Merge pull request 'MyFSIO v0.1.9 Release' (#10) from next into main

Reviewed-on: #10

README.md

@@ -8,7 +8,7 @@ MyFSIO is a batteries-included, Flask-based recreation of Amazon S3 and IAM work
 - **IAM + access keys:** Users, access keys, key rotation, and bucket-scoped actions (`list/read/write/delete/policy`) now live in `data/.myfsio.sys/config/iam.json` and are editable from the IAM dashboard.
 - **Bucket policies + hot reload:** `data/.myfsio.sys/config/bucket_policies.json` uses AWS' policy grammar (Version `2012-10-17`) with a built-in watcher, so editing the JSON file applies immediately. The UI also ships Public/Private/Custom presets for faster edits.
 - **Presigned URLs everywhere:** Signature Version 4 presigned URLs respect IAM + bucket policies and replace the now-removed "share link" feature for public access scenarios.
-- **Modern UI:** Responsive tables, quick filters, preview sidebar, object-level delete buttons, a presign modal, and an inline JSON policy editor that respects dark mode keep bucket management friendly.
+- **Modern UI:** Responsive tables, quick filters, preview sidebar, object-level delete buttons, a presign modal, and an inline JSON policy editor that respects dark mode keep bucket management friendly. The object browser supports folder navigation, infinite scroll pagination, bulk operations, and automatic retry on load failures.
 - **Tests & health:** `/healthz` for smoke checks and `pytest` coverage for IAM, CRUD, presign, and policy flows.
 
 ## Architecture at a Glance

app/__init__.py

@@ -45,7 +45,6 @@ def _migrate_config_file(active_path: Path, legacy_paths: List[Path]) -> Path:
             try:
                 shutil.move(str(legacy_path), str(active_path))
             except OSError:
-                # Fall back to copy + delete if move fails (e.g., cross-device)
                 shutil.copy2(legacy_path, active_path)
                 try:
                     legacy_path.unlink(missing_ok=True)
@@ -101,32 +100,28 @@ def create_app(
     bucket_policies = BucketPolicyStore(Path(app.config["BUCKET_POLICY_PATH"]))
     secret_store = EphemeralSecretStore(default_ttl=app.config.get("SECRET_TTL_SECONDS", 300))
     
-    # Initialize Replication components
-    # Store config files in the system config directory for consistency
     storage_root = Path(app.config["STORAGE_ROOT"])
     config_dir = storage_root / ".myfsio.sys" / "config"
     config_dir.mkdir(parents=True, exist_ok=True)
     
-    # Define paths with migration from legacy locations
     connections_path = _migrate_config_file(
         active_path=config_dir / "connections.json",
         legacy_paths=[
-            storage_root / ".myfsio.sys" / "connections.json",  # Previous location
+            storage_root / ".myfsio.sys" / "connections.json",
-            storage_root / ".connections.json",  # Original legacy location
+            storage_root / ".connections.json",
         ],
     )
     replication_rules_path = _migrate_config_file(
         active_path=config_dir / "replication_rules.json",
         legacy_paths=[
-            storage_root / ".myfsio.sys" / "replication_rules.json",  # Previous location
+            storage_root / ".myfsio.sys" / "replication_rules.json",
-            storage_root / ".replication_rules.json",  # Original legacy location
+            storage_root / ".replication_rules.json",
         ],
     )
     
     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"),
@@ -141,7 +136,6 @@ def create_app(
         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)
@@ -177,13 +171,22 @@ def create_app(
 
     @app.template_filter("timestamp_to_datetime")
     def timestamp_to_datetime(value: float) -> str:
-        """Format Unix timestamp as human-readable datetime."""
+        """Format Unix timestamp as human-readable datetime in configured timezone."""
-        from datetime import datetime
+        from datetime import datetime, timezone as dt_timezone
+        from zoneinfo import ZoneInfo
         if not value:
             return "Never"
         try:
-            dt = datetime.fromtimestamp(value)
+            dt_utc = datetime.fromtimestamp(value, dt_timezone.utc)
-            return dt.strftime("%Y-%m-%d %H:%M:%S")
+            display_tz = app.config.get("DISPLAY_TIMEZONE", "UTC")
+            if display_tz and display_tz != "UTC":
+                try:
+                    tz = ZoneInfo(display_tz)
+                    dt_local = dt_utc.astimezone(tz)
+                    return dt_local.strftime("%Y-%m-%d %H:%M:%S")
+                except (KeyError, ValueError):
+                    pass 
+            return dt_utc.strftime("%Y-%m-%d %H:%M:%S UTC")
         except (ValueError, OSError):
             return "Unknown"
 
@@ -244,7 +247,7 @@ def _configure_cors(app: Flask) -> None:
 class _RequestContextFilter(logging.Filter):
     """Inject request-specific attributes into log records."""
 
-    def filter(self, record: logging.LogRecord) -> bool:  # pragma: no cover - simple boilerplate
+    def filter(self, record: logging.LogRecord) -> bool:  
         if has_request_context():
             record.request_id = getattr(g, "request_id", "-")
             record.path = request.path

app/bucket_policies.py

@@ -188,7 +188,6 @@ class BucketPolicyStore:
         except FileNotFoundError:
             return None
 
-    # ------------------------------------------------------------------
     def evaluate(
         self,
         access_key: Optional[str],
@@ -229,7 +228,6 @@ class BucketPolicyStore:
         self._policies.pop(bucket, None)
         self._persist()
 
-    # ------------------------------------------------------------------
     def _load(self) -> None:
         try:
             content = self.policy_path.read_text(encoding='utf-8')

app/config.py

@@ -73,6 +73,7 @@ class AppConfig:
     kms_enabled: bool
     kms_keys_path: Path
     default_encryption_algorithm: str
+    display_timezone: str
 
     @classmethod
     def from_env(cls, overrides: Optional[Dict[str, Any]] = None) -> "AppConfig":
@@ -153,15 +154,15 @@ class AppConfig:
         cors_allow_headers = _csv(str(_get("CORS_ALLOW_HEADERS", "*")), ["*"])
         cors_expose_headers = _csv(str(_get("CORS_EXPOSE_HEADERS", "*")), ["*"])
         session_lifetime_days = int(_get("SESSION_LIFETIME_DAYS", 30))
-        bucket_stats_cache_ttl = int(_get("BUCKET_STATS_CACHE_TTL", 60))  # Default 60 seconds
+        bucket_stats_cache_ttl = int(_get("BUCKET_STATS_CACHE_TTL", 60)) 
         
-        # 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"))
+        display_timezone = str(_get("DISPLAY_TIMEZONE", "UTC"))
 
         return cls(storage_root=storage_root,
                    max_upload_size=max_upload_size,
@@ -196,7 +197,8 @@ class AppConfig:
                    encryption_master_key_path=encryption_master_key_path,
                    kms_enabled=kms_enabled,
                    kms_keys_path=kms_keys_path,
-                   default_encryption_algorithm=default_encryption_algorithm)
+                   default_encryption_algorithm=default_encryption_algorithm,
+                   display_timezone=display_timezone)
 
     def validate_and_report(self) -> list[str]:
         """Validate configuration and return a list of warnings/issues.
@@ -206,7 +208,6 @@ class AppConfig:
         """
         issues = []
         
-        # Check if storage_root is writable
         try:
             test_file = self.storage_root / ".write_test"
             test_file.touch()
@@ -214,24 +215,20 @@ class AppConfig:
         except (OSError, PermissionError) as e:
             issues.append(f"CRITICAL: STORAGE_ROOT '{self.storage_root}' is not writable: {e}")
         
-        # Check if storage_root looks like a temp directory
         storage_str = str(self.storage_root).lower()
         if "/tmp" in storage_str or "\\temp" in storage_str or "appdata\\local\\temp" in storage_str:
             issues.append(f"WARNING: STORAGE_ROOT '{self.storage_root}' appears to be a temporary directory. Data may be lost on reboot!")
         
-        # Check if IAM config path is under storage_root
         try:
             self.iam_config_path.relative_to(self.storage_root)
         except ValueError:
             issues.append(f"WARNING: IAM_CONFIG '{self.iam_config_path}' is outside STORAGE_ROOT '{self.storage_root}'. Consider setting IAM_CONFIG explicitly or ensuring paths are aligned.")
         
-        # Check if bucket policy path is under storage_root
         try:
             self.bucket_policy_path.relative_to(self.storage_root)
         except ValueError:
             issues.append(f"WARNING: BUCKET_POLICY_PATH '{self.bucket_policy_path}' is outside STORAGE_ROOT '{self.storage_root}'. Consider setting BUCKET_POLICY_PATH explicitly.")
         
-        # Check if log path is writable
         try:
             self.log_path.parent.mkdir(parents=True, exist_ok=True)
             test_log = self.log_path.parent / ".write_test"
@@ -240,26 +237,22 @@ class AppConfig:
         except (OSError, PermissionError) as e:
             issues.append(f"WARNING: Log directory '{self.log_path.parent}' is not writable: {e}")
         
-        # Check log path location
         log_str = str(self.log_path).lower()
         if "/tmp" in log_str or "\\temp" in log_str or "appdata\\local\\temp" in log_str:
             issues.append(f"WARNING: LOG_DIR '{self.log_path.parent}' appears to be a temporary directory. Logs may be lost on reboot!")
         
-        # Check if encryption keys path is under storage_root (when encryption is enabled)
         if self.encryption_enabled:
             try:
                 self.encryption_master_key_path.relative_to(self.storage_root)
             except ValueError:
                 issues.append(f"WARNING: ENCRYPTION_MASTER_KEY_PATH '{self.encryption_master_key_path}' is outside STORAGE_ROOT. Ensure proper backup procedures.")
         
-        # Check if KMS keys path is under storage_root (when KMS is enabled)
         if self.kms_enabled:
             try:
                 self.kms_keys_path.relative_to(self.storage_root)
             except ValueError:
                 issues.append(f"WARNING: KMS_KEYS_PATH '{self.kms_keys_path}' is outside STORAGE_ROOT. Ensure proper backup procedures.")
         
-        # Warn about production settings
         if self.secret_key == "dev-secret-key":
             issues.append("WARNING: Using default SECRET_KEY. Set SECRET_KEY environment variable for production.")
         
@@ -330,4 +323,5 @@ class AppConfig:
             "KMS_ENABLED": self.kms_enabled,
             "KMS_KEYS_PATH": str(self.kms_keys_path),
             "DEFAULT_ENCRYPTION_ALGORITHM": self.default_encryption_algorithm,
+            "DISPLAY_TIMEZONE": self.display_timezone,
         }

app/encrypted_storage.py

@@ -188,8 +188,11 @@ class EncryptedObjectStorage:
     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):
+    def list_objects(self, bucket_name: str, **kwargs):
-        return self.storage.list_objects(bucket_name)
+        return self.storage.list_objects(bucket_name, **kwargs)
+    
+    def list_objects_all(self, bucket_name: str):
+        return self.storage.list_objects_all(bucket_name)
     
     def get_object_path(self, bucket_name: str, object_key: str):
         return self.storage.get_object_path(bucket_name, object_key)

app/iam.py

@@ -6,7 +6,7 @@ import math
 import secrets
 from collections import deque
 from dataclasses import dataclass
-from datetime import datetime, timedelta
+from datetime import datetime, timedelta, timezone
 from pathlib import Path
 from typing import Any, Deque, Dict, Iterable, List, Optional, Sequence, Set
 
@@ -125,7 +125,6 @@ class IamService:
         except OSError:
             pass
 
-    # ---------------------- authz helpers ----------------------
     def authenticate(self, access_key: str, secret_key: str) -> Principal:
         self._maybe_reload()
         access_key = (access_key or "").strip()
@@ -149,7 +148,7 @@ class IamService:
             return
         attempts = self._failed_attempts.setdefault(access_key, deque())
         self._prune_attempts(attempts)
-        attempts.append(datetime.now())
+        attempts.append(datetime.now(timezone.utc))
 
     def _clear_failed_attempts(self, access_key: str) -> None:
         if not access_key:
@@ -157,7 +156,7 @@ class IamService:
         self._failed_attempts.pop(access_key, None)
 
     def _prune_attempts(self, attempts: Deque[datetime]) -> None:
-        cutoff = datetime.now() - self.auth_lockout_window
+        cutoff = datetime.now(timezone.utc) - self.auth_lockout_window
         while attempts and attempts[0] < cutoff:
             attempts.popleft()
 
@@ -178,7 +177,7 @@ class IamService:
         if len(attempts) < self.auth_max_attempts:
             return 0
         oldest = attempts[0]
-        elapsed = (datetime.now() - oldest).total_seconds()
+        elapsed = (datetime.now(timezone.utc) - oldest).total_seconds()
         return int(max(0, self.auth_lockout_window.total_seconds() - elapsed))
 
     def principal_for_key(self, access_key: str) -> Principal:
@@ -218,7 +217,6 @@ class IamService:
                 return True
         return False
 
-    # ---------------------- management helpers ----------------------
     def list_users(self) -> List[Dict[str, Any]]:
         listing: List[Dict[str, Any]] = []
         for access_key, record in self._users.items():
@@ -291,7 +289,6 @@ class IamService:
         self._save()
         self._load()
 
-    # ---------------------- config helpers ----------------------
     def _load(self) -> None:
         try:
             self._last_load_time = self.config_path.stat().st_mtime
@@ -337,7 +334,6 @@ class IamService:
         except (OSError, PermissionError) as e:
             raise IamError(f"Cannot save IAM config: {e}")
 
-    # ---------------------- insight helpers ----------------------
     def config_summary(self) -> Dict[str, Any]:
         return {
             "path": str(self.config_path),

app/kms_api.py

@@ -33,9 +33,6 @@ def _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():
@@ -65,7 +62,6 @@ def list_or_create_keys():
         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],
@@ -96,7 +92,6 @@ def get_or_delete_key(key_id: str):
         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)
@@ -149,9 +144,6 @@ def disable_key(key_id: str):
     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():
@@ -251,7 +243,6 @@ def generate_data_key():
     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]
         
@@ -322,10 +313,7 @@ def re_encrypt():
         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({
@@ -365,9 +353,6 @@ def generate_random():
     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():
@@ -427,9 +412,6 @@ def client_decrypt():
     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):

app/replication.py

@@ -9,7 +9,7 @@ import time
 from concurrent.futures import ThreadPoolExecutor
 from dataclasses import dataclass, field
 from pathlib import Path
-from typing import Dict, Optional
+from typing import Any, Dict, Optional
 
 import boto3
 from botocore.config import Config
@@ -22,18 +22,51 @@ from .storage import ObjectStorage, StorageError
 logger = logging.getLogger(__name__)
 
 REPLICATION_USER_AGENT = "S3ReplicationAgent/1.0"
+REPLICATION_CONNECT_TIMEOUT = 5
+REPLICATION_READ_TIMEOUT = 30
+STREAMING_THRESHOLD_BYTES = 10 * 1024 * 1024  # 10 MiB - use streaming for larger files
 
 REPLICATION_MODE_NEW_ONLY = "new_only"
 REPLICATION_MODE_ALL = "all"
 
