feat(nvidia-setup): add bcm service that aliases kernel headers (AICR #1093)#48
Merged
Conversation
… #1093) For service=bcm, the apply stage skips the kernel/EFA pipeline and runs a single step: symlink /usr/src/linux-$(uname -r) -> linux-headers-$(uname -r) so consumers reading /usr/src/linux-$(uname -r)/.config find it on Ubuntu. Without this, aicr validate stalls (DRA pod Init:0/1, driver daemonset unhealthy). Replaces the manual pdsh hardlink workaround and survives reboot and kernel upgrades. Bumps package version 0.2.2 -> 0.3.0.
ayuskauskas
requested review from
lockwobr,
mskalka and
rice-riley
as code owners
… a real dir A pre-existing real source tree at /usr/src/linux-$(uname -r) is not a failure for the bcm service — the consumer's .config lookup will still resolve. Log the situation and succeed instead of failing the step.
…RNING Pairs with the previous commit: the step now succeeds in this case, so the log prefix should match the exit code.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
bcmservice to thenvidia-setuppackage that creates a full-directory symlink/usr/src/linux-$(uname -r)->linux-headers-$(uname -r)so consumers (gpu-operatornvidia-driver-daemonset, NVIDIA DRA driver) which read/usr/src/linux-$(uname -r)/.configfind it on Ubuntu. Replaces the manualpdshhardlink workaround in AICR #1093 and survives reboot / kernel upgrade.service=bcmthe apply stage skips the kernel/EFA pipeline entirely and runs only this single step;apply-checkverifies the symlink resolves.bcm-*accelerator combinations: shipsbcm-h100andbcm-gb200defaults (intentionally empty — bcm uses no kernel/EFA/lustre versions).nvidia-setup0.2.2→0.3.0.Closes NVIDIA/aicr#1093.
Test plan
configMap: { service: bcm, accelerator: h100 }and confirm/usr/src/linux-$(uname -r)/.configresolvesaicr validate --phase deploymentand--phase conformanceend-to-end:nvidia-driver-daemonsetclean, DRA pod reachesRunning(noInit:0/1stall), health checks greeneks-h100/eks-gb200apply paths are unchanged