fix(storage): harden scanner and recovery edge cases (#5521)

* fix(ecstore): handle benign listing and GET disconnects

Co-Authored-By: heihutu <heihutu@gmail.com>

* fix(scanner): scope cache locks by set

Co-Authored-By: heihutu <heihutu@gmail.com>

* test(kms): stabilize Vault transport retry coverage

Co-Authored-By: heihutu <heihutu@gmail.com>

* fix(scanner): fence scoped cache locks by protocol

Co-Authored-By: heihutu <heihutu@gmail.com>

* test(ecstore): stabilize topology DNS fallback coverage

Co-Authored-By: heihutu <heihutu@gmail.com>

* fix(kms): remove stale local export test import

Co-Authored-By: heihutu <heihutu@gmail.com>

---------

Co-authored-by: heihutu <heihutu@gmail.com>
This commit is contained in:
houseme
2026-08-01 11:25:13 +08:00
committed by GitHub
parent 1524ed891f
commit b965bd6eef
13 changed files with 504 additions and 135 deletions
+117 -10
View File
@@ -1645,15 +1645,28 @@ impl Erasure {
}
Err(e) => {
record_get_stage_duration_if_enabled(GET_OBJECT_PATH_LEGACY_DUPLEX, GET_STAGE_EMIT, emit_stage_start);
error!(
block_offset,
block_length,
bytes_written = *written,
stage = GET_STAGE_EMIT,
reason = classify_io_error(&e).as_str(),
error = ?e,
"Erasure decode failed to emit reconstructed data"
);
let reason = classify_io_error(&e);
if reason == GetObjectFailureReason::DownstreamClosed {
debug!(
block_offset,
block_length,
bytes_written = *written,
stage = GET_STAGE_EMIT,
reason = reason.as_str(),
error = ?e,
"Erasure decode stopped after downstream closed"
);
} else {
error!(
block_offset,
block_length,
bytes_written = *written,
stage = GET_STAGE_EMIT,
reason = reason.as_str(),
error = ?e,
"Erasure decode failed to emit reconstructed data"
);
}
*ret_err = Some(e);
return StripeFlow::Stop;
}
@@ -1945,7 +1958,7 @@ mod tests {
use std::io::Cursor;
use std::pin::Pin;
use std::sync::{
Arc,
Arc, Mutex,
atomic::{AtomicUsize, Ordering},
};
use std::task::{Context, Poll};
@@ -2120,6 +2133,59 @@ mod tests {
}
}
struct DownstreamClosedWriter;
impl AsyncWrite for DownstreamClosedWriter {
fn poll_write(self: Pin<&mut Self>, _cx: &mut Context<'_>, _buf: &[u8]) -> Poll<io::Result<usize>> {
Poll::Ready(Err(crate::diagnostics::get::mark_get_object_downstream_closed(io::Error::new(
ErrorKind::BrokenPipe,
"injected downstream close",
))))
}
fn poll_flush(self: Pin<&mut Self>, _cx: &mut Context<'_>) -> Poll<io::Result<()>> {
Poll::Ready(Ok(()))
}
fn poll_shutdown(self: Pin<&mut Self>, _cx: &mut Context<'_>) -> Poll<io::Result<()>> {
Poll::Ready(Ok(()))
}
}
#[derive(Clone, Default)]
struct CapturedLogs(Arc<Mutex<Vec<u8>>>);
struct CapturedLogWriter(Arc<Mutex<Vec<u8>>>);
impl CapturedLogs {
fn contents(&self) -> String {
String::from_utf8(self.0.lock().expect("captured logs mutex should not be poisoned").clone())
.expect("captured logs should be valid UTF-8")
}
}
impl std::io::Write for CapturedLogWriter {
fn write(&mut self, buf: &[u8]) -> std::io::Result<usize> {
self.0
.lock()
.expect("captured logs mutex should not be poisoned")
.extend_from_slice(buf);
Ok(buf.len())
}
fn flush(&mut self) -> std::io::Result<()> {
Ok(())
}
}
impl<'a> tracing_subscriber::fmt::MakeWriter<'a> for CapturedLogs {
type Writer = CapturedLogWriter;
fn make_writer(&'a self) -> Self::Writer {
CapturedLogWriter(Arc::clone(&self.0))
}
}
#[test]
fn parallel_reader_constructor_variants_preserve_read_cost_and_verification_flags() {
let erasure = Erasure::new(2, 1, 64);
@@ -2215,6 +2281,47 @@ mod tests {
assert_eq!(err.to_string(), "injected emit failure");
}
#[tokio::test(flavor = "current_thread")]
async fn erasure_decode_logs_reconstructed_downstream_close_at_debug() {
let logs = CapturedLogs::default();
let subscriber = tracing_subscriber::fmt()
.with_max_level(tracing::Level::DEBUG)
.with_writer(logs.clone())
.with_ansi(false)
.without_time()
.finish();
let _guard = tracing::subscriber::set_default(subscriber);
let erasure = Erasure::new(2, 1, 64);
let data: Vec<u8> = (0..64).collect();
let shard_size = erasure.shard_size();
let encoded = erasure.encode_data(&data).expect("test data should encode");
let readers = vec![
None,
Some(BitrotReader::new(
Cursor::new(encoded[1].to_vec()),
shard_size,
HashAlgorithm::None,
false,
)),
Some(BitrotReader::new(
Cursor::new(encoded[2].to_vec()),
shard_size,
HashAlgorithm::None,
false,
)),
];
let mut writer = DownstreamClosedWriter;
let (written, err) = erasure.decode(&mut writer, readers, 0, data.len(), data.len()).await;
assert_eq!(written, 0);
assert_eq!(err.expect("downstream close must still terminate the GET").kind(), ErrorKind::BrokenPipe);
let captured = logs.contents();
assert!(captured.contains("Erasure decode stopped after downstream closed"));
assert!(!captured.contains("Erasure decode failed to emit reconstructed data"));
}
#[tokio::test]
async fn test_erasure_decode_rejects_reader_count_and_range_overflow() {
let erasure = Erasure::new(2, 1, 64);