 
+def _create_s3_client(connection: RemoteConnection, *, health_check: bool = False) -> Any:
+    """Create a boto3 S3 client for the given connection.
+
+    Args:
+        connection: Remote S3 connection configuration
+        health_check: If True, use minimal retries for quick health checks
+
+    Returns:
+        Configured boto3 S3 client
+    """
+    config = Config(
+        user_agent_extra=REPLICATION_USER_AGENT,
+        connect_timeout=REPLICATION_CONNECT_TIMEOUT,
+        read_timeout=REPLICATION_READ_TIMEOUT,
+        retries={'max_attempts': 1 if health_check else 2},
+        signature_version='s3v4',
+        s3={'addressing_style': 'path'},
+        request_checksum_calculation='when_required',
+        response_checksum_validation='when_required',
+    )
+    return boto3.client(
+        "s3",
+        endpoint_url=connection.endpoint_url,
+        aws_access_key_id=connection.access_key,
+        aws_secret_access_key=connection.secret_key,
+        region_name=connection.region or 'us-east-1',
+        config=config,
+    )
+
+
 @dataclass
 class ReplicationStats:
     """Statistics for replication operations - computed dynamically."""
-    objects_synced: int = 0          # Objects that exist in both source and destination
+    objects_synced: int = 0
-    objects_pending: int = 0         # Objects in source but not in destination
+    objects_pending: int = 0
-    objects_orphaned: int = 0        # Objects in destination but not in source (will be deleted)
+    objects_orphaned: int = 0
-    bytes_synced: int = 0            # Total bytes synced to destination
+    bytes_synced: int = 0      
     last_sync_at: Optional[float] = None
     last_sync_key: Optional[str] = None
     
@@ -83,7 +116,6 @@ class ReplicationRule:
     @classmethod
     def from_dict(cls, data: dict) -> "ReplicationRule":
         stats_data = data.pop("stats", {})
-        # Handle old rules without mode/created_at
         if "mode" not in data:
             data["mode"] = REPLICATION_MODE_NEW_ONLY
         if "created_at" not in data:
@@ -101,8 +133,19 @@ class ReplicationManager:
         self._rules: Dict[str, ReplicationRule] = {}
         self._stats_lock = threading.Lock()
         self._executor = ThreadPoolExecutor(max_workers=4, thread_name_prefix="ReplicationWorker")
+        self._shutdown = False
         self.reload_rules()
 
+    def shutdown(self, wait: bool = True) -> None:
+        """Shutdown the replication executor gracefully.
+
+        Args:
+            wait: If True, wait for pending tasks to complete
+        """
+        self._shutdown = True
+        self._executor.shutdown(wait=wait)
+        logger.info("Replication manager shut down")
+
     def reload_rules(self) -> None:
         if not self.rules_path.exists():
             self._rules = {}
@@ -121,6 +164,20 @@ class ReplicationManager:
         with open(self.rules_path, "w") as f:
             json.dump(data, f, indent=2)
 
+    def check_endpoint_health(self, connection: RemoteConnection) -> bool:
+        """Check if a remote endpoint is reachable and responsive.
+
+        Returns True if endpoint is healthy, False otherwise.
+        Uses short timeouts to prevent blocking.
+        """
+        try:
+            s3 = _create_s3_client(connection, health_check=True)
+            s3.list_buckets()
+            return True
+        except Exception as e:
+            logger.warning(f"Endpoint health check failed for {connection.name} ({connection.endpoint_url}): {e}")
+            return False
+
     def get_rule(self, bucket_name: str) -> Optional[ReplicationRule]:
         return self._rules.get(bucket_name)
 
@@ -151,21 +208,13 @@ class ReplicationManager:
         
         connection = self.connections.get(rule.target_connection_id)
         if not connection:
-            return rule.stats  # Return cached stats if connection unavailable
+            return rule.stats
         
         try:
-            # Get source objects
+            source_objects = self.storage.list_objects_all(bucket_name)
-            source_objects = self.storage.list_objects(bucket_name)
             source_keys = {obj.key: obj.size for obj in source_objects}
 
-            # Get destination objects
+            s3 = _create_s3_client(connection)
-            s3 = boto3.client(
-                "s3",
-                endpoint_url=connection.endpoint_url,
-                aws_access_key_id=connection.access_key,
-                aws_secret_access_key=connection.secret_key,
-                region_name=connection.region,
-            )
 
             dest_keys = set()
             bytes_synced = 0
@@ -178,24 +227,18 @@ class ReplicationManager:
                             bytes_synced += obj.get('Size', 0)
             except ClientError as e:
                 if e.response['Error']['Code'] == 'NoSuchBucket':
-                    # Destination bucket doesn't exist yet
                     dest_keys = set()
                 else:
                     raise
             
-            # Compute stats
+            synced = source_keys.keys() & dest_keys  
-            synced = source_keys.keys() & dest_keys  # Objects in both
+            orphaned = dest_keys - source_keys.keys() 
-            orphaned = dest_keys - source_keys.keys()  # In dest but not source
             
-            # For "new_only" mode, we can't determine pending since we don't know
-            # which objects existed before replication was enabled. Only "all" mode
-            # should show pending (objects that should be replicated but aren't yet).
             if rule.mode == REPLICATION_MODE_ALL:
-                pending = source_keys.keys() - dest_keys  # In source but not dest
+                pending = source_keys.keys() - dest_keys 
             else:
-                pending = set()  # New-only mode: don't show pre-existing as pending
+                pending = set()
             
-            # Update cached stats with computed values
             rule.stats.objects_synced = len(synced)
             rule.stats.objects_pending = len(pending)
             rule.stats.objects_orphaned = len(orphaned)
@@ -205,7 +248,7 @@ class ReplicationManager:
             
         except (ClientError, StorageError) as e:
             logger.error(f"Failed to compute sync status for {bucket_name}: {e}")
-            return rule.stats  # Return cached stats on error
+            return rule.stats
     
     def replicate_existing_objects(self, bucket_name: str) -> None:
         """Trigger replication for all existing objects in a bucket."""
@@ -218,8 +261,12 @@ class ReplicationManager:
             logger.warning(f"Cannot replicate existing objects: Connection {rule.target_connection_id} not found")
             return
         
+        if not self.check_endpoint_health(connection):
+            logger.warning(f"Cannot replicate existing objects: Endpoint {connection.name} ({connection.endpoint_url}) is not reachable")
+            return
+        
         try:
-            objects = self.storage.list_objects(bucket_name)
+            objects = self.storage.list_objects_all(bucket_name)
             logger.info(f"Starting replication of {len(objects)} existing objects from {bucket_name}")
             for obj in objects:
                 self._executor.submit(self._replicate_task, bucket_name, obj.key, rule, connection, "write")
@@ -233,13 +280,7 @@ class ReplicationManager:
             raise ValueError(f"Connection {connection_id} not found")
 
         try:
-            s3 = boto3.client(
+            s3 = _create_s3_client(connection)
-                "s3",
-                endpoint_url=connection.endpoint_url,
-                aws_access_key_id=connection.access_key,
-                aws_secret_access_key=connection.secret_key,
-                region_name=connection.region,
-            )
             s3.create_bucket(Bucket=bucket_name)
         except ClientError as e:
             logger.error(f"Failed to create remote bucket {bucket_name}: {e}")
@@ -255,9 +296,22 @@ class ReplicationManager:
             logger.warning(f"Replication skipped for {bucket_name}/{object_key}: Connection {rule.target_connection_id} not found")
             return
 
+        if not self.check_endpoint_health(connection):
+            logger.warning(f"Replication skipped for {bucket_name}/{object_key}: Endpoint {connection.name} ({connection.endpoint_url}) is not reachable")
+            return
+
         self._executor.submit(self._replicate_task, bucket_name, object_key, rule, connection, action)
 
     def _replicate_task(self, bucket_name: str, object_key: str, rule: ReplicationRule, conn: RemoteConnection, action: str) -> None:
+        if self._shutdown:
+            return
+
+        # Re-check if rule is still enabled (may have been paused after task was submitted)
+        current_rule = self.get_rule(bucket_name)
+        if not current_rule or not current_rule.enabled:
+            logger.debug(f"Replication skipped for {bucket_name}/{object_key}: rule disabled or removed")
+            return
+
         if ".." in object_key or object_key.startswith("/") or object_key.startswith("\\"):
             logger.error(f"Invalid object key in replication (path traversal attempt): {object_key}")
             return
@@ -269,17 +323,8 @@ class ReplicationManager:
             logger.error(f"Object key validation failed in replication: {e}")
             return
 
-        file_size = 0
         try:
-            config = Config(user_agent_extra=REPLICATION_USER_AGENT)
+            s3 = _create_s3_client(conn)
-            s3 = boto3.client(
-                "s3",
-                endpoint_url=conn.endpoint_url,
-                aws_access_key_id=conn.access_key,
-                aws_secret_access_key=conn.secret_key,
-                region_name=conn.region,
-                config=config,
-            )
 
             if action == "delete":
                 try:
@@ -296,54 +341,67 @@ class ReplicationManager:
                 logger.error(f"Source object not found: {bucket_name}/{object_key}")
                 return
 
-            metadata = self.storage.get_object_metadata(bucket_name, object_key)
-
-            extra_args = {}
-            if metadata:
-                extra_args["Metadata"] = metadata
-            
-            # Guess content type to prevent corruption/wrong handling
             content_type, _ = mimetypes.guess_type(path)
             file_size = path.stat().st_size
 
             logger.info(f"Replicating {bucket_name}/{object_key}: Size={file_size}, ContentType={content_type}")
 
-            try:
+            def do_upload() -> None:
-                with path.open("rb") as f:
+                """Upload object using appropriate method based on file size.
-                    s3.put_object(
+
-                        Bucket=rule.target_bucket,
+                For small files (< 10 MiB): Read into memory for simpler handling
-                        Key=object_key,
+                For large files: Use streaming upload to avoid memory issues
-                        Body=f,
+                """
-                        ContentLength=file_size,
+                extra_args = {}
-                        ContentType=content_type or "application/octet-stream",
+                if content_type:
-                        Metadata=metadata or {}
+                    extra_args["ContentType"] = content_type
+
+                if file_size >= STREAMING_THRESHOLD_BYTES:
+                    # Use multipart upload for large files
+                    s3.upload_file(
+                        str(path),
+                        rule.target_bucket,
+                        object_key,
+                        ExtraArgs=extra_args if extra_args else None,
                     )
+                else:
+                    # Read small files into memory
+                    file_content = path.read_bytes()
+                    put_kwargs = {
+                        "Bucket": rule.target_bucket,
+                        "Key": object_key,
+                        "Body": file_content,
+                        **extra_args,
+                    }
+                    s3.put_object(**put_kwargs)
+
+            try:
+                do_upload()
             except (ClientError, S3UploadFailedError) as e:
-                is_no_bucket = False
+                error_code = None
                 if isinstance(e, ClientError):
-                    if e.response['Error']['Code'] == 'NoSuchBucket':
+                    error_code = e.response['Error']['Code']
-                        is_no_bucket = True
                 elif isinstance(e, S3UploadFailedError):
                     if "NoSuchBucket" in str(e):
-                        is_no_bucket = True
+                        error_code = 'NoSuchBucket'
 
-                if is_no_bucket:
+                if error_code == 'NoSuchBucket':
                     logger.info(f"Target bucket {rule.target_bucket} not found. Attempting to create it.")
+                    bucket_ready = False
                     try:
                         s3.create_bucket(Bucket=rule.target_bucket)
-                        # Retry upload
+                        bucket_ready = True
-                        with path.open("rb") as f:
+                        logger.info(f"Created target bucket {rule.target_bucket}")
-                            s3.put_object(
+                    except ClientError as bucket_err:
-                                Bucket=rule.target_bucket,
+                        if bucket_err.response['Error']['Code'] in ('BucketAlreadyExists', 'BucketAlreadyOwnedByYou'):
-                                Key=object_key,
+                            logger.debug(f"Bucket {rule.target_bucket} already exists (created by another thread)")
-                                Body=f,
+                            bucket_ready = True
-                                ContentLength=file_size,
+                        else:
-                                ContentType=content_type or "application/octet-stream",
+                            logger.error(f"Failed to create target bucket {rule.target_bucket}: {bucket_err}")
-                                Metadata=metadata or {}
+                            raise e
-                            )
+
-                    except Exception as create_err:
+                    if bucket_ready:
-                        logger.error(f"Failed to create target bucket {rule.target_bucket}: {create_err}")
+                        do_upload()
-                        raise e # Raise original error
                 else:
                     raise e
 
@@ -354,3 +412,4 @@ class ReplicationManager:
             logger.error(f"Replication failed for {bucket_name}/{object_key}: {e}")
         except Exception:
             logger.exception(f"Unexpected error during replication for {bucket_name}/{object_key}")
+

app/s3_api.py

@@ -1,14 +1,16 @@
 """Flask blueprint exposing a subset of the S3 REST API."""
 from __future__ import annotations
 
+import base64
 import hashlib
 import hmac
+import logging
 import mimetypes
 import re
 import uuid
 from datetime import datetime, timedelta, timezone
-from typing import Any, Dict
+from typing import Any, Dict, Optional
-from urllib.parse import quote, urlencode, urlparse
+from urllib.parse import quote, urlencode, urlparse, unquote
 from xml.etree.ElementTree import Element, SubElement, tostring, fromstring, ParseError
 
 from flask import Blueprint, Response, current_app, jsonify, request, g
@@ -20,10 +22,10 @@ from .iam import IamError, Principal
 from .replication import ReplicationManager
 from .storage import ObjectStorage, StorageError, QuotaExceededError
 
+logger = logging.getLogger(__name__)
+
 s3_api_bp = Blueprint("s3_api", __name__)
 
-
-# ---------------------- helpers ----------------------
 def _storage() -> ObjectStorage:
     return current_app.extensions["object_storage"]
 
@@ -68,9 +70,26 @@ def _get_signature_key(key: str, date_stamp: str, region_name: str, service_name
     return k_signing
 
 
+def _get_canonical_uri(req: Any) -> str:
+    """Get the canonical URI for SigV4 signature verification.
+    
+    AWS SigV4 requires the canonical URI to be URL-encoded exactly as the client
+    sent it. Flask/Werkzeug automatically URL-decodes request.path, so we need
+    to get the raw path from the environ.
+    
+    The canonical URI should have each path segment URL-encoded (with '/' preserved),
+    and the encoding should match what the client used when signing.
+    """
+    raw_uri = req.environ.get('RAW_URI') or req.environ.get('REQUEST_URI')
+    
+    if raw_uri:
+        path = raw_uri.split('?')[0]
+        return path
+    
+    return quote(req.path, safe="/-_.~")
+
+
 def _verify_sigv4_header(req: Any, auth_header: str) -> Principal | None:
-    # Parse Authorization header
-    # AWS4-HMAC-SHA256 Credential=AKIA.../20230101/us-east-1/s3/aws4_request, SignedHeaders=host;x-amz-date, Signature=...
     match = re.match(
         r"AWS4-HMAC-SHA256 Credential=([^/]+)/([^/]+)/([^/]+)/([^/]+)/aws4_request, SignedHeaders=([^,]+), Signature=(.+)",
         auth_header,
@@ -79,17 +98,13 @@ def _verify_sigv4_header(req: Any, auth_header: str) -> Principal | None:
         return None
 
     access_key, date_stamp, region, service, signed_headers_str, signature = match.groups()
-
-    # Get secret key
     secret_key = _iam().get_secret_key(access_key)
     if not secret_key:
         raise IamError("Invalid access key")
 
