Skip to content

[Enhancement] modified template to automatically generate resources.claim - #195

Merged
kalantar merged 4 commits into
llm-d-incubation:mainfrom
ryojsb:modified-template
Jan 29, 2026
Merged

[Enhancement] modified template to automatically generate resources.claim#195
kalantar merged 4 commits into
llm-d-incubation:mainfrom
ryojsb:modified-template

Conversation

@ryojsb

@ryojsb ryojsb commented Jan 29, 2026

Copy link
Copy Markdown
Contributor

This PR let us release from the complicated specification when useing existing ResourceClaimTemplate.

Currently we have to specify spec.resourceClaims and resources.claims only when using existing ResourceClaimTemplate while accelerator's DRA parameters are automatically generated and merged.

accelerator:
  type: nvidia
  dra: true

decode:
  create: true
    resources:
      limits:
      ...
      claims:
      - name: llm-d-imex-channel-0
    mountModelVolume: true
  resourceClaims:
  - name: llm-d-imex-channel-0

In this PR, you'll only specify spec.resourceClaims when using existing ResourceClaimTemplate.

Then accelerator's resources.claims and the ones of existing ResourceClaimTemplate are automatically generated.

accelerator:
  type: nvidia
  dra: true

decode:
  create: true
    resources:
      limits:
      ...
    mountModelVolume: true
  resourceClaims:
  - name: llm-d-imex-channel-0

Signed-off-by: ryojsb <ryoh80213@gmail.com>
Signed-off-by: ryojsb <ryoh80213@gmail.com>
Signed-off-by: ryojsb <ryoh80213@gmail.com>
Signed-off-by: ryojsb <ryoh80213@gmail.com>

@kalantar kalantar left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

looks good. nice simplification

@kalantar
kalantar merged commit 5017eda into llm-d-incubation:main Jan 29, 2026
4 checks passed
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.

2 participants