Files
seaweedfs/weed/s3api/s3tables/handler.go
T
Chris Lu eef6f3d1e6 s3tables: add the maintenance configuration APIs (#10773)
* s3tables: add the maintenance configuration APIs

Stores the configuration verbatim as the wire shape under a new
s3tables.maintenance extended attribute, so Get hands back what Put took
and no translation layer can drift from the AWS model.

Nothing reads the configuration yet.

Put merges a single type into the stored map so configuring compaction
does not drop snapshot management, and asserts the attribute's prior value
so two concurrent Puts cannot silently clobber each other.

* iceberg: apply the maintenance configuration in the worker

The worker now reads the per-table and per-bucket maintenance
configuration written by the control plane, so the wildcard plugin config
is a default rather than the only setting a table can have.

Table properties still win by default, since a table declaring its own
layout is what every engine honours and the compactor has to agree with
whoever writes the files. Clearing table_properties_override makes the
maintenance configuration authoritative instead.

Status is not part of that contest: a disabled type drops its operations
and no property can re-enable them, so the operator's kill switch always
holds. Manifest and delete-file rewrites have no AWS equivalent and ride
with compaction.

Detection reads both attributes from entries it already lists.

* s3tables: report maintenance job status

The worker records the outcome of each run in its own extended attribute,
separate from the configuration so operator and worker writes do not
contend, and GetTableMaintenanceJobStatus reads it back.

Only the types a run touched are written, so a partial run cannot erase
what an earlier one recorded. The reader fills in the rest: Disabled when
the configuration switched a type off, Not_Yet_Run otherwise.

Status is advisory, so a lost race is logged rather than failing a job
whose work already committed.

* s3tables: route the maintenance APIs over REST

The five actions were only reachable by X-Amz-Target dispatch, which the
AWS CLI and SDK do not use for this service. They address the operations
by path, so the APIs were unreachable from any official client.

* s3tables: fix the table bucket ARN field name

GetTableBucketMaintenanceConfiguration emitted tableBucketArn where the
wire field is tableBucketARN, as every other response in this package
already spells it. Official SDK deserializers ignore the unknown key, so
the required field came back unset.

* s3tables: carry the compaction strategy through to the worker

IcebergCompactionSettings modelled only targetFileSizeMB, so a request
naming a strategy was accepted and then dropped on the way to storage.
The worker now maps binpack and sort onto its own rewrite strategy and
lets auto defer to the worker configuration.

z-order is rejected rather than accepted and quietly binpacked.

* s3tables: report bucket-level maintenance status

GetTableMaintenanceJobStatus read only the table's configuration, so
unreferenced file removal — which is configured on the bucket — reported
Not_Yet_Run or a stale success after an operator disabled it.

The merge helper now lives in this package and the worker shares it.

* iceberg: delete orphans only after the non-current window

AWS marks a file non-current once it has been unreferenced for
unreferencedDays, then deletes it a further nonCurrentDays later.
The cutoff was taken from unreferencedDays alone, so a 3/10 configuration
hard-deleted on day three and threw away the ten day recovery window.

remove_orphans deletes in one step rather than marking, so the cutoff is
now the sum of the two.

* s3tables: assert every attribute when rewriting an entry

UpdateEntry writes the whole entry back from the snapshot the caller
read, and its precondition only covers the keys the caller names. Both
maintenance writers named one key, so a job status write could revert a
maintenance configuration an operator had just disabled, turning an
advisory write into a silent re-enable.

Both now assert the entry's full attribute set, including the target key
when absent so a concurrent create also fails the precondition.

* s3tables: assert absent attributes when rewriting an entry

The precondition covered the attributes present when the writer read the
entry, so an attribute created between that read and the write was absent
from it. A first-time PutTableMaintenanceConfiguration disabling a type
therefore lands, passes the per-key checks, and is then deleted by the
stale whole-entry write.

Every attribute this package stores is now asserted, absent ones
included. The metadata commit and planning index writers rewrite the same
entries and had the same exposure, so both use the shared snapshot too.

* iceberg: implement the auto compaction strategy

auto was accepted, stored and read back, but left the worker on its own
default, so a sorted table configured as auto was compacted with binpack.

AWS defines auto as sorting tables that declare a sort order and
bin-packing the rest. That needs the table metadata, so the choice is made
where the rewrite plan is resolved: an unsorted table falls back to
binpack rather than failing the way an explicit sort request does.

* s3tables: validate the maintenance setting ranges

PUT accepted zero, negative and oversized values for every numeric
setting. The worker then ignores a non-positive value and saturates an
oversized one, so the configuration read back was not the one that ran.

