Server ignores SIGTERM, so every container stop waits 10s and ends in SIGKILL #81
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "fix/bb-91-sigterm"
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?
Handle SIGTERM alongside Ctrl-C so container shutdown starts promptly. Signal active SSE responses to close, allowing graceful shutdown to finish with connected clients.
Verified with
cargo build --workspace,cargo test --workspace,cargo clippy --workspace --all-targets -- -D warnings, andcargo run -p xtask -- check-docs.cargo fmt --all -- --checkreports existing formatting drift in untouched files.Closes #78
Implementation is available in PR #81 at
26c7e09d35108b3c602da26907151636aa7569fd. Local verification passed: workspace build, tests, Clippy, check-docs, and diff-check. Exact-head CI is pending and waiting to run; if it fails, move this card back to TODO.Not merged in this review pass. #82 (authenticate /events) and #79 (close SSE on broadcast lag) both edit the same
eventshandler and are now merged, so this branch conflicts incrates/breadboard-server/src/http.rs. Additionally, the newtests/shutdown.rsopens/eventswithout credentials and asserts success; after #82 that request returns 401, so the test also needs auth (e.g. a bearer header or the session cookie). Rebase on main, send auth in the SSE request, and this is good to merge.26c7e09d35707b758588