seaweedfs

12886 Commits

94 Branches

327 Tags

338 MiB

Tree: 49a64f50f1

Commit Graph

Author	SHA1	Message	Date
Chris Lu	c284e51d20	fix: multipart upload ETag calculation (#8238 ) * fix multipart etag * address comments * clean up * clean up * optimization * address comments * unquoted etag * dedup * upgrade * clean * etag * return quoted tag * quoted etag * debug * s3api: unify ETag retrieval and quoting across handlers Refactor newListEntry to take S3ApiServer and use getObjectETag, and update setResponseHeaders to use the same logic. This ensures consistent ETags are returned for both listing and direct access. s3api: implement ListObjects deduplication for versioned buckets Handle duplicate entries between the main path and the .versions directory by prioritizing the latest version when bucket versioning is enabled. * s3api: cleanup stale main file entries during versioned uploads Add explicit deletion of pre-existing "main" files when creating new versions in versioned buckets. This prevents stale entries from appearing in bucket listings and ensures consistency. * s3api: fix cleanup code placement in versioned uploads Correct the placement of rm calls in completeMultipartUpload and putVersionedObject to ensure stale main files are properly deleted during versioned uploads. * s3api: improve getObjectETag fallback for empty ExtETagKey Ensure that when ExtETagKey exists but contains an empty value, the function falls through to MD5/chunk-based calculation instead of returning an empty string. * s3api: fix test files for new newListEntry signature Update test files to use the new newListEntry signature where the first parameter is S3ApiServer. Created mockS3ApiServer to properly test owner display name lookup functionality. s3api: use filer.ETag for consistent Md5 handling in getEtagFromEntry Change getEtagFromEntry fallback to use filer.ETag(entry) instead of filer.ETagChunks to ensure legacy entries with Attributes.Md5 are handled consistently with the rest of the codebase. * s3api: optimize list logic and fix conditional header logging - Hoist bucket versioning check out of per-entry callback to avoid repeated getVersioningState calls - Extract appendOrDedup helper function to eliminate duplicate dedup/append logic across multiple code paths - Change If-Match mismatch logging from glog.Errorf to glog.V(3).Infof and remove DEBUG prefix for consistency * s3api: fix test mock to properly initialize IAM accounts Fixed nil pointer dereference in TestNewListEntryOwnerDisplayName by directly initializing the IdentityAccessManagement.accounts map in the test setup. This ensures newListEntry can properly look up account display names without panicking. * cleanup * s3api: remove premature main file cleanup in versioned uploads Removed incorrect cleanup logic that was deleting main files during versioned uploads. This was causing test failures because it deleted objects that should have been preserved as null versions when versioning was first enabled. The deduplication logic in listing is sufficient to handle duplicate entries without deleting files during upload. * s3api: add empty-value guard to getEtagFromEntry Added the same empty-value guard used in getObjectETag to prevent returning quoted empty strings. When ExtETagKey exists but is empty, the function now falls through to filer.ETag calculation instead of returning "". * s3api: fix listing of directory key objects with matching prefix Revert prefix handling logic to use strings.TrimPrefix instead of checking HasPrefix with empty string result. This ensures that when a directory key object exactly matches the prefix (e.g. prefix="dir/", object="dir/"), it is correctly handled as a regular entry instead of being skipped or incorrectly processed as a common prefix. Also fixed missing variable definition. * s3api: refactor list inline dedup to use appendOrDedup helper Refactored the inline deduplication logic in listFilerEntries to use the shared appendOrDedup helper function. This ensures consistent behavior and reduces code duplication. * test: fix port allocation race in s3tables integration test Updated startMiniCluster to find all required ports simultaneously using findAvailablePorts instead of sequentially. This prevents race conditions where the OS reallocates a port that was just released, causing multiple services (e.g. Filer and Volume) to be assigned the same port and fail to start.	2 weeks ago
