* pb: add multipart concurrency fields to RemoteConf and tier move requests
RemoteConf gains upload_concurrency/download_concurrency (0 = client
default); VolumeTierMoveDatToRemote/FromRemote requests gain a
concurrency field (0 = backend default).
* remote storage: honor RemoteConf upload/download concurrency in s3 and azure clients
s3 client: ReadFile passes conf download_concurrency to the downloader,
WriteFile uses upload_concurrency for the uploader; previously
hard-coded 1 upload / 5 download parts. 0 keeps defaults. Same for
azure client.
* storage: plumb concurrency through backend interface and tier upload/download
BackendStorage.CopyFile/DownloadFile take a concurrency hint (<=0 =
backend configured default); s3 backend reads
upload_concurrency/download_concurrency from scaffold config with
parseConcurrency fallback, rclone updated to the new signature. Tier
move gRPC handlers forward the request concurrency to the backend.
* shell: -upload_concurrency/-download_concurrency for remote.configure, -concurrent for volume.tier
remote.configure exposes upload/download concurrency persisted into
RemoteConf; volume.tier move/evict commands forward -concurrent to the
tier move requests. Documented in master-cloud.toml scaffold.
* test: cover concurrency propagation in remote tier integration test
* remote.configure: merge existing config on partial update
Load the stored RemoteConf before saving so a partial update (e.g. only
-upload_concurrency) preserves credentials, endpoints, and type instead
of replacing them with new-config defaults. Only treat a confirmed
ErrNotFound as a new configuration; propagate all other load errors so a
transient filer failure does not overwrite stored settings.
On a type transition, reset backend-specific fields to the destination
type's new-config defaults rather than inheriting the old backend's
empty values. Bound configured concurrency to a sane maximum.
* remote storage: honor configured download concurrency in S3 and Azure
ReadFileWithConcurrency now resolves a zero request override against the
client's configured download_concurrency (new downloadConcurrency()
helpers), so the remote-mount/cache read path honors
RemoteConf.DownloadConcurrency instead of the hard-coded default.
Azure also clamps the resolved value to math.MaxUint16 regardless of
whether the fallback was used, preventing uint16 wraparound when a
configured value exceeds 65535.
* shell: rename -concurrent to -concurrency and validate tier transfer bounds
Rename the -concurrent flag to -concurrency across volume.tier.upload,
volume.tier.download, and volume.tier.compact to match the proto field and
RemoteConf field names. Add validateTierConcurrency to reject values that
would wrap int32 or exceed a 1024 cap before constructing the request.
* server: clamp tier move concurrency in gRPC handlers
Add clampTierConcurrency to both VolumeTierMoveDatToRemote and
VolumeTierMoveDatFromRemote handlers so a direct gRPC caller cannot spawn
an unbounded number of network workers.
* trim verbose comments added with concurrency feature
Remove redundant doc comments on the backend interface, rclone backend,
s3_backend parseConcurrency, and test helpers that restated the obvious.
* remote.configure: apply type defaults before re-parse so explicit flags win
applyTypeDefaults ran after the second flag parse, overwriting explicit
destination flags (e.g. -s3.region=eu-west-1) with new-config defaults.
Move the type-transition default reset before the re-parse so user-supplied
flags override the destination defaults.
* remote.configure: only treat explicit -type as a type transition
The first parse defaults -type to s3, so a concurrency-only update on an
existing non-S3 config captured requestedType=s3 and wrongly triggered a
type transition, resetting the stored backend to S3. Use fs.Visit to
detect whether -type was explicitly supplied; an omitted -type keeps the
stored backend.
---------
Co-authored-by: Jack Meredith <9480542+jackusm@users.noreply.github.com>
Co-authored-by: Chris Lu <chris.lu@gmail.com>
* fix: report short S3 ReaderAt reads
Problem: S3BackendStorageFile.ReadAt returned a short buffer with a nil error, hiding truncated remote data.