-    # Canonical Request
     method = req.method
-    canonical_uri = quote(req.path, safe="/-_.~")
+    canonical_uri = _get_canonical_uri(req)
     
-    # Canonical Query String
     query_args = []
     for key, value in req.args.items(multi=True):
         query_args.append((key, value))
@@ -100,7 +115,6 @@ def _verify_sigv4_header(req: Any, auth_header: str) -> Principal | None:
         canonical_query_parts.append(f"{quote(k, safe='-_.~')}={quote(v, safe='-_.~')}")
     canonical_query_string = "&".join(canonical_query_parts)
 
-    # Canonical Headers
     signed_headers_list = signed_headers_str.split(";")
     canonical_headers_parts = []
     for header in signed_headers_list:
@@ -108,22 +122,20 @@ def _verify_sigv4_header(req: Any, auth_header: str) -> Principal | None:
         if header_val is None:
              header_val = ""
         
+        if header.lower() == 'expect' and header_val == "":
+            header_val = "100-continue"
+        
         header_val = " ".join(header_val.split())
         canonical_headers_parts.append(f"{header.lower()}:{header_val}\n")
     canonical_headers = "".join(canonical_headers_parts)
 
-    # Payload Hash
     payload_hash = req.headers.get("X-Amz-Content-Sha256")
     if not payload_hash:
         payload_hash = hashlib.sha256(req.get_data()).hexdigest()
 
     canonical_request = f"{method}\n{canonical_uri}\n{canonical_query_string}\n{canonical_headers}\n{signed_headers_str}\n{payload_hash}"
 
-    # String to Sign
+    amz_date = req.headers.get("X-Amz-Date") or req.headers.get("Date")
-    amz_date = req.headers.get("X-Amz-Date")
-    if not amz_date:
-        amz_date = req.headers.get("Date")
-    
     if not amz_date:
         raise IamError("Missing Date header")
 
@@ -134,13 +146,12 @@ def _verify_sigv4_header(req: Any, auth_header: str) -> Principal | None:
 
     now = datetime.now(timezone.utc)
     time_diff = abs((now - request_time).total_seconds())
-    if time_diff > 900:  # 15 minutes
+    if time_diff > 900:
         raise IamError("Request timestamp too old or too far in the future")
 
     required_headers = {'host', 'x-amz-date'}
     signed_headers_set = set(signed_headers_str.split(';'))
     if not required_headers.issubset(signed_headers_set):
-        # Some clients might sign 'date' instead of 'x-amz-date'
         if 'date' in signed_headers_set:
             required_headers.remove('x-amz-date')
             required_headers.add('date')
@@ -154,6 +165,18 @@ def _verify_sigv4_header(req: Any, auth_header: str) -> Principal | None:
     calculated_signature = hmac.new(signing_key, string_to_sign.encode("utf-8"), hashlib.sha256).hexdigest()
 
     if not hmac.compare_digest(calculated_signature, signature):
+        # Only log detailed signature debug info if DEBUG_SIGV4 is enabled
+        if current_app.config.get("DEBUG_SIGV4"):
+            logger.warning(
+                "SigV4 signature mismatch",
+                extra={
+                    "path": req.path,
+                    "method": method,
+                    "signed_headers": signed_headers_str,
+                    "content_type": req.headers.get("Content-Type"),
+                    "content_length": req.headers.get("Content-Length"),
+                }
+            )
         raise IamError("SignatureDoesNotMatch")
 
     return _iam().get_principal(access_key)
@@ -187,11 +210,9 @@ def _verify_sigv4_query(req: Any) -> Principal | None:
     if not secret_key:
         raise IamError("Invalid access key")
 
-    # Canonical Request
     method = req.method
-    canonical_uri = quote(req.path, safe="/-_.~")
+    canonical_uri = _get_canonical_uri(req)
     
-    # Canonical Query String
     query_args = []
     for key, value in req.args.items(multi=True):
         if key != "X-Amz-Signature":
@@ -203,16 +224,16 @@ def _verify_sigv4_query(req: Any) -> Principal | None:
         canonical_query_parts.append(f"{quote(k, safe='-_.~')}={quote(v, safe='-_.~')}")
     canonical_query_string = "&".join(canonical_query_parts)
     
-    # Canonical Headers
     signed_headers_list = signed_headers_str.split(";")
     canonical_headers_parts = []
     for header in signed_headers_list:
         val = req.headers.get(header, "").strip()
+        if header.lower() == 'expect' and val == "":
+            val = "100-continue"
         val = " ".join(val.split())
         canonical_headers_parts.append(f"{header}:{val}\n")
     canonical_headers = "".join(canonical_headers_parts)
     
-    # Payload Hash
     payload_hash = "UNSIGNED-PAYLOAD"
     
     canonical_request = "\n".join([
@@ -224,7 +245,6 @@ def _verify_sigv4_query(req: Any) -> Principal | None:
         payload_hash
     ])
     
-    # String to Sign
     algorithm = "AWS4-HMAC-SHA256"
     credential_scope = f"{date_stamp}/{region}/{service}/aws4_request"
     hashed_request = hashlib.sha256(canonical_request.encode('utf-8')).hexdigest()
@@ -235,7 +255,6 @@ def _verify_sigv4_query(req: Any) -> Principal | None:
         hashed_request
     ])
     
-    # Signature
     signing_key = _get_signature_key(secret_key, date_stamp, region, service)
     calculated_signature = hmac.new(signing_key, string_to_sign.encode("utf-8"), hashlib.sha256).hexdigest()
     
@@ -493,7 +512,6 @@ def _generate_presigned_url(
     }
     canonical_query = _encode_query_params(query_params)
 
-    # Determine host and scheme from config or request
     api_base = current_app.config.get("API_BASE_URL")
     if api_base:
         parsed = urlparse(api_base)
@@ -545,6 +563,28 @@ def _strip_ns(tag: str | None) -> str:
     return tag.split("}")[-1]
 
 
+def _find_element(parent: Element, name: str) -> Optional[Element]:
+    """Find a child element by name, trying both namespaced and non-namespaced variants.
+
+    This handles XML documents that may or may not include namespace prefixes.
+    """
+    el = parent.find(f"{{*}}{name}")
+    if el is None:
+        el = parent.find(name)
+    return el
+
+
+def _find_element_text(parent: Element, name: str, default: str = "") -> str:
+    """Find a child element and return its text content.
+
+    Returns the default value if element not found or has no text.
+    """
+    el = _find_element(parent, name)
+    if el is None or el.text is None:
+        return default
+    return el.text.strip()
+
+
 def _parse_tagging_document(payload: bytes) -> list[dict[str, str]]:
     try:
         root = fromstring(payload)
@@ -561,17 +601,11 @@ def _parse_tagging_document(payload: bytes) -> list[dict[str, str]]:
     for tag_el in list(tagset):
         if _strip_ns(tag_el.tag) != "Tag":
             continue
-        key_el = tag_el.find("{*}Key")
+        key = _find_element_text(tag_el, "Key")
-        if key_el is None:
-            key_el = tag_el.find("Key")
-        value_el = tag_el.find("{*}Value")
-        if value_el is None:
-            value_el = tag_el.find("Value")
-        key = (key_el.text or "").strip() if key_el is not None else ""
         if not key:
             continue
-        value = value_el.text if value_el is not None else ""
+        value = _find_element_text(tag_el, "Value")
-        tags.append({"Key": key, "Value": value or ""})
+        tags.append({"Key": key, "Value": value})
     return tags
 
 
@@ -853,7 +887,6 @@ def _bucket_versioning_handler(bucket_name: str) -> Response:
         current_app.logger.info("Bucket versioning updated", extra={"bucket": bucket_name, "status": status})
         return Response(status=200)
     
-    # GET
     try:
         enabled = storage.is_versioning_enabled(bucket_name)
     except StorageError as exc:
@@ -889,7 +922,7 @@ def _bucket_tagging_handler(bucket_name: str) -> Response:
             return _error_response("NoSuchBucket", str(exc), 404)
         current_app.logger.info("Bucket tags deleted", extra={"bucket": bucket_name})
         return Response(status=204)
-    # PUT
+
     payload = request.get_data(cache=False) or b""
     try:
         tags = _parse_tagging_document(payload)
@@ -914,7 +947,6 @@ def _object_tagging_handler(bucket_name: str, object_key: str) -> Response:
     if error:
         return error
     
-    # For tagging, we use read permission for GET, write for PUT/DELETE
     action = "read" if request.method == "GET" else "write"
     try:
         _authorize_action(principal, bucket_name, action, object_key=object_key)
@@ -1093,10 +1125,8 @@ def _bucket_location_handler(bucket_name: str) -> Response:
     if not storage.bucket_exists(bucket_name):
         return _error_response("NoSuchBucket", "Bucket does not exist", 404)
     
-    # Return the configured AWS_REGION
     region = current_app.config.get("AWS_REGION", "us-east-1")
     root = Element("LocationConstraint")
-    # AWS returns empty for us-east-1, but we'll be explicit
     root.text = region if region != "us-east-1" else None
     return _xml_response(root)
 
@@ -1116,13 +1146,11 @@ def _bucket_acl_handler(bucket_name: str) -> Response:
         return _error_response("NoSuchBucket", "Bucket does not exist", 404)
     
     if request.method == "PUT":
-        # We don't fully implement ACLs, but we accept the request for compatibility
+        # Accept canned ACL headers for S3 compatibility (not fully implemented)
-        # Check for canned ACL header
         canned_acl = request.headers.get("x-amz-acl", "private")
         current_app.logger.info("Bucket ACL set (canned)", extra={"bucket": bucket_name, "acl": canned_acl})
         return Response(status=200)
     
-    # GET - Return a basic ACL document showing full control for owner
     root = Element("AccessControlPolicy")
     owner = SubElement(root, "Owner")
     SubElement(owner, "ID").text = principal.access_key if principal else "anonymous"
@@ -1155,7 +1183,7 @@ def _bucket_list_versions_handler(bucket_name: str) -> Response:
     storage = _storage()
     
     try:
-        objects = storage.list_objects(bucket_name)
+        objects = storage.list_objects_all(bucket_name)
     except StorageError as exc:
         return _error_response("NoSuchBucket", str(exc), 404)
     
@@ -1170,7 +1198,6 @@ def _bucket_list_versions_handler(bucket_name: str) -> Response:
     if key_marker:
         objects = [obj for obj in objects if obj.key > key_marker]
     
-    # Build XML response
     root = Element("ListVersionsResult", xmlns="http://s3.amazonaws.com/doc/2006-03-01/")
     SubElement(root, "Name").text = bucket_name
     SubElement(root, "Prefix").text = prefix
@@ -1188,10 +1215,9 @@ def _bucket_list_versions_handler(bucket_name: str) -> Response:
             is_truncated = True
             break
         
-        # Current version
         version = SubElement(root, "Version")
         SubElement(version, "Key").text = obj.key
-        SubElement(version, "VersionId").text = "null"  # Current version ID
+        SubElement(version, "VersionId").text = "null"
         SubElement(version, "IsLatest").text = "true"
         SubElement(version, "LastModified").text = obj.last_modified.strftime("%Y-%m-%dT%H:%M:%S.000Z")
         SubElement(version, "ETag").text = f'"{obj.etag}"'
@@ -1205,7 +1231,6 @@ def _bucket_list_versions_handler(bucket_name: str) -> Response:
         version_count += 1
         next_key_marker = obj.key
         
-        # Get historical versions
         try:
             versions = storage.list_object_versions(bucket_name, obj.key)
             for v in versions:
@@ -1289,14 +1314,12 @@ def _render_lifecycle_config(config: list) -> Element:
         rule_el = SubElement(root, "Rule")
         SubElement(rule_el, "ID").text = rule.get("ID", "")
         
-        # Filter
         filter_el = SubElement(rule_el, "Filter")
         if rule.get("Prefix"):
             SubElement(filter_el, "Prefix").text = rule.get("Prefix", "")
         
         SubElement(rule_el, "Status").text = rule.get("Status", "Enabled")
         
-        # Expiration
         if "Expiration" in rule:
             exp = rule["Expiration"]
             exp_el = SubElement(rule_el, "Expiration")
@@ -1307,14 +1330,12 @@ def _render_lifecycle_config(config: list) -> Element:
             if exp.get("ExpiredObjectDeleteMarker"):
                 SubElement(exp_el, "ExpiredObjectDeleteMarker").text = "true"
         
-        # NoncurrentVersionExpiration
         if "NoncurrentVersionExpiration" in rule:
             nve = rule["NoncurrentVersionExpiration"]
             nve_el = SubElement(rule_el, "NoncurrentVersionExpiration")
             if "NoncurrentDays" in nve:
                 SubElement(nve_el, "NoncurrentDays").text = str(nve["NoncurrentDays"])
         
-        # AbortIncompleteMultipartUpload
         if "AbortIncompleteMultipartUpload" in rule:
             aimu = rule["AbortIncompleteMultipartUpload"]
             aimu_el = SubElement(rule_el, "AbortIncompleteMultipartUpload")
@@ -1338,29 +1359,24 @@ def _parse_lifecycle_config(payload: bytes) -> list:
     for rule_el in root.findall("{*}Rule") or root.findall("Rule"):
         rule: dict = {}
         
-        # ID
         id_el = rule_el.find("{*}ID") or rule_el.find("ID")
         if id_el is not None and id_el.text:
             rule["ID"] = id_el.text.strip()
         
-        # Filter/Prefix
         filter_el = rule_el.find("{*}Filter") or rule_el.find("Filter")
         if filter_el is not None:
             prefix_el = filter_el.find("{*}Prefix") or filter_el.find("Prefix")
             if prefix_el is not None and prefix_el.text:
                 rule["Prefix"] = prefix_el.text
         
-        # Legacy Prefix (outside Filter)
         if "Prefix" not in rule:
             prefix_el = rule_el.find("{*}Prefix") or rule_el.find("Prefix")
             if prefix_el is not None:
                 rule["Prefix"] = prefix_el.text or ""
         
-        # Status
         status_el = rule_el.find("{*}Status") or rule_el.find("Status")
         rule["Status"] = (status_el.text or "Enabled").strip() if status_el is not None else "Enabled"
         
-        # Expiration
         exp_el = rule_el.find("{*}Expiration") or rule_el.find("Expiration")
         if exp_el is not None:
             expiration: dict = {}
@@ -1376,7 +1392,6 @@ def _parse_lifecycle_config(payload: bytes) -> list:
             if expiration:
                 rule["Expiration"] = expiration
         
-        # NoncurrentVersionExpiration
         nve_el = rule_el.find("{*}NoncurrentVersionExpiration") or rule_el.find("NoncurrentVersionExpiration")
         if nve_el is not None:
             nve: dict = {}
@@ -1386,7 +1401,6 @@ def _parse_lifecycle_config(payload: bytes) -> list:
             if nve:
                 rule["NoncurrentVersionExpiration"] = nve
         
