don't use S3 CreateBucket and clean up

Makefile

@@ -2,7 +2,7 @@ SHELL := /bin/sh
 
 BUF ?= buf
 
-.PHONY: buf-lint buf-generate py-sync py-test go-test test compose-up compose-up-dev run-server
+.PHONY: buf-lint buf-generate py-sync py-test go-test test compose-up compose-up-dev s3-init run-server
 
 buf-lint:
 	$(BUF) lint
@@ -32,7 +32,21 @@ compose-up:
 compose-up-dev:
 	docker compose --env-file .env.example -f deploy/docker-compose.dev.yml up
 
-run-server:
+s3-init:
+	@set -a; \
+	if [ -f .env ]; then . ./.env; fi; \
+	set +a; \
+	endpoint="$${S3_ENDPOINT:-localhost:8333}"; \
+	case "$$endpoint" in seaweedfs:8333) endpoint=localhost:8333 ;; esac; \
+	bucket="$${S3_BUCKET:-officeconvert}"; \
+	access_key="$${S3_ACCESS_KEY:-minioadmin}"; \
+	secret_key="$${S3_SECRET_KEY:-minioadmin}"; \
+	port="$${endpoint#*:}"; \
+	docker run --rm --add-host=host.docker.internal:host-gateway minio/mc:latest /bin/sh -c " \
+	  mc alias set local http://host.docker.internal:$$port '$$access_key' '$$secret_key' && \
+	  mc mb local/$$bucket --ignore-existing"
+
+run-server: s3-init
 	@set -a; \
 	if [ -f .env ]; then . ./.env; fi; \
 	set +a; \

README.md

@@ -135,7 +135,7 @@ Use `.env.example` as your baseline env configuration.
 
 ## Storage Backend Notes
 
-- Local development defaults to **SeaweedFS** (S3-compatible) via Docker Compose.
+- Local development defaults to **SeaweedFS** (S3-compatible) via Docker Compose. Compose runs an `s3-init` step that creates the dev bucket before the server starts.
 - Production can use any S3-compatible provider; **AWS S3** is the expected choice.
 - The Python server uses the `minio` Python SDK against the S3 API.
 - Runtime configuration uses `S3_*` environment variables.
@@ -166,11 +166,11 @@ S3_SECRET_KEY=...
 
 Use your bucket's regional hostname for both endpoints unless you deliberately split internal vs client-facing access. `S3_PUBLIC_ENDPOINT` must be reachable by whatever uploads and downloads via presigned URLs (clients, not just the server).
 
-On startup the server calls `CreateBucket` if the bucket is missing. In AWS it is simpler to **pre-create the bucket** and grant object permissions only (see IAM below).
+On startup the server verifies the bucket exists via HeadBucket and fails fast if it is missing. **Pre-create the bucket** before deploying (see IAM below).
 
 **IAM permissions**
 
-Scope access to the single bucket. Object keys are per-conversion prefixes, so list/delete can target the whole bucket:
+Scope access to the single bucket. Object keys are per-conversion prefixes, so list/delete can target the whole bucket. Startup verification uses HeadBucket, which is satisfied by `s3:ListBucket` on the bucket ARN:
 
 ```json
 {
@@ -178,7 +178,7 @@ Scope access to the single bucket. Object keys are per-conversion prefixes, so l
   "Statement": [
     {
       "Effect": "Allow",
-      "Action": ["s3:ListBucket"],
+      "Action": ["s3:ListBucket", "s3:HeadBucket"],
       "Resource": "arn:aws:s3:::officeconvert-prod"
     },
     {
@@ -190,8 +190,6 @@ Scope access to the single bucket. Object keys are per-conversion prefixes, so l
 }
 ```
 
-Add `s3:CreateBucket` on `arn:aws:s3:::officeconvert-prod` only if you want the server to create the bucket on first boot.
-
 **CORS**
 
 Required only if uploads or downloads go **directly from a browser** to presigned URLs. Server-side clients (`curl`, the Go client) do not need CORS. Allow `PUT` and `GET` for your web origin on the bucket.