Root cause: Every terminal io.EOF was cleared regardless of how many bytes were read.
Fix: Clear io.EOF only when the requested buffer was completely filled.
Validation: go test ./weed/storage/backend/...
Co-authored-by: Codex <noreply@openai.com>
* fix: validate S3 ReaderAt requests
Co-authored-by: Codex <noreply@openai.com>
* s3 ReadAt: simplify negative offset error
* s3 ReadAt: stop reading once the buffer is full
Skips the extra Read that only collected the terminal EOF, and bounds
the loop if the server ignores the Range header and returns more data
than requested, where Read on an empty slice can spin forever.
* s3 ReadAt: cover full reads that arrive with io.EOF
---------
Co-authored-by: Codex <noreply@openai.com>
Co-authored-by: Chris Lu <chris.lu@gmail.com>
* tiering: stop a shared remote object being deleted while replicas still point at it
A remote-tiered volume's .dat content lives only in one cloud object that all
N replica .vif files point at. Deleting that object while destroying any one
replica, or before a downloaded replica is durable, bricks the survivors.
- volume.tier.move cleanup now deletes old replicas with keepRemoteData=true so
surviving replicas keep the shared object. Document why the alreadyPlaced
anchor needs no replica sync (same-object replicas are byte-identical).
- VolumeTierMoveDatFromRemote now fsyncs the downloaded .dat, fsyncs the
containing directory, trims the .vif (fsynced) and swaps to the local DiskFile
BEFORE deleting the remote object, on both the keep-remote and delete paths.
Only the final DeleteFile is gated by keep_remote_dat_file, so a keep-remote
download leaves the replica served from local disk rather than the shared
object, and a crash before delete merely leaks the object.
- volume.tier.download keeps the shared object for every replica except the
last, which deletes it.
- s3 and rclone download paths fsync the .dat before close.
* storage: swap the volume data backend under the data lock
The tier-download swap closed v.DataBackend and assigned the new local DiskFile
without holding dataFileAccessLock, racing concurrent reads/writes (use of a
closed file / nil deref). Add an exported Volume.SwapDataBackend that performs
the close-and-replace under the lock, and call it from the tier download.
* server: skip directory fsync on Windows in the tier download path
os.Open(dir).Sync() is unsupported on Windows and returns an error, which would
fail VolumeTierMoveDatFromRemote entirely there. Skip the directory fsync on
Windows, matching how the storage-side helper tolerates the unsupported case.
* shell: make multi-replica tier.download resilient to already-local replicas
If a multi-replica download is interrupted and retried, a replica made local
in the prior attempt returns "already on local disk", which aborted the whole
command and left the remaining remote replicas dangling. Treat that case as a
skip-and-continue so a retry completes the rest.
* server: assert downloaded .dat content, not just length, in the tier test
A length-only check passes even if the bytes are corrupted; compare the full
content of the local .dat against the original.
* chore(weed/storage/backend/s3_backend): remove unused function
* fix(s3_backend): cache session under the composite region|endpoint key
createSession looked up sessions by region|endpoint but stored them by
region alone, so the cache never hit and a new session was built every
call. With getSession gone the lock can also drop to a plain Mutex.
---------
Co-authored-by: Chris Lu <chris.lu@gmail.com>
* Replace removeDuplicateSlashes with NormalizeObjectKey
Use s3_constants.NormalizeObjectKey instead of removeDuplicateSlashes in most places
for consistency. NormalizeObjectKey handles both duplicate slash removal and ensures
the path starts with '/', providing more complete normalization.
* Fix double slash issues after NormalizeObjectKey
After using NormalizeObjectKey, object keys have a leading '/'. This commit ensures:
- getVersionedObjectDir strips leading slash before concatenation
- getEntry calls receive names without leading slash
- String concatenation with '/' doesn't create '//' paths
This prevents path construction errors like:
/buckets/bucket//object (wrong)
/buckets/bucket/object (correct)
* ensure object key leading "/"
* fix compilation
* fix: Strip leading slash from object keys in S3 API responses
After introducing NormalizeObjectKey, all internal object keys have a
leading slash. However, S3 API responses must return keys without
leading slashes to match AWS S3 behavior.
