kqjy/MyFSIO
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 13 Wiki Activity

Merge pull request 'Release v0.1.3' (#4) from next into main

Browse Source 
Reviewed-on: #4
This commit was merged in pull request #4.
This commit is contained in:
kqjy
2025-12-03 04:14:57 +00:00
parent 233780617f 114e684cb8
commit bae5009ec4
25 changed files with 5408 additions and 141 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
README.md
View File

@@ -86,7 +86,7 @@ Presigned URLs follow the AWS CLI playbook:
| `AWS_REGION` | `us-east-1` | Region used in Signature V4 scope |
| `AWS_SERVICE` | `s3` | Service used in Signature V4 scope |
> Buckets now live directly under `data/` while system metadata (versions, IAM, bucket policies, multipart uploads, etc.) lives in `data/.myfsio.sys`. Existing installs can keep their environment variables, but the defaults now match MinIO's `data/.system` pattern for easier bind-mounting.
> Buckets now live directly under `data/` while system metadata (versions, IAM, bucket policies, multipart uploads, etc.) lives in `data/.myfsio.sys`.
## API Cheatsheet (IAM headers required)

27
app/__init__.py
View File

@@ -18,8 +18,10 @@ from werkzeug.middleware.proxy_fix import ProxyFix
from .bucket_policies import BucketPolicyStore
from .config import AppConfig
from .connections import ConnectionStore
from .encryption import EncryptionManager
from .extensions import limiter, csrf
from .iam import IamService
from .kms import KMSManager
from .replication import ReplicationManager
from .secret_store import EphemeralSecretStore
from .storage import ObjectStorage
@@ -77,6 +79,26 @@ def create_app(
connections = ConnectionStore(connections_path)
replication = ReplicationManager(storage, connections, replication_rules_path)
# Initialize encryption and KMS
encryption_config = {
"encryption_enabled": app.config.get("ENCRYPTION_ENABLED", False),
"encryption_master_key_path": app.config.get("ENCRYPTION_MASTER_KEY_PATH"),
"default_encryption_algorithm": app.config.get("DEFAULT_ENCRYPTION_ALGORITHM", "AES256"),
}
encryption_manager = EncryptionManager(encryption_config)
kms_manager = None
if app.config.get("KMS_ENABLED", False):
kms_keys_path = Path(app.config.get("KMS_KEYS_PATH", ""))
kms_master_key_path = Path(app.config.get("ENCRYPTION_MASTER_KEY_PATH", ""))
kms_manager = KMSManager(kms_keys_path, kms_master_key_path)
encryption_manager.set_kms_provider(kms_manager)
# Wrap storage with encryption layer if encryption is enabled
if app.config.get("ENCRYPTION_ENABLED", False):
from .encrypted_storage import EncryptedObjectStorage
storage = EncryptedObjectStorage(storage, encryption_manager)
app.extensions["object_storage"] = storage
app.extensions["iam"] = iam
@@ -85,6 +107,8 @@ def create_app(
app.extensions["limiter"] = limiter
app.extensions["connections"] = connections
app.extensions["replication"] = replication
app.extensions["encryption"] = encryption_manager
app.extensions["kms"] = kms_manager
@app.errorhandler(500)
def internal_error(error):
@@ -119,9 +143,12 @@ def create_app(
if include_api:
from .s3_api import s3_api_bp
from .kms_api import kms_api_bp
app.register_blueprint(s3_api_bp)
app.register_blueprint(kms_api_bp)
csrf.exempt(s3_api_bp)
csrf.exempt(kms_api_bp)
if include_ui:
from .ui import ui_bp

38
app/bucket_policies.py
View File

@@ -11,17 +11,51 @@ from typing import Any, Dict, Iterable, List, Optional, Sequence
RESOURCE_PREFIX = "arn:aws:s3:::"
ACTION_ALIASES = {
"s3:getobject": "read",
"s3:getobjectversion": "read",
# List actions
"s3:listbucket": "list",
"s3:listallmybuckets": "list",
"s3:listbucketversions": "list",
"s3:listmultipartuploads": "list",
"s3:listparts": "list",
# Read actions
"s3:getobject": "read",
"s3:getobjectversion": "read",
"s3:getobjecttagging": "read",
"s3:getobjectversiontagging": "read",
"s3:getobjectacl": "read",
"s3:getbucketversioning": "read",
"s3:headobject": "read",
"s3:headbucket": "read",
# Write actions
"s3:putobject": "write",
"s3:createbucket": "write",
"s3:putobjecttagging": "write",
"s3:putbucketversioning": "write",
"s3:createmultipartupload": "write",
"s3:uploadpart": "write",
"s3:completemultipartupload": "write",
"s3:abortmultipartupload": "write",
"s3:copyobject": "write",
# Delete actions
"s3:deleteobject": "delete",
"s3:deleteobjectversion": "delete",
"s3:deletebucket": "delete",
"s3:deleteobjecttagging": "delete",
# Share actions (ACL)
"s3:putobjectacl": "share",
"s3:putbucketacl": "share",
"s3:getbucketacl": "share",
# Policy actions
"s3:putbucketpolicy": "policy",
"s3:getbucketpolicy": "policy",
"s3:deletebucketpolicy": "policy",
# Replication actions
"s3:getreplicationconfiguration": "replication",
"s3:putreplicationconfiguration": "replication",
"s3:deletereplicationconfiguration": "replication",
"s3:replicateobject": "replication",
"s3:replicatetags": "replication",
"s3:replicatedelete": "replication",
}

29
app/config.py
View File

@@ -50,6 +50,7 @@ class AppConfig:
aws_service: str
ui_enforce_bucket_policies: bool
log_level: str
log_to_file: bool
log_path: Path
log_max_bytes: int
log_backup_count: int
@@ -66,6 +67,11 @@ class AppConfig:
stream_chunk_size: int
multipart_min_part_size: int
bucket_stats_cache_ttl: int
encryption_enabled: bool
encryption_master_key_path: Path
kms_enabled: bool
kms_keys_path: Path
default_encryption_algorithm: str
@classmethod
def from_env(cls, overrides: Optional[Dict[str, Any]] = None) -> "AppConfig":
@@ -126,6 +132,7 @@ class AppConfig:
aws_service = str(_get("AWS_SERVICE", "s3"))
enforce_ui_policies = str(_get("UI_ENFORCE_BUCKET_POLICIES", "0")).lower() in {"1", "true", "yes", "on"}
log_level = str(_get("LOG_LEVEL", "INFO")).upper()
log_to_file = str(_get("LOG_TO_FILE", "1")).lower() in {"1", "true", "yes", "on"}
log_dir = Path(_get("LOG_DIR", PROJECT_ROOT / "logs")).resolve()
log_dir.mkdir(parents=True, exist_ok=True)
log_path = log_dir / str(_get("LOG_FILE", "app.log"))
@@ -155,6 +162,14 @@ class AppConfig:
])
session_lifetime_days = int(_get("SESSION_LIFETIME_DAYS", 30))
bucket_stats_cache_ttl = int(_get("BUCKET_STATS_CACHE_TTL", 60)) # Default 60 seconds
# Encryption settings
encryption_enabled = str(_get("ENCRYPTION_ENABLED", "0")).lower() in {"1", "true", "yes", "on"}
encryption_keys_dir = storage_root / ".myfsio.sys" / "keys"
encryption_master_key_path = Path(_get("ENCRYPTION_MASTER_KEY_PATH", encryption_keys_dir / "master.key")).resolve()
kms_enabled = str(_get("KMS_ENABLED", "0")).lower() in {"1", "true", "yes", "on"}
kms_keys_path = Path(_get("KMS_KEYS_PATH", encryption_keys_dir / "kms_keys.json")).resolve()
default_encryption_algorithm = str(_get("DEFAULT_ENCRYPTION_ALGORITHM", "AES256"))
return cls(storage_root=storage_root,
max_upload_size=max_upload_size,
@@ -167,6 +182,7 @@ class AppConfig:
aws_service=aws_service,
ui_enforce_bucket_policies=enforce_ui_policies,
log_level=log_level,
log_to_file=log_to_file,
log_path=log_path,
log_max_bytes=log_max_bytes,
log_backup_count=log_backup_count,
@@ -182,7 +198,12 @@ class AppConfig:
secret_ttl_seconds=secret_ttl_seconds,
stream_chunk_size=stream_chunk_size,
multipart_min_part_size=multipart_min_part_size,
bucket_stats_cache_ttl=bucket_stats_cache_ttl)
bucket_stats_cache_ttl=bucket_stats_cache_ttl,
encryption_enabled=encryption_enabled,
encryption_master_key_path=encryption_master_key_path,
kms_enabled=kms_enabled,
kms_keys_path=kms_keys_path,
default_encryption_algorithm=default_encryption_algorithm)
def to_flask_config(self) -> Dict[str, Any]:
return {
@@ -204,6 +225,7 @@ class AppConfig:
"MULTIPART_MIN_PART_SIZE": self.multipart_min_part_size,
"BUCKET_STATS_CACHE_TTL": self.bucket_stats_cache_ttl,
"LOG_LEVEL": self.log_level,
"LOG_TO_FILE": self.log_to_file,
"LOG_FILE": str(self.log_path),
"LOG_MAX_BYTES": self.log_max_bytes,
"LOG_BACKUP_COUNT": self.log_backup_count,
@@ -213,4 +235,9 @@ class AppConfig:
"CORS_METHODS": self.cors_methods,
"CORS_ALLOW_HEADERS": self.cors_allow_headers,
"SESSION_LIFETIME_DAYS": self.session_lifetime_days,
"ENCRYPTION_ENABLED": self.encryption_enabled,
"ENCRYPTION_MASTER_KEY_PATH": str(self.encryption_master_key_path),
"KMS_ENABLED": self.kms_enabled,
"KMS_KEYS_PATH": str(self.kms_keys_path),
"DEFAULT_ENCRYPTION_ALGORITHM": self.default_encryption_algorithm,
}

276
app/encrypted_storage.py Normal file
View File

@@ -0,0 +1,276 @@
"""Encrypted storage layer that wraps ObjectStorage with encryption support."""
from __future__ import annotations
import io
from pathlib import Path
from typing import Any, BinaryIO, Dict, Optional
from .encryption import EncryptionManager, EncryptionMetadata, EncryptionError
from .storage import ObjectStorage, ObjectMeta, StorageError
class EncryptedObjectStorage:
"""Object storage with transparent server-side encryption.
This class wraps ObjectStorage and provides transparent encryption/decryption
of objects based on bucket encryption configuration.
Encryption is applied when:
1. Bucket has default encryption configured (SSE-S3 or SSE-KMS)
2. Client explicitly requests encryption via headers
The encryption metadata is stored alongside object metadata.
"""
STREAMING_THRESHOLD = 64 * 1024
def __init__(self, storage: ObjectStorage, encryption_manager: EncryptionManager):
self.storage = storage
self.encryption = encryption_manager
@property
def root(self) -> Path:
return self.storage.root
def _should_encrypt(self, bucket_name: str,
server_side_encryption: str | None = None) -> tuple[bool, str, str | None]:
"""Determine if object should be encrypted.
Returns:
Tuple of (should_encrypt, algorithm, kms_key_id)
"""
if not self.encryption.enabled:
return False, "", None
if server_side_encryption:
if server_side_encryption == "AES256":
return True, "AES256", None
elif server_side_encryption.startswith("aws:kms"):
parts = server_side_encryption.split(":")
kms_key_id = parts[2] if len(parts) > 2 else None
return True, "aws:kms", kms_key_id
try:
encryption_config = self.storage.get_bucket_encryption(bucket_name)
if encryption_config and encryption_config.get("Rules"):
rule = encryption_config["Rules"][0]
# AWS format: Rules[].ApplyServerSideEncryptionByDefault.SSEAlgorithm
sse_default = rule.get("ApplyServerSideEncryptionByDefault", {})
algorithm = sse_default.get("SSEAlgorithm", "AES256")
kms_key_id = sse_default.get("KMSMasterKeyID")
return True, algorithm, kms_key_id
except StorageError:
pass
return False, "", None
def _is_encrypted(self, metadata: Dict[str, str]) -> bool:
"""Check if object is encrypted based on its metadata."""
return "x-amz-server-side-encryption" in metadata
def put_object(
self,
bucket_name: str,
object_key: str,
stream: BinaryIO,
*,
metadata: Optional[Dict[str, str]] = None,
server_side_encryption: Optional[str] = None,
kms_key_id: Optional[str] = None,
) -> ObjectMeta:
"""Store an object, optionally with encryption.
Args:
bucket_name: Name of the bucket
object_key: Key for the object
stream: Binary stream of object data
metadata: Optional user metadata
server_side_encryption: Encryption algorithm ("AES256" or "aws:kms")
kms_key_id: KMS key ID (for aws:kms encryption)
Returns:
ObjectMeta with object information
"""
should_encrypt, algorithm, detected_kms_key = self._should_encrypt(
bucket_name, server_side_encryption
)
if kms_key_id is None:
kms_key_id = detected_kms_key
if should_encrypt:
data = stream.read()
try:
ciphertext, enc_metadata = self.encryption.encrypt_object(
data,
algorithm=algorithm,
kms_key_id=kms_key_id,
context={"bucket": bucket_name, "key": object_key},
)
combined_metadata = metadata.copy() if metadata else {}
combined_metadata.update(enc_metadata.to_dict())
encrypted_stream = io.BytesIO(ciphertext)
result = self.storage.put_object(
bucket_name,
object_key,
encrypted_stream,
metadata=combined_metadata,
)
result.metadata = combined_metadata
return result
except EncryptionError as exc:
raise StorageError(f"Encryption failed: {exc}") from exc
else:
return self.storage.put_object(
bucket_name,
object_key,
stream,
metadata=metadata,
)
def get_object_data(self, bucket_name: str, object_key: str) -> tuple[bytes, Dict[str, str]]:
"""Get object data, decrypting if necessary.
Returns:
Tuple of (data, metadata)
"""
path = self.storage.get_object_path(bucket_name, object_key)
metadata = self.storage.get_object_metadata(bucket_name, object_key)
with path.open("rb") as f:
data = f.read()
enc_metadata = EncryptionMetadata.from_dict(metadata)
if enc_metadata:
try:
data = self.encryption.decrypt_object(
data,
enc_metadata,
context={"bucket": bucket_name, "key": object_key},
)
except EncryptionError as exc:
raise StorageError(f"Decryption failed: {exc}") from exc
clean_metadata = {
k: v for k, v in metadata.items()
if not k.startswith("x-amz-encryption")
and k != "x-amz-encrypted-data-key"
}
return data, clean_metadata
def get_object_stream(self, bucket_name: str, object_key: str) -> tuple[BinaryIO, Dict[str, str], int]:
"""Get object as a stream, decrypting if necessary.
Returns:
Tuple of (stream, metadata, original_size)
"""
data, metadata = self.get_object_data(bucket_name, object_key)
return io.BytesIO(data), metadata, len(data)
def list_buckets(self):
return self.storage.list_buckets()
def bucket_exists(self, bucket_name: str) -> bool:
return self.storage.bucket_exists(bucket_name)
def create_bucket(self, bucket_name: str) -> None:
return self.storage.create_bucket(bucket_name)
def delete_bucket(self, bucket_name: str) -> None:
return self.storage.delete_bucket(bucket_name)
def bucket_stats(self, bucket_name: str, cache_ttl: int = 60):
return self.storage.bucket_stats(bucket_name, cache_ttl)
def list_objects(self, bucket_name: str):
return self.storage.list_objects(bucket_name)
def get_object_path(self, bucket_name: str, object_key: str):
return self.storage.get_object_path(bucket_name, object_key)
def get_object_metadata(self, bucket_name: str, object_key: str):
return self.storage.get_object_metadata(bucket_name, object_key)
def delete_object(self, bucket_name: str, object_key: str) -> None:
return self.storage.delete_object(bucket_name, object_key)
def purge_object(self, bucket_name: str, object_key: str) -> None:
return self.storage.purge_object(bucket_name, object_key)
def is_versioning_enabled(self, bucket_name: str) -> bool:
return self.storage.is_versioning_enabled(bucket_name)
def set_bucket_versioning(self, bucket_name: str, enabled: bool) -> None:
return self.storage.set_bucket_versioning(bucket_name, enabled)
def get_bucket_tags(self, bucket_name: str):
return self.storage.get_bucket_tags(bucket_name)
def set_bucket_tags(self, bucket_name: str, tags):
return self.storage.set_bucket_tags(bucket_name, tags)
def get_bucket_cors(self, bucket_name: str):
return self.storage.get_bucket_cors(bucket_name)
def set_bucket_cors(self, bucket_name: str, rules):
return self.storage.set_bucket_cors(bucket_name, rules)
def get_bucket_encryption(self, bucket_name: str):
return self.storage.get_bucket_encryption(bucket_name)
def set_bucket_encryption(self, bucket_name: str, config_payload):
return self.storage.set_bucket_encryption(bucket_name, config_payload)
def get_bucket_lifecycle(self, bucket_name: str):
return self.storage.get_bucket_lifecycle(bucket_name)
def set_bucket_lifecycle(self, bucket_name: str, rules):
return self.storage.set_bucket_lifecycle(bucket_name, rules)
def get_object_tags(self, bucket_name: str, object_key: str):
return self.storage.get_object_tags(bucket_name, object_key)
def set_object_tags(self, bucket_name: str, object_key: str, tags):
return self.storage.set_object_tags(bucket_name, object_key, tags)
def delete_object_tags(self, bucket_name: str, object_key: str):
return self.storage.delete_object_tags(bucket_name, object_key)
def list_object_versions(self, bucket_name: str, object_key: str):
return self.storage.list_object_versions(bucket_name, object_key)
def restore_object_version(self, bucket_name: str, object_key: str, version_id: str):
return self.storage.restore_object_version(bucket_name, object_key, version_id)
def list_orphaned_objects(self, bucket_name: str):
return self.storage.list_orphaned_objects(bucket_name)
def initiate_multipart_upload(self, bucket_name: str, object_key: str, *, metadata=None) -> str:
return self.storage.initiate_multipart_upload(bucket_name, object_key, metadata=metadata)
def upload_multipart_part(self, bucket_name: str, upload_id: str, part_number: int, stream: BinaryIO) -> str:
return self.storage.upload_multipart_part(bucket_name, upload_id, part_number, stream)
def complete_multipart_upload(self, bucket_name: str, upload_id: str, ordered_parts):
return self.storage.complete_multipart_upload(bucket_name, upload_id, ordered_parts)
def abort_multipart_upload(self, bucket_name: str, upload_id: str) -> None:
return self.storage.abort_multipart_upload(bucket_name, upload_id)
def list_multipart_parts(self, bucket_name: str, upload_id: str):
return self.storage.list_multipart_parts(bucket_name, upload_id)
def get_bucket_quota(self, bucket_name: str):
return self.storage.get_bucket_quota(bucket_name)
def set_bucket_quota(self, bucket_name: str, *, max_bytes=None, max_objects=None):
return self.storage.set_bucket_quota(bucket_name, max_bytes=max_bytes, max_objects=max_objects)
def _compute_etag(self, path: Path) -> str:
return self.storage._compute_etag(path)