Chris Lu	b244bb58aa	s3tables: redesign Iceberg REST Catalog using iceberg-go and automate integration tests (#8197 ) * full integration with iceberg-go * Table Commit Operations (handleUpdateTable) * s3tables: fix Iceberg v2 compliance and namespace properties This commit ensures SeaweedFS Iceberg REST Catalog is compliant with Iceberg Format Version 2 by: - Using iceberg-go's table.NewMetadataWithUUID for strict v2 compliance. - Explicitly initializing namespace properties to empty maps. - Removing omitempty from required Iceberg response fields. - Fixing CommitTableRequest unmarshaling using table.Requirements and table.Updates. * s3tables: automate Iceberg integration tests - Added Makefile for local test execution and cluster management. - Added docker-compose for PyIceberg compatibility kit. - Added Go integration test harness for PyIceberg. - Updated GitHub CI to run Iceberg catalog tests automatically. * s3tables: update PyIceberg test suite for compatibility - Updated test_rest_catalog.py to use latest PyIceberg transaction APIs. - Updated Dockerfile to include pyarrow and pandas dependencies. - Improved namespace and table handling in integration tests. * s3tables: address review feedback on Iceberg Catalog - Implemented robust metadata version parsing and incrementing. - Ensured table metadata changes are persisted during commit (handleUpdateTable). - Standardized namespace property initialization for consistency. - Fixed unused variable and incorrect struct field build errors. * s3tables: finalize Iceberg REST Catalog and optimize tests - Implemented robust metadata versioning and persistence. - Standardized namespace property initialization. - Optimized integration tests using pre-built Docker image. - Added strict property persistence validation to test suite. - Fixed build errors from previous partial updates. * Address PR review: fix Table UUID stability, implement S3Tables UpdateTable, and support full metadata persistence individually * fix: Iceberg catalog stable UUIDs, metadata persistence, and file writing - Ensure table UUIDs are stable (do not regenerate on load). - Persist full table metadata (Iceberg JSON) in s3tables extended attributes. - Add `MetadataVersion` to explicitly track version numbers, replacing regex parsing. - Implement `saveMetadataFile` to persist metadata JSON files to the Filer on commit. - Update `CreateTable` and `UpdateTable` handlers to use the new logic. * test: bind weed mini to 0.0.0.0 in integration tests to fix Docker connectivity * Iceberg: fix metadata handling in REST catalog - Add nil guard in createTable - Fix updateTable to correctly load existing metadata from storage - Ensure full metadata persistence on updates - Populate loadTable result with parsed metadata * S3Tables: add auth checks and fix response fields in UpdateTable - Add CheckPermissionWithContext to UpdateTable handler - Include TableARN and MetadataLocation in UpdateTable response - Use ErrCodeConflict (409) for version token mismatches * Tests: improve Iceberg catalog test infrastructure and cleanup - Makefile: use PID file for precise process killing - test_rest_catalog.py: remove unused variables and fix f-strings * Iceberg: fix variable shadowing in UpdateTable - Rename inner loop variable `req` to `requirement` to avoid shadowing outer request variable * S3Tables: simplify MetadataVersion initialization - Use `max(req.MetadataVersion, 1)` instead of anonymous function * Tests: remove unicode characters from S3 tables integration test logs - Remove unicode checkmarks from test output for cleaner logs * Iceberg: improve metadata persistence robustness - Fix MetadataLocation in LoadTableResult to fallback to generated location - Improve saveMetadataFile to ensure directory hierarchy existence and robust error handling	3 weeks ago
Chris Lu	2bb21ea276	feat: Add Iceberg REST Catalog server and admin UI (#8175 ) * feat: Add Iceberg REST Catalog server Implement Iceberg REST Catalog API on a separate port (default 8181) that exposes S3 Tables metadata through the Apache Iceberg REST protocol. - Add new weed/s3api/iceberg package with REST handlers - Implement /v1/config endpoint returning catalog configuration - Implement namespace endpoints (list/create/get/head/delete) - Implement table endpoints (list/create/load/head/delete/update) - Add -port.iceberg flag to S3 standalone server (s3.go) - Add -s3.port.iceberg flag to combined server mode (server.go) - Add -s3.port.iceberg flag to mini cluster mode (mini.go) - Support prefix-based