-        # AbortIncompleteMultipartUpload
         aimu_el = rule_el.find("{*}AbortIncompleteMultipartUpload") or rule_el.find("AbortIncompleteMultipartUpload")
         if aimu_el is not None:
             aimu: dict = {}
@@ -1424,7 +1438,6 @@ def _bucket_quota_handler(bucket_name: str) -> Response:
         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,
@@ -1436,7 +1449,7 @@ def _bucket_quota_handler(bucket_name: str) -> Response:
     
     if request.method == "DELETE":
         try:
-            storage.set_bucket_quota(bucket_name, max_size_bytes=None, max_objects=None)
+            storage.set_bucket_quota(bucket_name, max_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})
@@ -1453,7 +1466,6 @@ def _bucket_quota_handler(bucket_name: str) -> Response:
     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)
@@ -1471,7 +1483,7 @@ def _bucket_quota_handler(bucket_name: str) -> Response:
             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)
+        storage.set_bucket_quota(bucket_name, max_bytes=max_size_bytes, max_objects=max_objects)
     except StorageError as exc:
         return _error_response("NoSuchBucket", str(exc), 404)
     
@@ -1564,7 +1576,6 @@ def _bulk_delete_handler(bucket_name: str) -> Response:
     return _xml_response(result, status=200)
 
 
-# ---------------------- routes ----------------------
 @s3_api_bp.get("/")
 @limiter.limit("60 per minute")
 def list_buckets() -> Response:
@@ -1642,7 +1653,6 @@ def bucket_handler(bucket_name: str) -> Response:
         current_app.logger.info("Bucket deleted", extra={"bucket": bucket_name})
         return Response(status=204)
 
-    # GET - list objects (supports both ListObjects and ListObjectsV2)
     principal, error = _require_principal()
     try:
         _authorize_action(principal, bucket_name, "list")
@@ -1650,18 +1660,12 @@ def bucket_handler(bucket_name: str) -> Response:
         if error:
             return error
         return _error_response("AccessDenied", str(exc), 403)
-    try:
-        objects = storage.list_objects(bucket_name)
-    except StorageError as exc:
-        return _error_response("NoSuchBucket", str(exc), 404)
 
-    # Check if this is ListObjectsV2 (list-type=2)
     list_type = request.args.get("list-type")
     prefix = request.args.get("prefix", "")
     delimiter = request.args.get("delimiter", "")
     max_keys = min(int(request.args.get("max-keys", current_app.config["UI_PAGE_SIZE"])), 1000)
     
-    # Pagination markers
     marker = request.args.get("marker", "")  # ListObjects v1
     continuation_token = request.args.get("continuation-token", "")  # ListObjectsV2
     start_after = request.args.get("start-after", "")  # ListObjectsV2
@@ -1671,7 +1675,6 @@ def bucket_handler(bucket_name: str) -> Response:
     effective_start = ""
     if list_type == "2":
         if continuation_token:
-            import base64
             try:
                 effective_start = base64.urlsafe_b64decode(continuation_token.encode()).decode("utf-8")
             except Exception:
@@ -1681,11 +1684,17 @@ def bucket_handler(bucket_name: str) -> Response:
     else:
         effective_start = marker
     
-    if prefix:
+    fetch_keys = max_keys * 10 if delimiter else max_keys
-        objects = [obj for obj in objects if obj.key.startswith(prefix)]
+    try:
-    
+        list_result = storage.list_objects(
-    if effective_start:
+            bucket_name,
-        objects = [obj for obj in objects if obj.key > effective_start]
+            max_keys=fetch_keys,
+            continuation_token=effective_start or None,
+            prefix=prefix or None,
+        )
+        objects = list_result.objects
+    except StorageError as exc:
+        return _error_response("NoSuchBucket", str(exc), 404)
     
     common_prefixes: list[str] = []
     filtered_objects: list = []
@@ -1694,7 +1703,6 @@ def bucket_handler(bucket_name: str) -> Response:
         for obj in objects:
             key_after_prefix = obj.key[len(prefix):] if prefix else obj.key
             if delimiter in key_after_prefix:
-                # This is a "folder" - extract the common prefix
                 common_prefix = prefix + key_after_prefix.split(delimiter)[0] + delimiter
                 if common_prefix not in seen_prefixes:
                     seen_prefixes.add(common_prefix)
@@ -1705,7 +1713,7 @@ def bucket_handler(bucket_name: str) -> Response:
         common_prefixes = sorted(common_prefixes)
     
     total_items = len(objects) + len(common_prefixes)
-    is_truncated = total_items > max_keys
+    is_truncated = total_items > max_keys or list_result.is_truncated
     
     if len(objects) >= max_keys:
         objects = objects[:max_keys]
@@ -1723,7 +1731,6 @@ def bucket_handler(bucket_name: str) -> Response:
             next_marker = common_prefixes[-1].rstrip(delimiter) if delimiter else common_prefixes[-1]
         
         if list_type == "2" and next_marker:
-            import base64
             next_continuation_token = base64.urlsafe_b64encode(next_marker.encode()).decode("utf-8")
 
     if list_type == "2":
@@ -1792,7 +1799,6 @@ def object_handler(bucket_name: str, object_key: str):
     if "tagging" in request.args:
         return _object_tagging_handler(bucket_name, object_key)
 
-    # Multipart Uploads
     if request.method == "POST":
         if "uploads" in request.args:
             return _initiate_multipart_upload(bucket_name, object_key)
@@ -1845,9 +1851,7 @@ def object_handler(bucket_name: str, object_key: str):
         response = Response(status=200)
         response.headers["ETag"] = f'"{meta.etag}"'
         
-        # Trigger replication if not a replication request
+        if "S3ReplicationAgent" not in request.headers.get("User-Agent", ""):
-        user_agent = request.headers.get("User-Agent", "")
-        if "S3ReplicationAgent" not in user_agent:
             _replication_manager().trigger_replication(bucket_name, object_key, action="write")
             
         return response
@@ -1866,31 +1870,25 @@ def object_handler(bucket_name: str, object_key: str):
         metadata = storage.get_object_metadata(bucket_name, object_key)
         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":
             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:
-            # 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)
@@ -1919,7 +1917,6 @@ def object_handler(bucket_name: str, object_key: str):
     storage.delete_object(bucket_name, object_key)
     current_app.logger.info("Object deleted", extra={"bucket": bucket_name, "key": object_key})
     
-    # Trigger replication if not a replication request
     user_agent = request.headers.get("User-Agent", "")
     if "S3ReplicationAgent" not in user_agent:
         _replication_manager().trigger_replication(bucket_name, object_key, action="delete")
@@ -2200,7 +2197,6 @@ class AwsChunkedDecoder:
                 self.chunk_remaining -= len(chunk)
                 
                 if self.chunk_remaining == 0:
-                    # Read CRLF after chunk data
                     crlf = self.stream.read(2)
                     if crlf != b"\r\n":
                         raise IOError("Malformed chunk: missing CRLF")
@@ -2219,7 +2215,6 @@ class AwsChunkedDecoder:
                 
                 try:
                     line_str = line.decode("ascii").strip()
-                    # Handle chunk-signature extension if present (e.g. "1000;chunk-signature=...")
                     if ";" in line_str:
                         line_str = line_str.split(";")[0]
                     chunk_size = int(line_str, 16)
@@ -2375,7 +2370,6 @@ def _abort_multipart_upload(bucket_name: str, object_key: str) -> Response:
     try:
         _storage().abort_multipart_upload(bucket_name, upload_id)
     except StorageError as exc:
-        # Abort is idempotent, but if bucket missing...
         if "Bucket does not exist" in str(exc):
             return _error_response("NoSuchBucket", str(exc), 404)
             
@@ -2385,7 +2379,6 @@ def _abort_multipart_upload(bucket_name: str, object_key: str) -> Response:
 @s3_api_bp.before_request
 def resolve_principal():
     g.principal = None
-    # Try SigV4
     try:
         if ("Authorization" in request.headers and request.headers["Authorization"].startswith("AWS4-HMAC-SHA256")) or \
            (request.args.get("X-Amz-Algorithm") == "AWS4-HMAC-SHA256"):
@@ -2394,7 +2387,6 @@ def resolve_principal():
     except Exception:
         pass
     
-    # Try simple auth headers (internal/testing)
     access_key = request.headers.get("X-Access-Key")
     secret_key = request.headers.get("X-Secret-Key")
     if access_key and secret_key:

app/storage.py

@@ -7,9 +7,11 @@ import os
 import re
 import shutil
 import stat
+import threading
 import time
 import unicodedata
 import uuid
+from collections import OrderedDict
 from contextlib import contextmanager
 from dataclasses import dataclass
 from datetime import datetime, timezone
@@ -99,6 +101,15 @@ class BucketMeta:
     created_at: datetime
 
 
+@dataclass
+class ListObjectsResult:
+    """Paginated result for object listing."""
+    objects: List[ObjectMeta]
+    is_truncated: bool
+    next_continuation_token: Optional[str]
+    total_count: Optional[int] = None  # Total objects in bucket (from stats cache)
+
+
 def _utcnow() -> datetime:
     return datetime.now(timezone.utc)
 
@@ -119,11 +130,18 @@ class ObjectStorage:
     BUCKET_VERSIONS_DIR = "versions"
     MULTIPART_MANIFEST = "manifest.json"
     BUCKET_CONFIG_FILE = ".bucket.json"
+    KEY_INDEX_CACHE_TTL = 30
+    OBJECT_CACHE_MAX_SIZE = 100  # Maximum number of buckets to cache
 
     def __init__(self, root: Path) -> None:
         self.root = Path(root)
         self.root.mkdir(parents=True, exist_ok=True)
         self._ensure_system_roots()
+        # LRU cache for object metadata with thread-safe access
+        self._object_cache: OrderedDict[str, tuple[Dict[str, ObjectMeta], float]] = OrderedDict()
+        self._cache_lock = threading.Lock()
+        # Cache version counter for detecting stale reads
+        self._cache_version: Dict[str, int] = {}
 
     def list_buckets(self) -> List[BucketMeta]:
         buckets: List[BucketMeta] = []
@@ -133,7 +151,7 @@ class ObjectStorage:
                 buckets.append(
                     BucketMeta(
                         name=bucket.name,
-                        created_at=datetime.fromtimestamp(stat.st_ctime),
+                        created_at=datetime.fromtimestamp(stat.st_ctime, timezone.utc),
                     )
                 )
         return buckets
@@ -181,7 +199,6 @@ class ObjectStorage:
         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)
@@ -193,7 +210,6 @@ class ObjectStorage:
                     object_count += 1
                     total_bytes += stat.st_size
 
-        # 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"):
@@ -207,8 +223,8 @@ class ObjectStorage:
             "bytes": total_bytes,
             "version_count": version_count,
             "version_bytes": version_bytes,
-            "total_objects": object_count + version_count,  # All objects including versions
+            "total_objects": object_count + version_count,
-            "total_bytes": total_bytes + version_bytes,  # All storage including versions
+            "total_bytes": total_bytes + version_bytes, 
         }
         
         try:
@@ -241,31 +257,76 @@ class ObjectStorage:
         self._remove_tree(self._system_bucket_root(bucket_path.name))
         self._remove_tree(self._multipart_bucket_root(bucket_path.name))
 
-    def list_objects(self, bucket_name: str) -> List[ObjectMeta]:
+    def list_objects(
+        self,
+        bucket_name: str,
+        *,
+        max_keys: int = 1000,
+        continuation_token: Optional[str] = None,
+        prefix: Optional[str] = None,
+    ) -> ListObjectsResult:
+        """List objects in a bucket with pagination support.
+        
+        Args:
+            bucket_name: Name of the bucket
+            max_keys: Maximum number of objects to return (default 1000)
+            continuation_token: Token from previous request for pagination
+            prefix: Filter objects by key prefix
+            
+        Returns:
+            ListObjectsResult with objects, truncation status, and continuation token
+        """
         bucket_path = self._bucket_path(bucket_name)
         if not bucket_path.exists():
             raise StorageError("Bucket does not exist")
         bucket_id = bucket_path.name
 
+        object_cache = self._get_object_cache(bucket_id, bucket_path)
+        
+        all_keys = sorted(object_cache.keys())
+        
+        if prefix:
+            all_keys = [k for k in all_keys if k.startswith(prefix)]
+        
+        total_count = len(all_keys)
+        start_index = 0
+        if continuation_token:
+            try:
+                import bisect
+                start_index = bisect.bisect_right(all_keys, continuation_token)
+                if start_index >= total_count:
+                    return ListObjectsResult(
+                        objects=[],
+                        is_truncated=False,
+                        next_continuation_token=None,
+                        total_count=total_count,
+                    )
+            except Exception:
+                pass  
+        
+        end_index = start_index + max_keys
+        keys_slice = all_keys[start_index:end_index]
+        is_truncated = end_index < total_count
+
         objects: List[ObjectMeta] = []
-        for path in bucket_path.rglob("*"):
+        for key in keys_slice:
-            if path.is_file():
+            obj = object_cache.get(key)
-                stat = path.stat()
+            if obj:
-                rel = path.relative_to(bucket_path)
+                objects.append(obj)
-                if rel.parts and rel.parts[0] in self.INTERNAL_FOLDERS:
+        
-                    continue
+        next_token = keys_slice[-1] if is_truncated and keys_slice else None
-                metadata = self._read_metadata(bucket_id, rel)
+        
-                objects.append(
+        return ListObjectsResult(
-                    ObjectMeta(
+            objects=objects,
-                        key=str(rel.as_posix()),
+            is_truncated=is_truncated,
-                        size=stat.st_size,
+            next_continuation_token=next_token,
-                        last_modified=datetime.fromtimestamp(stat.st_mtime),
+            total_count=total_count,
-                        etag=self._compute_etag(path),
-                        metadata=metadata or None,
         )
-                )
+
-        objects.sort(key=lambda meta: meta.key)
+    def list_objects_all(self, bucket_name: str) -> List[ObjectMeta]:
-        return objects
+        """List all objects in a bucket (no pagination). Use with caution for large buckets."""
+        result = self.list_objects(bucket_name, max_keys=100000)
+        return result.objects
 
     def put_object(
         self,
@@ -285,14 +346,12 @@ class ObjectStorage:
         destination = bucket_path / safe_key
         destination.parent.mkdir(parents=True, exist_ok=True)
 
-        # 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")
 
-        # 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"
@@ -304,9 +363,7 @@ class ObjectStorage:
             
             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
                 
@@ -322,29 +379,29 @@ class ObjectStorage:
                         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:
+        etag = checksum.hexdigest()
-            self._write_metadata(bucket_id, safe_key, metadata)
+        
-        else:
+        internal_meta = {"__etag__": etag, "__size__": str(stat.st_size)}
-            self._delete_metadata(bucket_id, safe_key)
+        combined_meta = {**internal_meta, **(metadata or {})}
+        self._write_metadata(bucket_id, safe_key, combined_meta)
         
         self._invalidate_bucket_stats_cache(bucket_id)
+        self._invalidate_object_cache(bucket_id)
         
         return ObjectMeta(
             key=safe_key.as_posix(),
             size=stat.st_size,
-            last_modified=datetime.fromtimestamp(stat.st_mtime),
+            last_modified=datetime.fromtimestamp(stat.st_mtime, timezone.utc),
-            etag=checksum.hexdigest(),
+            etag=etag,
             metadata=metadata,
         )
 
