mirror of
https://github.com/seaweedfs/seaweedfs.git
synced 2026-09-08 15:41:15 +02:00
* build(deps): bump github.com/twmb/avro from 1.7.2 to 1.8.0 Bumps [github.com/twmb/avro](https://github.com/twmb/avro) from 1.7.2 to 1.8.0. - [Commits](https://github.com/twmb/avro/compare/v1.7.2...v1.8.0) --- updated-dependencies: - dependency-name: github.com/twmb/avro dependency-version: 1.8.0 dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <support@github.com> * iceberg: adapt to twmb/avro v1.8.0 and iceberg-go defensive copies avro v1.8.0 changes Schema.Root() to return *SchemaNode, which breaks iceberg-go v0.6.0's internal avro_schemas.go. The fix (apache/iceberg-go#1843) is only on iceberg-go's main branch, unreleased, so bump iceberg-go to that commit (c210509) alongside the avro bump. That iceberg-go revision also changes two behaviors seaweedfs worked around: - It now infers a manifest list's format version from the embedded writer schema, so a list missing the "format-version" header entry (DuckDB's shape is read as v2, not v1. ReadManifestList's header patching is now a redundant safety net; tests updated to expect v2. - It returns defensive copies from DataFile.Partition(), so the ReadManifest shim's in-place partition normalization was silently discarded. Rebuild the entry through NewDataFileBuilder when any partition value is normalized, copying every other DataFile field so manifest round-trips are preserved. - It converts day-transform partitions to iceberg.Date on read (applyDayTransformDates), so the day-partition cases the shim and tests guarded now convert without help; tests updated to expect iceberg.Date from the raw read. EOF ) * iceberg: accept assert-ref-snapshot-id without snapshot-id iceberg-go's new nullableInt64 parser rejects an assert-ref-snapshot-id requirement whose "snapshot-id" field is absent from the JSON, even though the Iceberg REST spec makes it optional (null means the ref must not already exist). v0.6.0 used a plain *int64, so absent was nil and accepted. ClickHouse sends the requirement without snapshot-id when asserting a branch does not yet exist, so its writes fail with "missing required field \"snapshot-id\"". normalizeRequirements splices an explicit null into any assert-ref-snapshot-id requirement missing the field before handing the JSON to iceberg-go's parser, restoring the v0.6.0 behavior across both iceberg-go versions. * iceberg: fix v1 block_size_in_bytes default in rebuilt manifest entries rebuildManifestEntry set block_size_in_bytes to 0, but the v1 manifest schema requires the default of 64 MiB ("Always write default in v1"). The original value is not exposed on the DataFile interface, so use the spec default. Also clarify the fallback comment to note that empty (zero-record / zero-byte) files also trigger it, not just a nil spec. Add a round-trip test that writes a rebuilt entry as v1 and verifies block_size_in_bytes is 64 MiB via Avro decoding. --------- Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: Chris Lu <chris.lu@gmail.com>
378 lines
14 KiB
Go
378 lines
14 KiB
Go
package iceberg
|
|
|
|
import (
|
|
"encoding/json"
|
|
"errors"
|
|
"fmt"
|
|
"net/http"
|
|
"path"
|
|
"strings"
|
|
|
|
"github.com/apache/iceberg-go/table"
|
|
"github.com/google/uuid"
|
|
"github.com/gorilla/mux"
|
|
"github.com/seaweedfs/seaweedfs/weed/glog"
|
|
"github.com/seaweedfs/seaweedfs/weed/pb/filer_pb"
|
|
"github.com/seaweedfs/seaweedfs/weed/s3api/s3_constants"
|
|
"github.com/seaweedfs/seaweedfs/weed/s3api/s3tables"
|
|
)
|
|
|
|
// handleUpdateTable commits updates to a table.
|
|
// Implements the Iceberg REST Catalog commit protocol.
|
|
func (s *Server) handleUpdateTable(w http.ResponseWriter, r *http.Request) {
|
|
vars := mux.Vars(r)
|
|
namespace := parseNamespace(vars["namespace"])
|
|
tableName := vars["table"]
|
|
|
|
if len(namespace) == 0 || tableName == "" {
|
|
writeError(w, http.StatusBadRequest, "BadRequestException", "Namespace and table name are required")
|
|
return
|
|
}
|
|
|
|
bucketName := getBucketFromPrefix(r)
|
|
bucketARN := buildTableBucketARN(bucketName)
|
|
|
|
// Extract identity from context
|
|
identityName := s3_constants.GetIdentityNameFromContext(r)
|
|
|
|
// Parse commit request and keep statistics updates separate because iceberg-go v0.4.0
|
|
// does not decode set/remove-statistics update actions yet.