395
app/encryption.py Normal file
View File

@@ -0,0 +1,395 @@
"""Encryption providers for server-side and client-side encryption."""
from __future__ import annotations
import base64
import io
import json
import secrets
from dataclasses import dataclass
from pathlib import Path
from typing import Any, BinaryIO, Dict, Generator, Optional
from cryptography.hazmat.primitives.ciphers.aead import AESGCM
class EncryptionError(Exception):
"""Raised when encryption/decryption fails."""
@dataclass
class EncryptionResult:
"""Result of encrypting data."""
ciphertext: bytes
nonce: bytes
key_id: str
encrypted_data_key: bytes
@dataclass
class EncryptionMetadata:
"""Metadata stored with encrypted objects."""
algorithm: str
key_id: str
nonce: bytes
encrypted_data_key: bytes
def to_dict(self) -> Dict[str, str]:
return {
"x-amz-server-side-encryption": self.algorithm,
"x-amz-encryption-key-id": self.key_id,
"x-amz-encryption-nonce": base64.b64encode(self.nonce).decode(),
"x-amz-encrypted-data-key": base64.b64encode(self.encrypted_data_key).decode(),
}
@classmethod
def from_dict(cls, data: Dict[str, str]) -> Optional["EncryptionMetadata"]:
algorithm = data.get("x-amz-server-side-encryption")
if not algorithm:
return None
try:
return cls(
algorithm=algorithm,
key_id=data.get("x-amz-encryption-key-id", "local"),
nonce=base64.b64decode(data.get("x-amz-encryption-nonce", "")),
encrypted_data_key=base64.b64decode(data.get("x-amz-encrypted-data-key", "")),
)
except Exception:
return None
class EncryptionProvider:
"""Base class for encryption providers."""
def encrypt(self, plaintext: bytes, context: Dict[str, str] | None = None) -> EncryptionResult:
raise NotImplementedError
def decrypt(self, ciphertext: bytes, nonce: bytes, encrypted_data_key: bytes,
key_id: str, context: Dict[str, str] | None = None) -> bytes:
raise NotImplementedError
def generate_data_key(self) -> tuple[bytes, bytes]:
"""Generate a data key and its encrypted form.
Returns:
Tuple of (plaintext_key, encrypted_key)
"""
raise NotImplementedError
class LocalKeyEncryption(EncryptionProvider):
"""SSE-S3 style encryption using a local master key.
Uses envelope encryption:
1. Generate a unique data key for each object
2. Encrypt the data with the data key (AES-256-GCM)
3. Encrypt the data key with the master key
4. Store the encrypted data key alongside the ciphertext
"""
KEY_ID = "local"
def __init__(self, master_key_path: Path):
self.master_key_path = master_key_path
self._master_key: bytes | None = None
@property
def master_key(self) -> bytes:
if self._master_key is None:
self._master_key = self._load_or_create_master_key()
return self._master_key
def _load_or_create_master_key(self) -> bytes:
"""Load master key from file or generate a new one."""
if self.master_key_path.exists():
try:
return base64.b64decode(self.master_key_path.read_text().strip())
except Exception as exc:
raise EncryptionError(f"Failed to load master key: {exc}") from exc
key = secrets.token_bytes(32)
try:
self.master_key_path.parent.mkdir(parents=True, exist_ok=True)
self.master_key_path.write_text(base64.b64encode(key).decode())
except OSError as exc:
raise EncryptionError(f"Failed to save master key: {exc}") from exc
return key
def _encrypt_data_key(self, data_key: bytes) -> bytes:
"""Encrypt the data key with the master key."""
aesgcm = AESGCM(self.master_key)
nonce = secrets.token_bytes(12)
encrypted = aesgcm.encrypt(nonce, data_key, None)
return nonce + encrypted
def _decrypt_data_key(self, encrypted_data_key: bytes) -> bytes:
"""Decrypt the data key using the master key."""
if len(encrypted_data_key) < 12 + 32 + 16: # nonce + key + tag
raise EncryptionError("Invalid encrypted data key")
aesgcm = AESGCM(self.master_key)
nonce = encrypted_data_key[:12]
ciphertext = encrypted_data_key[12:]
try:
return aesgcm.decrypt(nonce, ciphertext, None)
except Exception as exc:
raise EncryptionError(f"Failed to decrypt data key: {exc}") from exc
def generate_data_key(self) -> tuple[bytes, bytes]:
"""Generate a data key and its encrypted form."""
plaintext_key = secrets.token_bytes(32)
encrypted_key = self._encrypt_data_key(plaintext_key)
return plaintext_key, encrypted_key
def encrypt(self, plaintext: bytes, context: Dict[str, str] | None = None) -> EncryptionResult:
"""Encrypt data using envelope encryption."""
data_key, encrypted_data_key = self.generate_data_key()
aesgcm = AESGCM(data_key)
nonce = secrets.token_bytes(12)
ciphertext = aesgcm.encrypt(nonce, plaintext, None)
return EncryptionResult(
ciphertext=ciphertext,
nonce=nonce,
key_id=self.KEY_ID,
encrypted_data_key=encrypted_data_key,
)
def decrypt(self, ciphertext: bytes, nonce: bytes, encrypted_data_key: bytes,
key_id: str, context: Dict[str, str] | None = None) -> bytes:
"""Decrypt data using envelope encryption."""
# Decrypt the data key
data_key = self._decrypt_data_key(encrypted_data_key)
# Decrypt the data
aesgcm = AESGCM(data_key)
try:
return aesgcm.decrypt(nonce, ciphertext, None)
except Exception as exc:
raise EncryptionError(f"Failed to decrypt data: {exc}") from exc
class StreamingEncryptor:
"""Encrypts/decrypts data in streaming fashion for large files.
For large files, we encrypt in chunks. Each chunk is encrypted with the
same data key but a unique nonce derived from the base nonce + chunk index.
"""
CHUNK_SIZE = 64 * 1024
HEADER_SIZE = 4
def __init__(self, provider: EncryptionProvider, chunk_size: int = CHUNK_SIZE):
self.provider = provider
self.chunk_size = chunk_size
def _derive_chunk_nonce(self, base_nonce: bytes, chunk_index: int) -> bytes:
"""Derive a unique nonce for each chunk."""
# XOR the base nonce with the chunk index
nonce_int = int.from_bytes(base_nonce, "big")
derived = nonce_int ^ chunk_index
return derived.to_bytes(12, "big")
def encrypt_stream(self, stream: BinaryIO,
context: Dict[str, str] | None = None) -> tuple[BinaryIO, EncryptionMetadata]:
"""Encrypt a stream and return encrypted stream + metadata."""
data_key, encrypted_data_key = self.provider.generate_data_key()
base_nonce = secrets.token_bytes(12)
aesgcm = AESGCM(data_key)
encrypted_chunks = []
chunk_index = 0
while True:
chunk = stream.read(self.chunk_size)
if not chunk:
break
chunk_nonce = self._derive_chunk_nonce(base_nonce, chunk_index)
encrypted_chunk = aesgcm.encrypt(chunk_nonce, chunk, None)
size_prefix = len(encrypted_chunk).to_bytes(self.HEADER_SIZE, "big")
encrypted_chunks.append(size_prefix + encrypted_chunk)
chunk_index += 1
header = chunk_index.to_bytes(4, "big")
encrypted_data = header + b"".join(encrypted_chunks)
metadata = EncryptionMetadata(
algorithm="AES256",
key_id=self.provider.KEY_ID if hasattr(self.provider, "KEY_ID") else "local",
nonce=base_nonce,
encrypted_data_key=encrypted_data_key,
)
return io.BytesIO(encrypted_data), metadata
def decrypt_stream(self, stream: BinaryIO, metadata: EncryptionMetadata) -> BinaryIO:
"""Decrypt a stream using the provided metadata."""
if isinstance(self.provider, LocalKeyEncryption):
data_key = self.provider._decrypt_data_key(metadata.encrypted_data_key)
else:
raise EncryptionError("Unsupported provider for streaming decryption")
aesgcm = AESGCM(data_key)
base_nonce = metadata.nonce
chunk_count_bytes = stream.read(4)
if len(chunk_count_bytes) < 4:
raise EncryptionError("Invalid encrypted stream: missing header")
chunk_count = int.from_bytes(chunk_count_bytes, "big")
decrypted_chunks = []
for chunk_index in range(chunk_count):
size_bytes = stream.read(self.HEADER_SIZE)
if len(size_bytes) < self.HEADER_SIZE:
raise EncryptionError(f"Invalid encrypted stream: truncated at chunk {chunk_index}")
chunk_size = int.from_bytes(size_bytes, "big")
encrypted_chunk = stream.read(chunk_size)
if len(encrypted_chunk) < chunk_size:
raise EncryptionError(f"Invalid encrypted stream: incomplete chunk {chunk_index}")
chunk_nonce = self._derive_chunk_nonce(base_nonce, chunk_index)
try:
decrypted_chunk = aesgcm.decrypt(chunk_nonce, encrypted_chunk, None)
decrypted_chunks.append(decrypted_chunk)
except Exception as exc:
raise EncryptionError(f"Failed to decrypt chunk {chunk_index}: {exc}") from exc
return io.BytesIO(b"".join(decrypted_chunks))
class EncryptionManager:
"""Manages encryption providers and operations."""
def __init__(self, config: Dict[str, Any]):
self.config = config
self._local_provider: LocalKeyEncryption | None = None
self._kms_provider: Any = None # Set by KMS module
self._streaming_encryptor: StreamingEncryptor | None = None
@property
def enabled(self) -> bool:
return self.config.get("encryption_enabled", False)
@property
def default_algorithm(self) -> str:
return self.config.get("default_encryption_algorithm", "AES256")
def get_local_provider(self) -> LocalKeyEncryption:
if self._local_provider is None:
key_path = Path(self.config.get("encryption_master_key_path", "data/.myfsio.sys/keys/master.key"))
self._local_provider = LocalKeyEncryption(key_path)
return self._local_provider
def set_kms_provider(self, kms_provider: Any) -> None:
"""Set the KMS provider (injected from kms module)."""
self._kms_provider = kms_provider
def get_provider(self, algorithm: str, kms_key_id: str | None = None) -> EncryptionProvider:
"""Get the appropriate encryption provider for the algorithm."""
if algorithm == "AES256":
return self.get_local_provider()
elif algorithm == "aws:kms":
if self._kms_provider is None:
raise EncryptionError("KMS is not configured")
return self._kms_provider.get_provider(kms_key_id)
else:
raise EncryptionError(f"Unsupported encryption algorithm: {algorithm}")
def get_streaming_encryptor(self) -> StreamingEncryptor:
if self._streaming_encryptor is None:
self._streaming_encryptor = StreamingEncryptor(self.get_local_provider())
return self._streaming_encryptor
def encrypt_object(self, data: bytes, algorithm: str = "AES256",
kms_key_id: str | None = None,
context: Dict[str, str] | None = None) -> tuple[bytes, EncryptionMetadata]:
"""Encrypt object data."""
provider = self.get_provider(algorithm, kms_key_id)
result = provider.encrypt(data, context)
metadata = EncryptionMetadata(
algorithm=algorithm,
key_id=result.key_id,
nonce=result.nonce,
encrypted_data_key=result.encrypted_data_key,
)
return result.ciphertext, metadata
def decrypt_object(self, ciphertext: bytes, metadata: EncryptionMetadata,
context: Dict[str, str] | None = None) -> bytes:
"""Decrypt object data."""
provider = self.get_provider(metadata.algorithm, metadata.key_id)
return provider.decrypt(
ciphertext,
metadata.nonce,
metadata.encrypted_data_key,
metadata.key_id,
context,
)
def encrypt_stream(self, stream: BinaryIO, algorithm: str = "AES256",
context: Dict[str, str] | None = None) -> tuple[BinaryIO, EncryptionMetadata]:
"""Encrypt a stream for large files."""
encryptor = self.get_streaming_encryptor()
return encryptor.encrypt_stream(stream, context)
def decrypt_stream(self, stream: BinaryIO, metadata: EncryptionMetadata) -> BinaryIO:
"""Decrypt a stream."""
encryptor = self.get_streaming_encryptor()
return encryptor.decrypt_stream(stream, metadata)
class ClientEncryptionHelper:
"""Helpers for client-side encryption.
Client-side encryption is performed by the client, but this helper
provides key generation and materials for clients that need them.
"""
@staticmethod
def generate_client_key() -> Dict[str, str]:
"""Generate a new client encryption key."""
from datetime import datetime, timezone
key = secrets.token_bytes(32)
return {
"key": base64.b64encode(key).decode(),
"algorithm": "AES-256-GCM",
"created_at": datetime.now(timezone.utc).isoformat(),
}
@staticmethod
def encrypt_with_key(plaintext: bytes, key_b64: str) -> Dict[str, str]:
"""Encrypt data with a client-provided key."""
key = base64.b64decode(key_b64)
if len(key) != 32:
raise EncryptionError("Key must be 256 bits (32 bytes)")
aesgcm = AESGCM(key)
nonce = secrets.token_bytes(12)
ciphertext = aesgcm.encrypt(nonce, plaintext, None)
return {
"ciphertext": base64.b64encode(ciphertext).decode(),
"nonce": base64.b64encode(nonce).decode(),
"algorithm": "AES-256-GCM",
}
@staticmethod
def decrypt_with_key(ciphertext_b64: str, nonce_b64: str, key_b64: str) -> bytes:
"""Decrypt data with a client-provided key."""
key = base64.b64decode(key_b64)
nonce = base64.b64decode(nonce_b64)
ciphertext = base64.b64decode(ciphertext_b64)
if len(key) != 32:
raise EncryptionError("Key must be 256 bits (32 bytes)")
aesgcm = AESGCM(key)
try:
return aesgcm.decrypt(nonce, ciphertext, None)
except Exception as exc:
raise EncryptionError(f"Decryption failed: {exc}") from exc

20
app/errors.py
View File

@@ -129,6 +129,25 @@ class EntityTooLargeError(AppError):
status_code: int = 413
@dataclass
class QuotaExceededAppError(AppError):
"""Bucket quota exceeded."""
code: str = "QuotaExceeded"
message: str = "The bucket quota has been exceeded"
status_code: int = 403
quota: Optional[Dict[str, Any]] = None
usage: Optional[Dict[str, int]] = None
def __post_init__(self):
if self.quota or self.usage:
self.details = {}
if self.quota:
self.details["quota"] = self.quota
if self.usage:
self.details["usage"] = self.usage
super().__post_init__()
def handle_app_error(error: AppError) -> Response:
"""Handle application errors with appropriate response format."""
log_extra = {"error_code": error.code}
@@ -163,5 +182,6 @@ def register_error_handlers(app):
ObjectNotFoundError, InvalidObjectKeyError,
AccessDeniedError, InvalidCredentialsError,
MalformedRequestError, InvalidArgumentError, EntityTooLargeError,
QuotaExceededAppError,
]:
app.register_error_handler(error_class, handle_app_error)

39
app/iam.py
View File

@@ -15,7 +15,7 @@ class IamError(RuntimeError):
"""Raised when authentication or authorization fails."""
S3_ACTIONS = {"list", "read", "write", "delete", "share", "policy"}
S3_ACTIONS = {"list", "read", "write", "delete", "share", "policy", "replication"}
IAM_ACTIONS = {
"iam:list_users",
"iam:create_user",
@@ -26,22 +26,59 @@ IAM_ACTIONS = {
ALLOWED_ACTIONS = (S3_ACTIONS | IAM_ACTIONS) | {"iam:*"}
ACTION_ALIASES = {
# List actions
"list": "list",
"s3:listbucket": "list",
"s3:listallmybuckets": "list",
"s3:listbucketversions": "list",
"s3:listmultipartuploads": "list",
"s3:listparts": "list",
# Read actions
"read": "read",
"s3:getobject": "read",
"s3:getobjectversion": "read",
"s3:getobjecttagging": "read",
"s3:getobjectversiontagging": "read",
"s3:getobjectacl": "read",
"s3:getbucketversioning": "read",
"s3:headobject": "read",
"s3:headbucket": "read",
# Write actions
"write": "write",
"s3:putobject": "write",
"s3:createbucket": "write",
"s3:putobjecttagging": "write",
"s3:putbucketversioning": "write",
"s3:createmultipartupload": "write",
"s3:uploadpart": "write",
"s3:completemultipartupload": "write",
"s3:abortmultipartupload": "write",
"s3:copyobject": "write",
# Delete actions
"delete": "delete",
"s3:deleteobject": "delete",
"s3:deleteobjectversion": "delete",
"s3:deletebucket": "delete",
"s3:deleteobjecttagging": "delete",
# Share actions (ACL)
"share": "share",
"s3:putobjectacl": "share",
"s3:putbucketacl": "share",
"s3:getbucketacl": "share",
# Policy actions
"policy": "policy",
"s3:putbucketpolicy": "policy",
"s3:getbucketpolicy": "policy",
"s3:deletebucketpolicy": "policy",
# Replication actions
"replication": "replication",
"s3:getreplicationconfiguration": "replication",
"s3:putreplicationconfiguration": "replication",
"s3:deletereplicationconfiguration": "replication",
"s3:replicateobject": "replication",
"s3:replicatetags": "replication",
"s3:replicatedelete": "replication",
# IAM actions
"iam:listusers": "iam:list_users",
"iam:createuser": "iam:create_user",
"iam:deleteuser": "iam:delete_user",

