Closed
Conversation
shivakunv
force-pushed
the
amazonlinuxsupport
branch
2 times, most recently
from
f6096c6 to
57fdf0b
Compare
shivakunv
commented
Oct 7, 2024
shivakunv
force-pushed
the
amazonlinuxsupport
branch
from
0a5abb1 to
7cb0b27
Compare
shivakunv
force-pushed
the
amazonlinuxsupport
branch
2 times, most recently
from
dedd781 to
31a2314
Compare
shivakunv
requested review from
ArangoGutierrez,
cdesiniotis,
elezar and
tariq1890
shivakunv
force-pushed
the
amazonlinuxsupport
branch
2 times, most recently
from
6d553ac to
bc5998b
Compare
shivakunv
force-pushed
the
amazonlinuxsupport
branch
3 times, most recently
from
426a2cf to
49b283c
Compare
tariq1890
reviewed
Oct 28, 2024
tariq1890
reviewed
Oct 28, 2024
cdesiniotis
reviewed
Oct 28, 2024
cdesiniotis
reviewed
Oct 28, 2024
shivakunv
force-pushed
the
amazonlinuxsupport
branch
2 times, most recently
from
eb0282a to
c46bcaf
Compare
shivakunv
commented
Oct 28, 2024
shivakunv
force-pushed
the
amazonlinuxsupport
branch
2 times, most recently
from
3d7b0bb to
d179a93
Compare
shivakunv
commented
Oct 29, 2024
shivakunv
force-pushed
the
amazonlinuxsupport
branch
2 times, most recently
from
74b2554 to
539af6c
Compare
Contributor
Author
|
@cdesiniotis and @tariq1890 PTAL |
tariq1890
reviewed
Oct 31, 2024
| # due to cuda repo cache issue , nvidia-fabric-manager refers to 565 version only | ||
| # install fabric-manager and nvidia-nscq | ||
| RUN if [ "$DRIVER_TYPE" != "vgpu" ] && [ "$TARGETARCH" != "arm64" ]; then \ | ||
| dnf install -y nvidia-fabric-manager libnvidia-nscq-${DRIVER_BRANCH}-${DRIVER_VERSION}-1; fi |
Contributor
There was a problem hiding this comment.
We can't commit this as-is. I just looked at the packages uploaded here, so the following should work
Suggested change
| dnf install -y nvidia-fabric-manager libnvidia-nscq-${DRIVER_BRANCH}-${DRIVER_VERSION}-1; fi | |
| dnf install -y nvidia-fabricmanager-${DRIVER_BRANCH}-${DRIVER_VERSION}-1 libnvidia-nscq-${DRIVER_BRANCH}-${DRIVER_VERSION}-1; fi |
Note that you have nvidia-fabric-manager, when it should be nvidia-fabricmanager*
Contributor
Author
There was a problem hiding this comment.
removed 560.35.03 . 560.35.03 fabric manager not available
dnf list *fabric*
24.23 cuda-drivers-fabricmanager.x86_64 565.57.01-1 cuda-amzn2023-x86_64
24.23 cuda-drivers-fabricmanager-555.x86_64 555.42.06-1 cuda-amzn2023-x86_64
24.23 cuda-drivers-fabricmanager-560.x86_64 560.35.03-1 cuda-amzn2023-x86_64
24.23 cuda-drivers-fabricmanager-565.x86_64 565.57.01-1 cuda-amzn2023-x86_64
24.23 libfabric.x86_64 1.14.0-2.amzn2023.0.2 amazonlinux
24.23 libfabric-devel.x86_64 1.14.0-2.amzn2023.0.2 amazonlinux
24.23 nvidia-fabric-manager.x86_64 565.57.01-1 cuda-amzn2023-x86_64
24.23 nvidia-fabric-manager-devel.x86_64 565.57.01-1 cuda-amzn2023-x86_64
added conditional check for both the packages and installation.
Contributor
Author
There was a problem hiding this comment.
removed if condition and added installation of nvidia-fabric-manager-${DRIVER_VERSION}-1
shivakunv
force-pushed
the
amazonlinuxsupport
branch
4 times, most recently
from
400da5c to
ed626af
Compare
Contributor
Author
|
PTAL @cdesiniotis @tariq1890 |
cdesiniotis
reviewed
Nov 5, 2024
| # Initialize the fabric manager package variable | ||
| FABRIC_PACKAGE=""; \ | ||
| if dnf list nvidia-fabric-manager-${DRIVER_VERSION}-1 &>/dev/null; then \ | ||
| FABRIC_PACKAGE="nvidia-fabric-manager-${DRIVER_VERSION}-1"; \ |
Contributor
There was a problem hiding this comment.
Looking at https://developer.download.nvidia.com/compute/cuda/repos/amzn2023/x86_64/, the only fabric manager packages available are named nvidia-fabric-manager-${DRIVER_VERSION}-1. Let's remove the conditional here and always use that package name. If the name ever changes, our builds will break and we will know right away.
Contributor
Author
There was a problem hiding this comment.
removed if condition and added installation of nvidia-fabric-manager-${DRIVER_VERSION}-1
Comment on lines
+402
to
+406
| if [ -f /sys/module/nvidia_fs/refcnt ]; then | ||
| nvidia_fs_refs=$(< /sys/module/nvidia_fs/refcnt) | ||
| rmmod_args+=("nvidia-fs") | ||
| ((++nvidia_deps)) | ||
| fi |
Contributor
There was a problem hiding this comment.
Is this change required? We have a separate sidecar container for loading / unloading nvidia-fs.
Signed-off-by: shiva kumar <shivaku@nvidia.com>
shivakunv
force-pushed
the
amazonlinuxsupport
branch
from
ed626af to
de77660
Compare
Contributor
|
@shivakunv do we need these changes? Or we are fine with closing this PR if these are no longer needed? |
Contributor
Author
|
Closing this since the release to amzn2023 is not planned. |
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
4 participants
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.
No description provided.