AWS bounds all five to 1..2147483647, which is now enforced. The fields
are pointers so an explicit zero is distinguishable from an omitted one
and can be rejected rather than silently ignored.

* s3tables: give every entry writer the same compare-and-swap

updateExtendedAttribute asserted the entry's attributes, but the helpers
behind the metadata, policy and tag handlers still wrote the whole entry
unconditionally. Any of them could land on a stale snapshot and delete a
maintenance configuration an operator had just written.

They all share one read-modify-write loop now, so the precondition and
the bounded retry apply wherever an entry is rewritten.

* s3tables: move the maintenance configuration with a renamed table

RenameTable carried the metadata, version, policy and tags to the new
name but left the maintenance configuration and job status behind. A
table with snapshot management disabled came back enabled under its new
name, and the stale configuration stayed on the old name where a table
created there would inherit it.

The decoupled-delete cleanup left the same two attributes behind.

* s3tables: accept every AWS partition in ARNs

The route regexes and the ARN patterns both hardcoded arn:aws, so valid
aws-cn and aws-us-gov ARNs never reached a handler. The router now shares
the partition-tolerant prefix with the parser, and a generated ARN uses
the partition its region belongs to so it parses back.

* s3tables: generate ARNs in the region's partition

The handler's own ARN generators still formatted arn:aws directly rather
than going through the partition-aware builder, so a China or GovCloud
deployment routed the request but then returned a commercial ARN and
matched IAM policies against it.

The round-trip test missed this because parsing accepts any partition, so
it now asserts the prefix the region implies.

* s3tables: complete the ARN partition table

aws-iso-e, aws-iso-f and aws-eusc were missing, so eu-isoe-*, us-isof-*
and eusc-* regions fell through to the commercial partition.

* s3tables: do not let a rename swallow a concurrent maintenance write

Rename copied the source attributes early and cleared the source at the
end, so a Put landing in between missed the copy to the destination and
was then deleted by the cleanup. It succeeded and vanished.

The cleanup now clears the source only while it still holds exactly what
was copied, and returns a conflict otherwise. Put checks the catalog
identity inside the same conditional mutation, so it also cannot write to
a name that a rename or delete has already soft-deleted.
2026-08-16 10:36:59 -07:00