@@ -370,16 +427,14 @@ class ObjectStorage:
         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
+                        break 
                 except OSError:
                     if attempt < 2:
-                        time.sleep(0.1)  # Brief delay before retry
+                        time.sleep(0.1)
-                    # Final attempt failed - continue to next parent
                     break
 
     def delete_object(self, bucket_name: str, object_key: str) -> None:
@@ -396,6 +451,7 @@ class ObjectStorage:
         self._delete_metadata(bucket_id, rel)
         
         self._invalidate_bucket_stats_cache(bucket_id)
+        self._invalidate_object_cache(bucket_id)
         self._cleanup_empty_parents(path, bucket_path)
 
     def purge_object(self, bucket_name: str, object_key: str) -> None:
@@ -416,8 +472,8 @@ class ObjectStorage:
         if legacy_version_dir.exists():
             shutil.rmtree(legacy_version_dir, ignore_errors=True)
         
-        # Invalidate bucket stats cache
         self._invalidate_bucket_stats_cache(bucket_id)
+        self._invalidate_object_cache(bucket_id)
         self._cleanup_empty_parents(target, bucket_path)
 
     def is_versioning_enabled(self, bucket_name: str) -> bool:
@@ -529,7 +585,6 @@ class ObjectStorage:
         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
         
@@ -571,9 +626,7 @@ class ObjectStorage:
                 "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))
         
@@ -683,8 +736,6 @@ class ObjectStorage:
         bucket_id = bucket_path.name
         safe_key = self._sanitize_object_key(object_key)
         version_dir = self._version_dir(bucket_id, safe_key)
-        if not version_dir.exists():
-            version_dir = self._legacy_version_dir(bucket_id, safe_key)
         if not version_dir.exists():
             version_dir = self._legacy_version_dir(bucket_id, safe_key)
             if not version_dir.exists():
@@ -734,7 +785,7 @@ class ObjectStorage:
         return ObjectMeta(
             key=safe_key.as_posix(),
             size=stat.st_size,
-            last_modified=datetime.fromtimestamp(stat.st_mtime),
+            last_modified=datetime.fromtimestamp(stat.st_mtime, timezone.utc),
             etag=self._compute_etag(destination),
             metadata=metadata or None,
         )
@@ -833,44 +884,72 @@ class ObjectStorage:
         part_number: int,
         stream: BinaryIO,
     ) -> str:
+        """Upload a part for a multipart upload.
+
+        Uses file locking to safely update the manifest and handle concurrent uploads.
+        """
         if part_number < 1:
             raise StorageError("part_number must be >= 1")
         bucket_path = self._bucket_path(bucket_name)
 
-        # Get the upload root directory
         upload_root = self._multipart_dir(bucket_path.name, upload_id)
         if not upload_root.exists():
             upload_root = self._legacy_multipart_dir(bucket_path.name, upload_id)
         if not upload_root.exists():
             raise StorageError("Multipart upload not found")
 
-        # Write the part data first (can happen concurrently)
+        # Write part to temporary file first, then rename atomically
         checksum = hashlib.md5()
         part_filename = f"part-{part_number:05d}.part"
         part_path = upload_root / part_filename
-        with part_path.open("wb") as target:
+        temp_path = upload_root / f".{part_filename}.tmp"
+
+        try:
+            with temp_path.open("wb") as target:
                 shutil.copyfileobj(_HashingReader(stream, checksum), target)
+
+            # Atomic rename (or replace on Windows)
+            temp_path.replace(part_path)
+        except OSError:
+            # Clean up temp file on failure
+            try:
+                temp_path.unlink(missing_ok=True)
+            except OSError:
+                pass
+            raise
+
         record = {
             "etag": checksum.hexdigest(),
             "size": part_path.stat().st_size,
             "filename": part_filename,
         }
 
-        # Update manifest with file locking to prevent race conditions
         manifest_path = upload_root / self.MULTIPART_MANIFEST
         lock_path = upload_root / ".manifest.lock"
 
+        # Retry loop for handling transient lock/read failures
+        max_retries = 3
+        for attempt in range(max_retries):
+            try:
                 with lock_path.open("w") as lock_file:
                     with _file_lock(lock_file):
-                # Re-read manifest under lock to get latest state
                         try:
                             manifest = json.loads(manifest_path.read_text(encoding="utf-8"))
                         except (OSError, json.JSONDecodeError) as exc:
+                            if attempt < max_retries - 1:
+                                time.sleep(0.1 * (attempt + 1))
+                                continue
                             raise StorageError("Multipart manifest unreadable") from exc
 
                         parts = manifest.setdefault("parts", {})
                         parts[str(part_number)] = record
                         manifest_path.write_text(json.dumps(manifest), encoding="utf-8")
+                break
+            except OSError as exc:
+                if attempt < max_retries - 1:
+                    time.sleep(0.1 * (attempt + 1))
+                    continue
+                raise StorageError(f"Failed to update multipart manifest: {exc}") from exc
 
         return record["etag"]
 
@@ -916,11 +995,9 @@ class ObjectStorage:
         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
@@ -977,12 +1054,13 @@ class ObjectStorage:
         shutil.rmtree(upload_root, ignore_errors=True)
 
         self._invalidate_bucket_stats_cache(bucket_id)
+        self._invalidate_object_cache(bucket_id)
 
         stat = destination.stat()
         return ObjectMeta(
             key=safe_key.as_posix(),
             size=stat.st_size,
-            last_modified=datetime.fromtimestamp(stat.st_mtime),
+            last_modified=datetime.fromtimestamp(stat.st_mtime, timezone.utc),
             etag=checksum.hexdigest(),
             metadata=metadata,
         )
@@ -1080,6 +1158,202 @@ class ObjectStorage:
     def _legacy_multipart_dir(self, bucket_name: str, upload_id: str) -> Path:
         return self._legacy_multipart_bucket_root(bucket_name) / upload_id
 
+    def _fast_list_keys(self, bucket_path: Path) -> List[str]:
+        """Fast directory walk using os.scandir instead of pathlib.rglob.
+        
+        This is significantly faster for large directories (10K+ files).
+        Returns just the keys (for backward compatibility).
+        """
+        return list(self._build_object_cache(bucket_path).keys())
+
+    def _build_object_cache(self, bucket_path: Path) -> Dict[str, ObjectMeta]:
+        """Build a complete object metadata cache for a bucket.
+        
+        Uses os.scandir for fast directory walking and a persistent etag index.
+        """
+        from concurrent.futures import ThreadPoolExecutor
+        
+        bucket_id = bucket_path.name
+        objects: Dict[str, ObjectMeta] = {}
+        bucket_str = str(bucket_path)
+        bucket_len = len(bucket_str) + 1
+        
+        etag_index_path = self._system_bucket_root(bucket_id) / "etag_index.json"
+        meta_cache: Dict[str, str] = {}
+        index_mtime: float = 0
+        
+        if etag_index_path.exists():
+            try:
+                index_mtime = etag_index_path.stat().st_mtime
+                with open(etag_index_path, 'r', encoding='utf-8') as f:
+                    meta_cache = json.load(f)
+            except (OSError, json.JSONDecodeError):
+                meta_cache = {}
+        
+        meta_root = self._bucket_meta_root(bucket_id)
+        needs_rebuild = False
+        
+        if meta_root.exists() and index_mtime > 0:
+            def check_newer(dir_path: str) -> bool:
+                try:
+                    with os.scandir(dir_path) as it:
+                        for entry in it:
+                            if entry.is_dir(follow_symlinks=False):
+                                if check_newer(entry.path):
+                                    return True
+                            elif entry.is_file(follow_symlinks=False) and entry.name.endswith('.meta.json'):
+                                if entry.stat().st_mtime > index_mtime:
+                                    return True
+                except OSError:
+                    pass
+                return False
+            needs_rebuild = check_newer(str(meta_root))
+        elif not meta_cache:
+            needs_rebuild = True
+            
+        if needs_rebuild and meta_root.exists():
+            meta_str = str(meta_root)
+            meta_len = len(meta_str) + 1
+            meta_files: list[tuple[str, str]] = []
+            
+            def collect_meta_files(dir_path: str) -> None:
+                try:
+                    with os.scandir(dir_path) as it:
+                        for entry in it:
+                            if entry.is_dir(follow_symlinks=False):
+                                collect_meta_files(entry.path)
+                            elif entry.is_file(follow_symlinks=False) and entry.name.endswith('.meta.json'):
+                                rel = entry.path[meta_len:]
+                                key = rel[:-10].replace(os.sep, '/')
+                                meta_files.append((key, entry.path))
+                except OSError:
+                    pass
+            
+            collect_meta_files(meta_str)
+            
+            def read_meta_file(item: tuple[str, str]) -> tuple[str, str | None]:
+                key, path = item
+                try:
+                    with open(path, 'rb') as f:
+                        content = f.read()
+                    etag_marker = b'"__etag__"'
+                    idx = content.find(etag_marker)
+                    if idx != -1:
+                        start = content.find(b'"', idx + len(etag_marker) + 1)
+                        if start != -1:
+                            end = content.find(b'"', start + 1)
+                            if end != -1:
+                                return key, content[start+1:end].decode('utf-8')
+                    return key, None
+                except (OSError, UnicodeDecodeError):
+                    return key, None
+            
+            if meta_files:
+                meta_cache = {}
+                with ThreadPoolExecutor(max_workers=min(64, len(meta_files))) as executor:
+                    for key, etag in executor.map(read_meta_file, meta_files):
+                        if etag:
+                            meta_cache[key] = etag
+                
+                try:
+                    etag_index_path.parent.mkdir(parents=True, exist_ok=True)
+                    with open(etag_index_path, 'w', encoding='utf-8') as f:
+                        json.dump(meta_cache, f)
+                except OSError:
+                    pass
+        
+        def scan_dir(dir_path: str) -> None:
+            try:
+                with os.scandir(dir_path) as it:
+                    for entry in it:
+                        if entry.is_dir(follow_symlinks=False):
+                            rel_start = entry.path[bucket_len:].split(os.sep)[0] if len(entry.path) > bucket_len else entry.name
+                            if rel_start in self.INTERNAL_FOLDERS:
+                                continue
+                            scan_dir(entry.path)
+                        elif entry.is_file(follow_symlinks=False):
+                            rel = entry.path[bucket_len:]
+                            first_part = rel.split(os.sep)[0] if os.sep in rel else rel
+                            if first_part in self.INTERNAL_FOLDERS:
+                                continue
+                            
+                            key = rel.replace(os.sep, '/')
+                            try:
+                                stat = entry.stat()
+                                
+                                etag = meta_cache.get(key)
+                                
+                                if not etag:
+                                    etag = f'"{stat.st_size}-{int(stat.st_mtime)}"'
+                                
+                                objects[key] = ObjectMeta(
+                                    key=key,
+                                    size=stat.st_size,
+                                    last_modified=datetime.fromtimestamp(stat.st_mtime, timezone.utc),
+                                    etag=etag,
+                                    metadata=None, 
+                                )
+                            except OSError:
+                                pass
+            except OSError:
+                pass
+        
+        scan_dir(bucket_str)
+        return objects
+
+    def _get_object_cache(self, bucket_id: str, bucket_path: Path) -> Dict[str, ObjectMeta]:
+        """Get cached object metadata for a bucket, refreshing if stale.
+
+        Uses LRU eviction to prevent unbounded cache growth.
+        Thread-safe with version tracking to detect concurrent invalidations.
+        """
+        now = time.time()
+
+        with self._cache_lock:
+            cached = self._object_cache.get(bucket_id)
+            cache_version = self._cache_version.get(bucket_id, 0)
+
+            if cached:
+                objects, timestamp = cached
+                if now - timestamp < self.KEY_INDEX_CACHE_TTL:
+                    # Move to end (most recently used)
+                    self._object_cache.move_to_end(bucket_id)
+                    return objects
+
+        # Build cache outside lock to avoid holding lock during I/O
+        objects = self._build_object_cache(bucket_path)
+
+        with self._cache_lock:
+            # Check if cache was invalidated while we were building
+            current_version = self._cache_version.get(bucket_id, 0)
+            if current_version != cache_version:
+                # Cache was invalidated, rebuild
+                objects = self._build_object_cache(bucket_path)
+
+            # Evict oldest entries if cache is full
+            while len(self._object_cache) >= self.OBJECT_CACHE_MAX_SIZE:
+                self._object_cache.popitem(last=False)
+
+            self._object_cache[bucket_id] = (objects, time.time())
+            self._object_cache.move_to_end(bucket_id)
+
+        return objects
+
+    def _invalidate_object_cache(self, bucket_id: str) -> None:
+        """Invalidate the object cache and etag index for a bucket.
+
+        Increments version counter to signal stale reads.
+        """
+        with self._cache_lock:
+            self._object_cache.pop(bucket_id, None)
+            self._cache_version[bucket_id] = self._cache_version.get(bucket_id, 0) + 1
+
+        etag_index_path = self._system_bucket_root(bucket_id) / "etag_index.json"
+        try:
+            etag_index_path.unlink(missing_ok=True)
+        except OSError:
+            pass
+
     def _ensure_system_roots(self) -> None:
         for path in (
             self._system_root_path(),

app/ui.py

@@ -189,7 +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
+        "can_view_metrics": can_manage, 
         "csrf_token": generate_csrf,
     }
 
@@ -294,7 +294,8 @@ def bucket_detail(bucket_name: str):
     storage = _storage()
     try:
         _authorize_ui(principal, bucket_name, "list")
-        objects = storage.list_objects(bucket_name)
+        if not storage.bucket_exists(bucket_name):
+            raise StorageError("Bucket does not exist")
     except (StorageError, IamError) as exc:
         flash(_friendly_error_message(exc), "danger")
         return redirect(url_for("ui.buckets_overview"))
@@ -341,7 +342,6 @@ def bucket_detail(bucket_name: str):
         except IamError:
             can_manage_versioning = False
 
-    # Check replication permission
     can_manage_replication = False
     if principal:
         try:
@@ -350,7 +350,6 @@ def bucket_detail(bucket_name: str):
         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:
@@ -359,12 +358,9 @@ def bucket_detail(bucket_name: str):
         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)
-    # 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 []
@@ -372,7 +368,6 @@ def bucket_detail(bucket_name: str):
     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
@@ -382,10 +377,12 @@ def bucket_detail(bucket_name: str):
     except IamError:
         pass
 
+    objects_api_url = url_for("ui.list_bucket_objects", bucket_name=bucket_name)
+
     return render_template(
         "bucket_detail.html",
         bucket_name=bucket_name,
-        objects=objects,
+        objects_api_url=objects_api_url,
         principal=principal,
         bucket_policy_text=policy_text,
         bucket_policy=bucket_policy,
@@ -408,6 +405,70 @@ def bucket_detail(bucket_name: str):
     )
 
