openclaw - 💡(How to fix) Fix New claude-cli live session fails immediately with FailoverError while session reuse succeeds [1 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
openclaw/openclaw#72206Fetched 2026-04-27 05:33:17
View on GitHub
Comments
0
Participants
1
Timeline
0
Reactions
0
Participants

When OpenClaw starts a new claude-cli live session (e.g. for systemEvent wake), all models in the fallback chain fail immediately (~300–330 ms) with FailoverError and reason=unknown. However, session reuse on the same model succeeds normally.

Error Message

claude live session turn failed: provider=claude-cli model=claude-sonnet-4-6 durationMs=2880 error=FailoverError

Root Cause

When OpenClaw starts a new claude-cli live session (e.g. for systemEvent wake), all models in the fallback chain fail immediately (~300–330 ms) with FailoverError and reason=unknown. However, session reuse on the same model succeeds normally.

RAW_BUFFERClick to expand / collapse

Summary

When OpenClaw starts a new claude-cli live session (e.g. for systemEvent wake), all models in the fallback chain fail immediately (~300–330 ms) with FailoverError and reason=unknown. However, session reuse on the same model succeeds normally.

Version

2026.4.25-beta.4 (6ecae22)

Reproduction

  1. Trigger a systemEvent that requires a new live session (e.g. openclaw system event --text "test" --mode now --expect-final)
  2. Observe: 6-model fallback chain exhausted in ~5s, all fail

Observed log pattern

``` claude live session start: provider=claude-cli model=claude-sonnet-4-6 activeSessions=2 claude live session turn failed: provider=claude-cli model=claude-sonnet-4-6 durationMs=2880 error=FailoverError model_fallback_decision: candidate_failed reason=unknown errorPreview="Claude CLI failed." ... (repeats for haiku-4-5, opus-4-6, opus-4-5, sonnet-4-5, opus-4-7) Embedded agent failed before reply: All models failed (6)

Immediately after, a Telegram direct message succeeds via reuse:

claude live session reuse: provider=claude-cli model=claude-sonnet-4-6 ← OK ```

Key observation

  • New session start → immediate failure (~330ms), reason=unknown (no stderr captured)
  • Session reuse → success
  • Identical model, same auth token

Related

  • #71803 (different: watchdog/idle timeout)
  • #71793 (closed: stdout buffer)

extent analysis

TL;DR

Investigate the difference in behavior between new session starts and session reuse to identify the root cause of the immediate FailoverError with reason=unknown.

Guidance

  • Review the authentication and authorization process for new sessions to ensure that the auth token is being properly validated and applied.
  • Compare the configuration and environment of new sessions versus reused sessions to identify any discrepancies that could be contributing to the failure.
  • Verify that the claude-cli provider is properly configured and functioning correctly for new sessions.
  • Check the error handling and logging mechanisms to ensure that any relevant error messages or exceptions are being captured and reported.

Example

No code snippet is provided as the issue does not contain sufficient information to create a relevant example.

Notes

The issue may be related to the way new sessions are handled by the claude-cli provider, and further investigation is needed to determine the root cause. The fact that session reuse succeeds suggests that the issue is specific to new session creation.

Recommendation

Apply a workaround to retry new session creation or to use an alternative provider if available, as the root cause of the issue is not yet clear. This will allow for temporary mitigation of the problem while further investigation is conducted.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING