Files
seaweedfs/weed/s3api/iceberg/handlers_namespace.go
T
Chris Lu f50917224a fix(iceberg): default namespace location so fresh CTAS does not race metadata write (#9074) (#9246)
* fix(iceberg): advertise default namespace location for REST clients

Trino's REST catalog has two code paths for CREATE TABLE depending on
whether a table location can be resolved before the catalog call:

    // TrinoRestCatalog.newCreateTableTransaction (Trino 479)
    if (location.isEmpty()) {
        return tableBuilder.create().newTransaction();   // EAGER: REST POST now
    }
    return tableBuilder.withLocation(...).createTransaction(); // DEFERRED

`tableLocation` resolves to null when the user does not pass
`location = '...'` AND `defaultTableLocation` returns null. The latter
happens whenever the namespace's `loadNamespaceMetadata` response has no
`location` property — and our handler returned exactly that.

In the eager branch Trino calls REST POST /v1/.../tables immediately, so
our handleCreateTable persists `<location>/metadata/v1.metadata.json` to
the filer. Trino's IcebergMetadata.beginCreateTable then runs
`fileSystem.listFiles(location).hasNext()` on the same path, finds the
metadata file we just wrote, and throws "Cannot create a table on a
non-empty location" — even on a brand-new schema and table name.

Synthesize a default `location` of `s3://<bucket>/<flattened-namespace>`
on Get/Create namespace responses when one is not stored. With a non-
null `defaultTableLocation`, Trino takes the deferred branch, picks a
unique `<namespace-location>/<table>-<UUID>` path (UUID added by the
standard `iceberg.unique-table-location=true` setting), and the empty-
location check passes. The actual REST POST is deferred to commit time,
so our metadata write lands alongside the data files Trino has already
produced.

Existing namespaces with an explicit `location` property are untouched —
the synthesis only kicks in when the property is absent.

Refs #9074

* test(trino): regression for fresh CREATE TABLE without explicit location

Exercises the follow-up scenario reported on issue #9074: a CREATE TABLE
on a brand-new schema and brand-new table name with NO `location = '...'`
clause, followed by a CTAS on top — exactly the SQL pattern from the
report. Before advertising a default namespace location, the first
CREATE TABLE failed with

    Cannot create a table on a non-empty location:
    s3://iceberg-tables/<schema>/<table>, set
    'iceberg.unique-table-location=true' in your Iceberg catalog
    properties to use unique table locations for every table.

even though the bucket was genuinely empty. The bug surfaced because our
handleCreateTable persists `<location>/metadata/v1.metadata.json` during
Trino's eager-create branch, and Trino's post-create listFiles
emptiness check then trips on the metadata file we just wrote.

The test asserts the CTAS succeeds AND the resulting table contains the
source rows, since the reporter saw the table get created with empty
data when the query failed.

Refs #9074
2026-04-27 16:37:33 -07:00

276 lines
9.2 KiB
Go

package iceberg
import (
"encoding/json"
"fmt"
"net/http"
"strings"
"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"
)
// handleConfig returns catalog configuration.
//
// When a client passes ?warehouse=s3://<bucket>/, the Iceberg REST spec
// expects the server to echo a catalog identifier back as overrides.prefix
// so subsequent calls use /v1/{prefix}/... and land on the right table
// bucket. Without this, clients like DuckDB's ATTACH flow fall back to an
// unprefixed path that resolves to the wrong bucket and report phantom
// "schema does not exist" errors. See issue #9103.
func (s *Server) handleConfig(w http.ResponseWriter, r *http.Request) {
w.Header().Set("Content-Type", "application/json")
config := CatalogConfig{
Defaults: map[string]string{},
Overrides: map[string]string{},
}
if warehouse := strings.TrimSpace(r.URL.Query().Get("warehouse")); warehouse != "" {
// Only the bucket portion of the warehouse URL is meaningful today —
// SeaweedFS table-bucket routing is bucket-scoped, so any sub-path
// (e.g. s3://bucket/prefix/) is ignored here and clients that try to
// scope a catalog under a sub-prefix will still land on the bucket.
if bucket, _, err := parseS3Location(warehouse); err == nil && bucket != "" {
config.Overrides["prefix"] = bucket
config.Defaults["warehouse"] = warehouse
}
}
if err := json.NewEncoder(w).Encode(config); err != nil {
glog.Warningf("handleConfig: Failed to encode config: %v", err)
}
}
// handleListNamespaces lists namespaces in a catalog.
func (s *Server) handleListNamespaces(w http.ResponseWriter, r *http.Request) {
bucketName := getBucketFromPrefix(r)
bucketARN := buildTableBucketARN(bucketName)
// Extract identity from context
identityName := s3_constants.GetIdentityNameFromContext(r)
pageToken, pageSize, err := parsePagination(r)
if err != nil {
writeError(w, http.StatusBadRequest, "BadRequestException", err.Error())
return
}
// The Iceberg REST spec allows a "parent" query parameter for hierarchical
// namespace listing. Convert it to the dot-separated prefix used by S3 Tables.
var prefix string
if parent := r.URL.Query().Get("parent"); parent != "" {
parentParts := parseNamespace(parent)
if len(parentParts) > 0 {
prefix = flattenNamespacePath(parentParts) + "."
}
}
// Use S3 Tables manager to list namespaces
var resp s3tables.ListNamespacesResponse
req := &s3tables.ListNamespacesRequest{
TableBucketARN: bucketARN,
Prefix: prefix,
ContinuationToken: pageToken,
MaxNamespaces: pageSize,
}
err = s.filerClient.WithFilerClient(false, func(client filer_pb.SeaweedFilerClient) error {
mgrClient := s3tables.NewManagerClient(client)
return s.tablesManager.Execute(r.Context(), mgrClient, "ListNamespaces", req, &resp, identityName)
})
if err != nil {
glog.Infof("Iceberg: ListNamespaces error: %v", err)
writeError(w, http.StatusInternalServerError, "InternalServerError", err.Error())
return
}
// Convert to Iceberg format
namespaces := make([]Namespace, 0, len(resp.Namespaces))
for _, ns := range resp.Namespaces {
namespaces = append(namespaces, Namespace(ns.Namespace))
}
result := ListNamespacesResponse{
NextPageToken: resp.ContinuationToken,
Namespaces: namespaces,
}
writeJSON(w, http.StatusOK, result)
}
// handleCreateNamespace creates a new namespace.
func (s *Server) handleCreateNamespace(w http.ResponseWriter, r *http.Request) {
bucketName := getBucketFromPrefix(r)
bucketARN := buildTableBucketARN(bucketName)
// Extract identity from context
identityName := s3_constants.GetIdentityNameFromContext(r)
var req CreateNamespaceRequest
if err := json.NewDecoder(r.Body).Decode(&req); err != nil {
writeError(w, http.StatusBadRequest, "BadRequestException", "Invalid request body")
return
}
if len(req.Namespace) == 0 {
writeError(w, http.StatusBadRequest, "BadRequestException", "Namespace is required")
return
}
// Use S3 Tables manager to create namespace
createReq := &s3tables.CreateNamespaceRequest{
TableBucketARN: bucketARN,
Namespace: req.Namespace,
Properties: normalizeNamespaceProperties(req.Properties),
}
var createResp s3tables.CreateNamespaceResponse
err := s.filerClient.WithFilerClient(false, func(client filer_pb.SeaweedFilerClient) error {
mgrClient := s3tables.NewManagerClient(client)
glog.V(2).Infof("Iceberg: handleCreateNamespace calling Execute with identityName=%s", identityName)
return s.tablesManager.Execute(r.Context(), mgrClient, "CreateNamespace", createReq, &createResp, identityName)
