fix(replication): forward single-part object checksums as headers (#7313)

This commit is contained in:
唐小鸭
2026-09-07 09:17:24 +08:00
committed by GitHub
parent ff00872922
commit 3b404e56c0
5 changed files with 282 additions and 20 deletions
@@ -3988,6 +3988,93 @@ async fn test_bucket_replication_version_purge_of_non_inline_object_releases_sou
Ok(())
}
/// Regression for rustfs/backlog#2340 (not Wasabi specific): a single-part
/// object uploaded with `x-amz-checksum-*` must reach the target with the same
/// checksum. The outbound options keyed the stored record by algorithm name,
/// which the target client sent as `x-amz-meta-*` user metadata, so a replica
/// never carried a checksum although the source HEAD returned one.
#[tokio::test]
async fn test_bucket_replication_forwards_single_part_object_checksums() -> TestResult {
init_logging();
let mut source_env = RustFSTestEnvironment::new().await?;
let mut source_env_vars = replication_fast_env();
source_env_vars.extend_from_slice(LOOPBACK_REPLICATION_TARGET_ENV);
source_env.start_rustfs_server_with_env(vec![], &source_env_vars).await?;
let mut target_env = RustFSTestEnvironment::new().await?;
target_env.start_rustfs_server_without_cleanup(vec![]).await?;
let source_bucket = "replication-checksum-src";
let target_bucket = "replication-checksum-dst";
let source_client = source_env.create_s3_client();
let target_client = target_env.create_s3_client();
source_client.create_bucket().bucket(source_bucket).send().await?;
target_client.create_bucket().bucket(target_bucket).send().await?;
enable_bucket_versioning(&source_env, source_bucket).await?;
enable_bucket_versioning(&target_env, target_bucket).await?;
let target_arn = set_replication_target(&source_env, source_bucket, &target_env, target_bucket).await?;
put_bucket_replication(&source_env, source_bucket, &target_arn).await?;
let body = b"123456789";
let crc32_key = "checksum-crc32.txt";
let sha256_key = "checksum-sha256.txt";
let crc32_put = source_client
.put_object()
.bucket(source_bucket)
.key(crc32_key)
.body(ByteStream::from_static(body))
.checksum_algorithm(aws_sdk_s3::types::ChecksumAlgorithm::Crc32)
.send()
.await?;
let expected_crc32 = crc32_put.checksum_crc32().ok_or("source PUT omitted CRC32")?.to_string();
let sha256_put = source_client
.put_object()
.bucket(source_bucket)
.key(sha256_key)
.body(ByteStream::from_static(body))
.checksum_algorithm(aws_sdk_s3::types::ChecksumAlgorithm::Sha256)
.send()
.await?;
let expected_sha256 = sha256_put.checksum_sha256().ok_or("source PUT omitted SHA256")?.to_string();
for key in [crc32_key, sha256_key] {
wait_for_source_replication_status(&source_client, source_bucket, key, "COMPLETED", false).await?;
}
let replica = target_client
.head_object()
.bucket(target_bucket)
.key(crc32_key)
.checksum_mode(aws_sdk_s3::types::ChecksumMode::Enabled)
.send()
.await?;
assert_eq!(replica.checksum_crc32(), Some(expected_crc32.as_str()), "replica lost the CRC32 checksum");
let replica = target_client
.head_object()
.bucket(target_bucket)
.key(sha256_key)
.checksum_mode(aws_sdk_s3::types::ChecksumMode::Enabled)
.send()
.await?;
assert_eq!(
replica.checksum_sha256(),
Some(expected_sha256.as_str()),
"replica lost the SHA256 checksum"
);
// The bare algorithm name must not leak as user metadata either.
assert!(
replica
.metadata()
.is_none_or(|meta| !meta.keys().any(|k| k.eq_ignore_ascii_case("sha256"))),
"replica carries the checksum as user metadata: {:?}",
replica.metadata()
);
Ok(())
}
#[tokio::test]
async fn test_bucket_replication_disabled_delete_marker_does_not_propagate() -> TestResult {
init_logging();
@@ -42,7 +42,8 @@ use crate::replication_extension_test::{
use aws_sdk_s3::Client;
use aws_sdk_s3::primitives::{ByteStream, DateTime};
use aws_sdk_s3::types::{
Checksum, CompletedMultipartUpload, CompletedPart, ObjectAttributes, ObjectLockLegalHoldStatus, ObjectLockMode,
Checksum, ChecksumAlgorithm, CompletedMultipartUpload, CompletedPart, ObjectAttributes, ObjectLockLegalHoldStatus,
ObjectLockMode,
};
use bytes::Bytes;
use std::error::Error;
@@ -114,10 +115,13 @@ enum ObjectShape {
LockedMultipart,
/// ODM stores two local parts while preserving a single-PUT source's MD5 ETag.
OdmPreservedMd5Multipart,
/// Single-part object uploaded with `x-amz-checksum-sha256`; the replica
/// must carry the same header (rustfs/backlog#2340).
Checksummed,
}
impl ObjectShape {
const ALL: [ObjectShape; 7] = [
const ALL: [ObjectShape; 8] = [
ObjectShape::Empty,
ObjectShape::Plain,
ObjectShape::Retention,
@@ -125,6 +129,7 @@ impl ObjectShape {
ObjectShape::Multipart,
ObjectShape::LockedMultipart,
ObjectShape::OdmPreservedMd5Multipart,
ObjectShape::Checksummed,
];
fn key(self) -> &'static str {
@@ -136,6 +141,16 @@ impl ObjectShape {
ObjectShape::Multipart => "matrix/multipart.bin",
ObjectShape::LockedMultipart => "matrix/locked-multipart.bin",
ObjectShape::OdmPreservedMd5Multipart => "matrix/odm-preserved-md5.bin",
ObjectShape::Checksummed => "matrix/checksummed.bin",
}
}
/// The `x-amz-checksum-*` header the source stored and every upload of
/// the replica must repeat.
fn forwarded_checksum_header(self) -> Option<&'static str> {
match self {
ObjectShape::Checksummed => Some("x-amz-checksum-sha256"),
_ => None,
}
}
@@ -198,6 +213,18 @@ impl ObjectShape {
ObjectShape::Multipart => multipart_put(client, bucket, key, 0x44, false).await,
ObjectShape::LockedMultipart => multipart_put(client, bucket, key, 0x55, true).await,
ObjectShape::OdmPreservedMd5Multipart => odm_preserved_md5_multipart(env, bucket, key).await,
ObjectShape::Checksummed => {
let body = payload(40 * 1024, 0x66);
client
.put_object()
.bucket(bucket)
.key(key)
.body(ByteStream::from(body.clone()))
.checksum_algorithm(ChecksumAlgorithm::Sha256)
.send()
.await?;
Ok(body)
}
}
}
}
@@ -614,6 +641,19 @@ async fn check_completed_cell(
}) {
return Err(format!("a locked PutObject went out without any integrity header (rustfs#7082): {bare:?}").into());
}
// rustfs/backlog#2340 contract: a source checksum reaches the target as
// the `x-amz-checksum-*` header, not as user metadata; every PutObject of
// the shape carries it.
if let Some(header) = shape.forwarded_checksum_header()
&& let Some(missing) = uploads.iter().find(|record| {
record.operation == FakeTargetOperation::PutObject
&& !record.transport.checksum_headers.iter().any(|name| name == header)
})
{
return Err(
format!("a PutObject went out without the source's {header} header (rustfs/backlog#2340): {missing:?}").into(),
);
}
Ok(())
}