344
app/kms.py Normal file
View File

@@ -0,0 +1,344 @@
"""Key Management Service (KMS) for encryption key management."""
from __future__ import annotations
import base64
import json
import secrets
import uuid
from dataclasses import dataclass, field
from datetime import datetime, timezone
from pathlib import Path
from typing import Any, Dict, List, Optional
from cryptography.hazmat.primitives.ciphers.aead import AESGCM
from .encryption import EncryptionError, EncryptionProvider, EncryptionResult
@dataclass
class KMSKey:
"""Represents a KMS encryption key."""
key_id: str
description: str
created_at: str
enabled: bool = True
key_material: bytes = field(default_factory=lambda: b"", repr=False)
@property
def arn(self) -> str:
return f"arn:aws:kms:local:000000000000:key/{self.key_id}"
def to_dict(self, include_key: bool = False) -> Dict[str, Any]:
data = {
"KeyId": self.key_id,
"Arn": self.arn,
"Description": self.description,
"CreationDate": self.created_at,
"Enabled": self.enabled,
"KeyState": "Enabled" if self.enabled else "Disabled",
"KeyUsage": "ENCRYPT_DECRYPT",
"KeySpec": "SYMMETRIC_DEFAULT",
}
if include_key:
data["KeyMaterial"] = base64.b64encode(self.key_material).decode()
return data
@classmethod
def from_dict(cls, data: Dict[str, Any]) -> "KMSKey":
key_material = b""
if "KeyMaterial" in data:
key_material = base64.b64decode(data["KeyMaterial"])
return cls(
key_id=data["KeyId"],
description=data.get("Description", ""),
created_at=data.get("CreationDate", datetime.now(timezone.utc).isoformat()),
enabled=data.get("Enabled", True),
key_material=key_material,
)
class KMSEncryptionProvider(EncryptionProvider):
"""Encryption provider using a specific KMS key."""
def __init__(self, kms: "KMSManager", key_id: str):
self.kms = kms
self.key_id = key_id
@property
def KEY_ID(self) -> str:
return self.key_id
def generate_data_key(self) -> tuple[bytes, bytes]:
"""Generate a data key encrypted with the KMS key."""
return self.kms.generate_data_key(self.key_id)
def encrypt(self, plaintext: bytes, context: Dict[str, str] | None = None) -> EncryptionResult:
"""Encrypt data using envelope encryption with KMS."""
data_key, encrypted_data_key = self.generate_data_key()
aesgcm = AESGCM(data_key)
nonce = secrets.token_bytes(12)
ciphertext = aesgcm.encrypt(nonce, plaintext,
json.dumps(context).encode() if context else None)
return EncryptionResult(
ciphertext=ciphertext,
nonce=nonce,
key_id=self.key_id,
encrypted_data_key=encrypted_data_key,
)
def decrypt(self, ciphertext: bytes, nonce: bytes, encrypted_data_key: bytes,
key_id: str, context: Dict[str, str] | None = None) -> bytes:
"""Decrypt data using envelope encryption with KMS."""
# Note: Data key is encrypted without context (AAD), so we decrypt without context
data_key = self.kms.decrypt_data_key(key_id, encrypted_data_key, context=None)
aesgcm = AESGCM(data_key)
try:
return aesgcm.decrypt(nonce, ciphertext,
json.dumps(context).encode() if context else None)
except Exception as exc:
raise EncryptionError(f"Failed to decrypt data: {exc}") from exc
class KMSManager:
"""Manages KMS keys and operations.
This is a local implementation that mimics AWS KMS functionality.
Keys are stored encrypted on disk.
"""
def __init__(self, keys_path: Path, master_key_path: Path):
self.keys_path = keys_path
self.master_key_path = master_key_path
self._keys: Dict[str, KMSKey] = {}
self._master_key: bytes | None = None
self._loaded = False
@property
def master_key(self) -> bytes:
"""Load or create the master key for encrypting KMS keys."""
if self._master_key is None:
if self.master_key_path.exists():
self._master_key = base64.b64decode(
self.master_key_path.read_text().strip()
)
else:
self._master_key = secrets.token_bytes(32)
self.master_key_path.parent.mkdir(parents=True, exist_ok=True)
self.master_key_path.write_text(
base64.b64encode(self._master_key).decode()
)
return self._master_key
def _load_keys(self) -> None:
"""Load keys from disk."""
if self._loaded:
return
if self.keys_path.exists():
try:
data = json.loads(self.keys_path.read_text(encoding="utf-8"))
for key_data in data.get("keys", []):
key = KMSKey.from_dict(key_data)
if key_data.get("EncryptedKeyMaterial"):
encrypted = base64.b64decode(key_data["EncryptedKeyMaterial"])
key.key_material = self._decrypt_key_material(encrypted)
self._keys[key.key_id] = key
except Exception:
pass
self._loaded = True
def _save_keys(self) -> None:
"""Save keys to disk (with encrypted key material)."""
keys_data = []
for key in self._keys.values():
data = key.to_dict(include_key=False)
encrypted = self._encrypt_key_material(key.key_material)
data["EncryptedKeyMaterial"] = base64.b64encode(encrypted).decode()
keys_data.append(data)
self.keys_path.parent.mkdir(parents=True, exist_ok=True)
self.keys_path.write_text(
json.dumps({"keys": keys_data}, indent=2),
encoding="utf-8"
)
def _encrypt_key_material(self, key_material: bytes) -> bytes:
"""Encrypt key material with the master key."""
aesgcm = AESGCM(self.master_key)
nonce = secrets.token_bytes(12)
ciphertext = aesgcm.encrypt(nonce, key_material, None)
return nonce + ciphertext
def _decrypt_key_material(self, encrypted: bytes) -> bytes:
"""Decrypt key material with the master key."""
aesgcm = AESGCM(self.master_key)
nonce = encrypted[:12]
ciphertext = encrypted[12:]
return aesgcm.decrypt(nonce, ciphertext, None)
def create_key(self, description: str = "", key_id: str | None = None) -> KMSKey:
"""Create a new KMS key."""
self._load_keys()
if key_id is None:
key_id = str(uuid.uuid4())
if key_id in self._keys:
raise EncryptionError(f"Key already exists: {key_id}")
key = KMSKey(
key_id=key_id,
description=description,
created_at=datetime.now(timezone.utc).isoformat(),
enabled=True,
key_material=secrets.token_bytes(32),
)
self._keys[key_id] = key
self._save_keys()
return key
def get_key(self, key_id: str) -> KMSKey | None:
"""Get a key by ID."""
self._load_keys()
return self._keys.get(key_id)
def list_keys(self) -> List[KMSKey]:
"""List all keys."""
self._load_keys()
return list(self._keys.values())
def enable_key(self, key_id: str) -> None:
"""Enable a key."""
self._load_keys()
key = self._keys.get(key_id)
if not key:
raise EncryptionError(f"Key not found: {key_id}")
key.enabled = True
self._save_keys()
def disable_key(self, key_id: str) -> None:
"""Disable a key."""
self._load_keys()
key = self._keys.get(key_id)
if not key:
raise EncryptionError(f"Key not found: {key_id}")
key.enabled = False
self._save_keys()
def delete_key(self, key_id: str) -> None:
"""Delete a key (schedule for deletion in real KMS)."""
self._load_keys()
if key_id not in self._keys:
raise EncryptionError(f"Key not found: {key_id}")
del self._keys[key_id]
self._save_keys()
def encrypt(self, key_id: str, plaintext: bytes,
context: Dict[str, str] | None = None) -> bytes:
"""Encrypt data directly with a KMS key."""
self._load_keys()
key = self._keys.get(key_id)
if not key:
raise EncryptionError(f"Key not found: {key_id}")
if not key.enabled:
raise EncryptionError(f"Key is disabled: {key_id}")
aesgcm = AESGCM(key.key_material)
nonce = secrets.token_bytes(12)
aad = json.dumps(context).encode() if context else None
ciphertext = aesgcm.encrypt(nonce, plaintext, aad)
key_id_bytes = key_id.encode("utf-8")
return len(key_id_bytes).to_bytes(2, "big") + key_id_bytes + nonce + ciphertext
def decrypt(self, ciphertext: bytes,
context: Dict[str, str] | None = None) -> tuple[bytes, str]:
"""Decrypt data directly with a KMS key.
Returns:
Tuple of (plaintext, key_id)
"""
self._load_keys()
key_id_len = int.from_bytes(ciphertext[:2], "big")
key_id = ciphertext[2:2 + key_id_len].decode("utf-8")
rest = ciphertext[2 + key_id_len:]
key = self._keys.get(key_id)
if not key:
raise EncryptionError(f"Key not found: {key_id}")
if not key.enabled:
raise EncryptionError(f"Key is disabled: {key_id}")
nonce = rest[:12]
encrypted = rest[12:]
aesgcm = AESGCM(key.key_material)
aad = json.dumps(context).encode() if context else None
try:
plaintext = aesgcm.decrypt(nonce, encrypted, aad)
return plaintext, key_id
except Exception as exc:
raise EncryptionError(f"Decryption failed: {exc}") from exc
def generate_data_key(self, key_id: str,
context: Dict[str, str] | None = None) -> tuple[bytes, bytes]:
"""Generate a data key and return both plaintext and encrypted versions.
Returns:
Tuple of (plaintext_key, encrypted_key)
"""
self._load_keys()
key = self._keys.get(key_id)
if not key:
raise EncryptionError(f"Key not found: {key_id}")
if not key.enabled:
raise EncryptionError(f"Key is disabled: {key_id}")
plaintext_key = secrets.token_bytes(32)
encrypted_key = self.encrypt(key_id, plaintext_key, context)
return plaintext_key, encrypted_key
def decrypt_data_key(self, key_id: str, encrypted_key: bytes,
context: Dict[str, str] | None = None) -> bytes:
"""Decrypt a data key."""
plaintext, _ = self.decrypt(encrypted_key, context)
return plaintext
def get_provider(self, key_id: str | None = None) -> KMSEncryptionProvider:
"""Get an encryption provider for a specific key."""
self._load_keys()
if key_id is None:
if not self._keys:
key = self.create_key("Default KMS Key")
key_id = key.key_id
else:
key_id = next(iter(self._keys.keys()))
if key_id not in self._keys:
raise EncryptionError(f"Key not found: {key_id}")
return KMSEncryptionProvider(self, key_id)
def re_encrypt(self, ciphertext: bytes, destination_key_id: str,
source_context: Dict[str, str] | None = None,
destination_context: Dict[str, str] | None = None) -> bytes:
"""Re-encrypt data with a different key."""
plaintext, source_key_id = self.decrypt(ciphertext, source_context)
return self.encrypt(destination_key_id, plaintext, destination_context)
def generate_random(self, num_bytes: int = 32) -> bytes:
"""Generate cryptographically secure random bytes."""
if num_bytes < 1 or num_bytes > 1024:
raise EncryptionError("Number of bytes must be between 1 and 1024")
return secrets.token_bytes(num_bytes)

463
app/kms_api.py Normal file
View File

@@ -0,0 +1,463 @@
"""KMS and encryption API endpoints."""
from __future__ import annotations
import base64
import uuid
from typing import Any, Dict
from flask import Blueprint, Response, current_app, jsonify, request
from .encryption import ClientEncryptionHelper, EncryptionError
from .extensions import limiter
from .iam import IamError
kms_api_bp = Blueprint("kms_api", __name__, url_prefix="/kms")
def _require_principal():
"""Require authentication for KMS operations."""
from .s3_api import _require_principal as s3_require_principal
return s3_require_principal()
def _kms():
"""Get KMS manager from app extensions."""
return current_app.extensions.get("kms")
def _encryption():
"""Get encryption manager from app extensions."""
return current_app.extensions.get("encryption")
def _error_response(code: str, message: str, status: int) -> tuple[Dict[str, Any], int]:
return {"__type": code, "message": message}, status
# ---------------------- Key Management ----------------------
@kms_api_bp.route("/keys", methods=["GET", "POST"])
@limiter.limit("30 per minute")
def list_or_create_keys():
"""List all KMS keys or create a new key."""
principal, error = _require_principal()
if error:
return error
kms = _kms()
if not kms:
return _error_response("KMSNotEnabled", "KMS is not configured", 400)
if request.method == "POST":
payload = request.get_json(silent=True) or {}
key_id = payload.get("KeyId") or payload.get("key_id")
description = payload.get("Description") or payload.get("description", "")
try:
key = kms.create_key(description=description, key_id=key_id)
current_app.logger.info(
"KMS key created",
extra={"key_id": key.key_id, "principal": principal.access_key},
)
return jsonify({
"KeyMetadata": key.to_dict(),
})
except EncryptionError as exc:
return _error_response("KMSInternalException", str(exc), 400)
# GET - List keys
keys = kms.list_keys()
return jsonify({
"Keys": [{"KeyId": k.key_id, "KeyArn": k.arn} for k in keys],
"Truncated": False,
})
@kms_api_bp.route("/keys/<key_id>", methods=["GET", "DELETE"])
@limiter.limit("30 per minute")
def get_or_delete_key(key_id: str):
"""Get or delete a specific KMS key."""
principal, error = _require_principal()
if error:
return error
kms = _kms()
if not kms:
return _error_response("KMSNotEnabled", "KMS is not configured", 400)
if request.method == "DELETE":
try:
kms.delete_key(key_id)
current_app.logger.info(
"KMS key deleted",
extra={"key_id": key_id, "principal": principal.access_key},
)
return Response(status=204)
except EncryptionError as exc:
return _error_response("NotFoundException", str(exc), 404)
# GET
key = kms.get_key(key_id)
if not key:
return _error_response("NotFoundException", f"Key not found: {key_id}", 404)
return jsonify({"KeyMetadata": key.to_dict()})
@kms_api_bp.route("/keys/<key_id>/enable", methods=["POST"])
@limiter.limit("30 per minute")
def enable_key(key_id: str):
"""Enable a KMS key."""
principal, error = _require_principal()
if error:
return error
kms = _kms()
if not kms:
return _error_response("KMSNotEnabled", "KMS is not configured", 400)
try:
kms.enable_key(key_id)
current_app.logger.info(
"KMS key enabled",
extra={"key_id": key_id, "principal": principal.access_key},
)
return Response(status=200)
except EncryptionError as exc:
return _error_response("NotFoundException", str(exc), 404)
@kms_api_bp.route("/keys/<key_id>/disable", methods=["POST"])
@limiter.limit("30 per minute")
def disable_key(key_id: str):
"""Disable a KMS key."""
principal, error = _require_principal()
if error:
return error
kms = _kms()
if not kms:
return _error_response("KMSNotEnabled", "KMS is not configured", 400)
try:
kms.disable_key(key_id)
current_app.logger.info(
"KMS key disabled",
extra={"key_id": key_id, "principal": principal.access_key},
)
return Response(status=200)
except EncryptionError as exc:
return _error_response("NotFoundException", str(exc), 404)
# ---------------------- Encryption Operations ----------------------
@kms_api_bp.route("/encrypt", methods=["POST"])
@limiter.limit("60 per minute")
def encrypt_data():
"""Encrypt data using a KMS key."""
principal, error = _require_principal()
if error:
return error
kms = _kms()
if not kms:
return _error_response("KMSNotEnabled", "KMS is not configured", 400)
payload = request.get_json(silent=True) or {}
key_id = payload.get("KeyId")
plaintext_b64 = payload.get("Plaintext")
context = payload.get("EncryptionContext")
if not key_id:
return _error_response("ValidationException", "KeyId is required", 400)
if not plaintext_b64:
return _error_response("ValidationException", "Plaintext is required", 400)
try:
plaintext = base64.b64decode(plaintext_b64)
except Exception:
return _error_response("ValidationException", "Plaintext must be base64 encoded", 400)
try:
ciphertext = kms.encrypt(key_id, plaintext, context)
return jsonify({
"CiphertextBlob": base64.b64encode(ciphertext).decode(),
"KeyId": key_id,
"EncryptionAlgorithm": "SYMMETRIC_DEFAULT",
})
except EncryptionError as exc:
return _error_response("KMSInternalException", str(exc), 400)
@kms_api_bp.route("/decrypt", methods=["POST"])
@limiter.limit("60 per minute")
def decrypt_data():
"""Decrypt data using a KMS key."""
principal, error = _require_principal()
if error:
return error
kms = _kms()
if not kms:
return _error_response("KMSNotEnabled", "KMS is not configured", 400)
payload = request.get_json(silent=True) or {}
ciphertext_b64 = payload.get("CiphertextBlob")
context = payload.get("EncryptionContext")
if not ciphertext_b64:
return _error_response("ValidationException", "CiphertextBlob is required", 400)
try:
ciphertext = base64.b64decode(ciphertext_b64)
except Exception:
return _error_response("ValidationException", "CiphertextBlob must be base64 encoded", 400)
try:
plaintext, key_id = kms.decrypt(ciphertext, context)
return jsonify({
"Plaintext": base64.b64encode(plaintext).decode(),
"KeyId": key_id,
"EncryptionAlgorithm": "SYMMETRIC_DEFAULT",
})
except EncryptionError as exc:
return _error_response("InvalidCiphertextException", str(exc), 400)
@kms_api_bp.route("/generate-data-key", methods=["POST"])
@limiter.limit("60 per minute")
def generate_data_key():
"""Generate a data encryption key."""
principal, error = _require_principal()
if error:
return error
kms = _kms()
if not kms:
return _error_response("KMSNotEnabled", "KMS is not configured", 400)
payload = request.get_json(silent=True) or {}
key_id = payload.get("KeyId")
context = payload.get("EncryptionContext")
key_spec = payload.get("KeySpec", "AES_256")
if not key_id:
return _error_response("ValidationException", "KeyId is required", 400)
if key_spec not in {"AES_256", "AES_128"}:
return _error_response("ValidationException", "KeySpec must be AES_256 or AES_128", 400)
try:
plaintext_key, encrypted_key = kms.generate_data_key(key_id, context)
# Trim key if AES_128 requested
if key_spec == "AES_128":
plaintext_key = plaintext_key[:16]
return jsonify({
"Plaintext": base64.b64encode(plaintext_key).decode(),
"CiphertextBlob": base64.b64encode(encrypted_key).decode(),
"KeyId": key_id,
})
except EncryptionError as exc:
return _error_response("KMSInternalException", str(exc), 400)
@kms_api_bp.route("/generate-data-key-without-plaintext", methods=["POST"])
@limiter.limit("60 per minute")
def generate_data_key_without_plaintext():
"""Generate a data encryption key without returning the plaintext."""
principal, error = _require_principal()
if error:
return error
kms = _kms()
if not kms:
return _error_response("KMSNotEnabled", "KMS is not configured", 400)
payload = request.get_json(silent=True) or {}
key_id = payload.get("KeyId")
context = payload.get("EncryptionContext")
if not key_id:
return _error_response("ValidationException", "KeyId is required", 400)
try:
_, encrypted_key = kms.generate_data_key(key_id, context)
return jsonify({
"CiphertextBlob": base64.b64encode(encrypted_key).decode(),
"KeyId": key_id,
})
except EncryptionError as exc:
return _error_response("KMSInternalException", str(exc), 400)
@kms_api_bp.route("/re-encrypt", methods=["POST"])
@limiter.limit("30 per minute")
def re_encrypt():
"""Re-encrypt data with a different key."""
principal, error = _require_principal()
if error:
return error
kms = _kms()
if not kms:
return _error_response("KMSNotEnabled", "KMS is not configured", 400)
payload = request.get_json(silent=True) or {}
ciphertext_b64 = payload.get("CiphertextBlob")
destination_key_id = payload.get("DestinationKeyId")
source_context = payload.get("SourceEncryptionContext")
destination_context = payload.get("DestinationEncryptionContext")
if not ciphertext_b64:
return _error_response("ValidationException", "CiphertextBlob is required", 400)
if not destination_key_id:
return _error_response("ValidationException", "DestinationKeyId is required", 400)
try:
ciphertext = base64.b64decode(ciphertext_b64)
except Exception:
return _error_response("ValidationException", "CiphertextBlob must be base64 encoded", 400)
try:
# First decrypt, get source key id
plaintext, source_key_id = kms.decrypt(ciphertext, source_context)
# Re-encrypt with destination key
new_ciphertext = kms.encrypt(destination_key_id, plaintext, destination_context)
return jsonify({
"CiphertextBlob": base64.b64encode(new_ciphertext).decode(),
"SourceKeyId": source_key_id,
"KeyId": destination_key_id,
})
except EncryptionError as exc:
return _error_response("KMSInternalException", str(exc), 400)
@kms_api_bp.route("/generate-random", methods=["POST"])
@limiter.limit("60 per minute")
def generate_random():
"""Generate random bytes."""
principal, error = _require_principal()
if error:
return error
kms = _kms()
if not kms:
return _error_response("KMSNotEnabled", "KMS is not configured", 400)
payload = request.get_json(silent=True) or {}
num_bytes = payload.get("NumberOfBytes", 32)
try:
num_bytes = int(num_bytes)
except (TypeError, ValueError):
return _error_response("ValidationException", "NumberOfBytes must be an integer", 400)
try:
random_bytes = kms.generate_random(num_bytes)
return jsonify({
"Plaintext": base64.b64encode(random_bytes).decode(),
})
except EncryptionError as exc:
return _error_response("ValidationException", str(exc), 400)
# ---------------------- Client-Side Encryption Helpers ----------------------
@kms_api_bp.route("/client/generate-key", methods=["POST"])
@limiter.limit("30 per minute")
def generate_client_key():
"""Generate a client-side encryption key."""
principal, error = _require_principal()
if error:
return error
key_info = ClientEncryptionHelper.generate_client_key()
return jsonify(key_info)
@kms_api_bp.route("/client/encrypt", methods=["POST"])
@limiter.limit("60 per minute")
def client_encrypt():
"""Encrypt data using client-side encryption."""
principal, error = _require_principal()
if error:
return error
payload = request.get_json(silent=True) or {}
plaintext_b64 = payload.get("Plaintext")
key_b64 = payload.get("Key")
if not plaintext_b64 or not key_b64:
return _error_response("ValidationException", "Plaintext and Key are required", 400)
try:
plaintext = base64.b64decode(plaintext_b64)
result = ClientEncryptionHelper.encrypt_with_key(plaintext, key_b64)
return jsonify(result)
except Exception as exc:
return _error_response("EncryptionError", str(exc), 400)
@kms_api_bp.route("/client/decrypt", methods=["POST"])
@limiter.limit("60 per minute")
def client_decrypt():
"""Decrypt data using client-side encryption."""
principal, error = _require_principal()
if error:
return error
payload = request.get_json(silent=True) or {}
ciphertext_b64 = payload.get("Ciphertext") or payload.get("ciphertext")
nonce_b64 = payload.get("Nonce") or payload.get("nonce")
key_b64 = payload.get("Key") or payload.get("key")
if not ciphertext_b64 or not nonce_b64 or not key_b64:
return _error_response("ValidationException", "Ciphertext, Nonce, and Key are required", 400)
try:
plaintext = ClientEncryptionHelper.decrypt_with_key(ciphertext_b64, nonce_b64, key_b64)
return jsonify({
"Plaintext": base64.b64encode(plaintext).decode(),
})
except Exception as exc:
return _error_response("DecryptionError", str(exc), 400)
# ---------------------- Encryption Materials for S3 Client-Side Encryption ----------------------
@kms_api_bp.route("/materials/<key_id>", methods=["POST"])
@limiter.limit("60 per minute")
def get_encryption_materials(key_id: str):
"""Get encryption materials for client-side S3 encryption.
This is used by S3 encryption clients that want to use KMS for
key management but perform encryption client-side.
"""
principal, error = _require_principal()
if error:
return error
kms = _kms()
if not kms:
return _error_response("KMSNotEnabled", "KMS is not configured", 400)
payload = request.get_json(silent=True) or {}
context = payload.get("EncryptionContext")
try:
plaintext_key, encrypted_key = kms.generate_data_key(key_id, context)
return jsonify({
"PlaintextKey": base64.b64encode(plaintext_key).decode(),
"EncryptedKey": base64.b64encode(encrypted_key).decode(),
"KeyId": key_id,
"Algorithm": "AES-256-GCM",
"KeyWrapAlgorithm": "kms",
})
except EncryptionError as exc:
return _error_response("KMSInternalException", str(exc), 400)

