Remove backwards symlinks - #1044
Merged
maherthomsi merged 2 commits intoSep 18, 2026
Merged
Conversation
maherthomsi
force-pushed
the
remove-backwards-symlinks
branch
from
September 17, 2026 22:04
bdd9f13 to
a843044
Compare
maherthomsi
marked this pull request as ready for review
September 17, 2026 22:10
maherthomsi
requested review from
ginglis13,
piyush-jena and
vigh-m
and removed request for
piyush-jena
September 18, 2026 18:29
ginglis13
reviewed
Sep 18, 2026
maherthomsi
force-pushed
the
remove-backwards-symlinks
branch
from
September 18, 2026 21:08
a843044 to
e2882fb
Compare
ginglis13
approved these changes
Sep 18, 2026
The FHS library-path normalization left generate-cdi-specs generating `/usr/lib/nvidia/tesla` backwards-compatibility symlinks for every variant. Starting with EKS k8s 1.37, nvidia variants should not create these symlinks. Make the clean behavior the default: generate-cdi-specs.service no longer passes `--additional-symlinks /usr/lib/nvidia/tesla`, and its ExecStart can be overridden by a rendered drop-in. Ship a generate-cdi-specs-exec-start-conf-compat template that restores the flag for k8s 1.36 and below, and order the unit after settings-applier so the drop-in is present before it runs. Signed-off-by: Maher Homsi <maherhom@amazon.com>
The default device-plugin exec-start template no longer passes `--cdi-enabled-hooks create-lib-symlinks`, so containers do not receive the legacy `/usr/lib/nvidia/tesla` symlinks on EKS k8s 1.37+ nvidia variants. Add a -compat exec-start template that retains the flag for k8s 1.36 and below. Signed-off-by: Maher Homsi <maherhom@amazon.com>
maherthomsi
force-pushed
the
remove-backwards-symlinks
branch
from
September 18, 2026 23:10
e2882fb to
dad951d
Compare
vigh-m
approved these changes
Sep 18, 2026
Merged
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
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.
Description of changes:
Make the NVIDIA /usr/lib/nvidia/tesla backwards-compatibility symlinks opt-in so EKS k8s 1.37+ nvidia variants can drop them, while k8s 1.36 and below keep them. The FHS library-path normalization previously created these legacy symlinks for every variant; we don't want to carry that forward on new variants, but must not change behavior for existing ones.
This changes the defaults to clean (no symlinks) and ships -compat templates that downstream variants render to restore the old behavior.
Testing done:
Built an
aws-k8s-1.36-nvidia x86_64AMI consuming this kit and booted it as a node on a live EKS 1.36 cluster (nvidia.com/gpu allocatable = 1). Verified both paths:device-plugin exec-start.confhad no create-lib-symlinks;/etc/cdi/nvidia.jsoncontained 0 nvidia/tesla references - confirming the intended 1.37+ behavior.device-plugin exec-start.conf= --cdi-enabled-hooks create-lib-symlinks; generate-cdi-specs ran with --additional-symlinks; /etc/cdi/nvidia.json contained 30 nvidia/tesla references; nvidia-smi worked (driver 580.178.04); host /usr/lib/nvidia/tesla present (49 symlinks).Terms of contribution:
By submitting this pull request, I agree that this contribution is dual-licensed under the terms of both the Apache License, version 2.0, and the MIT license.