|
|
var raw struct {
|
|
Identifier *TableIdentifier `json:"identifier,omitempty"`
|
|
Requirements json.RawMessage `json:"requirements"`
|
|
Updates []json.RawMessage `json:"updates"`
|
|
}
|
|
if err := json.NewDecoder(r.Body).Decode(&raw); err != nil {
|
|
writeError(w, http.StatusBadRequest, "BadRequestException", "Invalid request body: "+err.Error())
|
|
return
|
|
}
|
|
|
|
var req CommitTableRequest
|
|
req.Identifier = raw.Identifier
|
|
var statisticsUpdates []statisticsUpdate
|
|
if len(raw.Requirements) > 0 {
|
|
normalized, err := normalizeRequirements(raw.Requirements)
|
|
if err != nil {
|
|
writeError(w, http.StatusBadRequest, "BadRequestException", "Invalid requirements: "+err.Error())
|
|
return
|
|
}
|
|
if err := json.Unmarshal(normalized, &req.Requirements); err != nil {
|
|
writeError(w, http.StatusBadRequest, "BadRequestException", "Invalid requirements: "+err.Error())
|
|
return
|
|
}
|
|
}
|
|
if len(raw.Updates) > 0 {
|
|
var err error
|
|
req.Updates, statisticsUpdates, err = parseCommitUpdates(raw.Updates)
|
|
if err != nil {
|
|
writeError(w, http.StatusBadRequest, "BadRequestException", "Invalid updates: "+err.Error())
|
|
return
|
|
}
|
|
}
|
|
// Manifest repair runs once, as soon as the table location is known; on
|
|
// commit retries the updates already reference the repaired files. Repair
|
|
// is best effort end to end: the originals parsed already, so a repair
|
|
// that fails to re-parse is discarded rather than failing the commit.
|
|
manifestsRepaired := false
|
|
repairManifests := func(location string) {
|
|
if manifestsRepaired {
|
|
return
|
|
}
|
|
manifestsRepaired = true
|
|
repaired, changed := s.repairAddSnapshotManifests(r.Context(), location, raw.Updates)
|
|
if !changed {
|
|
return
|
|
}
|
|
repairedUpdates, repairedStatistics, err := parseCommitUpdates(repaired)
|
|
if err != nil {
|
|
glog.Warningf("Iceberg: repaired updates failed to parse, keeping originals: %v", err)
|
|
return
|
|
}
|
|
raw.Updates = repaired
|
|
req.Updates = repairedUpdates
|
|
statisticsUpdates = repairedStatistics
|
|
}
|
|
|
|
maxCommitAttempts := 3
|
|
generatedLegacyUUID := uuid.New()
|
|
stageCreateEnabled := isStageCreateEnabled()
|
|
for attempt := 1; attempt <= maxCommitAttempts; attempt++ {
|
|
getReq := &s3tables.GetTableRequest{
|
|
TableBucketARN: bucketARN,
|
|
Namespace: namespace,
|
|
Name: tableName,
|
|
}
|
|
var getResp s3tables.GetTableResponse
|
|
|
|
err := s.filerClient.WithFilerClient(false, func(client filer_pb.SeaweedFilerClient) error {
|
|
mgrClient := s3tables.NewManagerClient(client)
|
|
return s.tablesManager.Execute(r.Context(), mgrClient, "GetTable", getReq, &getResp, identityName)
|
|
})
|
|
if err != nil {
|
|
if isS3TablesNotFound(err) {
|
|
location := fmt.Sprintf("s3://%s/%s", bucketName, path.Join(flattenNamespacePath(namespace), tableName))
|
|
tableUUID := generatedLegacyUUID
|
|
baseMetadataVersion := 0
|
|
baseMetadataLocation := ""
|
|
var baseMetadata table.Metadata
|
|
|
|
var latestMarker *stageCreateMarker
|
|
if stageCreateEnabled {
|
|
var markerErr error
|
|
latestMarker, markerErr = s.loadLatestStageCreateMarker(r.Context(), bucketName, namespace, tableName)
|
|
if markerErr != nil {
|
|
writeError(w, http.StatusInternalServerError, "InternalServerError", "Failed to load stage-create marker: "+markerErr.Error())
|
|
return
|
|
}
|
|
}
|
|
if latestMarker != nil {
|
|