138
app/s3_api.py
View File

@@ -18,7 +18,7 @@ from .bucket_policies import BucketPolicyStore
from .extensions import limiter
from .iam import IamError, Principal
from .replication import ReplicationManager
from .storage import ObjectStorage, StorageError
from .storage import ObjectStorage, StorageError, QuotaExceededError
s3_api_bp = Blueprint("s3_api", __name__)
@@ -784,8 +784,9 @@ def _apply_object_headers(
metadata: Dict[str, str] | None,
etag: str,
) -> None:
response.headers["Content-Length"] = str(file_stat.st_size)
response.headers["Last-Modified"] = http_date(file_stat.st_mtime)
if file_stat is not None:
response.headers["Content-Length"] = str(file_stat.st_size)
response.headers["Last-Modified"] = http_date(file_stat.st_mtime)
response.headers["ETag"] = f'"{etag}"'
response.headers["Accept-Ranges"] = "bytes"
for key, value in (metadata or {}).items():
@@ -802,6 +803,7 @@ def _maybe_handle_bucket_subresource(bucket_name: str) -> Response | None:
"acl": _bucket_acl_handler,
"versions": _bucket_list_versions_handler,
"lifecycle": _bucket_lifecycle_handler,
"quota": _bucket_quota_handler,
}
requested = [key for key in handlers if key in request.args]
if not requested:
@@ -1399,6 +1401,87 @@ def _parse_lifecycle_config(payload: bytes) -> list:
return rules
def _bucket_quota_handler(bucket_name: str) -> Response:
"""Handle bucket quota configuration (GET/PUT/DELETE /<bucket>?quota)."""
if request.method not in {"GET", "PUT", "DELETE"}:
return _method_not_allowed(["GET", "PUT", "DELETE"])
principal, error = _require_principal()
if error:
return error
try:
_authorize_action(principal, bucket_name, "policy")
except IamError as exc:
return _error_response("AccessDenied", str(exc), 403)
storage = _storage()
if not storage.bucket_exists(bucket_name):
return _error_response("NoSuchBucket", "Bucket does not exist", 404)
if request.method == "GET":
quota = storage.get_bucket_quota(bucket_name)
if not quota:
return _error_response("NoSuchQuotaConfiguration", "No quota configuration found", 404)
# Return as JSON for simplicity (not a standard S3 API)
stats = storage.bucket_stats(bucket_name)
return jsonify({
"quota": quota,
"usage": {
"bytes": stats.get("bytes", 0),
"objects": stats.get("objects", 0),
}
})
if request.method == "DELETE":
try:
storage.set_bucket_quota(bucket_name, max_size_bytes=None, max_objects=None)
except StorageError as exc:
return _error_response("NoSuchBucket", str(exc), 404)
current_app.logger.info("Bucket quota deleted", extra={"bucket": bucket_name})
return Response(status=204)
# PUT
payload = request.get_json(silent=True)
if not payload:
return _error_response("MalformedRequest", "Request body must be JSON with quota limits", 400)
max_size_bytes = payload.get("max_size_bytes")
max_objects = payload.get("max_objects")
if max_size_bytes is None and max_objects is None:
return _error_response("InvalidArgument", "At least one of max_size_bytes or max_objects is required", 400)
# Validate types
if max_size_bytes is not None:
try:
max_size_bytes = int(max_size_bytes)
if max_size_bytes < 0:
raise ValueError("must be non-negative")
except (TypeError, ValueError) as exc:
return _error_response("InvalidArgument", f"max_size_bytes {exc}", 400)
if max_objects is not None:
try:
max_objects = int(max_objects)
if max_objects < 0:
raise ValueError("must be non-negative")
except (TypeError, ValueError) as exc:
return _error_response("InvalidArgument", f"max_objects {exc}", 400)
try:
storage.set_bucket_quota(bucket_name, max_size_bytes=max_size_bytes, max_objects=max_objects)
except StorageError as exc:
return _error_response("NoSuchBucket", str(exc), 404)
current_app.logger.info(
"Bucket quota updated",
extra={"bucket": bucket_name, "max_size_bytes": max_size_bytes, "max_objects": max_objects}
)
return Response(status=204)
def _bulk_delete_handler(bucket_name: str) -> Response:
principal, error = _require_principal()
if error:
@@ -1748,6 +1831,8 @@ def object_handler(bucket_name: str, object_key: str):
stream,
metadata=metadata or None,
)
except QuotaExceededError as exc:
return _error_response("QuotaExceeded", str(exc), 403)
except StorageError as exc:
message = str(exc)
if "Bucket" in message:
@@ -1779,19 +1864,48 @@ def object_handler(bucket_name: str, object_key: str):
except StorageError as exc:
return _error_response("NoSuchKey", str(exc), 404)
metadata = storage.get_object_metadata(bucket_name, object_key)
stat = path.stat()
mimetype = mimetypes.guess_type(path.name)[0] or "application/octet-stream"
etag = storage._compute_etag(path)
mimetype = mimetypes.guess_type(object_key)[0] or "application/octet-stream"
# Check if object is encrypted and needs decryption
is_encrypted = "x-amz-server-side-encryption" in metadata
if request.method == "GET":
response = Response(_stream_file(path), mimetype=mimetype, direct_passthrough=True)
logged_bytes = stat.st_size
if is_encrypted and hasattr(storage, 'get_object_data'):
# Use encrypted storage to decrypt
try:
data, clean_metadata = storage.get_object_data(bucket_name, object_key)
response = Response(data, mimetype=mimetype)
logged_bytes = len(data)
# Use decrypted size for Content-Length
response.headers["Content-Length"] = len(data)
etag = hashlib.md5(data).hexdigest()
except StorageError as exc:
return _error_response("InternalError", str(exc), 500)
else:
# Stream unencrypted file directly
stat = path.stat()
response = Response(_stream_file(path), mimetype=mimetype, direct_passthrough=True)
logged_bytes = stat.st_size
etag = storage._compute_etag(path)
else:
response = Response(status=200)
# HEAD request
if is_encrypted and hasattr(storage, 'get_object_data'):
# For encrypted objects, we need to report decrypted size
try:
data, _ = storage.get_object_data(bucket_name, object_key)
response = Response(status=200)
response.headers["Content-Length"] = len(data)
etag = hashlib.md5(data).hexdigest()
except StorageError as exc:
return _error_response("InternalError", str(exc), 500)
else:
stat = path.stat()
response = Response(status=200)
etag = storage._compute_etag(path)
response.headers["Content-Type"] = mimetype
logged_bytes = 0
_apply_object_headers(response, file_stat=stat, metadata=metadata, etag=etag)
_apply_object_headers(response, file_stat=path.stat() if not is_encrypted else None, metadata=metadata, etag=etag)
action = "Object read" if request.method == "GET" else "Object head"
current_app.logger.info(action, extra={"bucket": bucket_name, "key": object_key, "bytes": logged_bytes})
return response
@@ -2226,6 +2340,8 @@ def _complete_multipart_upload(bucket_name: str, object_key: str) -> Response:
try:
meta = _storage().complete_multipart_upload(bucket_name, upload_id, parts)
except QuotaExceededError as exc:
return _error_response("QuotaExceeded", str(exc), 403)
except StorageError as exc:
if "NoSuchBucket" in str(exc):
return _error_response("NoSuchBucket", str(exc), 404)

269
app/storage.py
View File

@@ -75,6 +75,15 @@ class StorageError(RuntimeError):
"""Raised when the storage layer encounters an unrecoverable problem."""
class QuotaExceededError(StorageError):
"""Raised when an operation would exceed bucket quota limits."""
def __init__(self, message: str, quota: Dict[str, Any], usage: Dict[str, int]):
super().__init__(message)
self.quota = quota
self.usage = usage
@dataclass
class ObjectMeta:
key: str
@@ -169,16 +178,38 @@ class ObjectStorage:
object_count = 0
total_bytes = 0
version_count = 0
version_bytes = 0
# Count current objects in the bucket folder
for path in bucket_path.rglob("*"):
if path.is_file():
rel = path.relative_to(bucket_path)
if rel.parts and rel.parts[0] in self.INTERNAL_FOLDERS:
if not rel.parts:
continue
stat = path.stat()
object_count += 1
total_bytes += stat.st_size
top_folder = rel.parts[0]
if top_folder not in self.INTERNAL_FOLDERS:
stat = path.stat()
object_count += 1
total_bytes += stat.st_size
stats = {"objects": object_count, "bytes": total_bytes}
# Count archived versions in the system folder
versions_root = self._bucket_versions_root(bucket_name)
if versions_root.exists():
for path in versions_root.rglob("*.bin"):
if path.is_file():
stat = path.stat()
version_count += 1
version_bytes += stat.st_size
stats = {
"objects": object_count,
"bytes": total_bytes,
"version_count": version_count,
"version_bytes": version_bytes,
"total_objects": object_count + version_count, # All objects including versions
"total_bytes": total_bytes + version_bytes, # All storage including versions
}
try:
cache_path.parent.mkdir(parents=True, exist_ok=True)
@@ -243,6 +274,7 @@ class ObjectStorage:
stream: BinaryIO,
*,
metadata: Optional[Dict[str, str]] = None,
enforce_quota: bool = True,
) -> ObjectMeta:
bucket_path = self._bucket_path(bucket_name)
if not bucket_path.exists():
@@ -253,12 +285,52 @@ class ObjectStorage:
destination = bucket_path / safe_key
destination.parent.mkdir(parents=True, exist_ok=True)
if self._is_versioning_enabled(bucket_path) and destination.exists():
# Check if this is an overwrite (won't add to object count)
is_overwrite = destination.exists()
existing_size = destination.stat().st_size if is_overwrite else 0
if self._is_versioning_enabled(bucket_path) and is_overwrite:
self._archive_current_version(bucket_id, safe_key, reason="overwrite")
checksum = hashlib.md5()
with destination.open("wb") as target:
shutil.copyfileobj(_HashingReader(stream, checksum), target)
# Write to temp file first to get actual size
tmp_dir = self._system_root_path() / self.SYSTEM_TMP_DIR
tmp_dir.mkdir(parents=True, exist_ok=True)
tmp_path = tmp_dir / f"{uuid.uuid4().hex}.tmp"
try:
checksum = hashlib.md5()
with tmp_path.open("wb") as target:
shutil.copyfileobj(_HashingReader(stream, checksum), target)
new_size = tmp_path.stat().st_size
# Check quota before finalizing
if enforce_quota:
# Calculate net change (new size minus size being replaced)
size_delta = new_size - existing_size
object_delta = 0 if is_overwrite else 1
quota_check = self.check_quota(
bucket_name,
additional_bytes=max(0, size_delta),
additional_objects=object_delta,
)
if not quota_check["allowed"]:
raise QuotaExceededError(
quota_check["message"] or "Quota exceeded",
quota_check["quota"],
quota_check["usage"],
)
# Move to final destination
shutil.move(str(tmp_path), str(destination))
finally:
# Clean up temp file if it still exists
try:
tmp_path.unlink(missing_ok=True)
except OSError:
pass
stat = destination.stat()
if metadata:
@@ -289,6 +361,27 @@ class ObjectStorage:
safe_key = self._sanitize_object_key(object_key)
return self._read_metadata(bucket_path.name, safe_key) or {}
def _cleanup_empty_parents(self, path: Path, stop_at: Path) -> None:
"""Remove empty parent directories up to (but not including) stop_at.
On Windows/OneDrive, directories may be locked briefly after file deletion.
This method retries with a small delay to handle that case.
"""
for parent in path.parents:
if parent == stop_at:
break
# Retry a few times with small delays for Windows/OneDrive
for attempt in range(3):
try:
if parent.exists() and not any(parent.iterdir()):
parent.rmdir()
break # Success, move to next parent
except OSError:
if attempt < 2:
time.sleep(0.1) # Brief delay before retry
# Final attempt failed - continue to next parent
break
def delete_object(self, bucket_name: str, object_key: str) -> None:
bucket_path = self._bucket_path(bucket_name)
path = self._object_path(bucket_name, object_key)
@@ -303,12 +396,7 @@ class ObjectStorage:
self._delete_metadata(bucket_id, rel)
self._invalidate_bucket_stats_cache(bucket_id)
for parent in path.parents:
if parent == bucket_path:
break
if parent.exists() and not any(parent.iterdir()):
parent.rmdir()
self._cleanup_empty_parents(path, bucket_path)
def purge_object(self, bucket_name: str, object_key: str) -> None:
bucket_path = self._bucket_path(bucket_name)
@@ -330,12 +418,7 @@ class ObjectStorage:
# Invalidate bucket stats cache
self._invalidate_bucket_stats_cache(bucket_id)
for parent in target.parents:
if parent == bucket_path:
break
if parent.exists() and not any(parent.iterdir()):
parent.rmdir()
self._cleanup_empty_parents(target, bucket_path)
def is_versioning_enabled(self, bucket_name: str) -> bool:
bucket_path = self._bucket_path(bucket_name)
@@ -413,6 +496,124 @@ class ObjectStorage:
bucket_path = self._require_bucket_path(bucket_name)
self._set_bucket_config_entry(bucket_path.name, "lifecycle", rules)
def get_bucket_quota(self, bucket_name: str) -> Dict[str, Any]:
"""Get quota configuration for bucket.
Returns:
Dict with 'max_bytes' and 'max_objects' (None if unlimited).
"""
bucket_path = self._require_bucket_path(bucket_name)
config = self._read_bucket_config(bucket_path.name)
quota = config.get("quota")
if isinstance(quota, dict):
return {
"max_bytes": quota.get("max_bytes"),
"max_objects": quota.get("max_objects"),
}
return {"max_bytes": None, "max_objects": None}
def set_bucket_quota(
self,
bucket_name: str,
*,
max_bytes: Optional[int] = None,
max_objects: Optional[int] = None,
) -> None:
"""Set quota limits for a bucket.
Args:
bucket_name: Name of the bucket
max_bytes: Maximum total size in bytes (None to remove limit)
max_objects: Maximum number of objects (None to remove limit)
"""
bucket_path = self._require_bucket_path(bucket_name)
if max_bytes is None and max_objects is None:
# Remove quota entirely
self._set_bucket_config_entry(bucket_path.name, "quota", None)
return
quota: Dict[str, Any] = {}
if max_bytes is not None:
if max_bytes < 0:
raise StorageError("max_bytes must be non-negative")
quota["max_bytes"] = max_bytes
if max_objects is not None:
if max_objects < 0:
raise StorageError("max_objects must be non-negative")
quota["max_objects"] = max_objects
self._set_bucket_config_entry(bucket_path.name, "quota", quota)
def check_quota(
self,
bucket_name: str,
additional_bytes: int = 0,
additional_objects: int = 0,
) -> Dict[str, Any]:
"""Check if an operation would exceed bucket quota.
Args:
bucket_name: Name of the bucket
additional_bytes: Bytes that would be added
additional_objects: Objects that would be added
Returns:
Dict with 'allowed' (bool), 'quota' (current limits),
'usage' (current usage), and 'message' (if not allowed).
"""
quota = self.get_bucket_quota(bucket_name)
if not quota:
return {
"allowed": True,
"quota": None,
"usage": None,
"message": None,
}
# Get current stats (uses cache when available)
stats = self.bucket_stats(bucket_name)
# Use totals which include versions for quota enforcement
current_bytes = stats.get("total_bytes", stats.get("bytes", 0))
current_objects = stats.get("total_objects", stats.get("objects", 0))
result = {
"allowed": True,
"quota": quota,
"usage": {
"bytes": current_bytes,
"objects": current_objects,
"version_count": stats.get("version_count", 0),
"version_bytes": stats.get("version_bytes", 0),
},
"message": None,
}
max_bytes_limit = quota.get("max_bytes")
max_objects = quota.get("max_objects")
if max_bytes_limit is not None:
projected_bytes = current_bytes + additional_bytes
if projected_bytes > max_bytes_limit:
result["allowed"] = False
result["message"] = (
f"Quota exceeded: adding {additional_bytes} bytes would result in "
f"{projected_bytes} bytes, exceeding limit of {max_bytes_limit} bytes"
)
return result
if max_objects is not None:
projected_objects = current_objects + additional_objects
if projected_objects > max_objects:
result["allowed"] = False
result["message"] = (
f"Quota exceeded: adding {additional_objects} objects would result in "
f"{projected_objects} objects, exceeding limit of {max_objects} objects"
)
return result
return result
def get_object_tags(self, bucket_name: str, object_key: str) -> List[Dict[str, str]]:
"""Get tags for an object."""
bucket_path = self._bucket_path(bucket_name)
@@ -529,6 +730,7 @@ class ObjectStorage:
else:
self._delete_metadata(bucket_id, safe_key)
stat = destination.stat()
self._invalidate_bucket_stats_cache(bucket_id)
return ObjectMeta(
key=safe_key.as_posix(),
size=stat.st_size,
@@ -677,6 +879,7 @@ class ObjectStorage:
bucket_name: str,
upload_id: str,
ordered_parts: List[Dict[str, Any]],
enforce_quota: bool = True,
) -> ObjectMeta:
if not ordered_parts:
raise StorageError("parts list required")
@@ -687,6 +890,7 @@ class ObjectStorage:
if not parts_map:
raise StorageError("No uploaded parts found")
validated: List[tuple[int, Dict[str, Any]]] = []
total_size = 0
for part in ordered_parts:
raw_number = part.get("part_number")
if raw_number is None:
@@ -706,10 +910,33 @@ class ObjectStorage:
if supplied_etag and record.get("etag") and supplied_etag.strip('"') != record["etag"]:
raise StorageError(f"ETag mismatch for part {number}")
validated.append((number, record))
total_size += record.get("size", 0)
validated.sort(key=lambda entry: entry[0])
safe_key = self._sanitize_object_key(manifest["object_key"])
destination = bucket_path / safe_key
# Check if this is an overwrite
is_overwrite = destination.exists()
existing_size = destination.stat().st_size if is_overwrite else 0
# Check quota before writing
if enforce_quota:
size_delta = total_size - existing_size
object_delta = 0 if is_overwrite else 1
quota_check = self.check_quota(
bucket_name,
additional_bytes=max(0, size_delta),
additional_objects=object_delta,
)
if not quota_check["allowed"]:
raise QuotaExceededError(
quota_check["message"] or "Quota exceeded",
quota_check["quota"],
quota_check["usage"],
)
destination.parent.mkdir(parents=True, exist_ok=True)
lock_file_path = self._system_bucket_root(bucket_id) / "locks" / f"{safe_key.as_posix().replace('/', '_')}.lock"