 
+@ui_bp.get("/buckets/<bucket_name>/objects")
+def list_bucket_objects(bucket_name: str):
+    """API endpoint for paginated object listing."""
+    principal = _current_principal()
+    storage = _storage()
+    try:
+        _authorize_ui(principal, bucket_name, "list")
+    except IamError as exc:
+        return jsonify({"error": str(exc)}), 403
+
+    max_keys = min(int(request.args.get("max_keys", 1000)), 100000)
+    continuation_token = request.args.get("continuation_token") or None
+    prefix = request.args.get("prefix") or None
+
+    try:
+        result = storage.list_objects(
+            bucket_name,
+            max_keys=max_keys,
+            continuation_token=continuation_token,
+            prefix=prefix,
+        )
+    except StorageError as exc:
+        return jsonify({"error": str(exc)}), 400
+
+    try:
+        versioning_enabled = storage.is_versioning_enabled(bucket_name)
+    except StorageError:
+        versioning_enabled = False
+
+    # Pre-compute URL templates once (not per-object) for performance
+    # Frontend will construct actual URLs by replacing KEY_PLACEHOLDER
+    preview_template = url_for("ui.object_preview", bucket_name=bucket_name, object_key="KEY_PLACEHOLDER")
+    delete_template = url_for("ui.delete_object", bucket_name=bucket_name, object_key="KEY_PLACEHOLDER")
+    presign_template = url_for("ui.object_presign", bucket_name=bucket_name, object_key="KEY_PLACEHOLDER")
+    versions_template = url_for("ui.object_versions", bucket_name=bucket_name, object_key="KEY_PLACEHOLDER")
+    restore_template = url_for("ui.restore_object_version", bucket_name=bucket_name, object_key="KEY_PLACEHOLDER", version_id="VERSION_ID_PLACEHOLDER")
+
+    objects_data = []
+    for obj in result.objects:
+        objects_data.append({
+            "key": obj.key,
+            "size": obj.size,
+            "last_modified": obj.last_modified.isoformat(),
+            "last_modified_display": obj.last_modified.strftime("%b %d, %Y %H:%M"),
+            "etag": obj.etag,
+        })
+
+    return jsonify({
+        "objects": objects_data,
+        "is_truncated": result.is_truncated,
+        "next_continuation_token": result.next_continuation_token,
+        "total_count": result.total_count,
+        "versioning_enabled": versioning_enabled,
+        "url_templates": {
+            "preview": preview_template,
+            "download": preview_template + "?download=1",
+            "presign": presign_template,
+            "delete": delete_template,
+            "versions": versions_template,
+            "restore": restore_template,
+        },
+    })
+
+
 @ui_bp.post("/buckets/<bucket_name>/upload")
 @limiter.limit("30 per minute")
 def upload_object(bucket_name: str):
@@ -678,41 +739,30 @@ def bulk_download_objects(bucket_name: str):
     unique_keys = list(dict.fromkeys(cleaned))
     storage = _storage()
     
-    # Check permissions for all keys first (or at least bucket read)
+    # Verify permission to read bucket contents
-    # We'll check bucket read once, then object read for each if needed?
-    # _authorize_ui checks bucket level if object_key is None, but we need to check each object if fine-grained policies exist.
-    # For simplicity/performance, we check bucket list/read.
     try:
         _authorize_ui(principal, bucket_name, "read")
     except IamError as exc:
         return jsonify({"error": str(exc)}), 403
 
-    # Create ZIP
+    # Create ZIP archive of selected objects
     buffer = io.BytesIO()
     with zipfile.ZipFile(buffer, "w", zipfile.ZIP_DEFLATED) as zf:
         for key in unique_keys:
             try:
-                # Verify individual object permission if needed? 
-                # _authorize_ui(principal, bucket_name, "read", object_key=key) 
-                # This might be slow for many objects. Assuming bucket read is enough for now or we accept the overhead.
-                # Let's skip individual check for bulk speed, assuming bucket read implies object read unless denied.
-                # But strictly we should check. Let's check.
                 _authorize_ui(principal, bucket_name, "read", object_key=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
+                # Skip objects that can't be accessed
                 continue
     
     buffer.seek(0)
@@ -1017,7 +1067,6 @@ def update_bucket_encryption(bucket_name: str):
     action = request.form.get("action", "enable")
     
     if action == "disable":
-        # Disable encryption
         try:
             _storage().set_bucket_encryption(bucket_name, None)
             flash("Default encryption disabled", "info")
@@ -1025,16 +1074,14 @@ def update_bucket_encryption(bucket_name: str):
             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
+    # Build encryption configuration in AWS S3 format
     encryption_config: dict[str, Any] = {
         "Rules": [
             {
@@ -1210,7 +1257,6 @@ def delete_iam_user(access_key: str):
         return redirect(url_for("ui.iam_dashboard"))
 
     if access_key == principal.access_key:
-        # Self-deletion
         try:
             _iam().delete_user(access_key)
             session.pop("credentials", None)
@@ -1292,6 +1338,9 @@ def create_connection():
 
 @ui_bp.post("/connections/test")
 def test_connection():
+    from botocore.config import Config as BotoConfig
+    from botocore.exceptions import ConnectTimeoutError, EndpointConnectionError, ReadTimeoutError
+    
     principal = _current_principal()
     try:
         _iam().authorize(principal, None, "iam:list_users")
@@ -1308,18 +1357,32 @@ def test_connection():
         return jsonify({"status": "error", "message": "Missing credentials"}), 400
 
     try:
+        config = BotoConfig(
+            connect_timeout=5,
+            read_timeout=10,
+            retries={'max_attempts': 1}
+        )
         s3 = boto3.client(
             "s3",
             endpoint_url=endpoint,
             aws_access_key_id=access_key,
             aws_secret_access_key=secret_key,
             region_name=region,
+            config=config,
         )
-        # Try to list buckets to verify credentials and endpoint
+
         s3.list_buckets()
         return jsonify({"status": "ok", "message": "Connection successful"})
+    except (ConnectTimeoutError, ReadTimeoutError):
+        return jsonify({"status": "error", "message": f"Connection timed out - endpoint may be down or unreachable: {endpoint}"}), 400
+    except EndpointConnectionError:
+        return jsonify({"status": "error", "message": f"Could not connect to endpoint: {endpoint}"}), 400
+    except ClientError as e:
+        error_code = e.response.get('Error', {}).get('Code', 'Unknown')
+        error_msg = e.response.get('Error', {}).get('Message', str(e))
+        return jsonify({"status": "error", "message": f"Connection failed ({error_code}): {error_msg}"}), 400
     except Exception as e:
-        return jsonify({"status": "error", "message": str(e)}), 400
+        return jsonify({"status": "error", "message": f"Connection failed: {str(e)}"}), 400
 
 
 @ui_bp.post("/connections/<connection_id>/update")
@@ -1380,7 +1443,6 @@ def update_bucket_replication(bucket_name: str):
         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")
@@ -1391,14 +1453,12 @@ def update_bucket_replication(bucket_name: str):
     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 configuration removed", "info")
     elif action == "pause":
-        # Users can pause - just set enabled=False
         rule = _replication().get_rule(bucket_name)
         if rule:
             rule.enabled = False
@@ -1407,16 +1467,20 @@ def update_bucket_replication(bucket_name: str):
         else:
             flash("No replication configuration to pause", "warning")
     elif action == "resume":
-        # Users can resume - just set enabled=True
+        from .replication import REPLICATION_MODE_ALL
         rule = _replication().get_rule(bucket_name)
         if rule:
             rule.enabled = True
             _replication().set_rule(rule)
+            # When resuming, sync any pending objects that accumulated while paused
+            if rule.mode == REPLICATION_MODE_ALL:
+                _replication().replicate_existing_objects(bucket_name)
+                flash("Replication resumed. Syncing pending objects in background.", "success")
+            else:
                 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"))
@@ -1441,7 +1505,6 @@ def update_bucket_replication(bucket_name: str):
             )
             _replication().set_rule(rule)
             
-            # If mode is "all", trigger replication of existing objects
             if replication_mode == REPLICATION_MODE_ALL:
                 _replication().replicate_existing_objects(bucket_name)
                 flash("Replication configured. Existing objects are being replicated in the background.", "success")
@@ -1466,10 +1529,31 @@ def get_replication_status(bucket_name: str):
     if not rule:
         return jsonify({"error": "No replication rule"}), 404
     
-    # This is the slow operation - compute sync status by comparing buckets
+    connection = _connections().get(rule.target_connection_id)
+    endpoint_healthy = False
+    endpoint_error = None
+    if connection:
+        endpoint_healthy = _replication().check_endpoint_health(connection)
+        if not endpoint_healthy:
+            endpoint_error = f"Cannot reach endpoint: {connection.endpoint_url}"
+    else:
+        endpoint_error = "Target connection not found"
+    
+    stats = None
+    if endpoint_healthy:
         stats = _replication().get_sync_status(bucket_name)
+    
     if not stats:
-        return jsonify({"error": "Failed to compute status"}), 500
+        return jsonify({
+            "objects_synced": 0,
+            "objects_pending": 0,
+            "objects_orphaned": 0,
+            "bytes_synced": 0,
+            "last_sync_at": rule.stats.last_sync_at if rule.stats else None,
+            "last_sync_key": rule.stats.last_sync_key if rule.stats else None,
+            "endpoint_healthy": endpoint_healthy,
+            "endpoint_error": endpoint_error,
+        })
     
     return jsonify({
         "objects_synced": stats.objects_synced,
@@ -1478,6 +1562,28 @@ def get_replication_status(bucket_name: str):
         "bytes_synced": stats.bytes_synced,
         "last_sync_at": stats.last_sync_at,
         "last_sync_key": stats.last_sync_key,
+        "endpoint_healthy": endpoint_healthy,
+        "endpoint_error": endpoint_error,
+    })
+
+
+@ui_bp.get("/connections/<connection_id>/health")
+def check_connection_health(connection_id: str):
+    """Check if a connection endpoint is reachable."""
+    principal = _current_principal()
+    try:
+        _iam().authorize(principal, None, "iam:list_users")
+    except IamError:
+        return jsonify({"error": "Access denied"}), 403
+    
+    conn = _connections().get(connection_id)
+    if not conn:
+        return jsonify({"healthy": False, "error": "Connection not found"}), 404
+    
+    healthy = _replication().check_endpoint_health(conn)
+    return jsonify({
+        "healthy": healthy,
+        "error": None if healthy else f"Cannot reach endpoint: {conn.endpoint_url}"
     })
 
 
@@ -1498,7 +1604,6 @@ 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:
@@ -1522,16 +1627,13 @@ def metrics_dashboard():
     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)
-        # 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)
     
-    # Calculate system uptime
     boot_time = psutil.boot_time()
     uptime_seconds = time.time() - boot_time
     uptime_days = int(uptime_seconds / 86400)

app/version.py

@@ -1,7 +1,7 @@
 """Central location for the application version string."""
 from __future__ import annotations
 
-APP_VERSION = "0.1.6"
+APP_VERSION = "0.1.9"
 
 
 def get_version() -> str:

docs.md

@@ -341,6 +341,7 @@ Before upgrading across major versions, verify compatibility:
 | From Version | To Version | Breaking Changes | Migration Required |
 |--------------|------------|------------------|-------------------|
 | 0.1.x | 0.2.x | None expected | No |
+| 0.1.6 | 0.1.7 | None | No |
 | < 0.1.0 | >= 0.1.0 | New IAM config format | Yes - run migration script |
 
 **Automatic compatibility detection:**
@@ -634,6 +635,48 @@ curl -X PUT http://127.0.0.1:5000/bucket-policy/test \
 
 The UI will reflect this change as soon as the request completes thanks to the hot reload.
 
+### UI Object Browser
+
+The bucket detail page includes a powerful object browser with the following features:
+
+#### Folder Navigation
+
+Objects with forward slashes (`/`) in their keys are displayed as a folder hierarchy. Click a folder row to navigate into it. A breadcrumb navigation bar shows your current path and allows quick navigation back to parent folders or the root.
+
+#### Pagination & Infinite Scroll
+
+- Objects load in configurable batches (50, 100, 150, 200, or 250 per page)
+- Scroll to the bottom to automatically load more objects (infinite scroll)
+- A **Load more** button is available as a fallback for touch devices or when infinite scroll doesn't trigger
+- The footer shows the current load status (e.g., "Showing 100 of 500 objects")
+
+#### Bulk Operations
+
+- Select multiple objects using checkboxes
+- **Bulk Delete**: Delete multiple objects at once
+- **Bulk Download**: Download selected objects as individual files
+
+#### Search & Filter
+
+Use the search box to filter objects by name in real-time. The filter applies to the currently loaded objects.
+
+#### Error Handling
+
+If object loading fails (e.g., network error), a friendly error message is displayed with a **Retry** button to attempt loading again.
+
+#### Object Preview
+
+Click any object row to view its details in the preview sidebar:
+- File size and last modified date
+- ETag (content hash)
+- Custom metadata (if present)
+- Download and presign (share link) buttons
+- Version history (when versioning is enabled)
+
+#### Drag & Drop Upload
+
+Drag files directly onto the objects table to upload them to the current bucket and folder path.
+
 ## 6. Presigned URLs
 
 - Trigger from the UI using the **Presign** button after selecting an object.

pytest.ini

@@ -0,0 +1,3 @@
+[pytest]
+testpaths = tests
+norecursedirs = data .git __pycache__ .venv

requirements.txt

@@ -1,10 +1,10 @@
 Flask>=3.1.2
-Flask-Limiter>=4.1.0
+Flask-Limiter>=4.1.1
-Flask-Cors>=6.0.1
+Flask-Cors>=6.0.2
 Flask-WTF>=1.2.2
-pytest>=9.0.1
+pytest>=9.0.2
 requests>=2.32.5
-boto3>=1.42.1
+boto3>=1.42.14
 waitress>=3.0.2
 psutil>=7.1.3
 cryptography>=46.0.3

static/css/main.css

@@ -66,8 +66,28 @@ html {
 	color: var(--myfsio-muted) !important;
 }
 
-.table-responsive { border-radius: 0.5rem; overflow: hidden; }
+.table-responsive { 
+	border-radius: 0.5rem; 
+	overflow-x: auto; 
+	-webkit-overflow-scrolling: touch;
+}
 .message-stack { position: sticky; top: 1rem; z-index: 100; }
+
+/* Mobile-friendly table improvements */
+.table-responsive table {
+	min-width: 600px;
+}
+
+.table-responsive table th,
+.table-responsive table td {
+	white-space: nowrap;
+}
+
+/* Allow text wrapping for description columns */
+.table-responsive table td.text-wrap {
+	white-space: normal;
+	min-width: 200px;
+}
 code { font-size: 0.85rem; }
 
 code {
@@ -389,8 +409,22 @@ code {
 .bucket-table th:last-child { white-space: nowrap; }
 
 .object-key {
-	word-break: break-word;
+	max-width: 0;
-	max-width: 32rem;
+	width: 100%;
+	overflow: hidden;
+	text-overflow: ellipsis;
+}
+
+.object-key .fw-medium {
+	white-space: nowrap;
+	overflow: hidden;
+	text-overflow: ellipsis;
+}
+
+.object-key .text-muted {
+	white-space: nowrap;
+	overflow: hidden;
+	text-overflow: ellipsis;
 }
 
 .preview-card { top: 1rem; }
@@ -517,6 +551,22 @@ code {
 	overflow-y: auto;
 }
 
+.objects-table-container thead {
+	position: sticky;
+	top: 0;
+	z-index: 10;
+}
+
+.objects-table-container thead th {
+	background-color: #f8f9fa;
+	border-bottom: 1px solid var(--myfsio-card-border);
+	box-shadow: 0 1px 2px rgba(0, 0, 0, 0.05);
+}
+
+[data-theme='dark'] .objects-table-container thead th {
+	background-color: #1e293b;
+}
+
 .btn-group form { display: inline; }
 
