lablup / lablup/backend.ai

Handle session creation failures

Open
#14,323 0 comments 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
670
Forks
183
Avg merge
21h 49m
Merged PRs (30d)
404

Description

A session that reaches CREATING has no failure path. No lifecycle handler targets SessionStatus.CREATING, and the CHECK_CREATING_PROGRESS promotion spec only produces successes, so SessionHandlerOptions timeout / max_retry_count are never consulted. When kernel creation fails on the agent side without emitting a termination event, the session stays in CREATING forever and the user sees no error.

Scope: detect and surface creation failures so a session leaves CREATING with a meaningful status and reason.

- Propagate agent-side kernel creation errors to the manager.
- Apply a phase timeout / retry budget to the CREATING phase.
- Expose the failure reason to the user instead of an indefinite CREATING state.

JIRA Issue: BA-7725

Contributor guide

Open the contributing guide

Research direction

Start at the SessionStatus.CREATING lifecycle handling and the CHECK_CREATING_PROGRESS promotion spec, then trace SessionHandlerOptions timeout and max_retry_count between the agent and manager. Use the promotion spec and related lifecycle tests to verify that failed creation exits CREATING with a user-visible status and reason.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
backend, distributed-systems
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Active
Clarity
Mostly clear
Newbie friendliness
48/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.