293
app/ui.py
View File

@@ -6,7 +6,7 @@ import uuid
import psutil
import shutil
from typing import Any
from urllib.parse import urlparse
from urllib.parse import quote, urlparse
import boto3
import requests
@@ -30,6 +30,7 @@ from .bucket_policies import BucketPolicyStore
from .connections import ConnectionStore, RemoteConnection
from .extensions import limiter
from .iam import IamError
from .kms import KMSManager
from .replication import ReplicationManager, ReplicationRule
from .secret_store import EphemeralSecretStore
from .storage import ObjectStorage, StorageError
@@ -50,6 +51,9 @@ def _iam():
return current_app.extensions["iam"]
def _kms() -> KMSManager | None:
return current_app.extensions.get("kms")
def _bucket_policies() -> BucketPolicyStore:
store: BucketPolicyStore = current_app.extensions["bucket_policies"]
@@ -185,6 +189,7 @@ def inject_nav_state() -> dict[str, Any]:
return {
"principal": principal,
"can_manage_iam": can_manage,
"can_view_metrics": can_manage, # Only admins can view metrics
"csrf_token": generate_csrf,
}
@@ -255,9 +260,9 @@ def buckets_overview():
visible_buckets.append({
"meta": bucket,
"summary": {
"objects": stats["objects"],
"total_bytes": stats["bytes"],
"human_size": _format_bytes(stats["bytes"]),
"objects": stats["total_objects"],
"total_bytes": stats["total_bytes"],
"human_size": _format_bytes(stats["total_bytes"]),
},
"access_label": access_label,
"access_badge": access_badge,
@@ -336,9 +341,46 @@ def bucket_detail(bucket_name: str):
except IamError:
can_manage_versioning = False
# Check replication permission
can_manage_replication = False
if principal:
try:
_iam().authorize(principal, bucket_name, "replication")
can_manage_replication = True
except IamError:
can_manage_replication = False
# Check if user is admin (can configure replication settings, not just toggle)
is_replication_admin = False
if principal:
try:
_iam().authorize(principal, None, "iam:list_users")
is_replication_admin = True
except IamError:
is_replication_admin = False
# Replication info - don't compute sync status here (it's slow), let JS fetch it async
replication_rule = _replication().get_rule(bucket_name)
connections = _connections().list()
# Load connections for admin, or for non-admin if there's an existing rule (to show target name)
connections = _connections().list() if (is_replication_admin or replication_rule) else []
# Encryption settings
encryption_config = storage.get_bucket_encryption(bucket_name)
kms_manager = _kms()
kms_keys = kms_manager.list_keys() if kms_manager else []
kms_enabled = current_app.config.get("KMS_ENABLED", False)
encryption_enabled = current_app.config.get("ENCRYPTION_ENABLED", False)
can_manage_encryption = can_manage_versioning # Same as other bucket properties
# Quota settings (admin only)
bucket_quota = storage.get_bucket_quota(bucket_name)
bucket_stats = storage.bucket_stats(bucket_name)
can_manage_quota = False
try:
_iam().authorize(principal, None, "iam:list_users")
can_manage_quota = True
except IamError:
pass
return render_template(
"bucket_detail.html",
@@ -349,10 +391,20 @@ def bucket_detail(bucket_name: str):
bucket_policy=bucket_policy,
can_edit_policy=can_edit_policy,
can_manage_versioning=can_manage_versioning,
can_manage_replication=can_manage_replication,
can_manage_encryption=can_manage_encryption,
is_replication_admin=is_replication_admin,
default_policy=default_policy,
versioning_enabled=versioning_enabled,
replication_rule=replication_rule,
connections=connections,
encryption_config=encryption_config,
kms_keys=kms_keys,
kms_enabled=kms_enabled,
encryption_enabled=encryption_enabled,
bucket_quota=bucket_quota,
bucket_stats=bucket_stats,
can_manage_quota=can_manage_quota,
)
@@ -647,9 +699,18 @@ def bulk_download_objects(bucket_name: str):
# But strictly we should check. Let's check.
_authorize_ui(principal, bucket_name, "read", object_key=key)
path = storage.get_object_path(bucket_name, key)
# Use the key as the filename in the zip
zf.write(path, arcname=key)
# Check if object is encrypted
metadata = storage.get_object_metadata(bucket_name, key)
is_encrypted = "x-amz-server-side-encryption" in metadata
if is_encrypted and hasattr(storage, 'get_object_data'):
# Decrypt and add to zip
data, _ = storage.get_object_data(bucket_name, key)
zf.writestr(key, data)
else:
# Add unencrypted file directly
path = storage.get_object_path(bucket_name, key)
zf.write(path, arcname=key)
except (StorageError, IamError):
# Skip files we can't read or don't exist
continue
@@ -691,13 +752,34 @@ def purge_object_versions(bucket_name: str, object_key: str):
@ui_bp.get("/buckets/<bucket_name>/objects/<path:object_key>/preview")
def object_preview(bucket_name: str, object_key: str) -> Response:
principal = _current_principal()
storage = _storage()
try:
_authorize_ui(principal, bucket_name, "read", object_key=object_key)
path = _storage().get_object_path(bucket_name, object_key)
path = storage.get_object_path(bucket_name, object_key)
metadata = storage.get_object_metadata(bucket_name, object_key)
except (StorageError, IamError) as exc:
status = 403 if isinstance(exc, IamError) else 404
return Response(str(exc), status=status)
download = request.args.get("download") == "1"
# Check if object is encrypted and needs decryption
is_encrypted = "x-amz-server-side-encryption" in metadata
if is_encrypted and hasattr(storage, 'get_object_data'):
try:
data, _ = storage.get_object_data(bucket_name, object_key)
import io
import mimetypes
mimetype = mimetypes.guess_type(object_key)[0] or "application/octet-stream"
return send_file(
io.BytesIO(data),
mimetype=mimetype,
as_attachment=download,
download_name=path.name
)
except StorageError as exc:
return Response(f"Decryption failed: {exc}", status=500)
return send_file(path, as_attachment=download, download_name=path.name)
@@ -712,12 +794,16 @@ def object_presign(bucket_name: str, object_key: str):
except IamError as exc:
return jsonify({"error": str(exc)}), 403
connection_url = "http://127.0.0.1:5000"
url = f"{connection_url}/presign/{bucket_name}/{object_key}"
api_base = current_app.config.get("API_BASE_URL") or "http://127.0.0.1:5000"
api_base = api_base.rstrip("/")
encoded_key = quote(object_key, safe="")
url = f"{api_base}/presign/{bucket_name}/{encoded_key}"
# Use API base URL for forwarded headers so presigned URLs point to API, not UI
parsed_api = urlparse(api_base)
headers = _api_headers()
headers["X-Forwarded-Host"] = request.host
headers["X-Forwarded-Proto"] = request.scheme
headers["X-Forwarded-Host"] = parsed_api.netloc or "127.0.0.1:5000"
headers["X-Forwarded-Proto"] = parsed_api.scheme or "http"
headers["X-Forwarded-For"] = request.remote_addr or "127.0.0.1"
try:
@@ -853,6 +939,127 @@ def update_bucket_versioning(bucket_name: str):
return redirect(url_for("ui.bucket_detail", bucket_name=bucket_name, tab="properties"))
@ui_bp.post("/buckets/<bucket_name>/quota")
def update_bucket_quota(bucket_name: str):
"""Update bucket quota configuration (admin only)."""
principal = _current_principal()
# Quota management is admin-only
is_admin = False
try:
_iam().authorize(principal, None, "iam:list_users")
is_admin = True
except IamError:
pass
if not is_admin:
flash("Only administrators can manage bucket quotas", "danger")
return redirect(url_for("ui.bucket_detail", bucket_name=bucket_name, tab="properties"))
action = request.form.get("action", "set")
if action == "remove":
try:
_storage().set_bucket_quota(bucket_name, max_bytes=None, max_objects=None)
flash("Bucket quota removed", "info")
except StorageError as exc:
flash(_friendly_error_message(exc), "danger")
return redirect(url_for("ui.bucket_detail", bucket_name=bucket_name, tab="properties"))
# Parse quota values
max_mb_str = request.form.get("max_mb", "").strip()
max_objects_str = request.form.get("max_objects", "").strip()
max_bytes = None
max_objects = None
if max_mb_str:
try:
max_mb = int(max_mb_str)
if max_mb < 1:
raise ValueError("Size must be at least 1 MB")
max_bytes = max_mb * 1024 * 1024 # Convert MB to bytes
except ValueError as exc:
flash(f"Invalid size value: {exc}", "danger")
return redirect(url_for("ui.bucket_detail", bucket_name=bucket_name, tab="properties"))
if max_objects_str:
try:
max_objects = int(max_objects_str)
if max_objects < 0:
raise ValueError("Object count must be non-negative")
except ValueError as exc:
flash(f"Invalid object count: {exc}", "danger")
return redirect(url_for("ui.bucket_detail", bucket_name=bucket_name, tab="properties"))
try:
_storage().set_bucket_quota(bucket_name, max_bytes=max_bytes, max_objects=max_objects)
if max_bytes is None and max_objects is None:
flash("Bucket quota removed", "info")
else:
flash("Bucket quota updated", "success")
except StorageError as exc:
flash(_friendly_error_message(exc), "danger")
return redirect(url_for("ui.bucket_detail", bucket_name=bucket_name, tab="properties"))
@ui_bp.post("/buckets/<bucket_name>/encryption")
def update_bucket_encryption(bucket_name: str):
"""Update bucket default encryption configuration."""
principal = _current_principal()
try:
_authorize_ui(principal, bucket_name, "write")
except IamError as exc:
flash(_friendly_error_message(exc), "danger")
return redirect(url_for("ui.bucket_detail", bucket_name=bucket_name, tab="properties"))
action = request.form.get("action", "enable")
if action == "disable":
# Disable encryption
try:
_storage().set_bucket_encryption(bucket_name, None)
flash("Default encryption disabled", "info")
except StorageError as exc:
flash(_friendly_error_message(exc), "danger")
return redirect(url_for("ui.bucket_detail", bucket_name=bucket_name, tab="properties"))
# Enable or update encryption
algorithm = request.form.get("algorithm", "AES256")
kms_key_id = request.form.get("kms_key_id", "").strip() or None
# Validate algorithm
if algorithm not in ("AES256", "aws:kms"):
flash("Invalid encryption algorithm", "danger")
return redirect(url_for("ui.bucket_detail", bucket_name=bucket_name, tab="properties"))
# Build encryption config following AWS format
encryption_config: dict[str, Any] = {
"Rules": [
{
"ApplyServerSideEncryptionByDefault": {
"SSEAlgorithm": algorithm,
}
}
]
}
if algorithm == "aws:kms" and kms_key_id:
encryption_config["Rules"][0]["ApplyServerSideEncryptionByDefault"]["KMSMasterKeyID"] = kms_key_id
try:
_storage().set_bucket_encryption(bucket_name, encryption_config)
if algorithm == "aws:kms":
flash("Default KMS encryption enabled", "success")
else:
flash("Default AES-256 encryption enabled", "success")
except StorageError as exc:
flash(_friendly_error_message(exc), "danger")
return redirect(url_for("ui.bucket_detail", bucket_name=bucket_name, tab="properties"))
@ui_bp.get("/iam")
def iam_dashboard():
principal = _current_principal()
@@ -1168,17 +1375,52 @@ def delete_connection(connection_id: str):
def update_bucket_replication(bucket_name: str):
principal = _current_principal()
try:
_authorize_ui(principal, bucket_name, "write")
_authorize_ui(principal, bucket_name, "replication")
except IamError as exc:
flash(str(exc), "danger")
return redirect(url_for("ui.bucket_detail", bucket_name=bucket_name, tab="replication"))
# Check if user is admin (required for create/delete operations)
is_admin = False
try:
_iam().authorize(principal, None, "iam:list_users")
is_admin = True
except IamError:
is_admin = False
action = request.form.get("action")
if action == "delete":
# Admin only - remove configuration entirely
if not is_admin:
flash("Only administrators can remove replication configuration", "danger")
return redirect(url_for("ui.bucket_detail", bucket_name=bucket_name, tab="replication"))
_replication().delete_rule(bucket_name)
flash("Replication disabled", "info")
else:
flash("Replication configuration removed", "info")
elif action == "pause":
# Users can pause - just set enabled=False
rule = _replication().get_rule(bucket_name)
if rule:
rule.enabled = False
_replication().set_rule(rule)
flash("Replication paused", "info")
else:
flash("No replication configuration to pause", "warning")
elif action == "resume":
# Users can resume - just set enabled=True
rule = _replication().get_rule(bucket_name)
if rule:
rule.enabled = True
_replication().set_rule(rule)
flash("Replication resumed", "success")
else:
flash("No replication configuration to resume", "warning")
elif action == "create":
# Admin only - create new configuration
if not is_admin:
flash("Only administrators can configure replication settings", "danger")
return redirect(url_for("ui.bucket_detail", bucket_name=bucket_name, tab="replication"))
from .replication import REPLICATION_MODE_NEW_ONLY, REPLICATION_MODE_ALL
import time
@@ -1205,6 +1447,8 @@ def update_bucket_replication(bucket_name: str):
flash("Replication configured. Existing objects are being replicated in the background.", "success")
else:
flash("Replication configured. Only new uploads will be replicated.", "success")
else:
flash("Invalid action", "danger")
return redirect(url_for("ui.bucket_detail", bucket_name=bucket_name, tab="replication"))
@@ -1214,7 +1458,7 @@ def get_replication_status(bucket_name: str):
"""Async endpoint to fetch replication sync status without blocking page load."""
principal = _current_principal()
try:
_authorize_ui(principal, bucket_name, "read")
_authorize_ui(principal, bucket_name, "replication")
except IamError:
return jsonify({"error": "Access denied"}), 403
@@ -1254,6 +1498,13 @@ def connections_dashboard():
def metrics_dashboard():
principal = _current_principal()
# Metrics are restricted to admin users
try:
_iam().authorize(principal, None, "iam:list_users")
except IamError:
flash("Access denied: Metrics require admin permissions", "danger")
return redirect(url_for("ui.buckets_overview"))
cpu_percent = psutil.cpu_percent(interval=0.1)
memory = psutil.virtual_memory()
@@ -1266,13 +1517,16 @@ def metrics_dashboard():
total_objects = 0
total_bytes_used = 0
total_versions = 0
# Note: Uses cached stats from storage layer to improve performance
cache_ttl = current_app.config.get("BUCKET_STATS_CACHE_TTL", 60)
for bucket in buckets:
stats = storage.bucket_stats(bucket.name, cache_ttl=cache_ttl)
total_objects += stats["objects"]
total_bytes_used += stats["bytes"]
# Use totals which include archived versions
total_objects += stats.get("total_objects", stats.get("objects", 0))
total_bytes_used += stats.get("total_bytes", stats.get("bytes", 0))
total_versions += stats.get("version_count", 0)
return render_template(
"metrics.html",
@@ -1293,6 +1547,7 @@ def metrics_dashboard():
app={
"buckets": total_buckets,
"objects": total_objects,
"versions": total_versions,
"storage_used": _format_bytes(total_bytes_used),
"storage_raw": total_bytes_used,
}

2
app/version.py
View File

@@ -1,7 +1,7 @@
"""Central location for the application version string."""
from __future__ import annotations
APP_VERSION = "0.1.2"
APP_VERSION = "0.1.3"
def get_version() -> str:

