Skip to content

M #-: OneSlurm guide in the Slurm chapter - #1019

Merged
rsmontero merged 12 commits into
masterfrom
feature/oneslurm-docs
Oct 13, 2026
Merged

rsmontero merged 12 commits into
masterfrom
feature/oneslurm-docs

Conversation

@pablodelarco

@pablodelarco pablodelarco commented Oct 7, 2026 •

Copy link
Copy Markdown
Contributor

Description

This moves the OneSlurm guide from the one-apps wiki (slurm_intro, slurm_quick and slurm_feature) to the Slurm chapter, with the structure proposed on 5 October.

  • Getting Started: Overview, Requirements and Preparation, Quick Start and Core Concepts. The old slurm/overview page moves here, and its URL redirects with a Hugo alias.
  • Management: Cluster Lifecycle (including scaling), Cluster Configuration, Worker Profiles and Accelerators, Identity, Shared Storage, InfiniBand and Troubleshooting.
  • References: Service Architecture, Configuration Parameters, CLI and API. The API page shows the OneFlow REST calls for a OneSlurm cluster.

I checked every statement against appliances/OneSlurm on one-apps master and the marketplace YAML. On OpenNebula 7.4 I tested the Quick Start from the CLI and from Sunstone, scaling up and down, the API calls, the local LDAP example and NFS /home. These tests corrected a few points from the wiki:

  • Workers take the VM name as hostname.
  • The CLI asks for the inputs before the network.
  • With the default root_squash, the home directories must be created on the NFS server.
  • slurm.conf is written again at every boot.
  • The API needs all the service inputs, and force when scaling.

GPU and InfiniBand are migrated as written in the wiki, because I had no hardware to test them.

The Integrations section with the Open OnDemand guide comes from #1013, and the two branches merge without conflicts. EESSI, CernVM-FS, External Identity and Storage and the multi-tenant content are not included yet.

Branches to which this PR applies

  • master
  • one-7.4
  • one-7.4-maintenance

@rsmontero
rsmontero merged commit 9f40707 into master Oct 13, 2026
1 check passed
@rsmontero
rsmontero deleted the feature/oneslurm-docs branch October 13, 2026 10:40
rsmontero pushed a commit that referenced this pull request Oct 13, 2026
This moves the OneSlurm guide from the one-apps wiki (slurm_intro,
slurm_quick and slurm_feature) to the Slurm chapter, with the structure
proposed on 5 October.

- Getting Started: Overview, Requirements and Preparation, Quick Start
and Core Concepts. The old `slurm/overview` page moves here, and its URL
redirects with a Hugo alias.
- Management: Cluster Lifecycle (including scaling), Cluster
Configuration, Worker Profiles and Accelerators, Identity, Shared
Storage, InfiniBand and Troubleshooting.
- References: Service Architecture, Configuration Parameters, CLI and
API. The API page shows the OneFlow REST calls for a OneSlurm cluster.

I checked every statement against `appliances/OneSlurm` on one-apps
master and the marketplace YAML. On OpenNebula 7.4 I tested the Quick
Start from the CLI and from Sunstone, scaling up and down, the API
calls, the local LDAP example and NFS `/home`. These tests corrected a
few points from the wiki:

- Workers take the VM name as hostname.
- The CLI asks for the inputs before the network.
- With the default `root_squash`, the home directories must be created
on the NFS server.
- `slurm.conf` is written again at every boot.
- The API needs all the service inputs, and `force` when scaling.

GPU and InfiniBand are migrated as written in the wiki, because I had no
hardware to test them.

The Integrations section with the Open OnDemand guide comes from #1013,
and the two branches merge without conflicts. EESSI, CernVM-FS, External
Identity and Storage and the multi-tenant content are not included yet.

Co-authored-by: mattrowe-opennebula <mrowe@opennebula.io>
(cherry picked from commit 9f40707)
rsmontero pushed a commit that referenced this pull request Oct 13, 2026
This moves the OneSlurm guide from the one-apps wiki (slurm_intro,
slurm_quick and slurm_feature) to the Slurm chapter, with the structure
proposed on 5 October.

- Getting Started: Overview, Requirements and Preparation, Quick Start
and Core Concepts. The old `slurm/overview` page moves here, and its URL
redirects with a Hugo alias.
- Management: Cluster Lifecycle (including scaling), Cluster
Configuration, Worker Profiles and Accelerators, Identity, Shared
Storage, InfiniBand and Troubleshooting.
- References: Service Architecture, Configuration Parameters, CLI and
API. The API page shows the OneFlow REST calls for a OneSlurm cluster.

I checked every statement against `appliances/OneSlurm` on one-apps
master and the marketplace YAML. On OpenNebula 7.4 I tested the Quick
Start from the CLI and from Sunstone, scaling up and down, the API
calls, the local LDAP example and NFS `/home`. These tests corrected a
few points from the wiki:

- Workers take the VM name as hostname.
- The CLI asks for the inputs before the network.
- With the default `root_squash`, the home directories must be created
on the NFS server.
- `slurm.conf` is written again at every boot.
- The API needs all the service inputs, and `force` when scaling.

GPU and InfiniBand are migrated as written in the wiki, because I had no
hardware to test them.

The Integrations section with the Open OnDemand guide comes from #1013,
and the two branches merge without conflicts. EESSI, CernVM-FS, External
Identity and Storage and the multi-tenant content are not included yet.

Co-authored-by: mattrowe-opennebula <mrowe@opennebula.io>
(cherry picked from commit 9f40707)
mattrowe-opennebula added a commit that referenced this pull request Oct 22, 2026
This moves the OneSlurm guide from the one-apps wiki (slurm_intro,
slurm_quick and slurm_feature) to the Slurm chapter, with the structure
proposed on 5 October.

- Getting Started: Overview, Requirements and Preparation, Quick Start
and Core Concepts. The old `slurm/overview` page moves here, and its URL
redirects with a Hugo alias.
- Management: Cluster Lifecycle (including scaling), Cluster
Configuration, Worker Profiles and Accelerators, Identity, Shared
Storage, InfiniBand and Troubleshooting.
- References: Service Architecture, Configuration Parameters, CLI and
API. The API page shows the OneFlow REST calls for a OneSlurm cluster.

I checked every statement against `appliances/OneSlurm` on one-apps
master and the marketplace YAML. On OpenNebula 7.4 I tested the Quick
Start from the CLI and from Sunstone, scaling up and down, the API
calls, the local LDAP example and NFS `/home`. These tests corrected a
few points from the wiki:

- Workers take the VM name as hostname.
- The CLI asks for the inputs before the network.
- With the default `root_squash`, the home directories must be created
on the NFS server.
- `slurm.conf` is written again at every boot.
- The API needs all the service inputs, and `force` when scaling.

GPU and InfiniBand are migrated as written in the wiki, because I had no
hardware to test them.

The Integrations section with the Open OnDemand guide comes from #1013,
and the two branches merge without conflicts. EESSI, CernVM-FS, External
Identity and Storage and the multi-tenant content are not included yet.

Co-authored-by: mattrowe-opennebula <mrowe@opennebula.io>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants