmd…
-
Hi Margus,
the hadd step does not seem to work if the .root files produced by running the analysis code are stored on /hdfs .The error message is attached below [*] (from /hdfs/local/ttH_2tau/veelk…
-
- name: restart and enable slurmdbd
service: name=slurmdbd state=restarted enabled=yes
Above task in dbd.yml playbook needs a "slurm" user to be successful.
Slurm user is created in service.yml whi…
A1ve5 updated
8 years ago
-
Scenario - first time a cluster is setup.
Tasks:
1. main.yml which includes:
2. common.yml - goes fine (installs slurm, templates in some files)
3. dbd.yml tries to copy fgci-ansible/files/munge.k…
-
The accounting service appears to start too early and stops because it does not detect the cluster properly.
When restarted manually later, it works normally.
-
We must have one single module for all slurm components : slurmd, slurmcommons, slurmclient, slurmctld must be merged.
Slurmdbd and munge remain separates modules.
-
Some reasoning behind the tasks. for example: we from inside the service.yml start and enable the slurmdbd and not in dbd.yml.
-
I'm trying to install SLURM DRMAA on a CentOS image and I'm running into problems installing it from source. I'm aware that CentOS doesn't have the nice slurm-drmaa package. I'm not sure why slurm-drm…
-
We should only restart if something changes.
https://github.com/CSC-IT-Center-for-Science/ansible-role-slurm/blob/master/tasks/compute.yml#L19
-
I just installed slurm to /usr/local:
-rwxr-xr-x 1 root root 3605314 Nov 21 21:53 /usr/local/bin/sacct
-rwxr-xr-x 1 root root 3921674 Nov 21 21:53 /usr/local/bin/sacctmgr
-rwxr-xr-x 1 root root 35…