272
docs.md
View File

@@ -80,6 +80,10 @@ The repo now tracks a human-friendly release string inside `app/version.py` (see
| `API_BASE_URL` | `None` | Used by the UI to hit API endpoints (presign/policy). If unset, the UI will auto-detect the host or use `X-Forwarded-*` headers. |
| `AWS_REGION` | `us-east-1` | Region embedded in SigV4 credential scope. |
| `AWS_SERVICE` | `s3` | Service string for SigV4. |
| `ENCRYPTION_ENABLED` | `false` | Enable server-side encryption support. |
| `KMS_ENABLED` | `false` | Enable KMS key management for encryption. |
| `KMS_KEYS_PATH` | `data/kms_keys.json` | Path to store KMS key metadata. |
| `ENCRYPTION_MASTER_KEY_PATH` | `data/master.key` | Path to the master encryption key file. |
Set env vars (or pass overrides to `create_app`) to point the servers at custom paths.
@@ -102,6 +106,46 @@ The application automatically trusts these headers to generate correct presigned
The API expects every request to include `X-Access-Key` and `X-Secret-Key` headers. The UI persists them in the Flask session after login.
### Available IAM Actions
| Action | Description | AWS Aliases |
| --- | --- | --- |
| `list` | List buckets and objects | `s3:ListBucket`, `s3:ListAllMyBuckets`, `s3:ListBucketVersions`, `s3:ListMultipartUploads`, `s3:ListParts` |
| `read` | Download objects | `s3:GetObject`, `s3:GetObjectVersion`, `s3:GetObjectTagging`, `s3:HeadObject`, `s3:HeadBucket` |
| `write` | Upload objects, create buckets | `s3:PutObject`, `s3:CreateBucket`, `s3:CreateMultipartUpload`, `s3:UploadPart`, `s3:CompleteMultipartUpload`, `s3:AbortMultipartUpload`, `s3:CopyObject` |
| `delete` | Remove objects and buckets | `s3:DeleteObject`, `s3:DeleteObjectVersion`, `s3:DeleteBucket` |
| `share` | Manage ACLs | `s3:PutObjectAcl`, `s3:PutBucketAcl`, `s3:GetBucketAcl` |
| `policy` | Manage bucket policies | `s3:PutBucketPolicy`, `s3:GetBucketPolicy`, `s3:DeleteBucketPolicy` |
| `replication` | Configure and manage replication | `s3:GetReplicationConfiguration`, `s3:PutReplicationConfiguration`, `s3:ReplicateObject`, `s3:ReplicateTags`, `s3:ReplicateDelete` |
| `iam:list_users` | View IAM users | `iam:ListUsers` |
| `iam:create_user` | Create IAM users | `iam:CreateUser` |
| `iam:delete_user` | Delete IAM users | `iam:DeleteUser` |
| `iam:rotate_key` | Rotate user secrets | `iam:RotateAccessKey` |
| `iam:update_policy` | Modify user policies | `iam:PutUserPolicy` |
| `iam:*` | All IAM actions (admin wildcard) | — |
### Example Policies
**Full Control (admin):**
```json
[{"bucket": "*", "actions": ["list", "read", "write", "delete", "share", "policy", "replication", "iam:*"]}]
```
**Read-Only:**
```json
[{"bucket": "*", "actions": ["list", "read"]}]
```
**Single Bucket Access (no listing other buckets):**
```json
[{"bucket": "user-bucket", "actions": ["read", "write", "delete"]}]
```
**Bucket Access with Replication:**
```json
[{"bucket": "my-bucket", "actions": ["list", "read", "write", "delete", "replication"]}]
```
## 5. Bucket Policies & Presets
- **Storage**: Policies are persisted in `data/.myfsio.sys/config/bucket_policies.json` under `{"policies": {"bucket": {...}}}`.
@@ -173,9 +217,207 @@ s3.complete_multipart_upload(
)
```
## 6. Site Replication
## 7. Encryption
MyFSIO supports **Site Replication**, allowing you to automatically copy new objects from one MyFSIO instance (Source) to another (Target). This is useful for disaster recovery, data locality, or backups.
MyFSIO supports **server-side encryption at rest** to protect your data. When enabled, objects are encrypted using AES-256-GCM before being written to disk.
### Encryption Types
| Type | Description |
|------|-------------|
| **AES-256 (SSE-S3)** | Server-managed encryption using a local master key |
| **KMS (SSE-KMS)** | Encryption using customer-managed keys via the built-in KMS |
### Enabling Encryption
#### 1. Set Environment Variables
```powershell
# PowerShell
$env:ENCRYPTION_ENABLED = "true"
$env:KMS_ENABLED = "true" # Optional, for KMS key management
python run.py
```
```bash
# Bash
export ENCRYPTION_ENABLED=true
export KMS_ENABLED=true
python run.py
```
#### 2. Configure Bucket Default Encryption (UI)
1. Navigate to your bucket in the UI
2. Click the **Properties** tab
3. Find the **Default Encryption** card
4. Click **Enable Encryption**
5. Choose algorithm:
- **AES-256**: Uses the server's master key
- **aws:kms**: Uses a KMS-managed key (select from dropdown)
6. Save changes
Once enabled, all **new objects** uploaded to the bucket will be automatically encrypted.
### KMS Key Management
When `KMS_ENABLED=true`, you can manage encryption keys via the KMS API:
```bash
# Create a new KMS key
curl -X POST http://localhost:5000/kms/keys \
-H "Content-Type: application/json" \
-H "X-Access-Key: ..." -H "X-Secret-Key: ..." \
-d '{"alias": "my-key", "description": "Production encryption key"}'
# List all keys
curl http://localhost:5000/kms/keys \
-H "X-Access-Key: ..." -H "X-Secret-Key: ..."
# Get key details
curl http://localhost:5000/kms/keys/{key-id} \
-H "X-Access-Key: ..." -H "X-Secret-Key: ..."
# Rotate a key (creates new key material)
curl -X POST http://localhost:5000/kms/keys/{key-id}/rotate \
-H "X-Access-Key: ..." -H "X-Secret-Key: ..."
# Disable/Enable a key
curl -X POST http://localhost:5000/kms/keys/{key-id}/disable \
-H "X-Access-Key: ..." -H "X-Secret-Key: ..."
curl -X POST http://localhost:5000/kms/keys/{key-id}/enable \
-H "X-Access-Key: ..." -H "X-Secret-Key: ..."
# Schedule key deletion (30-day waiting period)
curl -X DELETE http://localhost:5000/kms/keys/{key-id}?waiting_period_days=30 \
-H "X-Access-Key: ..." -H "X-Secret-Key: ..."
```
### How It Works
1. **Envelope Encryption**: Each object is encrypted with a unique Data Encryption Key (DEK)
2. **Key Wrapping**: The DEK is encrypted (wrapped) by the master key or KMS key
3. **Storage**: The encrypted DEK is stored alongside the encrypted object
4. **Decryption**: On read, the DEK is unwrapped and used to decrypt the object
### Client-Side Encryption
For additional security, you can use client-side encryption. The `ClientEncryptionHelper` class provides utilities:
```python
from app.encryption import ClientEncryptionHelper
# Generate a client-side key
key = ClientEncryptionHelper.generate_key()
key_b64 = ClientEncryptionHelper.key_to_base64(key)
# Encrypt before upload
plaintext = b"sensitive data"
encrypted, metadata = ClientEncryptionHelper.encrypt_for_upload(plaintext, key)
# Upload with metadata headers
# x-amz-meta-x-amz-key: <wrapped-key>
# x-amz-meta-x-amz-iv: <iv>
# x-amz-meta-x-amz-matdesc: <material-description>
# Decrypt after download
decrypted = ClientEncryptionHelper.decrypt_from_download(encrypted, metadata, key)
```
### Important Notes
- **Existing objects are NOT encrypted** - Only new uploads after enabling encryption are encrypted
- **Master key security** - The master key file (`master.key`) should be backed up securely and protected
- **Key rotation** - Rotating a KMS key creates new key material; existing objects remain encrypted with the old material
- **Disabled keys** - Objects encrypted with a disabled key cannot be decrypted until the key is re-enabled
- **Deleted keys** - Once a key is deleted (after the waiting period), objects encrypted with it are permanently inaccessible
### Verifying Encryption
To verify an object is encrypted:
1. Check the raw file in `data/<bucket>/` - it should be unreadable binary
2. Look for `.meta` files containing encryption metadata
3. Download via the API/UI - the object should be automatically decrypted
## 8. Bucket Quotas
MyFSIO supports **storage quotas** to limit how much data a bucket can hold. Quotas are enforced on uploads and multipart completions.
### Quota Types
| Limit | Description |
|-------|-------------|
| **Max Size (MB)** | Maximum total storage in megabytes (includes current objects + archived versions) |
| **Max Objects** | Maximum number of objects (includes current objects + archived versions) |
### Managing Quotas (Admin Only)
Quota management is restricted to administrators (users with `iam:*` or `iam:list_users` permissions).
#### Via UI
1. Navigate to your bucket in the UI
2. Click the **Properties** tab
3. Find the **Storage Quota** card
4. Enter limits:
- **Max Size (MB)**: Leave empty for unlimited
- **Max Objects**: Leave empty for unlimited
5. Click **Update Quota**
To remove a quota, click **Remove Quota**.
#### Via API
```bash
# Set quota (max 100MB, max 1000 objects)
curl -X PUT "http://localhost:5000/bucket/<bucket>?quota" \
-H "Content-Type: application/json" \
-H "X-Access-Key: ..." -H "X-Secret-Key: ..." \
-d '{"max_bytes": 104857600, "max_objects": 1000}'
# Get current quota
curl "http://localhost:5000/bucket/<bucket>?quota" \
-H "X-Access-Key: ..." -H "X-Secret-Key: ..."
# Remove quota
curl -X PUT "http://localhost:5000/bucket/<bucket>?quota" \
-H "Content-Type: application/json" \
-H "X-Access-Key: ..." -H "X-Secret-Key: ..." \
-d '{"max_bytes": null, "max_objects": null}'
```
### Quota Behavior
- **Version Counting**: When versioning is enabled, archived versions count toward the quota
- **Enforcement Points**: Quotas are checked during `PUT` object and `CompleteMultipartUpload` operations
- **Error Response**: When quota is exceeded, the API returns `HTTP 400` with error code `QuotaExceeded`
- **Visibility**: All users can view quota usage in the bucket detail page, but only admins can modify quotas
### Example Error
```xml
<Error>
<Code>QuotaExceeded</Code>
<Message>Bucket quota exceeded: storage limit reached</Message>
<BucketName>my-bucket</BucketName>
</Error>
```
## 9. Site Replication
### Permission Model
Replication uses a two-tier permission system:
| Role | Capabilities |
|------|--------------|
| **Admin** (users with `iam:*` permissions) | Create/delete replication rules, configure connections and target buckets |
| **Users** (with `replication` permission) | Enable/disable (pause/resume) existing replication rules |
> **Note:** The Replication tab is hidden for users without the `replication` permission on the bucket.
This separation allows administrators to pre-configure where data should replicate, while allowing authorized users to toggle replication on/off without accessing connection credentials.
### Architecture
@@ -253,13 +495,15 @@ Now, configure the primary instance to replicate to the target.
- **Secret Key**: The secret you generated on the Target.
- Click **Add Connection**.
3. **Enable Replication**:
3. **Enable Replication** (Admin):
- Navigate to **Buckets** and select the source bucket.
- Switch to the **Replication** tab.
- Select the `Secondary Site` connection.
- Enter the target bucket name (`backup-bucket`).
- Click **Enable Replication**.
Once configured, users with `replication` permission on this bucket can pause/resume replication without needing access to connection details.
### Verification
1. Upload a file to the source bucket.
@@ -270,6 +514,18 @@ Now, configure the primary instance to replicate to the target.
aws --endpoint-url http://target-server:5002 s3 ls s3://backup-bucket
```
### Pausing and Resuming Replication
Users with the `replication` permission (but not admin rights) can pause and resume existing replication rules:
1. Navigate to the bucket's **Replication** tab.
2. If replication is **Active**, click **Pause Replication** to temporarily stop syncing.
3. If replication is **Paused**, click **Resume Replication** to continue syncing.
When paused, new objects uploaded to the source will not replicate until replication is resumed. Objects uploaded while paused will be replicated once resumed.
> **Note:** Only admins can create new replication rules, change the target connection/bucket, or delete rules entirely.
### Bidirectional Replication (Active-Active)
To set up two-way replication (Server A ↔ Server B):
@@ -285,7 +541,7 @@ To set up two-way replication (Server A ↔ Server B):
**Note**: Deleting a bucket will automatically remove its associated replication configuration.
## 7. Running Tests
## 11. Running Tests
```bash
pytest -q
@@ -295,7 +551,7 @@ The suite now includes a boto3 integration test that spins up a live HTTP server
The suite covers bucket CRUD, presigned downloads, bucket policy enforcement, and regression tests for anonymous reads when a Public policy is attached.
## 8. Troubleshooting
## 12. Troubleshooting
| Symptom | Likely Cause | Fix |
| --- | --- | --- |
@@ -304,7 +560,7 @@ The suite covers bucket CRUD, presigned downloads, bucket policy enforcement, an
| Presign modal errors with 403 | IAM user lacks `read/write/delete` for target bucket or bucket policy denies | Update IAM inline policies or remove conflicting deny statements. |
| Large upload rejected immediately | File exceeds `MAX_UPLOAD_SIZE` | Increase env var or shrink object. |
## 9. API Matrix
## 13. API Matrix
```
GET / # List buckets
@@ -318,9 +574,11 @@ POST /presign/<bucket>/<key> # Generate SigV4 URL
GET /bucket-policy/<bucket> # Fetch policy
PUT /bucket-policy/<bucket> # Upsert policy
DELETE /bucket-policy/<bucket> # Delete policy
GET /<bucket>?quota # Get bucket quota
PUT /<bucket>?quota # Set bucket quota (admin only)
```
## 10. Next Steps
## 14. Next Steps
- Tailor IAM + policy JSON files for team-ready presets.
- Wrap `run_api.py` with gunicorn or another WSGI server for long-running workloads.

19
requirements.txt
View File

@@ -1,9 +1,10 @@
Flask>=3.0.2
Flask-Limiter>=3.5.0
Flask-Cors>=4.0.0
Flask-WTF>=1.2.1
pytest>=7.4
requests>=2.31
boto3>=1.34
waitress>=2.1.2
psutil>=5.9.0
Flask>=3.1.2
Flask-Limiter>=4.1.0
Flask-Cors>=6.0.1
Flask-WTF>=1.2.2
pytest>=9.0.1
requests>=2.32.5
boto3>=1.42.1
waitress>=3.0.2
psutil>=7.1.3
cryptography>=46.0.3

28
static/css/main.css
View File

@@ -396,12 +396,25 @@ code {
.preview-card { top: 1rem; }
.preview-stage {
min-height: 260px;
background-color: var(--myfsio-preview-bg);
overflow: hidden;
border-color: var(--myfsio-card-border) !important;
}
.preview-stage:has(#preview-placeholder:not(.d-none)) {
min-height: 0;
}
.preview-stage:has(#preview-image:not(.d-none)),
.preview-stage:has(#preview-video:not(.d-none)),
.preview-stage:has(#preview-iframe:not(.d-none)) {
min-height: 200px;
}
#preview-placeholder {
padding: 2rem 1rem;
}
.upload-progress-stack {
display: flex;
flex-direction: column;
@@ -928,6 +941,19 @@ pre code {
background-color: var(--myfsio-hover-bg) !important;
}
.folder-row {
background-color: var(--myfsio-section-bg);
transition: background-color 0.15s ease;
}
.folder-row:hover {
background-color: var(--myfsio-hover-bg) !important;
}
.folder-row td:first-child {
padding-left: 0.5rem;
}
.btn-group-sm .btn {
padding: 0.25rem 0.6rem;
font-size: 0.875rem;

12
templates/base.html
View File

@@ -51,22 +51,18 @@
<li class="nav-item">
<a class="nav-link" href="{{ url_for('ui.buckets_overview') }}">Buckets</a>
</li>
{% if can_manage_iam %}
<li class="nav-item">
<a class="nav-link {% if not can_manage_iam %}nav-link-muted{% endif %}" href="{{ url_for('ui.iam_dashboard') }}">
IAM
{% if not can_manage_iam %}<span class="badge ms-2 text-bg-warning">Restricted</span>{% endif %}
</a>
<a class="nav-link" href="{{ url_for('ui.iam_dashboard') }}">IAM</a>
</li>
<li class="nav-item">
<a class="nav-link {% if not can_manage_iam %}nav-link-muted{% endif %}" href="{{ url_for('ui.connections_dashboard') }}">
Connections
{% if not can_manage_iam %}<span class="badge ms-2 text-bg-warning">Restricted</span>{% endif %}
</a>
<a class="nav-link" href="{{ url_for('ui.connections_dashboard') }}">Connections</a>
</li>
<li class="nav-item">
<a class="nav-link" href="{{ url_for('ui.metrics_dashboard') }}">Metrics</a>
</li>
{% endif %}
{% endif %}
{% if principal %}
<li class="nav-item">
<a class="nav-link" href="{{ url_for('ui.docs_page') }}">Docs</a>

1155
templates/bucket_detail.html
View File

File diff suppressed because it is too large Load Diff

176
templates/docs.html
View File

@@ -83,6 +83,16 @@ python run.py --mode ui
<td><code>5000</code></td>
<td>Listen port.</td>
</tr>
<tr>
<td><code>ENCRYPTION_ENABLED</code></td>
<td><code>false</code></td>
<td>Enable server-side encryption support.</td>
</tr>
<tr>
<td><code>KMS_ENABLED</code></td>
<td><code>false</code></td>
<td>Enable KMS key management for encryption.</td>
</tr>
</tbody>
</table>
</div>
@@ -408,10 +418,172 @@ s3.complete_multipart_upload(
</p>
</div>
</article>
<article id="quotas" class="card shadow-sm docs-section">
<div class="card-body">
<div class="d-flex align-items-center gap-2 mb-3">
<span class="docs-section-kicker">10</span>
<h2 class="h4 mb-0">Bucket Quotas</h2>
</div>
<p class="text-muted">Limit how much data a bucket can hold using storage quotas. Quotas are enforced on uploads and multipart completions.</p>
<h3 class="h6 text-uppercase text-muted mt-4">Quota Types</h3>
<div class="table-responsive mb-3">
<table class="table table-sm table-bordered small">
<thead class="table-light">
<tr>
<th>Limit</th>
<th>Description</th>
</tr>
</thead>
<tbody>
<tr>
<td><strong>Max Size (MB)</strong></td>
<td>Maximum total storage in megabytes (includes current objects + archived versions)</td>
</tr>
<tr>
<td><strong>Max Objects</strong></td>
<td>Maximum number of objects (includes current objects + archived versions)</td>
</tr>
</tbody>
</table>
</div>
<h3 class="h6 text-uppercase text-muted mt-4">Managing Quotas (Admin Only)</h3>
<p class="small text-muted">Quota management is restricted to administrators (users with <code>iam:*</code> permissions).</p>
<ol class="docs-steps mb-3">
<li>Navigate to your bucket → <strong>Properties</strong> tab → <strong>Storage Quota</strong> card.</li>
<li>Enter limits: <strong>Max Size (MB)</strong> and/or <strong>Max Objects</strong>. Leave empty for unlimited.</li>
<li>Click <strong>Update Quota</strong> to save, or <strong>Remove Quota</strong> to clear limits.</li>
</ol>
<h3 class="h6 text-uppercase text-muted mt-4">API Usage</h3>
<pre class="mb-3"><code class="language-bash"># Set quota (max 100MB, max 1000 objects)
curl -X PUT "{{ api_base }}/bucket/&lt;bucket&gt;?quota" \
-H "Content-Type: application/json" \
-H "X-Access-Key: &lt;key&gt;" -H "X-Secret-Key: &lt;secret&gt;" \
-d '{"max_bytes": 104857600, "max_objects": 1000}'
# Get current quota
curl "{{ api_base }}/bucket/&lt;bucket&gt;?quota" \
-H "X-Access-Key: &lt;key&gt;" -H "X-Secret-Key: &lt;secret&gt;"
# Remove quota
curl -X PUT "{{ api_base }}/bucket/&lt;bucket&gt;?quota" \
-H "Content-Type: application/json" \
-H "X-Access-Key: &lt;key&gt;" -H "X-Secret-Key: &lt;secret&gt;" \
-d '{"max_bytes": null, "max_objects": null}'</code></pre>
<div class="alert alert-light border mb-0">
<div class="d-flex gap-2">
<svg xmlns="http://www.w3.org/2000/svg" width="16" height="16" fill="currentColor" class="bi bi-info-circle text-muted mt-1" viewBox="0 0 16 16">
<path d="M8 15A7 7 0 1 1 8 1a7 7 0 0 1 0 14zm0 1A8 8 0 1 0 8 0a8 8 0 0 0 0 16z"/>
<path d="m8.93 6.588-2.29.287-.082.38.45.083c.294.07.352.176.288.469l-.738 3.468c-.194.897.105 1.319.808 1.319.545 0 1.178-.252 1.465-.598l.088-.416c-.2.176-.492.246-.686.246-.275 0-.375-.193-.304-.533L8.93 6.588zM9 4.5a1 1 0 1 1-2 0 1 1 0 0 1 2 0z"/>
</svg>
<div>
<strong>Version Counting:</strong> When versioning is enabled, archived versions count toward the quota. The quota is checked against total storage, not just current objects.
</div>
</div>
</div>
</div>
</article>
<article id="encryption" class="card shadow-sm docs-section">
<div class="card-body">
<div class="d-flex align-items-center gap-2 mb-3">
<span class="docs-section-kicker">11</span>
<h2 class="h4 mb-0">Encryption</h2>
</div>
<p class="text-muted">Protect data at rest with server-side encryption using AES-256-GCM. Objects are encrypted before being written to disk and decrypted transparently on read.</p>
<h3 class="h6 text-uppercase text-muted mt-4">Encryption Types</h3>
<div class="table-responsive mb-3">
<table class="table table-sm table-bordered small">
<thead class="table-light">
<tr>
<th>Type</th>
<th>Description</th>
</tr>
</thead>
<tbody>
<tr>
<td><strong>AES-256 (SSE-S3)</strong></td>
<td>Server-managed encryption using a local master key</td>
</tr>
<tr>
<td><strong>KMS (SSE-KMS)</strong></td>
<td>Encryption using customer-managed keys via the built-in KMS</td>
</tr>
</tbody>
</table>
</div>
<h3 class="h6 text-uppercase text-muted mt-4">Enabling Encryption</h3>
<ol class="docs-steps mb-3">
<li>
<strong>Set environment variables:</strong>
<pre class="mb-2"><code class="language-bash"># PowerShell
$env:ENCRYPTION_ENABLED = "true"
$env:KMS_ENABLED = "true" # Optional
python run.py
# Bash
export ENCRYPTION_ENABLED=true
export KMS_ENABLED=true
python run.py</code></pre>
</li>
<li>
<strong>Configure bucket encryption:</strong> Navigate to your bucket → <strong>Properties</strong> tab → <strong>Default Encryption</strong> card → Click <strong>Enable Encryption</strong>.
</li>
<li>
<strong>Choose algorithm:</strong> Select <strong>AES-256</strong> for server-managed keys or <strong>aws:kms</strong> to use a KMS-managed key.
</li>
</ol>
<div class="alert alert-warning border-warning bg-warning-subtle mb-3">
<div class="d-flex gap-2">
<svg xmlns="http://www.w3.org/2000/svg" width="16" height="16" fill="currentColor" class="bi bi-exclamation-triangle mt-1" viewBox="0 0 16 16">
<path d="M7.938 2.016A.13.13 0 0 1 8.002 2a.13.13 0 0 1 .063.016.146.146 0 0 1 .054.057l6.857 11.667c.036.06.035.124.002.183a.163.163 0 0 1-.054.06.116.116 0 0 1-.066.017H1.146a.115.115 0 0 1-.066-.017.163.163 0 0 1-.054-.06.176.176 0 0 1 .002-.183L7.884 2.073a.147.147 0 0 1 .054-.057zm1.044-.45a1.13 1.13 0 0 0-1.96 0L.165 13.233c-.457.778.091 1.767.98 1.767h13.713c.889 0 1.438-.99.98-1.767L8.982 1.566z"/>
<path d="M7.002 12a1 1 0 1 1 2 0 1 1 0 0 1-2 0zM7.1 5.995a.905.905 0 1 1 1.8 0l-.35 3.507a.552.552 0 0 1-1.1 0L7.1 5.995z"/>
</svg>
<div>
<strong>Important:</strong> Only <em>new uploads</em> after enabling encryption will be encrypted. Existing objects remain unencrypted.
</div>
</div>
</div>
<h3 class="h6 text-uppercase text-muted mt-4">KMS Key Management</h3>