})
if err != nil {
if strings.Contains(err.Error(), "already exists") {
writeError(w, http.StatusConflict, "AlreadyExistsException", err.Error())
return
}
glog.Errorf("Iceberg: CreateNamespace error: %v", err)
writeError(w, http.StatusInternalServerError, "InternalServerError", err.Error())
return
}
result := CreateNamespaceResponse{
Namespace: Namespace(createResp.Namespace),
Properties: withDefaultNamespaceLocation(createResp.Properties, bucketName, createResp.Namespace),
}
writeJSON(w, http.StatusOK, result)
}
// handleGetNamespace gets namespace metadata.
func (s *Server) handleGetNamespace(w http.ResponseWriter, r *http.Request) {
vars := mux.Vars(r)
namespace := parseNamespace(vars["namespace"])
if len(namespace) == 0 {
writeError(w, http.StatusBadRequest, "BadRequestException", "Namespace is required")
return
}
bucketName := getBucketFromPrefix(r)
bucketARN := buildTableBucketARN(bucketName)
// Extract identity from context
identityName := s3_constants.GetIdentityNameFromContext(r)
// Use S3 Tables manager to get namespace
getReq := &s3tables.GetNamespaceRequest{
TableBucketARN: bucketARN,
Namespace: namespace,
}
var getResp s3tables.GetNamespaceResponse
err := s.filerClient.WithFilerClient(false, func(client filer_pb.SeaweedFilerClient) error {
mgrClient := s3tables.NewManagerClient(client)
return s.tablesManager.Execute(r.Context(), mgrClient, "GetNamespace", getReq, &getResp, identityName)
})
if err != nil {
if strings.Contains(err.Error(), "not found") {
writeError(w, http.StatusNotFound, "NoSuchNamespaceException", fmt.Sprintf("Namespace does not exist: %v", namespace))
return
}
glog.V(1).Infof("Iceberg: GetNamespace error: %v", err)
writeError(w, http.StatusInternalServerError, "InternalServerError", err.Error())
return
}
result := GetNamespaceResponse{
Namespace: Namespace(getResp.Namespace),
Properties: withDefaultNamespaceLocation(getResp.Properties, bucketName, getResp.Namespace),
}
writeJSON(w, http.StatusOK, result)
}
// handleNamespaceExists checks if a namespace exists.
func (s *Server) handleNamespaceExists(w http.ResponseWriter, r *http.Request) {
vars := mux.Vars(r)
namespace := parseNamespace(vars["namespace"])
if len(namespace) == 0 {
w.WriteHeader(http.StatusBadRequest)
return
}
bucketName := getBucketFromPrefix(r)
bucketARN := buildTableBucketARN(bucketName)
// Extract identity from context
identityName := s3_constants.GetIdentityNameFromContext(r)
getReq := &s3tables.GetNamespaceRequest{
TableBucketARN: bucketARN,
Namespace: namespace,
}
var getResp s3tables.GetNamespaceResponse
err := s.filerClient.WithFilerClient(false, func(client filer_pb.SeaweedFilerClient) error {
mgrClient := s3tables.NewManagerClient(client)
return s.tablesManager.Execute(r.Context(), mgrClient, "GetNamespace", getReq, &getResp, identityName)
})
if err != nil {
if strings.Contains(err.Error(), "not found") {
w.WriteHeader(http.StatusNotFound)
return
}
glog.V(1).Infof("Iceberg: NamespaceExists error: %v", err)
w.WriteHeader(http.StatusInternalServerError)
return
}
w.WriteHeader(http.StatusNoContent)
}
// handleDropNamespace deletes a namespace.
func (s *Server) handleDropNamespace(w http.ResponseWriter, r *http.Request) {
vars := mux.Vars(r)
namespace := parseNamespace(vars["namespace"])
if len(namespace) == 0 {
writeError(w, http.StatusBadRequest, "BadRequestException", "Namespace is required")
return
}
bucketName := getBucketFromPrefix(r)
bucketARN := buildTableBucketARN(bucketName)
// Extract identity from context
identityName := s3_constants.GetIdentityNameFromContext(r)
deleteReq := &s3tables.DeleteNamespaceRequest{
TableBucketARN: bucketARN,
Namespace: namespace,
}
err := s.filerClient.WithFilerClient(false, func(client filer_pb.SeaweedFilerClient) error {
mgrClient := s3tables.NewManagerClient(client)
return s.tablesManager.Execute(r.Context(), mgrClient, "DeleteNamespace", deleteReq, nil, identityName)
})
if err != nil {
if strings.Contains(err.Error(), "not found") {
writeError(w, http.StatusNotFound, "NoSuchNamespaceException", fmt.Sprintf("Namespace does not exist: %v", namespace))
return
}
if strings.Contains(err.Error(), "not empty") {
writeError(w, http.StatusConflict, "NamespaceNotEmptyException", "Namespace is not empty")
return
}
glog.V(1).Infof("Iceberg: DropNamespace error: %v", err)
writeError(w, http.StatusInternalServerError, "InternalServerError", err.Error())
return
}
w.WriteHeader(http.StatusNoContent)
}