routing for multiple catalogs The Iceberg REST server reuses S3 Tables metadata storage under /table-buckets and enables DuckDB, Spark, and other Iceberg clients to connect to SeaweedFS as a catalog. * feat: Add Iceberg Catalog pages to admin UI Add admin UI pages to browse Iceberg catalogs, namespaces, and tables. - Add Iceberg Catalog menu item under Object Store navigation - Create iceberg_catalog.templ showing catalog overview with REST info - Create iceberg_namespaces.templ listing namespaces in a catalog - Create iceberg_tables.templ listing tables in a namespace - Add handlers and routes in admin_handlers.go - Add Iceberg data provider methods in s3tables_management.go - Add Iceberg data types in types.go The Iceberg Catalog pages provide visibility into the same S3 Tables data through an Iceberg-centric lens, including REST endpoint examples for DuckDB and PyIceberg. * test: Add Iceberg catalog integration tests and reorg s3tables tests - Reorganize existing s3tables tests to test/s3tables/table-buckets/ - Add new test/s3tables/catalog/ for Iceberg REST catalog tests - Add TestIcebergConfig to verify /v1/config endpoint - Add TestIcebergNamespaces to verify namespace listing - Add TestDuckDBIntegration for DuckDB connectivity (requires Docker) - Update CI workflow to use new test paths * fix: Generate proper random UUIDs for Iceberg tables Address code review feedback: - Replace placeholder UUID with crypto/rand-based UUID v4 generation - Add detailed TODO comments for handleUpdateTable stub explaining the required atomic metadata swap implementation * fix: Serve Iceberg on localhost listener when binding to different interface Address code review feedback: properly serve the localhost listener when the Iceberg server is bound to a non-localhost interface. * ci: Add Iceberg catalog integration tests to CI Add new job to run Iceberg catalog tests in CI, along with: - Iceberg package build verification - Iceberg unit tests - Iceberg go vet checks - Iceberg format checks * fix: Address code review feedback for Iceberg implementation - fix: Replace hardcoded account ID with s3_constants.AccountAdminId in buildTableBucketARN() - fix: Improve UUID generation error handling with deterministic fallback (timestamp + PID + counter) - fix: Update handleUpdateTable to return HTTP 501 Not Implemented instead of fake success - fix: Better error handling in handleNamespaceExists to distinguish 404 from 500 errors - fix: Use relative URL in template instead of hardcoded localhost:8181 - fix: Add HTTP timeout to test's waitForService function to avoid hangs - fix: Use dynamic ephemeral ports in integration tests to avoid flaky parallel failures - fix: Add Iceberg port to final port configuration logging in mini.go * fix: Address critical issues in Iceberg implementation - fix: Cache table UUIDs to ensure persistence across LoadTable calls The UUID now remains stable for the lifetime of the server session. TODO: For production, UUIDs should be persisted in S3 Tables metadata. - fix: Remove redundant URL-encoded namespace parsing mux router already decodes %1F to \x1F before passing to handlers. Redundant ReplaceAll call could cause bugs with literal %1F in namespace. * fix: Improve test robustness and reduce code duplication - fix: Make DuckDB test more robust by failing on unexpected errors Instead of silently logging errors, now explicitly check for expected conditions (extension not available) and skip the test appropriately. - fix: Extract username helper method to reduce duplication Created getUsername() helper in AdminHandlers to avoid duplicating the username retrieval logic across Iceberg page handlers. * fix: Add mutex protection to table UUID cache Protects concurrent access to the tableUUIDs map with sync.RWMutex. Uses read-lock for fast path when UUID already cached, and write-lock for generating new UUIDs. Includes double-check pattern to handle race condition between read-unlock and write-lock. * style: fix go fmt errors * feat(iceberg): persist table UUID in S3 Tables metadata * feat(admin): configure Iceberg port in Admin UI and commands * refactor: address review comments (flags, tests, handlers) - command/mini: fix tracking of explicit s3.port.iceberg flag - command/admin: add explicit -iceberg.port flag - admin/handlers: reuse getUsername helper - tests: use 127.0.0.1 for ephemeral ports and os.Stat for file size check * test: check error from FileStat in verify_gc_empty_test	3 weeks ago