if latestMarker.Location != "" {
|
|
location = strings.TrimSuffix(latestMarker.Location, "/")
|
|
}
|
|
if latestMarker.TableUUID != "" {
|
|
if parsedUUID, parseErr := uuid.Parse(latestMarker.TableUUID); parseErr == nil {
|
|
tableUUID = parsedUUID
|
|
}
|
|
}
|
|
|
|
stagedMetadataLocation := latestMarker.StagedMetadataLocation
|
|
if stagedMetadataLocation == "" {
|
|
stagedMetadataLocation = fmt.Sprintf("%s/metadata/v1.metadata.json", strings.TrimSuffix(location, "/"))
|
|
}
|
|
stagedLocation := tableLocationFromMetadataLocation(stagedMetadataLocation)
|
|
stagedFileName := path.Base(stagedMetadataLocation)
|
|
stagedBucket, stagedPath, parseLocationErr := parseS3Location(stagedLocation)
|
|
if parseLocationErr != nil {
|
|
writeError(w, http.StatusInternalServerError, "InternalServerError", "Invalid staged metadata location: "+parseLocationErr.Error())
|
|
return
|
|
}
|
|
stagedMetadataBytes, loadErr := s.loadMetadataFile(r.Context(), stagedBucket, stagedPath, stagedFileName)
|
|
if loadErr != nil {
|
|
if !errors.Is(loadErr, filer_pb.ErrNotFound) {
|
|
writeError(w, http.StatusInternalServerError, "InternalServerError", "Failed to load staged metadata: "+loadErr.Error())
|
|
return
|
|
}
|
|
} else if len(stagedMetadataBytes) > 0 {
|
|
stagedMetadata, parseErr := table.ParseMetadataBytes(stagedMetadataBytes)
|
|
if parseErr != nil {
|
|
writeError(w, http.StatusInternalServerError, "InternalServerError", "Failed to parse staged metadata: "+parseErr.Error())
|
|
return
|
|
}
|
|
// Staged metadata is only a template for table creation; commit starts from version 1.
|
|
baseMetadata = stagedMetadata
|
|
baseMetadataLocation = ""
|
|
baseMetadataVersion = 0
|
|
if stagedMetadata.TableUUID() != uuid.Nil {
|
|
tableUUID = stagedMetadata.TableUUID()
|
|
}
|
|
}
|
|
}
|
|
|
|
hasAssertCreate := hasAssertCreateRequirement(req.Requirements)
|
|
hasStagedTemplate := baseMetadata != nil
|
|
if !(stageCreateEnabled && (hasAssertCreate || hasStagedTemplate)) {
|
|
writeError(w, http.StatusNotFound, "NoSuchTableException", fmt.Sprintf("Table does not exist: %s", tableName))
|
|
return
|
|
}
|
|
// From here the commit creates the table, writing its metadata
|
|
// file before the create that authorizes it.
|
|
if authErr := s.authorizeCreateTable(r.Context(), bucketARN, namespace, tableName, identityName); authErr != nil {
|
|
writeManagerError(w, authErr)
|
|
return
|
|
}
|
|
|
|
for _, requirement := range req.Requirements {
|
|
validateAgainst := table.Metadata(nil)
|
|
if hasStagedTemplate && requirement.GetType() != requirementAssertCreate {
|
|
validateAgainst = baseMetadata
|
|
}
|
|
if requirementErr := requirement.Validate(validateAgainst); requirementErr != nil {
|
|
writeError(w, http.StatusConflict, "CommitFailedException", "Requirement failed: "+requirementErr.Error())
|
|
return
|
|
}
|
|
}
|
|
|
|
if baseMetadata == nil {
|
|
var buildErr error
|
|
if baseMetadata, buildErr = newTableMetadata(tableUUID, location, nil, nil, nil, nil); buildErr != nil {
|
|
glog.Errorf("Iceberg: CommitTable placeholder metadata for %s: %v", tableName, buildErr)
|
|
writeError(w, http.StatusInternalServerError, "InternalServerError", "Failed to build current metadata")
|
|
return
|
|
}
|
|
}
|
|
|
|
repairManifests(location)
|
|
|
|
result, reqErr := s.finalizeCreateOnCommit(r.Context(), createOnCommitInput{
|
|
bucketARN: bucketARN,
|
|
markerBucket: bucketName,
|
|
