CritasWang opened a new pull request, #11:
URL: https://github.com/apache/iotdb-client-rust/pull/11
## Summary
One combined PR for the session-pool / connection-lifecycle / redirect
review findings, split into
issues #4–#10 for tracking. The root cause (#4) is fixed by a new
socket-level I/O timeout, and the
other fixes are the follow-on work each issue describes.
## What changed
- **#4 — client-side time bound after the TCP handshake.** New
`socket_timeout` on
`ConnectionOptions`/`SessionConfig` (`Some(60s)` by default, `None`
restores the old
behaviour, zero treated as `None`). It is applied via `socket2` right
after connect and
**before** the TLS handshake, so the TLS handshake, every RPC read, and
the best-effort
`closeSession` on `Drop` are all bounded. `SO_KEEPALIVE` is now enabled on
the socket.
Wedged reads surface as `Error::Thrift`, so `with_retry`/reconnect
actually engages.
- **#5 — handshake-level failover in `Session::open`.** The endpoint loop
now retries
connect + `openSession` + `requestStatementId` together, mirroring
`reconnect()`.
- **#6 — pool accounting.** `acquire()` spends its `acquire_timeout` budget
on growth /
hand-out failures instead of failing instantly; `live` is now decremented
under the state lock
(no lost Condvar wakeups); `close()` wakes waiters before the blocking
`closeSession` calls;
`acquire_timeout = Duration::MAX` means "wait without a deadline" via
`checked_add`.
- **#7 — per-endpoint connect budget.** `connect_stream` shares one total
`connect_timeout`
across every resolved address of an endpoint.
- **#8 — desynchronized connections.** A transport-level failure (including
frame-too-large
rejection and socket timeouts, and `fetch_results` which cannot be
retried) marks the
connection broken; `is_open()` then reports false and pools discard the
session. Note: the
16,384,000-byte frame cap itself cannot be raised without a thrift 0.23
API change
(`TFramedReadTransport` exposes no config setter), so this PR fixes the
desync/reuse half.
- **#9 — redirect/pool refinements.** Endpoint matching now normalizes
case/brackets/whitespace
and treats loopback spellings as equivalent (general hostname-vs-IP DNS
resolution is deliberately
not done on the acquire path); the newest hint wins when idle sessions
hold conflicting hints
(the cache insertion seq is now process-wide); `SessionConfig` gains
`redirect_cache_ttl` / `redirect_cache_max_entries`;
`TableSessionPool::acquire_for_device`
is exposed.
- **#10 — reconnect slot hold.** Pooled sessions skip the between-attempt
pacing sleeps during
reconnect, so a pool slot is no longer held for the full C#-style
reconnect walk.
## Tests
New regression tests use fake/silent listeners:
- `connection::tests::socket_timeout_bounds_reads_after_handshake`
- `connection::tls_tests::tls_handshake_times_out_against_silent_peer`
-
`session::tests::open_fails_over_to_next_endpoint_when_authentication_fails`
(live-server gated)
- `no_reconnect_when_disabled` now asserts the broken-session behaviour
- `pool::tests::growth_failure_spends_acquire_timeout_budget`,
`close_wakes_waiters_promptly`,
`acquire_timeout_max_waits_until_closed_without_panicking`,
`acquire_for_device_prefers_newest_conflicting_hint`
- `connection::tests::endpoint_equivalent_normalizes_and_matches_loopback`
Verified locally: `cargo fmt --check`, `./tools/check-license.sh`,
`cargo clippy --all-targets -- -D warnings` (default and `--features tls`),
`cargo test` 120 passed, `cargo test --features tls` 132 passed
(live-server tests ran against a local IoTDB instance).
Closes #4, closes #5, closes #6, closes #7, closes #8, closes #9, closes #10.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
