You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Headless opencode run intermittently hangs at startup (log frozen at exactly 1,472 bytes, before any session/API activity) #36763
opencode run sometimes freezes at startup before creating a session row or making any API call. Without --print-logs we observed it repeatedly (~12 launches; one prompt went 3-for-3). With --print-logs it is much rarer but not eliminated — five occurrences to date where the log froze right after init + cleanup prune=7.days: no session row, no spend, 2+ minutes of nothing.
Consistent signature across the three most recent occurrences:
The frozen log is exactly 1,472 bytes.
The cleanup prune=7.days line lands ~60 s after init before the freeze (healthy runs: <1 s gap) — looks like a timeout being hit.
4 of 5 occurred within minutes after a previous opencode process was killed (SIGTERM/pkill) — stale-state suspicion, though we found no lock files or leftover processes.
Kill and relaunch with the identical command clears it every time.
Expected: startup either proceeds or exits with an error. Suspected non-TTY output race or stale-state wait during init.
Not the same as #30337 (node_modules scanning) or #25668 (plugin loading, Windows) — this workspace has no .opencode/node_modules and no plugins. Possibly adjacent to #29395 (init deadlock), but here there is no concurrent instance running at launch time.
Workaround we use: always --print-logs; treat "no session row + no spend within ~60 s" as hung; kill and relaunch.
Plugins
None
OpenCode version
1.17.18
Steps to reproduce
Not deterministic. Shape: opencode run --print-logs --log-level INFO "<prompt>" > log 2>&1 in a fresh non-TTY shell, most reliably within a few minutes of killing a previous opencode process. Occasionally the log stops after the cleanup line at exactly 1,472 bytes.
If this is the same root cause, please add a comment there rather than opening a new issue. If you believe this is distinct (e.g. the exact-1,472-byte log size and ~60 s cleanup delay point to a different code path), please clarify what differentiates it from #35870.
Activity
github-actions commentedon Jul 13, 2026
This issue might be a duplicate of existing issues. Please check:
If this is the same root cause, please add a comment there rather than opening a new issue. If you believe this is distinct (e.g. the exact-1,472-byte log size and ~60 s cleanup delay point to a different code path), please clarify what differentiates it from #35870.
github-actions commentedon Sep 13, 2026
To stay organized issues are automatically closed after 60 days of no activity. If the issue is still relevant please open a new one.