Fixed in three functions:
- addVersion: Strip slash for version list entries
- processRegularFile: Strip slash for regular file entries
- processExplicitDirectory: Strip slash for directory entries
This ensures ListObjectVersions and similar APIs return keys like
'bar' instead of '/bar', matching S3 API specifications.
* fix: Normalize keyMarker for consistent pagination comparison
The S3 API provides keyMarker without a leading slash (e.g., 'object-001'),
but after introducing NormalizeObjectKey, all internal object keys have
leading slashes (e.g., '/object-001').
When comparing keyMarker < normalizedObjectKey in shouldSkipObjectForMarker,
the ASCII value of '/' (47) is less than 'o' (111), causing all objects
to be incorrectly skipped during pagination. This resulted in page 2 and
beyond returning 0 results.
Fix: Normalize the keyMarker when creating versionCollector so comparisons
work correctly with normalized object keys.
Fixes pagination tests:
- TestVersioningPaginationOver1000Versions
- TestVersioningPaginationMultipleObjectsManyVersions
* refactor: Change NormalizeObjectKey to return keys without leading slash
BREAKING STRATEGY CHANGE:
Previously, NormalizeObjectKey added a leading slash to all object keys,
which required stripping it when returning keys to S3 API clients and
caused complexity in marker normalization for pagination.
NEW STRATEGY:
- NormalizeObjectKey now returns keys WITHOUT leading slash (e.g., 'foo/bar' not '/foo/bar')
- This matches the S3 API format directly
- All path concatenations now explicitly add '/' between bucket and object
- No need to strip slashes in responses or normalize markers
Changes:
1. Modified NormalizeObjectKey to strip leading slash instead of adding it
2. Fixed all path concatenations to use:
- BucketsPath + '/' + bucket + '/' + object
instead of:
- BucketsPath + '/' + bucket + object
3. Reverted response key stripping in:
- addVersion()
- processRegularFile()
- processExplicitDirectory()
4. Reverted keyMarker normalization in findVersionsRecursively()
5. Updated matchesPrefixFilter() to work with keys without leading slash
6. Fixed paths in handlers:
- s3api_object_handlers.go (GetObject, HeadObject, cacheRemoteObjectForStreaming)
- s3api_object_handlers_postpolicy.go
- s3api_object_handlers_tagging.go
- s3api_object_handlers_acl.go
- s3api_version_id.go (getVersionedObjectDir, getVersionIdFormat)
- s3api_object_versioning.go (getObjectVersionList, updateLatestVersionAfterDeletion)
All versioning tests pass including pagination stress tests.
* adjust format
* Update post policy tests to match new NormalizeObjectKey behavior
- Update TestPostPolicyKeyNormalization to expect keys without leading slashes
- Update TestNormalizeObjectKey to expect keys without leading slashes
- Update TestPostPolicyFilenameSubstitution to expect keys without leading slashes
- Update path construction in tests to use new pattern: BucketsPath + '/' + bucket + '/' + object
* Fix ListObjectVersions prefix filtering
Remove leading slash addition to prefix parameter to allow correct filtering
of .versions directories when listing object versions with a specific prefix.
The prefix parameter should match entry paths relative to bucket root.
Adding a leading slash was breaking the prefix filter for paginated requests.
Fixes pagination issue where second page returned 0 versions instead of
continuing with remaining versions.
* no leading slash
* Fix urlEscapeObject to add leading slash for filer paths
NormalizeObjectKey now returns keys without leading slashes to match S3 API format.
However, urlEscapeObject is used for filer paths which require leading slashes.
Add leading slash back after normalization to ensure filer paths are correct.
