medium riskPatching

Kernel / OS security patch

Apply OS/kernel security patches with a rolling reboot, keeping the driver aligned to the kernel.

Change scope

Rolling across nodes

Maintenance impact

Node reboots; GPU workloads drained per node.

Prerequisites
  • ·Patched image or package set validated with the GPU driver
  • ·Advisory reviewed (see Security / KnowledgeOps)
Pre-checks
  • ·Driver builds against the target kernel
  • ·Spare capacity for one node at a time
Procedure
  1. 1

    Cordon + drain, patch, reboot the node.

  2. 2

    The GPU Operator rebuilds/loads the driver for the new kernel; validate.

    bash
    ssh $NODE nvidia-smi
  3. 3

    Uncordon and proceed to the next node.

Validation
  • ✓Kernel at the patched version
  • ✓Driver loaded; GPU test pod runs
  • ✓No secure-boot/module signing errors
Rollback

Boot the previous kernel entry and re-pin; re-validate the driver.

Communication

Announce rolling reboots and per-node capacity reduction.