deploy/docker-compose.yml

@@ -14,12 +14,27 @@ services:
     volumes:
       - seaweedfs_data:/data
 
+  s3-init:
+    image: minio/mc:latest
+    depends_on:
+      - seaweedfs
+    environment:
+      AWS_ACCESS_KEY_ID: ${S3_ACCESS_KEY:-minioadmin}
+      AWS_SECRET_ACCESS_KEY: ${S3_SECRET_KEY:-minioadmin}
+      S3_BUCKET: ${S3_BUCKET:-officeconvert}
+    entrypoint: >
+      /bin/sh -c "
+      until mc alias set local http://seaweedfs:8333 $$AWS_ACCESS_KEY_ID $$AWS_SECRET_ACCESS_KEY; do sleep 1; done &&
+      mc mb local/$$S3_BUCKET --ignore-existing
+      "
+
   server:
     build:
       context: ..
       dockerfile: Dockerfile.server
     depends_on:
-      - seaweedfs
+      s3-init:
+        condition: service_completed_successfully
     environment:
       S3_ENDPOINT: ${S3_ENDPOINT:-seaweedfs:8333}
       S3_PUBLIC_ENDPOINT: ${S3_PUBLIC_ENDPOINT:-localhost:8333}

docker-compose.yml

@@ -15,10 +15,25 @@ services:
     volumes:
       - seaweedfs_data:/data
 
+  s3-init:
+    image: minio/mc:latest
+    depends_on:
+      - seaweedfs
+    environment:
+      AWS_ACCESS_KEY_ID: ${S3_ACCESS_KEY:-minioadmin}
+      AWS_SECRET_ACCESS_KEY: ${S3_SECRET_KEY:-minioadmin}
+      S3_BUCKET: ${S3_BUCKET:-officeconvert}
+    entrypoint: >
+      /bin/sh -c "
+      until mc alias set local http://seaweedfs:8333 $$AWS_ACCESS_KEY_ID $$AWS_SECRET_ACCESS_KEY; do sleep 1; done &&
+      mc mb local/$$S3_BUCKET --ignore-existing
+      "
+
   server:
     image: gitea.auvem.com/end/officeconvert-server:${OFFICECONVERT_IMAGE_TAG:-latest}
     depends_on:
-      - seaweedfs
+      s3-init:
+        condition: service_completed_successfully
     environment:
       S3_ENDPOINT: ${S3_ENDPOINT:-seaweedfs:8333}
       S3_PUBLIC_ENDPOINT: ${S3_PUBLIC_ENDPOINT:-localhost:8333}

gen/go/officeconvertapi/v1/conversion.pb.go

@@ -777,7 +777,6 @@ type CreateConversionResponse struct {
 	state protoimpl.MessageState `protogen:"open.v1"`
 	// Session identifier: KSUID in standard base62 text form. Well-formed values are at most 27 characters (see https://github.com/segmentio/ksuid).
 	ConversionId    string                 `protobuf:"bytes,1,opt,name=conversion_id,json=conversionId,proto3" json:"conversion_id,omitempty"`
-	UploadBucket    string                 `protobuf:"bytes,2,opt,name=upload_bucket,json=uploadBucket,proto3" json:"upload_bucket,omitempty"`
 	UploadObjectKey string                 `protobuf:"bytes,3,opt,name=upload_object_key,json=uploadObjectKey,proto3" json:"upload_object_key,omitempty"`
 	UploadUrl       string                 `protobuf:"bytes,4,opt,name=upload_url,json=uploadUrl,proto3" json:"upload_url,omitempty"`
 	ExpiresAt       *timestamppb.Timestamp `protobuf:"bytes,5,opt,name=expires_at,json=expiresAt,proto3" json:"expires_at,omitempty"`
@@ -822,13 +821,6 @@ func (x *CreateConversionResponse) GetConversionId() string {
 	return ""
 }
 