Fixes TestS3ApiServer_toFilerPath test failures.
* adjust tests
* normalize
* Fix: Normalize prefixes and markers in LIST operations using NormalizeObjectKey
Ensure consistent key normalization across all S3 operations (GET, PUT, LIST).
Previously, LIST operations were not applying the same normalization rules
(handling backslashes, duplicate slashes, leading slashes) as GET/PUT operations.
Changes:
- Updated normalizePrefixMarker() to call NormalizeObjectKey for both prefix and marker
- This ensures prefixes with leading slashes, backslashes, or duplicate slashes are
handled consistently with how object keys are normalized
- Fixes Parquet test failures where pads.write_dataset creates implicit directory
structures that couldn't be discovered by subsequent LIST operations
- Added TestPrefixNormalizationInList and TestListPrefixConsistency tests
All existing LIST tests continue to pass with the normalization improvements.
* Add debugging logging to LIST operations to track prefix normalization
* Fix: Remove leading slash addition from GetPrefix to work with NormalizeObjectKey
The NormalizeObjectKey function removes leading slashes to match S3 API format
(e.g., 'foo/bar' not '/foo/bar'). However, GetPrefix was adding a leading slash
back, which caused LIST operations to fail with incorrect path handling.
Now GetPrefix only normalizes duplicate slashes without adding a leading slash,
which allows NormalizeObjectKey changes to work correctly for S3 LIST operations.
All Parquet integration tests now pass (20/20).
* Fix: Handle object paths without leading slash in checkDirectoryObject
NormalizeObjectKey() removes the leading slash to match S3 API format.
However, checkDirectoryObject() was assuming the object path has a leading
slash when processing directory markers (paths ending with '/').
Now we ensure the object has a leading slash before processing it for
filer operations.
Fixes implicit directory marker test (explicit_dir/) while keeping
Parquet integration tests passing (20/20).
All tests pass:
- Implicit directory tests: 6/6
- Parquet integration tests: 20/20
* Fix: Handle explicit directory markers with trailing slashes
Explicit directory markers created with put_object(Key='dir/', ...) are stored
in the filer with the trailing slash as part of the name. The checkDirectoryObject()
function now checks for both:
1. Explicit directories: lookup with trailing slash preserved (e.g., 'explicit_dir/')
2. Implicit directories: lookup without trailing slash (e.g., 'implicit_dir')
This ensures both types of directory markers are properly recognized.
All tests pass:
- Implicit directory tests: 6/6 (including explicit directory marker test)
- Parquet integration tests: 20/20
* Fix: Preserve trailing slash in NormalizeObjectKey
NormalizeObjectKey now preserves trailing slashes when normalizing object keys.
This is important for explicit directory markers like 'explicit_dir/' which rely
on the trailing slash to be recognized as directory objects.
The normalization process:
1. Notes if trailing slash was present
2. Removes duplicate slashes and converts backslashes
3. Removes leading slash for S3 API format
4. Restores trailing slash if it was in the original
This ensures explicit directory markers created with put_object(Key='dir/', ...)
are properly normalized and can be looked up by their exact name.
All tests pass:
- Implicit directory tests: 6/6
- Parquet integration tests: 20/20
* clean object
* Fix: Don't restore trailing slash if result is empty
When normalizing paths that are only slashes (e.g., '///', '/'), the function
should return an empty string, not a single slash. The fix ensures we only
restore the trailing slash if the result is non-empty.
This fixes the 'just_slashes' test case:
- Input: '///'
- Expected: ''
- Previous: '/'
- Fixed: ''
All tests now pass:
- Unit tests: TestNormalizeObjectKey (13/13)
- Implicit directory tests: 6/6
- Parquet integration tests: 20/20
* prefixEndsOnDelimiter
* Update s3api_object_handlers_list.go
* Update s3api_object_handlers_list.go
* handle create directory
Chris LuandGitHub
chaletandGitHub