443 lines
14 KiB
Go

package s3tables
import (
"context"
"encoding/json"
"errors"
"fmt"
"io"
"net/http"
"reflect"
"strings"
"github.com/seaweedfs/seaweedfs/weed/glog"
"github.com/seaweedfs/seaweedfs/weed/pb/filer_pb"
"github.com/seaweedfs/seaweedfs/weed/s3api/s3_constants"
)
const (
TablesPath = s3_constants.DefaultBucketsPath
DefaultAccountID = "000000000000"
DefaultRegion = "us-east-1"
// Extended entry attributes for metadata storage
ExtendedKeyTableBucket = "s3tables.tableBucket"
ExtendedKeyMetadata = "s3tables.metadata"
ExtendedKeyMetadataVersion = "s3tables.metadataVersion"
ExtendedKeyPolicy = "s3tables.policy"
ExtendedKeyTags = "s3tables.tags"
ExtendedKeyMaintenance = "s3tables.maintenance"
// Written by the maintenance worker, read by GetTableMaintenanceJobStatus.
// Separate from ExtendedKeyMaintenance so worker and operator writes do not
// contend on the same attribute.
ExtendedKeyMaintenanceStatus = "s3tables.maintenanceStatus"
ExtendedKeyEntryType = "s3tables.entryType"
// Entry-type marker values for ExtendedKeyEntryType. Absent or "table" means
// a table; views are stored like tables but tagged "view".
EntryTypeTable = "table"
EntryTypeView = "view"
// Maximum request body size (10MB)
maxRequestBodySize = 10 * 1024 * 1024
)
var (
ErrVersionTokenMismatch = errors.New("version token mismatch")
ErrAccessDenied = errors.New("access denied")
ErrTableAlreadyExists = errors.New("table already exists")
)
type ResourceType string
const (
ResourceTypeBucket ResourceType = "bucket"
ResourceTypeTable ResourceType = "table"
)
// S3TablesHandler handles S3 Tables API requests
type S3TablesHandler struct {
region string
accountID string
defaultAllow bool // Whether to allow access by default (for zero-config IAM)
trusted bool // Trusted local tooling (shell/admin) bypasses authorization
iamAuthorizer IAMAuthorizer
}
// NewS3TablesHandler creates a new S3 Tables handler
func NewS3TablesHandler() *S3TablesHandler {
return &S3TablesHandler{
region: DefaultRegion,
accountID: DefaultAccountID,
defaultAllow: false,
}
}
// SetRegion sets the AWS region for ARN generation
func (h *S3TablesHandler) SetRegion(region string) {
if region != "" {
h.region = region
}
}
// SetAccountID sets the AWS account ID for ARN generation
func (h *S3TablesHandler) SetAccountID(accountID string) {
if accountID != "" {
h.accountID = accountID
}
}
// SetDefaultAllow sets whether to allow access by default
func (h *S3TablesHandler) SetDefaultAllow(allow bool) {
h.defaultAllow = allow
}
// SetTrusted lets local tooling that talks to the filer directly (shell, admin
// console) bypass authorization. HTTP-facing callers must not set it.
func (h *S3TablesHandler) SetTrusted(trusted bool) {
h.trusted = trusted
}
// FilerClient interface for filer operations
type FilerClient interface {
WithFilerClient(streamingMode bool, fn func(client filer_pb.SeaweedFilerClient) error) error
}
// HandleRequest is the main entry point for S3 Tables API requests
func (h *S3TablesHandler) HandleRequest(w http.ResponseWriter, r *http.Request, filerClient FilerClient) {
operation := r.Header.Get("X-Amz-Target")
if operation != "" {
if idx := strings.LastIndex(operation, "."); idx != -1 {
operation = operation[idx+1:]
}
}
if operation == "" {
glog.V(1).Infof("S3Tables: missing X-Amz-Target header")
h.writeError(w, http.StatusBadRequest, ErrCodeInvalidRequest, "Missing X-Amz-Target header")
return
}
glog.V(3).Infof("S3Tables: handling operation %s", operation)
var err error
switch operation {
// Table Bucket operations
case "CreateTableBucket":
err = h.handleCreateTableBucket(w, r, filerClient)
case "GetTableBucket":
err = h.handleGetTableBucket(w, r, filerClient)
case "ListTableBuckets":
err = h.handleListTableBuckets(w, r, filerClient)
case "DeleteTableBucket":
err = h.handleDeleteTableBucket(w, r, filerClient)
// Table Bucket Policy operations
case "PutTableBucketPolicy":
err = h.handlePutTableBucketPolicy(w, r, filerClient)
case "GetTableBucketPolicy":
err = h.handleGetTableBucketPolicy(w, r, filerClient)
case "DeleteTableBucketPolicy":
err = h.handleDeleteTableBucketPolicy(w, r, filerClient)
// Namespace operations
case "CreateNamespace":
err = h.handleCreateNamespace(w, r, filerClient)
case "GetNamespace":
err = h.handleGetNamespace(w, r, filerClient)
case "UpdateNamespace":
err = h.handleUpdateNamespace(w, r, filerClient)
case "ListNamespaces":
err = h.handleListNamespaces(w, r, filerClient)
case "DeleteNamespace":
err = h.handleDeleteNamespace(w, r, filerClient)
// Table operations
case "CreateTable":
err = h.handleCreateTable(w, r, filerClient)
case "RegisterTable":
err = h.handleRegisterTable(w, r, filerClient)
case "GetTable":