-func (x *CreateConversionResponse) GetUploadBucket() string {
-	if x != nil {
-		return x.UploadBucket
-	}
-	return ""
-}
-
 func (x *CreateConversionResponse) GetUploadObjectKey() string {
 	if x != nil {
 		return x.UploadObjectKey
@@ -1330,15 +1322,14 @@ const file_officeconvertapi_v1_conversion_proto_rawDesc = "" +
 	"\x0fsource_filename\x18\x01 \x01(\tR\x0esourceFilename\x12;\n" +
 	"\x04full\x18\x02 \x01(\v2'.officeconvertapi.v1.SlideRasterOptionsR\x04full\x12E\n" +
 	"\tthumbnail\x18\x03 \x01(\v2'.officeconvertapi.v1.SlideRasterOptionsR\tthumbnail\x127\n" +
-	"\x05notes\x18\x04 \x01(\v2!.officeconvertapi.v1.NotesOptionsR\x05notes\"\xea\x01\n" +
+	"\x05notes\x18\x04 \x01(\v2!.officeconvertapi.v1.NotesOptionsR\x05notes\"\xda\x01\n" +
 	"\x18CreateConversionResponse\x12#\n" +
-	"\rconversion_id\x18\x01 \x01(\tR\fconversionId\x12#\n" +
+	"\rconversion_id\x18\x01 \x01(\tR\fconversionId\x12*\n" +
-	"\rupload_bucket\x18\x02 \x01(\tR\fuploadBucket\x12*\n" +
 	"\x11upload_object_key\x18\x03 \x01(\tR\x0fuploadObjectKey\x12\x1d\n" +
 	"\n" +
 	"upload_url\x18\x04 \x01(\tR\tuploadUrl\x129\n" +
 	"\n" +
-	"expires_at\x18\x05 \x01(\v2\x1a.google.protobuf.TimestampR\texpiresAt\"=\n" +
+	"expires_at\x18\x05 \x01(\v2\x1a.google.protobuf.TimestampR\texpiresAtJ\x04\b\x02\x10\x03R\rupload_bucket\"=\n" +
 	"\x16StartConversionRequest\x12#\n" +
 	"\rconversion_id\x18\x01 \x01(\tR\fconversionId\"}\n" +
 	"\x17StartConversionResponse\x12#\n" +

gen/python/officeconvertapi/v1/conversion_pb2.pyi

@@ -149,18 +149,16 @@ class CreateConversionRequest(_message.Message):
     def __init__(self, source_filename: _Optional[str] = ..., full: _Optional[_Union[SlideRasterOptions, _Mapping]] = ..., thumbnail: _Optional[_Union[SlideRasterOptions, _Mapping]] = ..., notes: _Optional[_Union[NotesOptions, _Mapping]] = ...) -> None: ...
 
 class CreateConversionResponse(_message.Message):
-    __slots__ = ("conversion_id", "upload_bucket", "upload_object_key", "upload_url", "expires_at")
+    __slots__ = ("conversion_id", "upload_object_key", "upload_url", "expires_at")
     CONVERSION_ID_FIELD_NUMBER: _ClassVar[int]
-    UPLOAD_BUCKET_FIELD_NUMBER: _ClassVar[int]
     UPLOAD_OBJECT_KEY_FIELD_NUMBER: _ClassVar[int]
     UPLOAD_URL_FIELD_NUMBER: _ClassVar[int]
     EXPIRES_AT_FIELD_NUMBER: _ClassVar[int]
     conversion_id: str
-    upload_bucket: str
     upload_object_key: str
     upload_url: str
     expires_at: _timestamp_pb2.Timestamp
-    def __init__(self, conversion_id: _Optional[str] = ..., upload_bucket: _Optional[str] = ..., upload_object_key: _Optional[str] = ..., upload_url: _Optional[str] = ..., expires_at: _Optional[_Union[datetime.datetime, _timestamp_pb2.Timestamp, _Mapping]] = ...) -> None: ...
+    def __init__(self, conversion_id: _Optional[str] = ..., upload_object_key: _Optional[str] = ..., upload_url: _Optional[str] = ..., expires_at: _Optional[_Union[datetime.datetime, _timestamp_pb2.Timestamp, _Mapping]] = ...) -> None: ...
 
