Add DGXCloud Ray backend and improve DGXCloudExecutor workload management#480
Draft
rapaul-nv wants to merge 1 commit intoNVIDIA-NeMo:mainfrom
Draft
Add DGXCloud Ray backend and improve DGXCloudExecutor workload management#480rapaul-nv wants to merge 1 commit intoNVIDIA-NeMo:mainfrom
rapaul-nv wants to merge 1 commit intoNVIDIA-NeMo:mainfrom
Conversation
- Renamed app_id and app_secret to client_id and client_secret for clarity. - Introduced new methods for deleting workloads and checking workspace status. - Enhanced data movement functionality to use a tarball when within character limits, falling back to individual file deployment otherwise. - Updated RayCluster and RayJob to integrate DGXCloudExecutor and its corresponding classes. Fixes NVIDIA-NeMo#478 Signed-off-by: Rakesh Paul <rapaul@nvidia.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
nemo_run/run/ray/dgxcloud.py): NewDGXCloudRayClusterandDGXCloudRayJobclasses that enable Ray orchestration on DGX Cloud via distributed workloads. Pods self-organise into a Ray head + worker topology using hostname-derived rank and a shared PVC for head-IP discovery.DGXCloudExecutorworkload management (nemo_run/core/execution/dgxcloud.py): Migrate auth fromapp_tokentoclient_credentialsgrant type; add generic_run_workspace_and_waithelper for polling workspace workloads; support large data transfers via chunked per-file fallback when the tarball exceeds the API arg limit; improvefetch_logswith terminal-state detection and non-cluster log polling; adddeploy_script_to_pvcandlargeShmRequestsupport.RayClusterandRayJobfactory maps (run/ray/cluster.py,run/ray/job.py).Fixes: #478
Test plan
client_credentialsgrant typeMAX_ARGS_CHARSlimit (chunked fallback)fetch_logsfor both cluster-launched and non-cluster-launched jobsdeploy_script_to_pvcdeploys and sets executable permissions correctly