Author

SHA1

Message

Date

Chris Lu

c284e51d20

fix: multipart upload ETag calculation (#8238 )

* fix multipart etag

* address comments

* clean up

* clean up

* optimization

* address comments

* unquoted etag

* dedup

* upgrade

* clean

* etag

* return quoted tag

* quoted etag

* debug

* s3api: unify ETag retrieval and quoting across handlers

Refactor newListEntry to take *S3ApiServer and use getObjectETag,
and update setResponseHeaders to use the same logic. This ensures
consistent ETags are returned for both listing and direct access.

* s3api: implement ListObjects deduplication for versioned buckets

Handle duplicate entries between the main path and the .versions
directory by prioritizing the latest version when bucket versioning
is enabled.

* s3api: cleanup stale main file entries during versioned uploads

Add explicit deletion of pre-existing "main" files when creating new
versions in versioned buckets. This prevents stale entries from
appearing in bucket listings and ensures consistency.

* s3api: fix cleanup code placement in versioned uploads

Correct the placement of rm calls in completeMultipartUpload and
putVersionedObject to ensure stale main files are properly deleted
during versioned uploads.

* s3api: improve getObjectETag fallback for empty ExtETagKey

Ensure that when ExtETagKey exists but contains an empty value,
the function falls through to MD5/chunk-based calculation instead
of returning an empty string.

* s3api: fix test files for new newListEntry signature

Update test files to use the new newListEntry signature where the
first parameter is *S3ApiServer. Created mockS3ApiServer to properly
test owner display name lookup functionality.

* s3api: use filer.ETag for consistent Md5 handling in getEtagFromEntry

Change getEtagFromEntry fallback to use filer.ETag(entry) instead of
filer.ETagChunks to ensure legacy entries with Attributes.Md5 are
handled consistently with the rest of the codebase.

* s3api: optimize list logic and fix conditional header logging

- Hoist bucket versioning check out of per-entry callback to avoid
  repeated getVersioningState calls
- Extract appendOrDedup helper function to eliminate duplicate
  dedup/append logic across multiple code paths
- Change If-Match mismatch logging from glog.Errorf to glog.V(3).Infof
  and remove DEBUG prefix for consistency

* s3api: fix test mock to properly initialize IAM accounts

Fixed nil pointer dereference in TestNewListEntryOwnerDisplayName by
directly initializing the IdentityAccessManagement.accounts map in the
test setup. This ensures newListEntry can properly look up account
display names without panicking.

* cleanup

* s3api: remove premature main file cleanup in versioned uploads

Removed incorrect cleanup logic that was deleting main files during
versioned uploads. This was causing test failures because it deleted
objects that should have been preserved as null versions when
versioning was first enabled. The deduplication logic in listing is
sufficient to handle duplicate entries without deleting files during
upload.

* s3api: add empty-value guard to getEtagFromEntry

Added the same empty-value guard used in getObjectETag to prevent
returning quoted empty strings. When ExtETagKey exists but is empty,
the function now falls through to filer.ETag calculation instead of
returning "".

* s3api: fix listing of directory key objects with matching prefix

Revert prefix handling logic to use strings.TrimPrefix instead of
checking HasPrefix with empty string result. This ensures that when a
directory key object exactly matches the prefix (e.g. prefix="dir/",
object="dir/"), it is correctly handled as a regular entry instead of
being skipped or incorrectly processed as a common prefix. Also fixed
missing variable definition.

* s3api: refactor list inline dedup to use appendOrDedup helper

Refactored the inline deduplication logic in listFilerEntries to use the
shared appendOrDedup helper function. This ensures consistent behavior
and reduces code duplication.

* test: fix port allocation race in s3tables integration test

Updated startMiniCluster to find all required ports simultaneously using
findAvailablePorts instead of sequentially. This prevents race conditions
where the OS reallocates a port that was just released, causing multiple
services (e.g. Filer and Volume) to be assigned the same port and fail
to start.

2 weeks ago

Chris Lu

b244bb58aa

s3tables: redesign Iceberg REST Catalog using iceberg-go and automate integration tests (#8197 )

* full integration with iceberg-go

* Table Commit Operations (handleUpdateTable)

* s3tables: fix Iceberg v2 compliance and namespace properties

This commit ensures SeaweedFS Iceberg REST Catalog is compliant with
Iceberg Format Version 2 by:
- Using iceberg-go's table.NewMetadataWithUUID for strict v2 compliance.
- Explicitly initializing namespace properties to empty maps.
- Removing omitempty from required Iceberg response fields.
- Fixing CommitTableRequest unmarshaling using table.Requirements and table.Updates.