namespace: namespace,
|
|
tableName: tableName,
|
|
identityName: identityName,
|
|
location: location,
|
|
tableUUID: tableUUID,
|
|
baseMetadata: baseMetadata,
|
|
baseMetadataLoc: baseMetadataLocation,
|
|
baseMetadataVer: baseMetadataVersion,
|
|
updates: req.Updates,
|
|
statisticsUpdates: statisticsUpdates,
|
|
})
|
|
if reqErr != nil {
|
|
writeError(w, reqErr.status, reqErr.errType, reqErr.message)
|
|
return
|
|
}
|
|
writeJSON(w, http.StatusOK, result)
|
|
return
|
|
}
|
|
glog.V(1).Infof("Iceberg: CommitTable GetTable error: %v", err)
|
|
writeManagerError(w, err)
|
|
return
|
|
}
|
|
|
|
location := tableLocationFromMetadataLocation(getResp.MetadataLocation)
|
|
if location == "" {
|
|
location = fmt.Sprintf("s3://%s/%s", bucketName, path.Join(flattenNamespacePath(namespace), tableName))
|
|
}
|
|
tableUUID := uuid.Nil
|
|
if getResp.Metadata != nil && getResp.Metadata.Iceberg != nil && getResp.Metadata.Iceberg.TableUUID != "" {
|
|
if parsed, parseErr := uuid.Parse(getResp.Metadata.Iceberg.TableUUID); parseErr == nil {
|
|
tableUUID = parsed
|
|
}
|
|
}
|
|
if tableUUID == uuid.Nil {
|
|
tableUUID = generatedLegacyUUID
|
|
}
|
|
|
|
var currentMetadata table.Metadata
|
|
if getResp.Metadata != nil && len(getResp.Metadata.FullMetadata) > 0 {
|
|
currentMetadata, err = table.ParseMetadataBytes(getResp.Metadata.FullMetadata)
|
|
if err != nil {
|
|
glog.Errorf("Iceberg: Failed to parse current metadata for %s: %v", tableName, err)
|
|
writeError(w, http.StatusInternalServerError, "InternalServerError", "Failed to parse current metadata")
|
|
return
|
|
}
|
|
} else {
|
|
currentMetadata, err = newTableMetadata(tableUUID, location, nil, nil, nil, nil)
|
|
if err != nil {
|
|
glog.Errorf("Iceberg: CommitTable placeholder metadata for %s: %v", tableName, err)
|
|
writeError(w, http.StatusInternalServerError, "InternalServerError", "Failed to build current metadata")
|
|
return
|
|
}
|
|
}
|
|
|
|
for _, requirement := range req.Requirements {
|
|
if err := requirement.Validate(currentMetadata); err != nil {
|
|
writeError(w, http.StatusConflict, "CommitFailedException", "Requirement failed: "+err.Error())
|
|
return
|
|
}
|
|
}
|
|
|
|
repairManifests(location)
|
|
|
|
builder, err := table.MetadataBuilderFromBase(currentMetadata, getResp.MetadataLocation)
|
|
if err != nil {
|
|
writeError(w, http.StatusInternalServerError, "InternalServerError", "Failed to create metadata builder: "+err.Error())
|
|
return
|
|
}
|
|
for _, update := range req.Updates {
|
|
if err := update.Apply(builder); err != nil {
|
|
writeError(w, http.StatusBadRequest, "BadRequestException", "Failed to apply update: "+err.Error())
|
|
return
|
|
}
|
|
}
|
|
|
|
newMetadata, err := builder.Build()
|
|
if err != nil {
|
|
writeError(w, http.StatusBadRequest, "BadRequestException", "Failed to build new metadata: "+err.Error())
|
|
return
|
|
}
|
|
|
|
metadataVersion := getResp.MetadataVersion + 1
|
|
metadataFileName := fmt.Sprintf("v%d.metadata.json", metadataVersion)
|
|
newMetadataLocation := fmt.Sprintf("%s/metadata/%s", strings.TrimSuffix(location, "/"), metadataFileName)
|
|
|
|
metadataBytes, err := json.Marshal(newMetadata)
|
|
if err != nil {
|
|
writeError(w, http.StatusInternalServerError, "InternalServerError", "Failed to serialize metadata: "+err.Error())
|
|
return
|
|
}
|
|
// iceberg-go does not currently support set/remove-statistics updates in MetadataBuilder.