<p class="small text-muted">When <code>KMS_ENABLED=true</code>, manage encryption keys via the API:</p>
<pre class="mb-3"><code class="language-bash"># Create a new KMS key
curl -X POST {{ api_base }}/kms/keys \
-H "Content-Type: application/json" \
-H "X-Access-Key: &lt;key&gt;" -H "X-Secret-Key: &lt;secret&gt;" \
-d '{"alias": "my-key", "description": "Production key"}'
# List all keys
curl {{ api_base }}/kms/keys \
-H "X-Access-Key: &lt;key&gt;" -H "X-Secret-Key: &lt;secret&gt;"
# Rotate a key (creates new key material)
curl -X POST {{ api_base }}/kms/keys/{key-id}/rotate \
-H "X-Access-Key: &lt;key&gt;" -H "X-Secret-Key: &lt;secret&gt;"
# Disable/Enable a key
curl -X POST {{ api_base }}/kms/keys/{key-id}/disable \
-H "X-Access-Key: &lt;key&gt;" -H "X-Secret-Key: &lt;secret&gt;"
# Schedule key deletion (30-day waiting period)
curl -X DELETE "{{ api_base }}/kms/keys/{key-id}?waiting_period_days=30" \
-H "X-Access-Key: &lt;key&gt;" -H "X-Secret-Key: &lt;secret&gt;"</code></pre>
<h3 class="h6 text-uppercase text-muted mt-4">How It Works</h3>
<p class="small text-muted mb-0">
<strong>Envelope Encryption:</strong> Each object is encrypted with a unique Data Encryption Key (DEK). The DEK is then encrypted (wrapped) by the master key or KMS key and stored alongside the ciphertext. On read, the DEK is unwrapped and used to decrypt the object transparently.
</p>
</div>
</article>
<article id="troubleshooting" class="card shadow-sm docs-section">
<div class="card-body">
<div class="d-flex align-items-center gap-2 mb-3">
<span class="docs-section-kicker">09</span>
<span class="docs-section-kicker">12</span>
<h2 class="h4 mb-0">Troubleshooting &amp; tips</h2>
</div>
<div class="table-responsive">
@@ -468,6 +640,8 @@ s3.complete_multipart_upload(
<li><a href="#api">REST endpoints</a></li>
<li><a href="#examples">API Examples</a></li>
<li><a href="#replication">Site Replication</a></li>
<li><a href="#quotas">Bucket Quotas</a></li>
<li><a href="#encryption">Encryption</a></li>
<li><a href="#troubleshooting">Troubleshooting</a></li>
</ul>
<div class="docs-sidebar-callouts">

8
templates/iam.html
View File

@@ -509,7 +509,7 @@
full: [
{
bucket: '*',
actions: ['list', 'read', 'write', 'delete', 'share', 'policy', 'iam:list_users', 'iam:*'],
actions: ['list', 'read', 'write', 'delete', 'share', 'policy', 'replication', 'iam:list_users', 'iam:*'],
},
],
readonly: [
@@ -543,7 +543,7 @@
full: [
{
bucket: '*',
actions: ['list', 'read', 'write', 'delete', 'share', 'policy', 'iam:list_users', 'iam:*'],
actions: ['list', 'read', 'write', 'delete', 'share', 'policy', 'replication', 'iam:list_users', 'iam:*'],
},
],
readonly: [
@@ -688,7 +688,9 @@
rotateDoneBtn.classList.remove('d-none');
} catch (err) {
alert(err.message);
if (window.showToast) {
window.showToast(err.message, 'Error', 'danger');
}
rotateSecretModal.hide();
} finally {
confirmRotateBtn.disabled = false;

763
tests/test_encryption.py Normal file
View File

@@ -0,0 +1,763 @@
"""Tests for encryption functionality."""
from __future__ import annotations
import base64
import io
import json
import os
import secrets
import tempfile
from pathlib import Path
import pytest
class TestLocalKeyEncryption:
"""Tests for LocalKeyEncryption provider."""
def test_create_master_key(self, tmp_path):
"""Test that master key is created if it doesn't exist."""
from app.encryption import LocalKeyEncryption
key_path = tmp_path / "keys" / "master.key"
provider = LocalKeyEncryption(key_path)
# Access master key to trigger creation
key = provider.master_key
assert key_path.exists()
assert len(key) == 32 # 256-bit key
def test_load_existing_master_key(self, tmp_path):
"""Test loading an existing master key."""
from app.encryption import LocalKeyEncryption
key_path = tmp_path / "master.key"
original_key = secrets.token_bytes(32)
key_path.write_text(base64.b64encode(original_key).decode())
provider = LocalKeyEncryption(key_path)
loaded_key = provider.master_key
assert loaded_key == original_key
def test_encrypt_decrypt_roundtrip(self, tmp_path):
"""Test that data can be encrypted and decrypted correctly."""
from app.encryption import LocalKeyEncryption
key_path = tmp_path / "master.key"
provider = LocalKeyEncryption(key_path)
plaintext = b"Hello, World! This is a test message."
# Encrypt
result = provider.encrypt(plaintext)
assert result.ciphertext != plaintext
assert result.key_id == "local"
assert len(result.nonce) == 12
assert len(result.encrypted_data_key) > 0
# Decrypt
decrypted = provider.decrypt(
result.ciphertext,
result.nonce,
result.encrypted_data_key,
result.key_id,
)
assert decrypted == plaintext
def test_different_data_keys_per_encryption(self, tmp_path):
"""Test that each encryption uses a different data key."""
from app.encryption import LocalKeyEncryption
key_path = tmp_path / "master.key"
provider = LocalKeyEncryption(key_path)
plaintext = b"Same message"
result1 = provider.encrypt(plaintext)
result2 = provider.encrypt(plaintext)
# Different encrypted data keys
assert result1.encrypted_data_key != result2.encrypted_data_key
# Different nonces
assert result1.nonce != result2.nonce
# Different ciphertexts
assert result1.ciphertext != result2.ciphertext
def test_generate_data_key(self, tmp_path):
"""Test data key generation."""
from app.encryption import LocalKeyEncryption
key_path = tmp_path / "master.key"
provider = LocalKeyEncryption(key_path)
plaintext_key, encrypted_key = provider.generate_data_key()
assert len(plaintext_key) == 32
assert len(encrypted_key) > 32 # nonce + ciphertext + tag
# Verify we can decrypt the key
decrypted_key = provider._decrypt_data_key(encrypted_key)
assert decrypted_key == plaintext_key
def test_decrypt_with_wrong_key_fails(self, tmp_path):
"""Test that decryption fails with wrong master key."""
from app.encryption import LocalKeyEncryption, EncryptionError
# Create two providers with different keys
key_path1 = tmp_path / "master1.key"
key_path2 = tmp_path / "master2.key"
provider1 = LocalKeyEncryption(key_path1)
provider2 = LocalKeyEncryption(key_path2)
# Encrypt with provider1
plaintext = b"Secret message"
result = provider1.encrypt(plaintext)
# Try to decrypt with provider2
with pytest.raises(EncryptionError):
provider2.decrypt(
result.ciphertext,
result.nonce,
result.encrypted_data_key,
result.key_id,
)
class TestEncryptionMetadata:
"""Tests for EncryptionMetadata class."""
def test_to_dict(self):
"""Test converting metadata to dictionary."""
from app.encryption import EncryptionMetadata
nonce = secrets.token_bytes(12)
encrypted_key = secrets.token_bytes(60)
metadata = EncryptionMetadata(
algorithm="AES256",
key_id="local",
nonce=nonce,
encrypted_data_key=encrypted_key,
)
result = metadata.to_dict()
assert result["x-amz-server-side-encryption"] == "AES256"
assert result["x-amz-encryption-key-id"] == "local"
assert base64.b64decode(result["x-amz-encryption-nonce"]) == nonce
assert base64.b64decode(result["x-amz-encrypted-data-key"]) == encrypted_key
def test_from_dict(self):
"""Test creating metadata from dictionary."""
from app.encryption import EncryptionMetadata
nonce = secrets.token_bytes(12)
encrypted_key = secrets.token_bytes(60)
data = {
"x-amz-server-side-encryption": "AES256",
"x-amz-encryption-key-id": "local",
"x-amz-encryption-nonce": base64.b64encode(nonce).decode(),
"x-amz-encrypted-data-key": base64.b64encode(encrypted_key).decode(),
}
metadata = EncryptionMetadata.from_dict(data)
assert metadata is not None
assert metadata.algorithm == "AES256"
assert metadata.key_id == "local"
assert metadata.nonce == nonce
assert metadata.encrypted_data_key == encrypted_key
def test_from_dict_returns_none_for_unencrypted(self):
"""Test that from_dict returns None for unencrypted objects."""
from app.encryption import EncryptionMetadata
data = {"some-other-key": "value"}
metadata = EncryptionMetadata.from_dict(data)
assert metadata is None
class TestStreamingEncryptor:
"""Tests for streaming encryption."""
def test_encrypt_decrypt_stream(self, tmp_path):
"""Test streaming encryption and decryption."""
from app.encryption import LocalKeyEncryption, StreamingEncryptor
key_path = tmp_path / "master.key"
provider = LocalKeyEncryption(key_path)
encryptor = StreamingEncryptor(provider, chunk_size=1024)
# Create test data
original_data = b"A" * 5000 + b"B" * 5000 + b"C" * 5000 # 15KB
stream = io.BytesIO(original_data)
# Encrypt
encrypted_stream, metadata = encryptor.encrypt_stream(stream)
encrypted_data = encrypted_stream.read()
assert encrypted_data != original_data
assert metadata.algorithm == "AES256"
# Decrypt
encrypted_stream = io.BytesIO(encrypted_data)
decrypted_stream = encryptor.decrypt_stream(encrypted_stream, metadata)
decrypted_data = decrypted_stream.read()
assert decrypted_data == original_data
def test_encrypt_small_data(self, tmp_path):
"""Test encrypting data smaller than chunk size."""
from app.encryption import LocalKeyEncryption, StreamingEncryptor
key_path = tmp_path / "master.key"
provider = LocalKeyEncryption(key_path)
encryptor = StreamingEncryptor(provider, chunk_size=1024)
original_data = b"Small data"
stream = io.BytesIO(original_data)
encrypted_stream, metadata = encryptor.encrypt_stream(stream)
encrypted_stream.seek(0)
decrypted_stream = encryptor.decrypt_stream(encrypted_stream, metadata)
decrypted_data = decrypted_stream.read()
assert decrypted_data == original_data
def test_encrypt_empty_data(self, tmp_path):
"""Test encrypting empty data."""
from app.encryption import LocalKeyEncryption, StreamingEncryptor
key_path = tmp_path / "master.key"
provider = LocalKeyEncryption(key_path)
encryptor = StreamingEncryptor(provider)
stream = io.BytesIO(b"")
encrypted_stream, metadata = encryptor.encrypt_stream(stream)
encrypted_stream.seek(0)
decrypted_stream = encryptor.decrypt_stream(encrypted_stream, metadata)
decrypted_data = decrypted_stream.read()
assert decrypted_data == b""
class TestEncryptionManager:
"""Tests for EncryptionManager."""
def test_encryption_disabled_by_default(self, tmp_path):
"""Test that encryption is disabled by default."""
from app.encryption import EncryptionManager
config = {
"encryption_enabled": False,
"encryption_master_key_path": str(tmp_path / "master.key"),
}
manager = EncryptionManager(config)
assert not manager.enabled
def test_encryption_enabled(self, tmp_path):
"""Test enabling encryption."""
from app.encryption import EncryptionManager
config = {
"encryption_enabled": True,
"encryption_master_key_path": str(tmp_path / "master.key"),
"default_encryption_algorithm": "AES256",
}
manager = EncryptionManager(config)
assert manager.enabled
assert manager.default_algorithm == "AES256"
def test_encrypt_decrypt_object(self, tmp_path):
"""Test encrypting and decrypting an object."""
from app.encryption import EncryptionManager
config = {
"encryption_enabled": True,
"encryption_master_key_path": str(tmp_path / "master.key"),
}
manager = EncryptionManager(config)
plaintext = b"Object data to encrypt"
ciphertext, metadata = manager.encrypt_object(plaintext)
assert ciphertext != plaintext
assert metadata.algorithm == "AES256"
decrypted = manager.decrypt_object(ciphertext, metadata)
assert decrypted == plaintext
class TestClientEncryptionHelper:
"""Tests for client-side encryption helpers."""
def test_generate_client_key(self):
"""Test generating a client encryption key."""
from app.encryption import ClientEncryptionHelper
key_info = ClientEncryptionHelper.generate_client_key()
assert "key" in key_info
assert key_info["algorithm"] == "AES-256-GCM"
assert "created_at" in key_info
# Verify key is 256 bits
key = base64.b64decode(key_info["key"])
assert len(key) == 32
def test_encrypt_with_key(self):
"""Test encrypting data with a client key."""
from app.encryption import ClientEncryptionHelper
key = base64.b64encode(secrets.token_bytes(32)).decode()
plaintext = b"Client-side encrypted data"
result = ClientEncryptionHelper.encrypt_with_key(plaintext, key)
assert "ciphertext" in result
assert "nonce" in result
assert result["algorithm"] == "AES-256-GCM"
def test_encrypt_decrypt_with_key(self):
"""Test round-trip client-side encryption."""
from app.encryption import ClientEncryptionHelper
key = base64.b64encode(secrets.token_bytes(32)).decode()
plaintext = b"Client-side encrypted data"
encrypted = ClientEncryptionHelper.encrypt_with_key(plaintext, key)
decrypted = ClientEncryptionHelper.decrypt_with_key(
encrypted["ciphertext"],
encrypted["nonce"],
key,
)
assert decrypted == plaintext
def test_wrong_key_fails(self):
"""Test that decryption with wrong key fails."""
from app.encryption import ClientEncryptionHelper, EncryptionError
key1 = base64.b64encode(secrets.token_bytes(32)).decode()
key2 = base64.b64encode(secrets.token_bytes(32)).decode()
plaintext = b"Secret data"
encrypted = ClientEncryptionHelper.encrypt_with_key(plaintext, key1)
with pytest.raises(EncryptionError):
ClientEncryptionHelper.decrypt_with_key(
encrypted["ciphertext"],
encrypted["nonce"],
key2,
)
class TestKMSManager:
"""Tests for KMS key management."""
def test_create_key(self, tmp_path):
"""Test creating a KMS key."""
from app.kms import KMSManager
keys_path = tmp_path / "kms_keys.json"
master_key_path = tmp_path / "master.key"
kms = KMSManager(keys_path, master_key_path)
key = kms.create_key("Test key", key_id="test-key-1")
assert key.key_id == "test-key-1"
assert key.description == "Test key"
assert key.enabled
assert keys_path.exists()
def test_list_keys(self, tmp_path):
"""Test listing KMS keys."""
from app.kms import KMSManager
keys_path = tmp_path / "kms_keys.json"
master_key_path = tmp_path / "master.key"
kms = KMSManager(keys_path, master_key_path)
kms.create_key("Key 1", key_id="key-1")
kms.create_key("Key 2", key_id="key-2")
keys = kms.list_keys()
assert len(keys) == 2
key_ids = {k.key_id for k in keys}
assert "key-1" in key_ids
assert "key-2" in key_ids
def test_get_key(self, tmp_path):
"""Test getting a specific key."""
from app.kms import KMSManager
keys_path = tmp_path / "kms_keys.json"
master_key_path = tmp_path / "master.key"
kms = KMSManager(keys_path, master_key_path)
kms.create_key("Test key", key_id="test-key")
key = kms.get_key("test-key")
assert key is not None
assert key.key_id == "test-key"
# Non-existent key
assert kms.get_key("non-existent") is None
def test_enable_disable_key(self, tmp_path):
"""Test enabling and disabling keys."""
from app.kms import KMSManager
keys_path = tmp_path / "kms_keys.json"
master_key_path = tmp_path / "master.key"
kms = KMSManager(keys_path, master_key_path)
kms.create_key("Test key", key_id="test-key")
# Initially enabled
assert kms.get_key("test-key").enabled
# Disable
kms.disable_key("test-key")
assert not kms.get_key("test-key").enabled
# Enable
kms.enable_key("test-key")
assert kms.get_key("test-key").enabled
def test_delete_key(self, tmp_path):
"""Test deleting a key."""
from app.kms import KMSManager
keys_path = tmp_path / "kms_keys.json"
master_key_path = tmp_path / "master.key"
kms = KMSManager(keys_path, master_key_path)
kms.create_key("Test key", key_id="test-key")
assert kms.get_key("test-key") is not None
kms.delete_key("test-key")
assert kms.get_key("test-key") is None
def test_encrypt_decrypt(self, tmp_path):
"""Test KMS encrypt and decrypt."""
from app.kms import KMSManager
keys_path = tmp_path / "kms_keys.json"
master_key_path = tmp_path / "master.key"
kms = KMSManager(keys_path, master_key_path)
key = kms.create_key("Test key", key_id="test-key")
plaintext = b"Secret data to encrypt"
ciphertext = kms.encrypt("test-key", plaintext)
assert ciphertext != plaintext
decrypted, key_id = kms.decrypt(ciphertext)
assert decrypted == plaintext
assert key_id == "test-key"
def test_encrypt_with_context(self, tmp_path):
"""Test encryption with encryption context."""
from app.kms import KMSManager, EncryptionError
keys_path = tmp_path / "kms_keys.json"
master_key_path = tmp_path / "master.key"
kms = KMSManager(keys_path, master_key_path)
kms.create_key("Test key", key_id="test-key")
plaintext = b"Secret data"
context = {"bucket": "test-bucket", "key": "test-key"}
ciphertext = kms.encrypt("test-key", plaintext, context)
# Decrypt with same context succeeds
decrypted, _ = kms.decrypt(ciphertext, context)
assert decrypted == plaintext
# Decrypt with different context fails
with pytest.raises(EncryptionError):
kms.decrypt(ciphertext, {"different": "context"})
def test_generate_data_key(self, tmp_path):
"""Test generating a data key."""
from app.kms import KMSManager
keys_path = tmp_path / "kms_keys.json"
master_key_path = tmp_path / "master.key"
kms = KMSManager(keys_path, master_key_path)
kms.create_key("Test key", key_id="test-key")
plaintext_key, encrypted_key = kms.generate_data_key("test-key")
assert len(plaintext_key) == 32
assert len(encrypted_key) > 0
# Decrypt the encrypted key
decrypted_key = kms.decrypt_data_key("test-key", encrypted_key)
assert decrypted_key == plaintext_key
def test_disabled_key_cannot_encrypt(self, tmp_path):
"""Test that disabled keys cannot be used for encryption."""
from app.kms import KMSManager, EncryptionError
keys_path = tmp_path / "kms_keys.json"
master_key_path = tmp_path / "master.key"
kms = KMSManager(keys_path, master_key_path)
kms.create_key("Test key", key_id="test-key")
kms.disable_key("test-key")
with pytest.raises(EncryptionError, match="disabled"):
kms.encrypt("test-key", b"data")
def test_re_encrypt(self, tmp_path):
"""Test re-encrypting data with a different key."""
from app.kms import KMSManager
keys_path = tmp_path / "kms_keys.json"
master_key_path = tmp_path / "master.key"
kms = KMSManager(keys_path, master_key_path)
kms.create_key("Key 1", key_id="key-1")
kms.create_key("Key 2", key_id="key-2")
plaintext = b"Data to re-encrypt"
# Encrypt with key-1
ciphertext1 = kms.encrypt("key-1", plaintext)
# Re-encrypt with key-2
ciphertext2 = kms.re_encrypt(ciphertext1, "key-2")
# Decrypt with key-2
decrypted, key_id = kms.decrypt(ciphertext2)
assert decrypted == plaintext
assert key_id == "key-2"
def test_generate_random(self, tmp_path):
"""Test generating random bytes."""
from app.kms import KMSManager
keys_path = tmp_path / "kms_keys.json"
master_key_path = tmp_path / "master.key"
kms = KMSManager(keys_path, master_key_path)
random1 = kms.generate_random(32)
random2 = kms.generate_random(32)
assert len(random1) == 32
assert len(random2) == 32
assert random1 != random2 # Very unlikely to be equal
def test_keys_persist_across_instances(self, tmp_path):
"""Test that keys persist and can be loaded by new instances."""
from app.kms import KMSManager
keys_path = tmp_path / "kms_keys.json"
master_key_path = tmp_path / "master.key"
# Create key with first instance
kms1 = KMSManager(keys_path, master_key_path)
kms1.create_key("Test key", key_id="test-key")
plaintext = b"Persistent encryption test"
ciphertext = kms1.encrypt("test-key", plaintext)
# Create new instance and verify key works
kms2 = KMSManager(keys_path, master_key_path)
decrypted, key_id = kms2.decrypt(ciphertext)
assert decrypted == plaintext
assert key_id == "test-key"
class TestKMSEncryptionProvider:
"""Tests for KMS encryption provider."""
def test_kms_encryption_provider(self, tmp_path):
"""Test using KMS as an encryption provider."""
from app.kms import KMSManager
keys_path = tmp_path / "kms_keys.json"
master_key_path = tmp_path / "master.key"
kms = KMSManager(keys_path, master_key_path)
kms.create_key("Test key", key_id="test-key")
provider = kms.get_provider("test-key")
plaintext = b"Data encrypted with KMS provider"
result = provider.encrypt(plaintext)
assert result.key_id == "test-key"
assert result.ciphertext != plaintext
decrypted = provider.decrypt(
result.ciphertext,
result.nonce,
result.encrypted_data_key,
result.key_id,
)
assert decrypted == plaintext
class TestEncryptedStorage:
"""Tests for encrypted storage layer."""
def test_put_and_get_encrypted_object(self, tmp_path):
"""Test storing and retrieving an encrypted object."""
from app.storage import ObjectStorage
from app.encryption import EncryptionManager
from app.encrypted_storage import EncryptedObjectStorage
storage_root = tmp_path / "storage"
storage = ObjectStorage(storage_root)
config = {
"encryption_enabled": True,
"encryption_master_key_path": str(tmp_path / "master.key"),
"default_encryption_algorithm": "AES256",
}
encryption = EncryptionManager(config)
encrypted_storage = EncryptedObjectStorage(storage, encryption)
# Create bucket with encryption config
storage.create_bucket("test-bucket")
storage.set_bucket_encryption("test-bucket", {
"Rules": [{"SSEAlgorithm": "AES256"}]
})
# Put object
original_data = b"This is secret data that should be encrypted"
stream = io.BytesIO(original_data)
meta = encrypted_storage.put_object(
"test-bucket",
"secret.txt",
stream,
)
assert meta is not None
# Verify file on disk is encrypted (not plaintext)
file_path = storage_root / "test-bucket" / "secret.txt"
stored_data = file_path.read_bytes()
assert stored_data != original_data
# Get object - should be decrypted
data, metadata = encrypted_storage.get_object_data("test-bucket", "secret.txt")
assert data == original_data
def test_no_encryption_without_config(self, tmp_path):
"""Test that objects are not encrypted without bucket config."""
from app.storage import ObjectStorage
from app.encryption import EncryptionManager
from app.encrypted_storage import EncryptedObjectStorage
storage_root = tmp_path / "storage"
storage = ObjectStorage(storage_root)
config = {
"encryption_enabled": True,
"encryption_master_key_path": str(tmp_path / "master.key"),
}
encryption = EncryptionManager(config)
encrypted_storage = EncryptedObjectStorage(storage, encryption)
storage.create_bucket("test-bucket")
# No encryption config
original_data = b"Unencrypted data"
stream = io.BytesIO(original_data)
encrypted_storage.put_object("test-bucket", "plain.txt", stream)
# Verify file on disk is NOT encrypted
file_path = storage_root / "test-bucket" / "plain.txt"
stored_data = file_path.read_bytes()
assert stored_data == original_data
def test_explicit_encryption_request(self, tmp_path):
"""Test explicitly requesting encryption."""
from app.storage import ObjectStorage
