NVIDIA-NeMo / NVIDIA-NeMo/Curator

Resuming the job on slurm after it gets cancelled.

Open
#297 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

community-request enhancement jira
Dominant language
Python
Stars
1.8k
Forks
328
Avg merge
4d 5h
Merged PRs (30d)
30

Description

Initial discussion happened with @VibhuJawa

Is your feature request related to a problem? Please describe.
While running a workflow on slurm with large files, if it needs to be cancelled or stopped due to 4 hr slurm job restriction, we will have partial results. Can we have robust resume feature which can handle even if files are partially processed.

Describe the solution you'd like
Merging output file parts (0.part, 1.part, ...) into 1 file and compare with input file for checking whether complete processing has happened or not, if not then while job resumes the remaining file part will also be counted in set which will go for processing.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

No files, tests, or entry points are named. Start by locating the Slurm workflow execution and output-part handling, then trace how cancellation and restart currently behave; done should include safely combining completed parts, detecting incomplete input processing, and resuming only the remaining parts.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
data-engineering
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.