Skip to content

Conversation

@Pavani-Panakanti
Copy link
Contributor

Issue

Update daemonset config changes required for neuron instances

Description of changes

  1. Dual DaemonSet Architecture
  • Split into two DaemonSets: regular and privileged
  • Regular DaemonSet: targets standard EFA instances with restricted security context
  • Privileged DaemonSet: targets inf/trn instances with privileged security context
  • Mutually exclusive node targeting ensures no conflicts
  1. Security Context Differentiation
  • Added privilegedSecurityContext in values.yaml for neuron instances
  • privileged: true, allowPrivilegeEscalation: true, runAsUser: 0
  • Existing securityContext unchanged for backward compatibility
  1. Instance Type Lists
  • Added privilegedSupportedInstanceLabels for neuron instances (inf1.24xlarge, trn1.32xlarge, etc.)
  • Existing supportedInstanceLabels excludes privileged instances
  • Lists are mutually exclusive by design
  1. Neuron-Specific Volume Mounts
  • Added /opt/aws/neuron/ volume mount only to privileged DaemonSet
  • Required for neuron tooling access on inf/trn instances

Backward compatible

Testing

Will update

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.

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.

1 participant