|
|
// Patch the encoded metadata JSON and parse it back to keep the response object consistent.
|
|
metadataBytes, err = applyStatisticsUpdates(metadataBytes, statisticsUpdates)
|
|
if err != nil {
|
|
writeError(w, http.StatusBadRequest, "BadRequestException", "Failed to apply statistics updates: "+err.Error())
|
|
return
|
|
}
|
|
metadataBytes = refreshDefaultNameMapping(metadataBytes, newMetadata)
|
|
// Same spec-compliance fixup we apply on create-table; ensures
|
|
// v{N}.metadata.json files written during commit are also readable by
|
|
// strict Iceberg clients reading directly from S3, and that the
|
|
// FullMetadata persisted in S3Tables stays consistent.
|
|
metadataBytes = ensureMetadataSpecCompliance(metadataBytes)
|
|
newMetadata, err = table.ParseMetadataBytes(metadataBytes)
|
|
if err != nil {
|
|
writeError(w, http.StatusInternalServerError, "InternalServerError", "Failed to parse committed metadata: "+err.Error())
|
|
return
|
|
}
|
|
|
|
metadataBucket, metadataPath, err := parseS3Location(location)
|
|
if err != nil {
|
|
writeError(w, http.StatusInternalServerError, "InternalServerError", "Invalid table location: "+err.Error())
|
|
return
|
|
}
|
|
metadataFileName, newMetadataLocation, err = s.stageCommitMetadata(r.Context(), metadataBucket, metadataPath, location, metadataFileName, metadataBytes)
|
|
if err != nil {
|
|
writeError(w, http.StatusInternalServerError, "InternalServerError", "Failed to save metadata file: "+err.Error())
|
|
return
|
|
}
|
|
|
|
updateReq := &s3tables.UpdateTableRequest{
|
|
TableBucketARN: bucketARN,
|
|
Namespace: namespace,
|
|
Name: tableName,
|
|
VersionToken: getResp.VersionToken,
|
|
Metadata: &s3tables.TableMetadata{
|
|
Iceberg: &s3tables.IcebergMetadata{
|
|
TableUUID: tableUUID.String(),
|
|
},
|
|
FullMetadata: metadataBytes,
|
|
},
|
|
MetadataVersion: metadataVersion,
|
|
MetadataLocation: newMetadataLocation,
|
|
}
|
|
|
|
err = s.filerClient.WithFilerClient(false, func(client filer_pb.SeaweedFilerClient) error {
|
|
mgrClient := s3tables.NewManagerClient(client)
|
|
return s.tablesManager.Execute(r.Context(), mgrClient, "UpdateTable", updateReq, nil, identityName)
|
|
})
|
|
if err == nil {
|
|
result := CommitTableResponse{
|
|
MetadataLocation: newMetadataLocation,
|
|
Metadata: newMetadata,
|
|
}
|
|
writeJSON(w, http.StatusOK, result)
|
|
return
|
|
}
|
|
|
|
if isS3TablesConflict(err) {
|
|
if cleanupErr := s.deleteMetadataFile(r.Context(), metadataBucket, metadataPath, metadataFileName); cleanupErr != nil {
|
|
glog.V(1).Infof("Iceberg: failed to cleanup metadata file %s on conflict: %v", newMetadataLocation, cleanupErr)
|
|
}
|
|
if attempt < maxCommitAttempts {
|
|
glog.V(1).Infof("Iceberg: CommitTable conflict for %s (attempt %d/%d), retrying", tableName, attempt, maxCommitAttempts)
|
|
sleepBeforeCommitRetry(attempt)
|
|
continue
|
|
}
|
|
writeError(w, http.StatusConflict, "CommitFailedException", "Version token mismatch")
|
|
return
|
|
}
|
|
|
|
if cleanupErr := s.deleteMetadataFile(r.Context(), metadataBucket, metadataPath, metadataFileName); cleanupErr != nil {
|
|
glog.V(1).Infof("Iceberg: failed to cleanup metadata file %s after update failure: %v", newMetadataLocation, cleanupErr)
|
|
}
|
|
glog.Errorf("Iceberg: CommitTable UpdateTable error: %v", err)
|
|
writeError(w, http.StatusInternalServerError, "InternalServerError", "Failed to commit table update: "+err.Error())
|
|
return
|
|
}
|
|
}
|