 .font-monospace { font-family: 'SFMono-Regular', Consolas, 'Liberation Mono', Menlo, monospace; }
@@ -1537,6 +1587,41 @@ pre code {
 		position: relative !important;
 		top: 0 !important;
 	}
+	
+	/* Ensure tables are scrollable on mobile */
+	.card-body .table-responsive {
+		margin: -1rem;
+		padding: 0;
+		width: calc(100% + 2rem);
+	}
+	
+	.card-body .table-responsive table {
+		margin-bottom: 0;
+	}
+	
+	/* IAM users table mobile adjustments */
+	.table th,
+	.table td {
+		padding: 0.5rem 0.75rem;
+	}
+	
+	/* Better touch scrolling indicator */
+	.table-responsive::after {
+		content: '';
+		position: absolute;
+		top: 0;
+		right: 0;
+		bottom: 0;
+		width: 20px;
+		background: linear-gradient(to left, var(--myfsio-card-bg), transparent);
+		pointer-events: none;
+		opacity: 0;
+		transition: opacity 0.3s;
+	}
+	
+	.table-responsive:not(:hover)::after {
+		opacity: 0.8;
+	}
 }
 
 *:focus-visible {

templates/base.html

@@ -199,7 +199,7 @@
       })();
     </script>
     <script>
-      // Toast utility
+
       window.showToast = function(message, title = 'Notification', type = 'info') {
         const toastEl = document.getElementById('liveToast');
         const toastTitle = document.getElementById('toastTitle');
@@ -208,7 +208,6 @@
         toastTitle.textContent = title;
         toastMessage.textContent = message;
 
-        // Reset classes
         toastEl.classList.remove('text-bg-primary', 'text-bg-success', 'text-bg-danger', 'text-bg-warning');
         
         if (type === 'success') toastEl.classList.add('text-bg-success');
@@ -221,13 +220,11 @@
     </script>
     <script>
       (function () {
-        // Show flashed messages as toasts
+
         {% with messages = get_flashed_messages(with_categories=true) %}
         {% if messages %}
           {% for category, message in messages %}
-            // Map Flask categories to Toast types
+
-            // Flask: success, danger, warning, info
-            // Toast: success, error, warning, info
             var type = "{{ category }}";
             if (type === "danger") type = "error";
             window.showToast({{ message | tojson | safe }}, "Notification", type);

templates/buckets.html

@@ -131,7 +131,7 @@
 {{ super() }}
 <script>
   (function () {
-    // Search functionality
+
     const searchInput = document.getElementById('bucket-search');
     const bucketItems = document.querySelectorAll('.bucket-item');
     const noBucketsMsg = document.querySelector('.text-center.py-5'); // The "No buckets found" empty state
@@ -153,7 +153,6 @@
       });
     }
 
-    // View toggle functionality
     const viewGrid = document.getElementById('view-grid');
     const viewList = document.getElementById('view-list');
     const container = document.getElementById('buckets-container');
@@ -168,8 +167,7 @@
         });
         cards.forEach(card => {
           card.classList.remove('h-100');
-          // Optional: Add flex-row to card-body content if we want a horizontal layout
+
-          // For now, full-width stacked cards is a good list view
         });
         localStorage.setItem('bucket-view-pref', 'list');
       } else {
@@ -188,7 +186,6 @@
       viewGrid.addEventListener('change', () => setView('grid'));
       viewList.addEventListener('change', () => setView('list'));
 
-      // Restore preference
       const pref = localStorage.getItem('bucket-view-pref');
       if (pref === 'list') {
         viewList.checked = true;

templates/connections.html

@@ -104,6 +104,7 @@
                     <table class="table table-hover align-middle mb-0">
                         <thead class="table-light">
                             <tr>
+                                <th scope="col" style="width: 50px;">Status</th>
                                 <th scope="col">Name</th>
                                 <th scope="col">Endpoint</th>
                                 <th scope="col">Region</th>
@@ -113,7 +114,12 @@
                         </thead>
                         <tbody>
                             {% for conn in connections %}
-                            <tr>
+                            <tr data-connection-id="{{ conn.id }}">
+                                <td class="text-center">
+                                    <span class="connection-status" data-status="checking" title="Checking...">
+                                        <span class="spinner-border spinner-border-sm text-muted" role="status" style="width: 12px; height: 12px;"></span>
+                                    </span>
+                                </td>
                                 <td>
                                     <div class="d-flex align-items-center gap-2">
                                         <div class="connection-icon">
@@ -181,7 +187,6 @@
     </div>
 </div>
 
-<!-- Edit Connection Modal -->
 <div class="modal fade" id="editConnectionModal" tabindex="-1" aria-hidden="true">
     <div class="modal-dialog modal-dialog-centered">
         <div class="modal-content">
@@ -247,7 +252,6 @@
     </div>
 </div>
 
-<!-- Delete Connection Modal -->
 <div class="modal fade" id="deleteConnectionModal" tabindex="-1" aria-hidden="true">
     <div class="modal-dialog modal-dialog-centered">
         <div class="modal-content">
@@ -297,14 +301,17 @@
         }
     }
 
-    // Test Connection Logic
     async function testConnection(formId, resultId) {
         const form = document.getElementById(formId);
         const resultDiv = document.getElementById(resultId);
         const formData = new FormData(form);
         const data = Object.fromEntries(formData.entries());
         
-        resultDiv.innerHTML = '<div class="text-info"><span class="spinner-border spinner-border-sm" role="status" aria-hidden="true"></span> Testing...</div>';
+        resultDiv.innerHTML = '<div class="text-info"><span class="spinner-border spinner-border-sm" role="status" aria-hidden="true"></span> Testing connection...</div>';
+        
+        // Use AbortController to timeout client-side after 20 seconds
+        const controller = new AbortController();
+        const timeoutId = setTimeout(() => controller.abort(), 20000);
         
         try {
             const response = await fetch("{{ url_for('ui.test_connection') }}", {
@@ -313,17 +320,44 @@
                     "Content-Type": "application/json",
                     "X-CSRFToken": "{{ csrf_token() }}"
                 },
-                body: JSON.stringify(data)
+                body: JSON.stringify(data),
+                signal: controller.signal
             });
+            clearTimeout(timeoutId);
             
             const result = await response.json();
             if (response.ok) {
-                resultDiv.innerHTML = `<div class="text-success"><i class="bi bi-check-circle"></i> ${result.message}</div>`;
+                resultDiv.innerHTML = `<div class="text-success">
+                    <svg xmlns="http://www.w3.org/2000/svg" width="16" height="16" fill="currentColor" class="me-1" viewBox="0 0 16 16">
+                        <path d="M16 8A8 8 0 1 1 0 8a8 8 0 0 1 16 0zm-3.97-3.03a.75.75 0 0 0-1.08.022L7.477 9.417 5.384 7.323a.75.75 0 0 0-1.06 1.06L6.97 11.03a.75.75 0 0 0 1.079-.02l3.992-4.99a.75.75 0 0 0-.01-1.05z"/>
+                    </svg>
+                    ${result.message}
+                </div>`;
             } else {
-                resultDiv.innerHTML = `<div class="text-danger"><i class="bi bi-exclamation-circle"></i> ${result.message}</div>`;
+                resultDiv.innerHTML = `<div class="text-danger">
+                    <svg xmlns="http://www.w3.org/2000/svg" width="16" height="16" fill="currentColor" class="me-1" viewBox="0 0 16 16">
+                        <path d="M16 8A8 8 0 1 1 0 8a8 8 0 0 1 16 0zM5.354 4.646a.5.5 0 1 0-.708.708L7.293 8l-2.647 2.646a.5.5 0 0 0 .708.708L8 8.707l2.646 2.647a.5.5 0 0 0 .708-.708L8.707 8l2.647-2.646a.5.5 0 0 0-.708-.708L8 7.293 5.354 4.646z"/>
+                    </svg>
+                    ${result.message}
+                </div>`;
             }
         } catch (error) {
-            resultDiv.innerHTML = `<div class="text-danger"><i class="bi bi-exclamation-circle"></i> Connection failed</div>`;
+            clearTimeout(timeoutId);
+            if (error.name === 'AbortError') {
+                resultDiv.innerHTML = `<div class="text-danger">
+                    <svg xmlns="http://www.w3.org/2000/svg" width="16" height="16" fill="currentColor" class="me-1" viewBox="0 0 16 16">
+                        <path d="M16 8A8 8 0 1 1 0 8a8 8 0 0 1 16 0zM5.354 4.646a.5.5 0 1 0-.708.708L7.293 8l-2.647 2.646a.5.5 0 0 0 .708.708L8 8.707l2.646 2.647a.5.5 0 0 0 .708-.708L8.707 8l2.647-2.646a.5.5 0 0 0-.708-.708L8 7.293 5.354 4.646z"/>
+                    </svg>
+                    Connection test timed out - endpoint may be unreachable
+                </div>`;
+            } else {
+                resultDiv.innerHTML = `<div class="text-danger">
+                    <svg xmlns="http://www.w3.org/2000/svg" width="16" height="16" fill="currentColor" class="me-1" viewBox="0 0 16 16">
+                        <path d="M16 8A8 8 0 1 1 0 8a8 8 0 0 1 16 0zM5.354 4.646a.5.5 0 1 0-.708.708L7.293 8l-2.647 2.646a.5.5 0 0 0 .708.708L8 8.707l2.646 2.647a.5.5 0 0 0 .708-.708L8.707 8l2.647-2.646a.5.5 0 0 0-.708-.708L8 7.293 5.354 4.646z"/>
+                    </svg>
+                    Connection failed: Network error
+                </div>`;
+            }
         }
     }
 
@@ -335,7 +369,6 @@
         testConnection('editConnectionForm', 'editTestResult');
     });
 
-    // Modal Event Listeners
     const editModal = document.getElementById('editConnectionModal');
     editModal.addEventListener('show.bs.modal', event => {
         const button = event.relatedTarget;
@@ -362,5 +395,54 @@
         const form = document.getElementById('deleteConnectionForm');
         form.action = "{{ url_for('ui.delete_connection', connection_id='CONN_ID') }}".replace('CONN_ID', id);
     });
+
+    // Check connection health for each connection in the table
+    // Uses staggered requests to avoid overwhelming the server
+    async function checkConnectionHealth(connectionId, statusEl) {
+        try {
+            const controller = new AbortController();
+            const timeoutId = setTimeout(() => controller.abort(), 15000);
+            
+            const response = await fetch(`/ui/connections/${connectionId}/health`, {
+                signal: controller.signal
+            });
+            clearTimeout(timeoutId);
+            
+            const data = await response.json();
+            if (data.healthy) {
+                statusEl.innerHTML = `
+                    <svg xmlns="http://www.w3.org/2000/svg" width="16" height="16" fill="currentColor" class="text-success" viewBox="0 0 16 16">
+                        <path d="M16 8A8 8 0 1 1 0 8a8 8 0 0 1 16 0zm-3.97-3.03a.75.75 0 0 0-1.08.022L7.477 9.417 5.384 7.323a.75.75 0 0 0-1.06 1.06L6.97 11.03a.75.75 0 0 0 1.079-.02l3.992-4.99a.75.75 0 0 0-.01-1.05z"/>
+                    </svg>`;
+                statusEl.setAttribute('data-status', 'healthy');
+                statusEl.setAttribute('title', 'Connected');
+            } else {
+                statusEl.innerHTML = `
+                    <svg xmlns="http://www.w3.org/2000/svg" width="16" height="16" fill="currentColor" class="text-danger" viewBox="0 0 16 16">
+                        <path d="M16 8A8 8 0 1 1 0 8a8 8 0 0 1 16 0zM5.354 4.646a.5.5 0 1 0-.708.708L7.293 8l-2.647 2.646a.5.5 0 0 0 .708.708L8 8.707l2.646 2.647a.5.5 0 0 0 .708-.708L8.707 8l2.647-2.646a.5.5 0 0 0-.708-.708L8 7.293 5.354 4.646z"/>
+                    </svg>`;
+                statusEl.setAttribute('data-status', 'unhealthy');
+                statusEl.setAttribute('title', data.error || 'Unreachable');
+            }
+        } catch (error) {
+            statusEl.innerHTML = `
+                <svg xmlns="http://www.w3.org/2000/svg" width="16" height="16" fill="currentColor" class="text-warning" viewBox="0 0 16 16">
+                    <path d="M8.982 1.566a1.13 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 1.566zM8 5c.535 0 .954.462.9.995l-.35 3.507a.552.552 0 0 1-1.1 0L7.1 5.995A.905.905 0 0 1 8 5zm.002 6a1 1 0 1 1 0 2 1 1 0 0 1 0-2z"/>
+                </svg>`;
+            statusEl.setAttribute('data-status', 'unknown');
+            statusEl.setAttribute('title', 'Could not check status');
+        }
+    }
+
+    // Stagger health checks to avoid all requests at once
+    const connectionRows = document.querySelectorAll('tr[data-connection-id]');
+    connectionRows.forEach((row, index) => {
+        const connectionId = row.getAttribute('data-connection-id');
+        const statusEl = row.querySelector('.connection-status');
+        if (statusEl) {
+            // Stagger requests by 200ms each
+            setTimeout(() => checkConnectionHealth(connectionId, statusEl), index * 200);
+        }
+    });
 </script>
 {% endblock %}

templates/docs.html

@@ -47,9 +47,9 @@ python run.py --mode ui
           <table class="table table-sm table-bordered small mb-0">
             <thead class="table-light">
               <tr>
-                <th>Variable</th>
+                <th style="min-width: 180px;">Variable</th>
-                <th>Default</th>
+                <th style="min-width: 120px;">Default</th>
-                <th>Description</th>
+                <th class="text-wrap" style="min-width: 250px;">Description</th>
               </tr>
             </thead>
             <tbody>
@@ -255,6 +255,15 @@ sudo journalctl -u myfsio -f   # View logs</code></pre>
               <li>Progress rows highlight retries, throughput, and completion even if you close the modal.</li>
             </ul>
           </div>
+          <div>
+            <h3 class="h6 text-uppercase text-muted">Object browser</h3>
+            <ul>
+              <li>Navigate folder hierarchies using breadcrumbs. Objects with <code>/</code> in keys display as folders.</li>
+              <li>Infinite scroll loads more objects automatically. Choose batch size (50–250) from the footer dropdown.</li>
+              <li>Bulk select objects for multi-delete or multi-download. Filter by name using the search box.</li>
+              <li>If loading fails, click <strong>Retry</strong> to attempt again—no page refresh needed.</li>
+            </ul>
+          </div>
           <div>
             <h3 class="h6 text-uppercase text-muted">Object details</h3>
             <ul>
@@ -398,10 +407,62 @@ curl -X POST {{ api_base }}/presign/demo/notes.txt \
           <span class="docs-section-kicker">07</span>
           <h2 class="h4 mb-0">API Examples</h2>
         </div>
-        <p class="text-muted">Common operations using boto3.</p>
+        <p class="text-muted">Common operations using popular SDKs and tools.</p>
 
