langchain-ai/langgraph
1
0
Fork 0
mirror of https://github.com/langchain-ai/langgraph.git synced 2026-07-15 12:35:43 -04:00
Code Issues 314 Packages Projects Releases 17 Wiki Activity

[PR #6223] fix(langgraph): fix nested subgraph checkpoint replay #4913

New Issue
Closed
opened 2026-02-20 17:50:56 -05:00 by yindo · 0 comments
yindo commented 2026-02-20 17:50:56 -05:00
Owner
Copy Link
Copy Source

Original Pull Request: https://github.com/langchain-ai/langgraph/pull/6223

State: closed
Merged: No

Overview

Subgraphs do not have a checkpoint ID at config.configurable.checkpoint_id. In PregelLoop.__enter__ when the most recent checkpoint tuple is retrieved, if no checkpoint ID exists, checkpointer.get_tuple gets the maximum checkpoint in the namespace. This causes issues when time-traveling to a checkpoint preceding an already-executed subgraph, because the subgraph will replay the global latest checkpoint (in the subgraph namespace) instead of re-executing.

This PR adds some protections to ensure that when time-travelling back to a node that executes a subgraph, we ignore checkpoints newer than the boundary checkpoint so the subgraph reruns instead of replaying future state.

Tests

  • tests/test_pregel.py::test_subgraph_resume_reexecutes_from_valid_checkpoint
  • tests/test_pregel_async.py::test_subgraph_resume_reexecutes_from_valid_checkpoint

These are modeled after https://github.com/langchain-ai/langgraph/issues/6165, which this PR solves

**Original Pull Request:** https://github.com/langchain-ai/langgraph/pull/6223 **State:** closed **Merged:** No --- Overview Subgraphs do not have a checkpoint ID at `config.configurable.checkpoint_id`. In `PregelLoop.__enter__` when the most recent checkpoint tuple is retrieved, if no checkpoint ID exists, `checkpointer.get_tuple` gets the maximum checkpoint in the namespace. This causes issues when time-traveling to a checkpoint preceding an already-executed subgraph, because the subgraph will replay the global latest checkpoint (in the subgraph namespace) instead of re-executing. This PR adds some protections to ensure that when time-travelling back to a node that executes a subgraph, we ignore checkpoints newer than the boundary checkpoint so the subgraph reruns instead of replaying future state. Tests - tests/test_pregel.py::test_subgraph_resume_reexecutes_from_valid_checkpoint - tests/test_pregel_async.py::test_subgraph_resume_reexecutes_from_valid_checkpoint These are modeled after https://github.com/langchain-ai/langgraph/issues/6165, which this PR solves
yindo added the pull-request label 2026-02-20 17:50:56 -05:00
yindo closed this issue 2026-02-20 17:50:56 -05:00
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
agents
awaiting author revision
bug
change
chat models
checkpointer
ci
code quality
code quality
dependencies
documentation
drawing
duplicate
enhancement
good first issue
help wanted
invalid
maintainer
meta
missing information
not maintained by langchain
open-swe-auto
open-swe-max
open-swe-max-auto
pending
performance
pull-request
Mirrored from GitHub Pull Request
 question
requires design
sdk
server
stale
studio
tests
type checking
usage question
v0.5
v1.1
No Label pull-request
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
yindo
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: langchain-ai/langgraph#4913
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?