from app.encryption import EncryptionManager
from app.encrypted_storage import EncryptedObjectStorage
storage_root = tmp_path / "storage"
storage = ObjectStorage(storage_root)
config = {
"encryption_enabled": True,
"encryption_master_key_path": str(tmp_path / "master.key"),
}
encryption = EncryptionManager(config)
encrypted_storage = EncryptedObjectStorage(storage, encryption)
storage.create_bucket("test-bucket")
original_data = b"Explicitly encrypted data"
stream = io.BytesIO(original_data)
# Request encryption explicitly
encrypted_storage.put_object(
"test-bucket",
"encrypted.txt",
stream,
server_side_encryption="AES256",
)
# Verify file is encrypted
file_path = storage_root / "test-bucket" / "encrypted.txt"
stored_data = file_path.read_bytes()
assert stored_data != original_data
# Get object - should be decrypted
data, _ = encrypted_storage.get_object_data("test-bucket", "encrypted.txt")
assert data == original_data

506
tests/test_kms_api.py Normal file
View File

@@ -0,0 +1,506 @@
"""Tests for KMS API endpoints."""
from __future__ import annotations
import base64
import json
import secrets
import pytest
@pytest.fixture
def kms_client(tmp_path):
"""Create a test client with KMS enabled."""
from app import create_app
app = create_app({
"TESTING": True,
"STORAGE_ROOT": str(tmp_path / "storage"),
"IAM_CONFIG": str(tmp_path / "iam.json"),
"BUCKET_POLICY_PATH": str(tmp_path / "policies.json"),
"ENCRYPTION_ENABLED": True,
"KMS_ENABLED": True,
"ENCRYPTION_MASTER_KEY_PATH": str(tmp_path / "master.key"),
"KMS_KEYS_PATH": str(tmp_path / "kms_keys.json"),
})
# Create default IAM config with admin user
iam_config = {
"users": [
{
"access_key": "test-access-key",
"secret_key": "test-secret-key",
"display_name": "Test User",
"permissions": ["*"]
}
]
}
(tmp_path / "iam.json").write_text(json.dumps(iam_config))
return app.test_client()
@pytest.fixture
def auth_headers():
"""Get authentication headers."""
return {
"X-Access-Key": "test-access-key",
"X-Secret-Key": "test-secret-key",
}
class TestKMSKeyManagement:
"""Tests for KMS key management endpoints."""
def test_create_key(self, kms_client, auth_headers):
"""Test creating a KMS key."""
response = kms_client.post(
"/kms/keys",
json={"Description": "Test encryption key"},
headers=auth_headers,
)
assert response.status_code == 200
data = response.get_json()
assert "KeyMetadata" in data
assert data["KeyMetadata"]["Description"] == "Test encryption key"
assert data["KeyMetadata"]["Enabled"] is True
assert "KeyId" in data["KeyMetadata"]
def test_create_key_with_custom_id(self, kms_client, auth_headers):
"""Test creating a key with a custom ID."""
response = kms_client.post(
"/kms/keys",
json={"KeyId": "my-custom-key", "Description": "Custom key"},
headers=auth_headers,
)
assert response.status_code == 200
data = response.get_json()
assert data["KeyMetadata"]["KeyId"] == "my-custom-key"
def test_list_keys(self, kms_client, auth_headers):
"""Test listing KMS keys."""
# Create some keys
kms_client.post("/kms/keys", json={"Description": "Key 1"}, headers=auth_headers)
kms_client.post("/kms/keys", json={"Description": "Key 2"}, headers=auth_headers)
response = kms_client.get("/kms/keys", headers=auth_headers)
assert response.status_code == 200
data = response.get_json()
assert "Keys" in data
assert len(data["Keys"]) == 2
def test_get_key(self, kms_client, auth_headers):
"""Test getting a specific key."""
# Create a key
create_response = kms_client.post(
"/kms/keys",
json={"KeyId": "test-key", "Description": "Test key"},
headers=auth_headers,
)
response = kms_client.get("/kms/keys/test-key", headers=auth_headers)
assert response.status_code == 200
data = response.get_json()
assert data["KeyMetadata"]["KeyId"] == "test-key"
assert data["KeyMetadata"]["Description"] == "Test key"
def test_get_nonexistent_key(self, kms_client, auth_headers):
"""Test getting a key that doesn't exist."""
response = kms_client.get("/kms/keys/nonexistent", headers=auth_headers)
assert response.status_code == 404
def test_delete_key(self, kms_client, auth_headers):
"""Test deleting a key."""
# Create a key
kms_client.post("/kms/keys", json={"KeyId": "test-key"}, headers=auth_headers)
# Delete it
response = kms_client.delete("/kms/keys/test-key", headers=auth_headers)
assert response.status_code == 204
# Verify it's gone
get_response = kms_client.get("/kms/keys/test-key", headers=auth_headers)
assert get_response.status_code == 404
def test_enable_disable_key(self, kms_client, auth_headers):
"""Test enabling and disabling a key."""
# Create a key
kms_client.post("/kms/keys", json={"KeyId": "test-key"}, headers=auth_headers)
# Disable
response = kms_client.post("/kms/keys/test-key/disable", headers=auth_headers)
assert response.status_code == 200
# Verify disabled
get_response = kms_client.get("/kms/keys/test-key", headers=auth_headers)
assert get_response.get_json()["KeyMetadata"]["Enabled"] is False
# Enable
response = kms_client.post("/kms/keys/test-key/enable", headers=auth_headers)
assert response.status_code == 200
# Verify enabled
get_response = kms_client.get("/kms/keys/test-key", headers=auth_headers)
assert get_response.get_json()["KeyMetadata"]["Enabled"] is True
class TestKMSEncryption:
"""Tests for KMS encryption operations."""
def test_encrypt_decrypt(self, kms_client, auth_headers):
"""Test encrypting and decrypting data."""
# Create a key
kms_client.post("/kms/keys", json={"KeyId": "test-key"}, headers=auth_headers)
plaintext = b"Hello, World!"
plaintext_b64 = base64.b64encode(plaintext).decode()
# Encrypt
encrypt_response = kms_client.post(
"/kms/encrypt",
json={"KeyId": "test-key", "Plaintext": plaintext_b64},
headers=auth_headers,
)
assert encrypt_response.status_code == 200
encrypt_data = encrypt_response.get_json()
assert "CiphertextBlob" in encrypt_data
assert encrypt_data["KeyId"] == "test-key"
# Decrypt
decrypt_response = kms_client.post(
"/kms/decrypt",
json={"CiphertextBlob": encrypt_data["CiphertextBlob"]},
headers=auth_headers,
)
assert decrypt_response.status_code == 200
decrypt_data = decrypt_response.get_json()
decrypted = base64.b64decode(decrypt_data["Plaintext"])
assert decrypted == plaintext
def test_encrypt_with_context(self, kms_client, auth_headers):
"""Test encryption with encryption context."""
kms_client.post("/kms/keys", json={"KeyId": "test-key"}, headers=auth_headers)
plaintext = b"Contextualized data"
plaintext_b64 = base64.b64encode(plaintext).decode()
context = {"purpose": "testing", "bucket": "my-bucket"}
# Encrypt with context
encrypt_response = kms_client.post(
"/kms/encrypt",
json={
"KeyId": "test-key",
"Plaintext": plaintext_b64,
"EncryptionContext": context,
},
headers=auth_headers,
)
assert encrypt_response.status_code == 200
ciphertext = encrypt_response.get_json()["CiphertextBlob"]
# Decrypt with same context succeeds
decrypt_response = kms_client.post(
"/kms/decrypt",
json={
"CiphertextBlob": ciphertext,
"EncryptionContext": context,
},
headers=auth_headers,
)
assert decrypt_response.status_code == 200
# Decrypt with wrong context fails
wrong_context_response = kms_client.post(
"/kms/decrypt",
json={
"CiphertextBlob": ciphertext,
"EncryptionContext": {"wrong": "context"},
},
headers=auth_headers,
)
assert wrong_context_response.status_code == 400
def test_encrypt_missing_key_id(self, kms_client, auth_headers):
"""Test encryption without KeyId."""
response = kms_client.post(
"/kms/encrypt",
json={"Plaintext": base64.b64encode(b"data").decode()},
headers=auth_headers,
)
assert response.status_code == 400
assert "KeyId is required" in response.get_json()["message"]
def test_encrypt_missing_plaintext(self, kms_client, auth_headers):
"""Test encryption without Plaintext."""
kms_client.post("/kms/keys", json={"KeyId": "test-key"}, headers=auth_headers)
response = kms_client.post(
"/kms/encrypt",
json={"KeyId": "test-key"},
headers=auth_headers,
)
assert response.status_code == 400
assert "Plaintext is required" in response.get_json()["message"]
class TestKMSDataKey:
"""Tests for KMS data key generation."""
def test_generate_data_key(self, kms_client, auth_headers):
"""Test generating a data key."""
kms_client.post("/kms/keys", json={"KeyId": "test-key"}, headers=auth_headers)
response = kms_client.post(
"/kms/generate-data-key",
json={"KeyId": "test-key"},
headers=auth_headers,
)
assert response.status_code == 200
data = response.get_json()
assert "Plaintext" in data
assert "CiphertextBlob" in data
assert data["KeyId"] == "test-key"
# Verify plaintext key is 256 bits (32 bytes)
plaintext_key = base64.b64decode(data["Plaintext"])
assert len(plaintext_key) == 32
def test_generate_data_key_aes_128(self, kms_client, auth_headers):
"""Test generating an AES-128 data key."""
kms_client.post("/kms/keys", json={"KeyId": "test-key"}, headers=auth_headers)
response = kms_client.post(
"/kms/generate-data-key",
json={"KeyId": "test-key", "KeySpec": "AES_128"},
headers=auth_headers,
)
assert response.status_code == 200
data = response.get_json()
# Verify plaintext key is 128 bits (16 bytes)
plaintext_key = base64.b64decode(data["Plaintext"])
assert len(plaintext_key) == 16
def test_generate_data_key_without_plaintext(self, kms_client, auth_headers):
"""Test generating a data key without plaintext."""
kms_client.post("/kms/keys", json={"KeyId": "test-key"}, headers=auth_headers)
response = kms_client.post(
"/kms/generate-data-key-without-plaintext",
json={"KeyId": "test-key"},
headers=auth_headers,
)
assert response.status_code == 200
data = response.get_json()
assert "CiphertextBlob" in data
assert "Plaintext" not in data
class TestKMSReEncrypt:
"""Tests for KMS re-encryption."""
def test_re_encrypt(self, kms_client, auth_headers):
"""Test re-encrypting data with a different key."""
# Create two keys
kms_client.post("/kms/keys", json={"KeyId": "key-1"}, headers=auth_headers)
kms_client.post("/kms/keys", json={"KeyId": "key-2"}, headers=auth_headers)
# Encrypt with key-1
plaintext = b"Data to re-encrypt"
encrypt_response = kms_client.post(
"/kms/encrypt",
json={
"KeyId": "key-1",
"Plaintext": base64.b64encode(plaintext).decode(),
},
headers=auth_headers,
)
ciphertext = encrypt_response.get_json()["CiphertextBlob"]
# Re-encrypt with key-2
re_encrypt_response = kms_client.post(
"/kms/re-encrypt",
json={
"CiphertextBlob": ciphertext,
"DestinationKeyId": "key-2",
},
headers=auth_headers,
)
assert re_encrypt_response.status_code == 200
data = re_encrypt_response.get_json()
assert data["SourceKeyId"] == "key-1"
assert data["KeyId"] == "key-2"
# Verify new ciphertext can be decrypted
decrypt_response = kms_client.post(
"/kms/decrypt",
json={"CiphertextBlob": data["CiphertextBlob"]},
headers=auth_headers,
)
decrypted = base64.b64decode(decrypt_response.get_json()["Plaintext"])
assert decrypted == plaintext
class TestKMSRandom:
"""Tests for random number generation."""
def test_generate_random(self, kms_client, auth_headers):
"""Test generating random bytes."""
response = kms_client.post(
"/kms/generate-random",
json={"NumberOfBytes": 64},
headers=auth_headers,
)
assert response.status_code == 200
data = response.get_json()
random_bytes = base64.b64decode(data["Plaintext"])
assert len(random_bytes) == 64
def test_generate_random_default_size(self, kms_client, auth_headers):
"""Test generating random bytes with default size."""
response = kms_client.post(
"/kms/generate-random",
json={},
headers=auth_headers,
)
assert response.status_code == 200
data = response.get_json()
random_bytes = base64.b64decode(data["Plaintext"])
assert len(random_bytes) == 32 # Default is 32 bytes
class TestClientSideEncryption:
"""Tests for client-side encryption helpers."""
def test_generate_client_key(self, kms_client, auth_headers):
"""Test generating a client encryption key."""
response = kms_client.post(
"/kms/client/generate-key",
headers=auth_headers,
)
assert response.status_code == 200
data = response.get_json()
assert "key" in data
assert data["algorithm"] == "AES-256-GCM"
key = base64.b64decode(data["key"])
assert len(key) == 32
def test_client_encrypt_decrypt(self, kms_client, auth_headers):
"""Test client-side encryption and decryption."""
# Generate a key
key_response = kms_client.post("/kms/client/generate-key", headers=auth_headers)
key = key_response.get_json()["key"]
# Encrypt
plaintext = b"Client-side encrypted data"
encrypt_response = kms_client.post(
"/kms/client/encrypt",
json={
"Plaintext": base64.b64encode(plaintext).decode(),
"Key": key,
},
headers=auth_headers,
)
assert encrypt_response.status_code == 200
encrypted = encrypt_response.get_json()
# Decrypt
decrypt_response = kms_client.post(
"/kms/client/decrypt",
json={
"Ciphertext": encrypted["ciphertext"],
"Nonce": encrypted["nonce"],
"Key": key,
},
headers=auth_headers,
)
assert decrypt_response.status_code == 200
decrypted = base64.b64decode(decrypt_response.get_json()["Plaintext"])
assert decrypted == plaintext
class TestEncryptionMaterials:
"""Tests for S3 encryption materials endpoint."""
def test_get_encryption_materials(self, kms_client, auth_headers):
"""Test getting encryption materials for client-side S3 encryption."""
# Create a key
kms_client.post("/kms/keys", json={"KeyId": "s3-key"}, headers=auth_headers)
response = kms_client.post(
"/kms/materials/s3-key",
json={},
headers=auth_headers,
)
assert response.status_code == 200
data = response.get_json()
assert "PlaintextKey" in data
assert "EncryptedKey" in data
assert data["KeyId"] == "s3-key"
assert data["Algorithm"] == "AES-256-GCM"
# Verify key is 256 bits
key = base64.b64decode(data["PlaintextKey"])
assert len(key) == 32
class TestKMSAuthentication:
"""Tests for KMS authentication requirements."""
def test_unauthenticated_request_fails(self, kms_client):
"""Test that unauthenticated requests are rejected."""
response = kms_client.get("/kms/keys")
# Should fail with 403 (no credentials)
assert response.status_code == 403
def test_invalid_credentials_fail(self, kms_client):
"""Test that invalid credentials are rejected."""
response = kms_client.get(
"/kms/keys",
headers={
"X-Access-Key": "wrong-key",
"X-Secret-Key": "wrong-secret",
},
)
assert response.status_code == 403

7
tests/test_security.py
View File

@@ -116,7 +116,12 @@ def test_path_traversal_in_key(client, signer):
def test_storage_path_traversal(app):
storage = app.extensions["object_storage"]
from app.storage import StorageError
from app.storage import StorageError, ObjectStorage
from app.encrypted_storage import EncryptedObjectStorage
# Get the underlying ObjectStorage if wrapped
if isinstance(storage, EncryptedObjectStorage):
storage = storage.storage
with pytest.raises(StorageError, match="Object key contains parent directory references"):
storage._sanitize_object_key("folder/../file.txt")

268
tests/test_ui_encryption.py Normal file
View File

@@ -0,0 +1,268 @@
"""Tests for UI-based encryption configuration."""
import json
from pathlib import Path
import pytest
from app import create_app
def get_csrf_token(response):
"""Extract CSRF token from response HTML."""
html = response.data.decode("utf-8")
import re
match = re.search(r'name="csrf_token"\s+value="([^"]+)"', html)
return match.group(1) if match else None
def _make_encryption_app(tmp_path: Path, *, kms_enabled: bool = True):
"""Create an app with encryption enabled."""
storage_root = tmp_path / "data"
iam_config = tmp_path / "iam.json"
bucket_policies = tmp_path / "bucket_policies.json"
iam_payload = {
"users": [
{
"access_key": "test",
"secret_key": "secret",
"display_name": "Test User",
"policies": [{"bucket": "*", "actions": ["list", "read", "write", "delete", "policy"]}],
},
{
"access_key": "readonly",
"secret_key": "secret",
"display_name": "Read Only User",
"policies": [{"bucket": "*", "actions": ["list", "read"]}],
},
]
}
iam_config.write_text(json.dumps(iam_payload))
config = {
"TESTING": True,
"STORAGE_ROOT": storage_root,
"IAM_CONFIG": iam_config,
"BUCKET_POLICY_PATH": bucket_policies,
"API_BASE_URL": "http://testserver",
"SECRET_KEY": "testing",
"ENCRYPTION_ENABLED": True,
}
if kms_enabled:
config["KMS_ENABLED"] = True
config["KMS_KEYS_PATH"] = str(tmp_path / "kms_keys.json")
config["ENCRYPTION_MASTER_KEY_PATH"] = str(tmp_path / "master.key")
app = create_app(config)
storage = app.extensions["object_storage"]
storage.create_bucket("test-bucket")
return app
class TestUIBucketEncryption:
"""Test bucket encryption configuration via UI."""
def test_bucket_detail_shows_encryption_card(self, tmp_path):
"""Encryption card should be visible on bucket detail page."""
app = _make_encryption_app(tmp_path)
client = app.test_client()
# Login first
client.post("/ui/login", data={"access_key": "test", "secret_key": "secret"}, follow_redirects=True)
response = client.get("/ui/buckets/test-bucket?tab=properties")
assert response.status_code == 200
html = response.data.decode("utf-8")
assert "Default Encryption" in html
assert "Encryption Algorithm" in html or "Default encryption disabled" in html
def test_enable_aes256_encryption(self, tmp_path):
"""Should be able to enable AES-256 encryption."""
app = _make_encryption_app(tmp_path)
client = app.test_client()
# Login
client.post("/ui/login", data={"access_key": "test", "secret_key": "secret"}, follow_redirects=True)
# Get CSRF token
response = client.get("/ui/buckets/test-bucket?tab=properties")
csrf_token = get_csrf_token(response)
# Enable AES-256 encryption
response = client.post(
"/ui/buckets/test-bucket/encryption",
data={
"csrf_token": csrf_token,
"action": "enable",
"algorithm": "AES256",
},
follow_redirects=True,
)
assert response.status_code == 200
html = response.data.decode("utf-8")
# Should see success message or enabled state
assert "AES-256" in html or "encryption enabled" in html.lower()
def test_enable_kms_encryption(self, tmp_path):
"""Should be able to enable KMS encryption."""
app = _make_encryption_app(tmp_path, kms_enabled=True)
client = app.test_client()
# Create a KMS key first
with app.app_context():
kms = app.extensions.get("kms")
if kms:
key = kms.create_key("test-key")
key_id = key.key_id
else:
pytest.skip("KMS not available")
# Login
client.post("/ui/login", data={"access_key": "test", "secret_key": "secret"}, follow_redirects=True)
# Get CSRF token
response = client.get("/ui/buckets/test-bucket?tab=properties")
csrf_token = get_csrf_token(response)
# Enable KMS encryption
response = client.post(
"/ui/buckets/test-bucket/encryption",
data={
"csrf_token": csrf_token,
"action": "enable",
"algorithm": "aws:kms",
"kms_key_id": key_id,
},
follow_redirects=True,
)
assert response.status_code == 200
html = response.data.decode("utf-8")
assert "KMS" in html or "encryption enabled" in html.lower()
def test_disable_encryption(self, tmp_path):
"""Should be able to disable encryption."""
app = _make_encryption_app(tmp_path)
client = app.test_client()
# Login
client.post("/ui/login", data={"access_key": "test", "secret_key": "secret"}, follow_redirects=True)
# First enable encryption
response = client.get("/ui/buckets/test-bucket?tab=properties")
csrf_token = get_csrf_token(response)
client.post(
"/ui/buckets/test-bucket/encryption",
data={
"csrf_token": csrf_token,
"action": "enable",
"algorithm": "AES256",
},
)
# Now disable it
response = client.get("/ui/buckets/test-bucket?tab=properties")
csrf_token = get_csrf_token(response)
response = client.post(
"/ui/buckets/test-bucket/encryption",
data={
"csrf_token": csrf_token,
"action": "disable",
},
follow_redirects=True,
)
assert response.status_code == 200
html = response.data.decode("utf-8")
assert "disabled" in html.lower() or "Default encryption disabled" in html
def test_invalid_algorithm_rejected(self, tmp_path):
"""Invalid encryption algorithm should be rejected."""
app = _make_encryption_app(tmp_path)
client = app.test_client()
# Login
client.post("/ui/login", data={"access_key": "test", "secret_key": "secret"}, follow_redirects=True)
response = client.get("/ui/buckets/test-bucket?tab=properties")
csrf_token = get_csrf_token(response)
response = client.post(
"/ui/buckets/test-bucket/encryption",
data={
"csrf_token": csrf_token,
"action": "enable",
"algorithm": "INVALID",
},
follow_redirects=True,
)
assert response.status_code == 200
html = response.data.decode("utf-8")
assert "Invalid" in html or "danger" in html
def test_encryption_persists_in_config(self, tmp_path):
"""Encryption config should persist in bucket config."""
app = _make_encryption_app(tmp_path)
client = app.test_client()
# Login
client.post("/ui/login", data={"access_key": "test", "secret_key": "secret"}, follow_redirects=True)
# Enable encryption
response = client.get("/ui/buckets/test-bucket?tab=properties")
csrf_token = get_csrf_token(response)
client.post(
"/ui/buckets/test-bucket/encryption",
data={
"csrf_token": csrf_token,
"action": "enable",
"algorithm": "AES256",
},
)
# Verify it's stored
with app.app_context():
storage = app.extensions["object_storage"]
config = storage.get_bucket_encryption("test-bucket")
assert "Rules" in config
assert len(config["Rules"]) == 1
assert config["Rules"][0]["ApplyServerSideEncryptionByDefault"]["SSEAlgorithm"] == "AES256"
class TestUIEncryptionWithoutPermission:
"""Test encryption UI when user lacks permissions."""
def test_readonly_user_cannot_change_encryption(self, tmp_path):
"""Read-only user should not be able to change encryption settings."""
app = _make_encryption_app(tmp_path)
client = app.test_client()
# Login as readonly user
client.post("/ui/login", data={"access_key": "readonly", "secret_key": "secret"}, follow_redirects=True)
# This should fail or be rejected
response = client.get("/ui/buckets/test-bucket?tab=properties")
csrf_token = get_csrf_token(response)
response = client.post(
"/ui/buckets/test-bucket/encryption",
data={
"csrf_token": csrf_token,
"action": "enable",
"algorithm": "AES256",
},
follow_redirects=True,
)
# Should either redirect with error or show permission denied
assert response.status_code == 200
html = response.data.decode("utf-8")
# Should contain error about permission denied
assert "Access denied" in html or "permission" in html.lower() or "not authorized" in html.lower()