 class StartConversionRequest(_message.Message):
     __slots__ = ("conversion_id",)

proto/officeconvertapi/v1/conversion.proto

@@ -136,7 +136,8 @@ message CreateConversionRequest {
 message CreateConversionResponse {
   // Session identifier: KSUID in standard base62 text form. Well-formed values are at most 27 characters (see https://github.com/segmentio/ksuid).
   string conversion_id = 1;
-  string upload_bucket = 2;
+  reserved 2;
+  reserved "upload_bucket";
   string upload_object_key = 3;
   string upload_url = 4;
   google.protobuf.Timestamp expires_at = 5;

python/packages/server/src/officeconvert_server/app.py

@@ -46,6 +46,7 @@ def create_app() -> ConversionServiceASGIApplication:
     _configure_application_logging()
     config = load_server_config()
     store = S3Store(
+        bucket=config.s3_bucket,
         endpoint=config.s3_endpoint,
         access_key=config.s3_access_key,
         secret_key=config.s3_secret_key,
@@ -58,10 +59,10 @@ def create_app() -> ConversionServiceASGIApplication:
         store.enable_http_trace(sys.stderr)
         logger.warning("OFFICECONVERT_S3_TRACE enabled: S3 HTTP dumps on stderr")
     try:
-        store.ensure_bucket(config.s3_bucket)
+        store.require_bucket()
     except S3Error as exc:
         log_s3_error(
-            "ensure_bucket",
+            "require_bucket",
             endpoint=config.s3_endpoint,
             secure=config.s3_secure,
             exc=exc,

python/packages/server/src/officeconvert_server/service.py

@@ -128,7 +128,6 @@ class ConversionServiceImpl(conversion_connect.ConversionService):
 
         try:
             upload_url = self._store.presigned_put_url(
-                self._config.s3_bucket,
                 upload_key,
                 ttl_seconds=self._config.s3_session_ttl_seconds,
             )
@@ -158,7 +157,6 @@ class ConversionServiceImpl(conversion_connect.ConversionService):
 
         return conversion_pb2.CreateConversionResponse(
             conversion_id=conversion_id,
-            upload_bucket=self._config.s3_bucket,
             upload_object_key=upload_key,
             upload_url=upload_url,
             expires_at=_to_timestamp(expires_at),
@@ -257,7 +255,6 @@ class ConversionServiceImpl(conversion_connect.ConversionService):
         await self._cleanup_local_artifacts(session)
         await asyncio.to_thread(
             self._store.remove_prefix,
-            self._config.s3_bucket,
             session.object_prefix,
         )
         return conversion_pb2.DeleteConversionResponse(
@@ -289,7 +286,6 @@ class ConversionServiceImpl(conversion_connect.ConversionService):
         try:
             await asyncio.to_thread(
                 self._store.fget_object,
-                self._config.s3_bucket,
                 session.upload_object_key,
                 source_path,
             )
@@ -433,9 +429,8 @@ class ConversionServiceImpl(conversion_connect.ConversionService):
             object_key = (
                 f"{session.object_prefix}output/slide-{slide.index:04d}{slide.image_path.suffix}"
             )
-            self._store.fput_object(self._config.s3_bucket, object_key, slide.image_path)
+            self._store.fput_object(object_key, slide.image_path)
             image_url = self._store.presigned_get_url(
-                self._config.s3_bucket,
                 object_key,
                 ttl_seconds=self._config.s3_session_ttl_seconds,
             )
@@ -447,12 +442,10 @@ class ConversionServiceImpl(conversion_connect.ConversionService):
                 f"{slide.thumbnail_path.suffix}"
             )
             self._store.fput_object(
-                self._config.s3_bucket,
                 thumbnail_object_key,
                 slide.thumbnail_path,
             )
             thumbnail_image_url = self._store.presigned_get_url(
-                self._config.s3_bucket,
                 thumbnail_object_key,
                 ttl_seconds=self._config.s3_session_ttl_seconds,
             )
@@ -514,7 +507,6 @@ class ConversionServiceImpl(conversion_connect.ConversionService):
             await asyncio.sleep(self._config.conversion_cleanup_delay_seconds)
             await asyncio.to_thread(
                 self._store.remove_prefix,
-                self._config.s3_bucket,
                 session.object_prefix,
             )
         except asyncio.CancelledError:

python/packages/server/src/officeconvert_server/storage.py

@@ -6,7 +6,6 @@ import logging
 from datetime import timedelta
 from pathlib import Path
 from typing import TextIO
-from urllib.parse import urlparse
 
 from minio import Minio
 from minio.deleteobjects import DeleteObject
@@ -54,6 +53,7 @@ class S3Store:
     def __init__(
         self,
         *,
+        bucket: str,
         endpoint: str,
         access_key: str,
         secret_key: str,
@@ -63,6 +63,7 @@ class S3Store:
         public_secure: bool,
     ) -> None:
         """Initialize S3 clients for internal and public URL generation."""
+        self._bucket = bucket
         self._client = Minio(
             endpoint,
             access_key=access_key,
@@ -83,55 +84,44 @@ class S3Store:
         self._client.trace_on(stream)
         self._public_client.trace_on(stream)
 
