[pull] master from NVIDIA:master#80
Open
pull[bot] wants to merge 627 commits into
Open
Conversation
DeepOps Release 22.04
Update Ansible to match Kubespray supported versions
See https://developer.nvidia.com/blog/updating-the-cuda-linux-gpg-repository-key/ for details on the key changes. This commit: - Removes the nvidia-ml repo, which is deprecated and will not be updated - Updates the nvidia_cuda and nvidia_dcgm roles to use the new key and install workflow - Updates roles/requirements.txt to point to an updated version of nvidia.nvidia_driver
Update NVIDIA signing key
Update default Slurm version to 21.08.8
This specifies the env path for the sudo user.
Add a basic playbook for installing Mellanox OFED
Trident fix alt
…CL container image name and location
Share slurm.conf to have same of it - in case of configuring nodes and/or partitions. Locate slurm.conf at /sw/.slurm/slurm.conf of controller where the /sw is nfs mount point to all of nodes. Let /etc/slurm/slurm.conf be a soft link to it from nodes and controller. Signed-off-by: Seyong Um <seyong.um@hyundai.com>
The flag slurm_conf_symlink will allow modes to share slurm.conf via nfs. Signed-off-by: Seyong Um <seyong.um@hyundai.com>
Fix GPG key import for a large set of DeepOps roles
Signed-off-by: Seyong Um <seyong.um@hyundai.com>
Have same slurm.conf among nodes and controller
Bugs:
1. docker.slurm-exporter.service creates subdirectories
`/usr/loca/bin/{sinfo,sdiag,squeue} when the excutables are absent
(due to uninstall process). When deepops tries to install the executables
again, they are put under the unwanted subdirectories (e.g.
/usr/local/bin/sinfo/sinfo instead of /usr/local/bin/sinfo)
2. Slurm 'make uninstall' doesn't remove files under "{{ slurm_install_prefix
}}/lib/slurm". Therefore the directory needs to be removed. A safer solution
would be to run the correct 'make uninstall' command (at the correct location)
to uninstall the files. Without this handling, slurm fails to start because of
library version conflicts.
Updated NCCL results with DGX A100s; MPI commands; NCCL container image name and location
Checking for ansible_hostname is unneccessary. Signed-off-by: Vang Le-Quy <biocyberman@gmail.com>
Fix bugs preventing slurm reinstall or rebuild
Signed-off-by: Brandon Tuttle <tuttlebr@ncc-1701.biz>
Signed-off-by: Brandon Tuttle <tuttlebr@ncc-1701.biz>
release: 26.05 notes and README tag
feat(dgx): update DGX software stack role
Refresh CUDA example images
Refresh DCGM exporter
Add Ubuntu 24.04 NVIDIA Container Toolkit path
Retire legacy PXE provisioning paths
docs: clarify legacy CI and virtual lab status
Support Red Hat NVIDIA Container Toolkit path
Refresh Container Toolkit airgap docs
docs: clarify current OS validation targets
parse_versioning kept the last json.loads-parseable line and then indexed it with a string key. When that line parsed to a non-dict (e.g. a JSON array), the index raised an uncaught TypeError because only KeyError was handled, aborting the task. Guard that the parsed value is a dict before indexing and fall through to the existing 'no JSON could be loaded' path otherwise. Adds a regression test covering the non-dict input case.
scripts/common.sh printed an always-empty DEEPOPS_VERSION unless a user happened to set it in config/env.sh. Default it to the checkout's git tag description (git describe --tags --always) with an unknown fallback, keep the env.sh override, and document the variable in config.example/env.sh.
…or-6394703 fix: handle non-dict parsed JSON in firmware version manifest parser
Refresh workload, RoCE, and registry examples
chore(release): bump component versions for 26.07
Clarify Kubernetes storage helper support boundaries
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
19 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.
See Commits and Changes for more details.
Created by
pull[bot]
Can you help keep this open source service alive? 💖 Please sponsor : )