NVIDIA-NeMo / NVIDIA-NeMo/Gym

[checkpoint] Define checkpoint adapters for ns_tools sessions

Open
#3,243 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

area:environment feature
Dominant language
Python
Stars
1.2k
Forks
349
Avg merge
1d 21h
Merged PRs (30d)
318

Description

Parent

Child of #3219 and #3024.

Problem and state owner

resources_servers/ns_tools/app.py owns ToolManager-owned stateful tool sessions and per-session timing data used by verification.

Proposed adapter

Add a ToolManager snapshot/reference API for supported tools, preserve exact timing records, and reject arbitrary externally mutating tools without a safe contract.

Implement explicit checkpoint opt-in, route classification, export/restore/retire hooks, execution-to-session binding, revision reporting, atomic replacement under attempt N+1, and stale-attempt fencing.

Acceptance criteria

  • DirectPython state and timing totals survive process replacement; unsupported stateful tools fail checkpoint admission explicitly.
  • Checkpoint after multiple mutations, replace the server process, restore, and compare the next operation with an uninterrupted execution.
  • Corrupt or incomplete bulk state activates no sessions.
  • Attempt N cannot mutate or retire restored attempt N+1.
  • Terminal cleanup is idempotent and releases restored state exactly once.

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

Start with resources_servers/ns_tools/app.py and review the parent issues #3219 and #3024 to understand ToolManager-owned session state and timing data. Trace the checkpoint entry points and define completion against the listed acceptance criteria: safe export and restore, rejection of unsupported state, atomic replacement, stale-attempt fencing, and idempotent cleanup.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
backend
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Active
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.