-    def ensure_bucket(self, bucket_name: str) -> None:
+    def require_bucket(self) -> None:
-        """Create a bucket if it does not already exist.
+        """Verify the configured bucket exists before serving traffic."""
+        if not self._client.bucket_exists(self._bucket):
+            raise RuntimeError(
+                f"S3 bucket {self._bucket!r} does not exist; create it before starting the server"
+            )
 
-        Tries CreateBucket first (idempotent on SeaweedFS and when the caller
+    def presigned_put_url(self, object_key: str, *, ttl_seconds: int) -> str:
-        owns the bucket). AWS production IAM often grants object access only on
-        a pre-provisioned bucket; in that case CreateBucket returns
-        AccessDenied even though HeadBucket succeeds.
-        """
-        try:
-            self._client.make_bucket(bucket_name)
-        except S3Error as exc:
-            if exc.code in ("BucketAlreadyOwnedByYou", "BucketAlreadyExists"):
-                return
-            if exc.code in ("AccessDenied", "Forbidden"):
-                if self._client.bucket_exists(bucket_name):
-                    return
-            raise
-
-    def presigned_put_url(self, bucket_name: str, object_key: str, *, ttl_seconds: int) -> str:
         """Generate a presigned PUT URL for a single object upload."""
         return self._public_client.presigned_put_object(
-            bucket_name,
+            self._bucket,
             object_key,
             expires=timedelta(seconds=ttl_seconds),
         )
 
-    def presigned_get_url(self, bucket_name: str, object_key: str, *, ttl_seconds: int) -> str:
+    def presigned_get_url(self, object_key: str, *, ttl_seconds: int) -> str:
         """Generate a presigned GET URL for downloading one object."""
         return self._public_client.presigned_get_object(
-            bucket_name,
+            self._bucket,
             object_key,
             expires=timedelta(seconds=ttl_seconds),
         )
 
-    def fget_object(self, bucket_name: str, object_key: str, output_path: Path) -> None:
+    def fget_object(self, object_key: str, output_path: Path) -> None:
         """Download one object from storage to a local filesystem path."""
         output_path.parent.mkdir(parents=True, exist_ok=True)
-        self._client.fget_object(bucket_name, object_key, str(output_path))
+        self._client.fget_object(self._bucket, object_key, str(output_path))
 
-    def fput_object(self, bucket_name: str, object_key: str, source_path: Path) -> None:
+    def fput_object(self, object_key: str, source_path: Path) -> None:
         """Upload one local filesystem object to storage."""
-        self._client.fput_object(bucket_name, object_key, str(source_path))
+        self._client.fput_object(self._bucket, object_key, str(source_path))
 
-    def remove_prefix(self, bucket_name: str, prefix: str) -> None:
+    def remove_prefix(self, prefix: str) -> None:
-        """Remove all objects under a key prefix within a bucket."""
+        """Remove all objects under a key prefix within the configured bucket."""
         normalized_prefix = prefix if prefix.endswith("/") else f"{prefix}/"
         objects = list(
             self._client.list_objects(
-                bucket_name,
+                self._bucket,
                 prefix=normalized_prefix,
                 recursive=True,
             )
@@ -149,7 +139,7 @@ class S3Store:
             delete_requests.append(DeleteObject(object_name))
 
         errors = self._client.remove_objects(
-            bucket_name,
+            self._bucket,
             delete_requests,
         )
         for err in errors:
@@ -158,10 +148,3 @@ class S3Store:
             raise RuntimeError(
                 f"failed to delete object {object_name}: {message}"
             )
-
-
-def object_key_from_presigned_url(url: str) -> str:
-    """Extract object key from a presigned URL path for diagnostics."""
-    path = urlparse(url).path
-    path_parts = [part for part in path.split("/") if part]
-    return "/".join(path_parts[1:]) if len(path_parts) >= 2 else ""