Skip per-boot update-initramfs when nouveau blacklist is baked into the VHD#161
Draft
ganeshkumarashok wants to merge 1 commit into
Draft
Skip per-boot update-initramfs when nouveau blacklist is baked into the VHD#161ganeshkumarashok wants to merge 1 commit into
ganeshkumarashok wants to merge 1 commit into
Conversation
…he VHD install.sh blacklists nouveau and runs `update-initramfs -u` on every node boot (~10-30s). When AgentBaker has already baked the nouveau blacklist into the VHD initramfs for the running kernel, skip the rebuild. The fast path engages only when a kernel-gated marker (/opt/azure/aks-gpu/nouveau-blacklist-marker) matches the running kernel AND the on-disk blacklist content matches the image's copy, so any VHD/image version skew, kernel drift, or content mismatch falls back to the original copy + update-initramfs path. Backward/forward compatible with VHDs that don't write the marker. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Collaborator
Author
|
Paired AgentBaker PR that writes the VHD marker: Azure/AgentBaker#8614 |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What
install.shblacklists nouveau and runsupdate-initramfs -uon every node boot (~10-30s). This change skips that rebuild when AgentBaker has already baked the nouveau blacklist into the VHD initramfs for the running kernel.The fast path engages only when all of:
/opt/azure/aks-gpu/nouveau-blacklist-markerexists and itskernel=line matchesuname -r, and/etc/modprobe.d/blacklist-nouveau.confis byte-identical (cmp) to the image's/opt/gpu/blacklist-nouveau.conf.Any mismatch — older VHD without the marker, kernel drift between VHD build and node boot, or altered content — falls back to the original
cp+update-initramfspath. Backward/forward compatible in both directions.Why
GPU provisioning-time reduction. Removes a deterministic ~10-30s boot cost and makes nouveau blacklisted from first boot.
Cross-repo dependency
The win only materializes once the VHD writes the marker. Pairs with Azure/AgentBaker PR (bake nouveau blacklist into the Ubuntu VHD). Draft until that lands and is validated on a VHD build. This PR is safe to ship independently (no marker ⇒ unchanged behavior).