* s3tables: automate Iceberg integration tests

- Added Makefile for local test execution and cluster management.
- Added docker-compose for PyIceberg compatibility kit.
- Added Go integration test harness for PyIceberg.
- Updated GitHub CI to run Iceberg catalog tests automatically.

* s3tables: update PyIceberg test suite for compatibility

- Updated test_rest_catalog.py to use latest PyIceberg transaction APIs.
- Updated Dockerfile to include pyarrow and pandas dependencies.
- Improved namespace and table handling in integration tests.

* s3tables: address review feedback on Iceberg Catalog

- Implemented robust metadata version parsing and incrementing.
- Ensured table metadata changes are persisted during commit (handleUpdateTable).
- Standardized namespace property initialization for consistency.
- Fixed unused variable and incorrect struct field build errors.

* s3tables: finalize Iceberg REST Catalog and optimize tests

- Implemented robust metadata versioning and persistence.
- Standardized namespace property initialization.
- Optimized integration tests using pre-built Docker image.
- Added strict property persistence validation to test suite.
- Fixed build errors from previous partial updates.

* Address PR review: fix Table UUID stability, implement S3Tables UpdateTable, and support full metadata persistence individually

* fix: Iceberg catalog stable UUIDs, metadata persistence, and file writing

- Ensure table UUIDs are stable (do not regenerate on load).
- Persist full table metadata (Iceberg JSON) in s3tables extended attributes.
- Add `MetadataVersion` to explicitly track version numbers, replacing regex parsing.
- Implement `saveMetadataFile` to persist metadata JSON files to the Filer on commit.
- Update `CreateTable` and `UpdateTable` handlers to use the new logic.

* test: bind weed mini to 0.0.0.0 in integration tests to fix Docker connectivity

* Iceberg: fix metadata handling in REST catalog

- Add nil guard in createTable
- Fix updateTable to correctly load existing metadata from storage
- Ensure full metadata persistence on updates
- Populate loadTable result with parsed metadata

* S3Tables: add auth checks and fix response fields in UpdateTable

- Add CheckPermissionWithContext to UpdateTable handler
- Include TableARN and MetadataLocation in UpdateTable response
- Use ErrCodeConflict (409) for version token mismatches

* Tests: improve Iceberg catalog test infrastructure and cleanup

- Makefile: use PID file for precise process killing
- test_rest_catalog.py: remove unused variables and fix f-strings

* Iceberg: fix variable shadowing in UpdateTable

- Rename inner loop variable `req` to `requirement` to avoid shadowing outer request variable

* S3Tables: simplify MetadataVersion initialization

- Use `max(req.MetadataVersion, 1)` instead of anonymous function

* Tests: remove unicode characters from S3 tables integration test logs

- Remove unicode checkmarks from test output for cleaner logs

* Iceberg: improve metadata persistence robustness

- Fix MetadataLocation in LoadTableResult to fallback to generated location
- Improve saveMetadataFile to ensure directory hierarchy existence and robust error handling

3 weeks ago

Chris Lu

2bb21ea276

feat: Add Iceberg REST Catalog server and admin UI (#8175 )

* feat: Add Iceberg REST Catalog server

Implement Iceberg REST Catalog API on a separate port (default 8181)
that exposes S3 Tables metadata through the Apache Iceberg REST protocol.

- Add new weed/s3api/iceberg package with REST handlers
- Implement /v1/config endpoint returning catalog configuration
- Implement namespace endpoints (list/create/get/head/delete)
- Implement table endpoints (list/create/load/head/delete/update)
- Add -port.iceberg flag to S3 standalone server (s3.go)
- Add -s3.port.iceberg flag to combined server mode (server.go)
- Add -s3.port.iceberg flag to mini cluster mode (mini.go)
- Support prefix-based routing for multiple catalogs

The Iceberg REST server reuses S3 Tables metadata storage under
/table-buckets and enables DuckDB, Spark, and other Iceberg clients
to connect to SeaweedFS as a catalog.

* feat: Add Iceberg Catalog pages to admin UI

Add admin UI pages to browse Iceberg catalogs, namespaces, and tables.

