fix(langgraph): persist a failure record when recovery finds no checkpoint - #2
Open
SabreKeyZ wants to merge 1 commit into
Open
fix(langgraph): persist a failure record when recovery finds no checkpoint#2SabreKeyZ wants to merge 1 commit into
SabreKeyZ wants to merge 1 commit into
Conversation
…point A run that dies before its first durable put leaves the thread empty. Recovery via invoke(None) then raised EmptyInputError with no resumable state and no durable evidence the accepted run was lost. On that empty-resume path, persist a loop checkpoint plus an ERROR write. With durability=sync, also wait for the input checkpoint before the first user node so an accepted record exists before node execution. Fixes langchain-ai#8764 Co-authored-by: 梁朝伟 <SabreKeyZ@users.noreply.github.com>
SabreKeyZ
marked this pull request as ready for review
August 31, 2026 03:06
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Fixes langchain-ai#8764
A fire-and-forget run that dies before its first durable
put()left recovery withEmptyInputErrorand no checkpoint. Recovery now writes a loop checkpoint plus anERRORpending write for that empty-resume path, anddurability="sync"waits for the input checkpoint before the first user node.This PR is on SabreKeyZ/langgraph because the agent token cannot open a PR against langchain-ai/langgraph (
Resource not accessible by integration). Compare URL for the intended upstream PR:https://github.com/langchain-ai/langgraph/compare/main...SabreKeyZ:langgraph:cursor/crash-before-first-checkpoint-370a?expand=1
Assignment requested on the issue: langchain-ai#8764 (comment)
Release note
Recovery of a thread with no durable checkpoint now persists a failure record (
ERRORpending write) instead of raisingEmptyInputErrorwith an empty store. Withdurability="sync", the input checkpoint is flushed before the first user node runs.How did you verify your code works?
In
libs/langgraph:make format— 151 files unchangedmake lint— ruff, import order, andty checkpassedtest_crash_before_first_checkpoint_records_failure_on_recovery(sync + async) — passedtest_checkpoint_recovery,test_checkpoint_errors,test_pending_writes_resume,test_empty_invoke,test_checkpoint_recovery_async,test_drain_with_exit_durability_persists_resume_checkpoint— 33 passeddurabilityofsync,async, andexit— each persisted anERRORwrite