err = h.handleGetTable(w, r, filerClient)
case "ListTables":
err = h.handleListTables(w, r, filerClient)
case "UpdateTable":
err = h.handleUpdateTable(w, r, filerClient)
case "DeleteTable":
err = h.handleDeleteTable(w, r, filerClient)
case "RenameTable":
err = h.handleRenameTable(w, r, filerClient)
// View operations
case "CreateView":
err = h.handleCreateView(w, r, filerClient)
case "GetView":
err = h.handleGetView(w, r, filerClient)
case "ListViews":
err = h.handleListViews(w, r, filerClient)
case "UpdateView":
err = h.handleUpdateView(w, r, filerClient)
case "DeleteView":
err = h.handleDeleteView(w, r, filerClient)
// Table Policy operations
case "PutTablePolicy":
err = h.handlePutTablePolicy(w, r, filerClient)
case "GetTablePolicy":
err = h.handleGetTablePolicy(w, r, filerClient)
case "DeleteTablePolicy":
err = h.handleDeleteTablePolicy(w, r, filerClient)
// Maintenance configuration operations
case "PutTableBucketMaintenanceConfiguration":
err = h.handlePutTableBucketMaintenanceConfiguration(w, r, filerClient)
case "GetTableBucketMaintenanceConfiguration":
err = h.handleGetTableBucketMaintenanceConfiguration(w, r, filerClient)
case "PutTableMaintenanceConfiguration":
err = h.handlePutTableMaintenanceConfiguration(w, r, filerClient)
case "GetTableMaintenanceConfiguration":
err = h.handleGetTableMaintenanceConfiguration(w, r, filerClient)
case "GetTableMaintenanceJobStatus":
err = h.handleGetTableMaintenanceJobStatus(w, r, filerClient)
// Tagging operations
case "TagResource":
err = h.handleTagResource(w, r, filerClient)
case "ListTagsForResource":
err = h.handleListTagsForResource(w, r, filerClient)
case "UntagResource":
err = h.handleUntagResource(w, r, filerClient)
default:
h.writeError(w, http.StatusBadRequest, ErrCodeInvalidRequest, fmt.Sprintf("Unknown operation: %s", operation))
return
}
if err != nil {
glog.Errorf("S3Tables: error handling %s: %v", operation, err)
}
}
// Principal/authorization helpers
// getAccountID returns a stable caller identifier for ownership and permission checks.
// Reflection depends on the identity shape produced by JWT/STS auth (Account *struct{Id string},
// Claims map[string]interface{} containing string values for preferred_username/sub, and optional
// identity name/header values). Changing those fields without updating the reflection here will
// break the handler, so refactorers should replace this with a typed interface if needed.
func (h *S3TablesHandler) getAccountID(r *http.Request) string {
identityRaw := s3_constants.GetIdentityFromContext(r)
if identityRaw != nil {
// Use reflection to access identity fields and avoid import cycles.
val := reflect.ValueOf(identityRaw)
if val.Kind() == reflect.Ptr {
val = val.Elem()
}
if val.Kind() == reflect.Struct {
// Prefer stable claims from JWT/STS identities. Only "sub" is guaranteed durable per OIDC;
// preferred_username is ergonomic but can rotate and may orphan ownership data, while email
// is explicitly excluded to avoid storing PII in metadata.
claimsField := val.FieldByName("Claims")
if claimsField.IsValid() && claimsField.Kind() == reflect.Map && !claimsField.IsNil() && claimsField.Type().Key().Kind() == reflect.String {
for _, claimKey := range []string{"sub", "preferred_username"} {
claimVal := claimsField.MapIndex(reflect.ValueOf(claimKey))
if !claimVal.IsValid() {
continue
}
if claimVal.Kind() == reflect.Interface && !claimVal.IsNil() {
claimVal = claimVal.Elem()
}
if claimVal.Kind() == reflect.String {
if principal := normalizePrincipalID(claimVal.String()); principal != "" {
return principal
}
}
}
}
accountField := val.FieldByName("Account")
if accountField.IsValid() && !accountField.IsNil() {
accountVal := accountField.Elem()
if accountVal.Kind() == reflect.Struct {
idField := accountVal.FieldByName("Id")
if idField.IsValid() && idField.Kind() == reflect.String {
if principal := normalizePrincipalID(idField.String()); principal != "" {
// Account-less identities default to the admin account; only
// keep it for real admins, else use the unique identity name.
if principal != s3_constants.AccountAdminId || hasAdminAction(getIdentityActions(r)) {
return principal
}
}
}
}
}
}
}
if identityName := s3_constants.GetIdentityNameFromContext(r); identityName != "" {
if principal := normalizePrincipalID(identityName); principal != "" {
return principal
}
}
if accountID := r.Header.Get(s3_constants.AmzAccountId); accountID != "" {
if principal := normalizePrincipalID(accountID); principal != "" {
return principal
}
}
return h.accountID
}
// normalizePrincipalID collapses ARN and identity strings to a key that is stable within a single account.
// WARNING: this assumes identity names are unique per account; distinct principals such as