- Add Iceberg Catalog menu item under Object Store navigation
- Create iceberg_catalog.templ showing catalog overview with REST info
- Create iceberg_namespaces.templ listing namespaces in a catalog
- Create iceberg_tables.templ listing tables in a namespace
- Add handlers and routes in admin_handlers.go
- Add Iceberg data provider methods in s3tables_management.go
- Add Iceberg data types in types.go

The Iceberg Catalog pages provide visibility into the same S3 Tables
data through an Iceberg-centric lens, including REST endpoint examples
for DuckDB and PyIceberg.

* test: Add Iceberg catalog integration tests and reorg s3tables tests

- Reorganize existing s3tables tests to test/s3tables/table-buckets/
- Add new test/s3tables/catalog/ for Iceberg REST catalog tests
- Add TestIcebergConfig to verify /v1/config endpoint
- Add TestIcebergNamespaces to verify namespace listing
- Add TestDuckDBIntegration for DuckDB connectivity (requires Docker)
- Update CI workflow to use new test paths

* fix: Generate proper random UUIDs for Iceberg tables

Address code review feedback:
- Replace placeholder UUID with crypto/rand-based UUID v4 generation
- Add detailed TODO comments for handleUpdateTable stub explaining
  the required atomic metadata swap implementation

* fix: Serve Iceberg on localhost listener when binding to different interface

Address code review feedback: properly serve the localhost listener
when the Iceberg server is bound to a non-localhost interface.

* ci: Add Iceberg catalog integration tests to CI

Add new job to run Iceberg catalog tests in CI, along with:
- Iceberg package build verification
- Iceberg unit tests
- Iceberg go vet checks
- Iceberg format checks

* fix: Address code review feedback for Iceberg implementation

- fix: Replace hardcoded account ID with s3_constants.AccountAdminId in buildTableBucketARN()
- fix: Improve UUID generation error handling with deterministic fallback (timestamp + PID + counter)
- fix: Update handleUpdateTable to return HTTP 501 Not Implemented instead of fake success
- fix: Better error handling in handleNamespaceExists to distinguish 404 from 500 errors
- fix: Use relative URL in template instead of hardcoded localhost:8181
- fix: Add HTTP timeout to test's waitForService function to avoid hangs
- fix: Use dynamic ephemeral ports in integration tests to avoid flaky parallel failures
- fix: Add Iceberg port to final port configuration logging in mini.go

* fix: Address critical issues in Iceberg implementation

- fix: Cache table UUIDs to ensure persistence across LoadTable calls
  The UUID now remains stable for the lifetime of the server session.
  TODO: For production, UUIDs should be persisted in S3 Tables metadata.

- fix: Remove redundant URL-encoded namespace parsing
  mux router already decodes %1F to \x1F before passing to handlers.
  Redundant ReplaceAll call could cause bugs with literal %1F in namespace.

* fix: Improve test robustness and reduce code duplication

- fix: Make DuckDB test more robust by failing on unexpected errors
  Instead of silently logging errors, now explicitly check for expected
  conditions (extension not available) and skip the test appropriately.

- fix: Extract username helper method to reduce duplication
  Created getUsername() helper in AdminHandlers to avoid duplicating
  the username retrieval logic across Iceberg page handlers.

* fix: Add mutex protection to table UUID cache

Protects concurrent access to the tableUUIDs map with sync.RWMutex.
Uses read-lock for fast path when UUID already cached, and write-lock
for generating new UUIDs. Includes double-check pattern to handle race
condition between read-unlock and write-lock.

* style: fix go fmt errors

* feat(iceberg): persist table UUID in S3 Tables metadata

* feat(admin): configure Iceberg port in Admin UI and commands

* refactor: address review comments (flags, tests, handlers)

- command/mini: fix tracking of explicit s3.port.iceberg flag
- command/admin: add explicit -iceberg.port flag
- admin/handlers: reuse getUsername helper
- tests: use 127.0.0.1 for ephemeral ports and os.Stat for file size check

* test: check error from FileStat in verify_gc_empty_test

3 weeks ago

3 Commits (49a64f50f1f67f6d69133167cc1339052d3f4c92)