// arn:aws:iam::111:user/alice and arn:aws:iam::222:user/alice will both normalize to "alice".
// If future work adds multi-account support, revisit this function to include the account ID or full ARN
// so ownership checks remain correct.
func normalizePrincipalID(id string) string {
id = strings.TrimSpace(id)
if id == "" {
return ""
}
// If this is an ARN (common for assumed roles), use the trailing segment as a
// stable-ish principal key instead of embedding the full ARN in ownership fields.
if strings.HasPrefix(id, "arn:") {
if idx := strings.LastIndex(id, "/"); idx >= 0 && idx+1 < len(id) {
return strings.TrimSpace(id[idx+1:])
}
if idx := strings.LastIndex(id, ":"); idx >= 0 && idx+1 < len(id) {
return strings.TrimSpace(id[idx+1:])
}
return strings.TrimSpace(id)
}
return id
}
// getIdentityActions extracts the action list from the identity object in the request context.
// Uses reflection to avoid import cycles with s3api package.
func getIdentityActions(r *http.Request) []string {
val, ok := getIdentityStructValue(r)
if !ok {
return nil
}
actionsField := val.FieldByName("Actions")
if !actionsField.IsValid() || actionsField.Kind() != reflect.Slice {
return nil
}
// Convert actions to string slice
actions := make([]string, actionsField.Len())
for i := 0; i < actionsField.Len(); i++ {
action := actionsField.Index(i)
// Action is likely a custom type (e.g., type Action string)
// Convert to string using String() or direct string conversion
if action.Kind() == reflect.String {
actions[i] = action.String()
} else if action.CanInterface() {
// Try to convert via fmt.Sprint
actions[i] = fmt.Sprint(action.Interface())
}
}
return actions
}
// Request/Response helpers
func (h *S3TablesHandler) readRequestBody(r *http.Request, v interface{}) error {
defer r.Body.Close()
// Limit request body size to prevent unbounded reads
limitedReader := io.LimitReader(r.Body, maxRequestBodySize+1)
body, err := io.ReadAll(limitedReader)
if err != nil {
return fmt.Errorf("failed to read request body: %w", err)
}
// Check if body exceeds size limit
if len(body) > maxRequestBodySize {
return fmt.Errorf("request body too large: exceeds maximum size of %d bytes", maxRequestBodySize)
}
if len(body) == 0 {
return nil
}
if err := json.Unmarshal(body, v); err != nil {
return fmt.Errorf("failed to decode request: %w", err)
}
return nil
}
// Response writing helpers
func (h *S3TablesHandler) writeJSON(w http.ResponseWriter, status int, data interface{}) {
w.Header().Set("Content-Type", "application/x-amz-json-1.1")
w.WriteHeader(status)
if data != nil {
if err := json.NewEncoder(w).Encode(data); err != nil {
glog.Errorf("S3Tables: failed to encode response: %v", err)
}
}
}
func (h *S3TablesHandler) writeError(w http.ResponseWriter, status int, code, message string) {
w.Header().Set("Content-Type", "application/x-amz-json-1.1")
w.WriteHeader(status)
errorResponse := map[string]interface{}{
"__type": code,
"message": message,
}
if err := json.NewEncoder(w).Encode(errorResponse); err != nil {
glog.Errorf("S3Tables: failed to encode error response: %v", err)
}
}
// ARN generation helpers
func (h *S3TablesHandler) generateTableBucketARN(ownerAccountID, bucketName string) string {
return buildARN(h.region, ownerAccountID, fmt.Sprintf("bucket/%s", bucketName))
}
func (h *S3TablesHandler) generateTableARN(ownerAccountID, bucketName, tableID string) string {
return buildARN(h.region, ownerAccountID, fmt.Sprintf("bucket/%s/table/%s", bucketName, tableID))
}
func (h *S3TablesHandler) generateViewARN(ownerAccountID, bucketName, viewID string) string {
return buildARN(h.region, ownerAccountID, fmt.Sprintf("bucket/%s/view/%s", bucketName, viewID))
}
// generateS3BucketARN builds the plain S3 ARN used for IAM resource matching.
func (h *S3TablesHandler) generateS3BucketARN(bucketName string) string {
return fmt.Sprintf("arn:%s:s3:::%s", arnPartitionForRegion(h.region), bucketName)
}
func isAuthError(err error) bool {
var authErr *AuthError
return errors.As(err, &authErr) || errors.Is(err, ErrAccessDenied)
}
func (h *S3TablesHandler) readTags(ctx context.Context, client filer_pb.SeaweedFilerClient, path string) (map[string]string, error) {
data, err := h.getExtendedAttribute(ctx, client, path, ExtendedKeyTags)
if err != nil {
if errors.Is(err, ErrAttributeNotFound) {
return nil, nil
}
return nil, err
}
tags := make(map[string]string)
if err := json.Unmarshal(data, &tags); err != nil {
return nil, fmt.Errorf("failed to unmarshal tags: %w", err)
}
return tags, nil
}
func mapKeys(tags map[string]string) []string {
if len(tags) == 0 {
return nil
}
keys := make([]string, 0, len(tags))
for key := range tags {
keys = append(keys, key)
}
return keys
}