-        <h5 class="mt-4">Multipart Upload</h5>
+        <h3 class="h6 text-uppercase text-muted mt-4">Python (boto3)</h3>
-<pre><code class="language-python">import boto3
+<pre class="mb-4"><code class="language-python">import boto3
+
+s3 = boto3.client(
+    's3',
+    endpoint_url='{{ api_base }}',
+    aws_access_key_id='&lt;access_key&gt;',
+    aws_secret_access_key='&lt;secret_key&gt;'
+)
+
+# List buckets
+buckets = s3.list_buckets()['Buckets']
+
+# Create bucket
+s3.create_bucket(Bucket='mybucket')
+
+# Upload file
+s3.upload_file('local.txt', 'mybucket', 'remote.txt')
+
+# Download file
+s3.download_file('mybucket', 'remote.txt', 'downloaded.txt')
+
+# Generate presigned URL (valid 1 hour)
+url = s3.generate_presigned_url(
+    'get_object',
+    Params={'Bucket': 'mybucket', 'Key': 'remote.txt'},
+    ExpiresIn=3600
+)</code></pre>
+
+        <h3 class="h6 text-uppercase text-muted mt-4">JavaScript (AWS SDK v3)</h3>
+<pre class="mb-4"><code class="language-javascript">import { S3Client, ListBucketsCommand, PutObjectCommand } from '@aws-sdk/client-s3';
+
+const s3 = new S3Client({
+  endpoint: '{{ api_base }}',
+  region: 'us-east-1',
+  credentials: {
+    accessKeyId: '&lt;access_key&gt;',
+    secretAccessKey: '&lt;secret_key&gt;'
+  },
+  forcePathStyle: true  // Required for S3-compatible services
+});
+
+// List buckets
+const { Buckets } = await s3.send(new ListBucketsCommand({}));
+
+// Upload object
+await s3.send(new PutObjectCommand({
+  Bucket: 'mybucket',
+  Key: 'hello.txt',
+  Body: 'Hello, World!'
+}));</code></pre>
+
+        <h3 class="h6 text-uppercase text-muted mt-4">Multipart Upload (Python)</h3>
+<pre class="mb-4"><code class="language-python">import boto3
 
 s3 = boto3.client('s3', endpoint_url='{{ api_base }}')
 
@@ -409,9 +470,9 @@ s3 = boto3.client('s3', endpoint_url='{{ api_base }}')
 response = s3.create_multipart_upload(Bucket='mybucket', Key='large.bin')
 upload_id = response['UploadId']
 
-# Upload parts
+# Upload parts (minimum 5MB each, except last part)
 parts = []
-chunks = [b'chunk1', b'chunk2'] # Example data chunks
+chunks = [b'chunk1...', b'chunk2...']
 for part_number, chunk in enumerate(chunks, start=1):
     response = s3.upload_part(
         Bucket='mybucket',
@@ -429,6 +490,19 @@ s3.complete_multipart_upload(
     UploadId=upload_id,
     MultipartUpload={'Parts': parts}
 )</code></pre>
+
+        <h3 class="h6 text-uppercase text-muted mt-4">Presigned URLs for Sharing</h3>
+<pre class="mb-0"><code class="language-bash"># Generate a download link valid for 15 minutes
+curl -X POST "{{ api_base }}/presign/mybucket/photo.jpg" \
+  -H "Content-Type: application/json" \
+  -H "X-Access-Key: &lt;key&gt;" -H "X-Secret-Key: &lt;secret&gt;" \
+  -d '{"method": "GET", "expires_in": 900}'
+
+# Generate an upload link (PUT) valid for 1 hour
+curl -X POST "{{ api_base }}/presign/mybucket/upload.bin" \
+  -H "Content-Type: application/json" \
+  -H "X-Access-Key: &lt;key&gt;" -H "X-Secret-Key: &lt;secret&gt;" \
+  -d '{"method": "PUT", "expires_in": 3600}'</code></pre>
       </div>
     </article>
     <article id="replication" class="card shadow-sm docs-section">
@@ -478,6 +552,86 @@ s3.complete_multipart_upload(
         </p>
       </div>
     </article>
+    <article id="versioning" 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>
+          <h2 class="h4 mb-0">Object Versioning</h2>
+        </div>
+        <p class="text-muted">Keep multiple versions of objects to protect against accidental deletions and overwrites. Restore previous versions at any time.</p>
+
+        <h3 class="h6 text-uppercase text-muted mt-4">Enabling Versioning</h3>
+        <ol class="docs-steps mb-3">
+          <li>Navigate to your bucket's <strong>Properties</strong> tab.</li>
+          <li>Find the <strong>Versioning</strong> card and click <strong>Enable</strong>.</li>
+          <li>All subsequent uploads will create new versions instead of overwriting.</li>
+        </ol>
+
+        <h3 class="h6 text-uppercase text-muted mt-4">Version Operations</h3>
+        <div class="table-responsive mb-3">
+          <table class="table table-sm table-bordered small">
+            <thead class="table-light">
+              <tr>
+                <th>Operation</th>
+                <th>Description</th>
+              </tr>
+            </thead>
+            <tbody>
+              <tr>
+                <td><strong>View Versions</strong></td>
+                <td>Click the version icon on any object to see all historical versions with timestamps and sizes.</td>
+              </tr>
+              <tr>
+                <td><strong>Restore Version</strong></td>
+                <td>Click <strong>Restore</strong> on any version to make it the current version (creates a copy).</td>
+              </tr>
+              <tr>
+                <td><strong>Delete Current</strong></td>
+                <td>Deleting an object archives it. Previous versions remain accessible.</td>
+              </tr>
+              <tr>
+                <td><strong>Purge All</strong></td>
+                <td>Permanently delete an object and all its versions. This cannot be undone.</td>
+              </tr>
+            </tbody>
+          </table>
+        </div>
+
+        <h3 class="h6 text-uppercase text-muted mt-4">Archived Objects</h3>
+        <p class="small text-muted mb-3">When you delete a versioned object, it becomes "archived" - the current version is removed but historical versions remain. The <strong>Archived</strong> tab shows these objects so you can restore them.</p>
+
+        <h3 class="h6 text-uppercase text-muted mt-4">API Usage</h3>
+<pre class="mb-3"><code class="language-bash"># Enable versioning
+curl -X PUT "{{ api_base }}/&lt;bucket&gt;?versioning" \
+  -H "Content-Type: application/json" \
+  -H "X-Access-Key: &lt;key&gt;" -H "X-Secret-Key: &lt;secret&gt;" \
+  -d '{"Status": "Enabled"}'
+
+# Get versioning status
+curl "{{ api_base }}/&lt;bucket&gt;?versioning" \
+  -H "X-Access-Key: &lt;key&gt;" -H "X-Secret-Key: &lt;secret&gt;"
+
+# List object versions
+curl "{{ api_base }}/&lt;bucket&gt;?versions" \
+  -H "X-Access-Key: &lt;key&gt;" -H "X-Secret-Key: &lt;secret&gt;"
+
+# Get specific version
+curl "{{ api_base }}/&lt;bucket&gt;/&lt;key&gt;?versionId=&lt;version-id&gt;" \
+  -H "X-Access-Key: &lt;key&gt;" -H "X-Secret-Key: &lt;secret&gt;"</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>Storage Impact:</strong> Each version consumes storage. Enable quotas to limit total bucket size including all versions.
+            </div>
+          </div>
+        </div>
+      </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">
@@ -700,6 +854,7 @@ curl -X DELETE "{{ api_base }}/kms/keys/{key-id}?waiting_period_days=30" \
           <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="#versioning">Object Versioning</a></li>
           <li><a href="#quotas">Bucket Quotas</a></li>
           <li><a href="#encryption">Encryption</a></li>
           <li><a href="#troubleshooting">Troubleshooting</a></li>

templates/iam.html

@@ -203,7 +203,6 @@
   {% endif %}
 </div>
 
-<!-- Create User Modal -->
 <div class="modal fade" id="createUserModal" tabindex="-1" aria-hidden="true">
   <div class="modal-dialog modal-dialog-centered">
     <div class="modal-content">
@@ -252,7 +251,6 @@
   </div>
 </div>
 
-<!-- Policy Editor Modal -->
 <div class="modal fade" id="policyEditorModal" tabindex="-1" aria-hidden="true">
   <div class="modal-dialog modal-lg modal-dialog-centered">
     <div class="modal-content">
@@ -303,7 +301,6 @@
   </div>
 </div>
 
-<!-- Edit User Modal -->
 <div class="modal fade" id="editUserModal" tabindex="-1" aria-hidden="true">
   <div class="modal-dialog modal-dialog-centered">
     <div class="modal-content">
@@ -338,7 +335,6 @@
   </div>
 </div>
 
-<!-- Delete User Modal -->
 <div class="modal fade" id="deleteUserModal" tabindex="-1" aria-hidden="true">
   <div class="modal-dialog modal-dialog-centered">
     <div class="modal-content">
@@ -382,7 +378,6 @@
   </div>
 </div>
 
-<!-- Rotate Secret Modal -->
 <div class="modal fade" id="rotateSecretModal" tabindex="-1" aria-hidden="true">
   <div class="modal-dialog modal-dialog-centered">
     <div class="modal-content">
@@ -487,7 +482,6 @@
     const iamUsersData = document.getElementById('iamUsersJson');
     const users = iamUsersData ? JSON.parse(iamUsersData.textContent || '[]') : [];
 
-    // Policy Editor Logic
     const policyModalEl = document.getElementById('policyEditorModal');
     const policyModal = new bootstrap.Modal(policyModalEl);
     const userLabelEl = document.getElementById('policyEditorUserLabel');
@@ -534,7 +528,6 @@
       button.addEventListener('click', () => applyTemplate(button.dataset.policyTemplate));
     });
 
-    // Create User modal template buttons
     const createUserPoliciesEl = document.getElementById('createUserPolicies');
     const createTemplateButtons = document.querySelectorAll('[data-create-policy-template]');
     
@@ -591,7 +584,6 @@
       });
     });
 
-    // Edit User Logic
     const editUserModal = new bootstrap.Modal(document.getElementById('editUserModal'));
     const editUserForm = document.getElementById('editUserForm');
     const editUserDisplayName = document.getElementById('editUserDisplayName');
@@ -606,7 +598,6 @@
         });
     });
 
-    // Delete User Logic
     const deleteUserModal = new bootstrap.Modal(document.getElementById('deleteUserModal'));
     const deleteUserForm = document.getElementById('deleteUserForm');
     const deleteUserLabel = document.getElementById('deleteUserLabel');
@@ -628,7 +619,6 @@
         });
     });
 
-    // Rotate Secret Logic
     const rotateSecretModal = new bootstrap.Modal(document.getElementById('rotateSecretModal'));
     const rotateUserLabel = document.getElementById('rotateUserLabel');
     const confirmRotateBtn = document.getElementById('confirmRotateBtn');
@@ -645,7 +635,6 @@
             currentRotateKey = btn.dataset.rotateUser;
             rotateUserLabel.textContent = currentRotateKey;
 
-            // Reset Modal State
             rotateSecretConfirm.classList.remove('d-none');
             rotateSecretResult.classList.add('d-none');
             confirmRotateBtn.classList.remove('d-none');
@@ -680,7 +669,6 @@
             const data = await response.json();
             newSecretKeyInput.value = data.secret_key;
 
-            // Show Result
             rotateSecretConfirm.classList.add('d-none');
             rotateSecretResult.classList.remove('d-none');
             confirmRotateBtn.classList.add('d-none');

tests/test_edge_cases.py

@@ -38,7 +38,7 @@ def test_unicode_bucket_and_object_names(tmp_path: Path):
     assert storage.get_object_path("unicode-test", key).exists()
     
     # Verify listing
-    objects = storage.list_objects("unicode-test")
+    objects = storage.list_objects_all("unicode-test")
     assert any(o.key == key for o in objects)
 
 def test_special_characters_in_metadata(tmp_path: Path):

tests/test_storage_features.py

@@ -220,7 +220,7 @@ def test_bucket_config_filename_allowed(tmp_path):
     storage.create_bucket("demo")
     storage.put_object("demo", ".bucket.json", io.BytesIO(b"{}"))
 
-    objects = storage.list_objects("demo")
+    objects = storage.list_objects_all("demo")
     assert any(meta.key == ".bucket.json" for meta in objects)
 
 

tests/test_ui_bulk_delete.py

@@ -62,7 +62,7 @@ def test_bulk_delete_json_route(tmp_path: Path):
     assert set(payload["deleted"]) == {"first.txt", "missing.txt"}
     assert payload["errors"] == []
 
-    listing = storage.list_objects("demo")
+    listing = storage.list_objects_all("demo")
     assert {meta.key for meta in listing} == {"second.txt"}
 
 
@@ -92,5 +92,5 @@ def test_bulk_delete_validation(tmp_path: Path):
     assert limit_response.status_code == 400
     assert limit_response.get_json()["status"] == "error"
 
-    still_there = storage.list_objects("demo")
+    still_there = storage.list_objects_all("demo")
     assert {meta.key for meta in still_there} == {"keep.txt"}

tests/test_ui_pagination.py

@@ -0,0 +1,188 @@
+"""Tests for UI pagination of bucket objects."""
+import json
+from io import BytesIO
+from pathlib import Path
+
+import pytest
+
+from app import create_app
+
+
+def _make_app(tmp_path: Path):
+    """Create an app for testing."""
+    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"]}],
+            },
+        ]
+    }
+    iam_config.write_text(json.dumps(iam_payload))
+
+    flask_app = create_app(
+        {
+            "TESTING": True,
+            "WTF_CSRF_ENABLED": False,
+            "STORAGE_ROOT": storage_root,
+            "IAM_CONFIG": iam_config,
+            "BUCKET_POLICY_PATH": bucket_policies,
+        }
+    )
+    return flask_app
+
+
+class TestPaginatedObjectListing:
+    """Test paginated object listing API."""
+
+    def test_objects_api_returns_paginated_results(self, tmp_path):
+        """Objects API should return paginated results."""
+        app = _make_app(tmp_path)
+        storage = app.extensions["object_storage"]
+        storage.create_bucket("test-bucket")
+        
+        # Create 10 test objects
+        for i in range(10):
+            storage.put_object("test-bucket", f"file{i:02d}.txt", BytesIO(b"content"))
+        
+        with app.test_client() as client:
+            # Login first
+            client.post("/ui/login", data={"access_key": "test", "secret_key": "secret"}, follow_redirects=True)
+            
+            # Request first page of 3 objects
+            resp = client.get("/ui/buckets/test-bucket/objects?max_keys=3")
+            assert resp.status_code == 200
+            
+            data = resp.get_json()
+            assert len(data["objects"]) == 3
+            assert data["is_truncated"] is True
+            assert data["next_continuation_token"] is not None
+            assert data["total_count"] == 10
+    
+    def test_objects_api_pagination_continuation(self, tmp_path):
+        """Objects API should support continuation tokens."""
+        app = _make_app(tmp_path)
+        storage = app.extensions["object_storage"]
+        storage.create_bucket("test-bucket")
+        
+        # Create 5 test objects
+        for i in range(5):
+            storage.put_object("test-bucket", f"file{i:02d}.txt", BytesIO(b"content"))
+        
+        with app.test_client() as client:
+            client.post("/ui/login", data={"access_key": "test", "secret_key": "secret"}, follow_redirects=True)
+            
+            # Get first page
+            resp = client.get("/ui/buckets/test-bucket/objects?max_keys=2")
+            assert resp.status_code == 200
+            data = resp.get_json()
+            
+            first_page_keys = [obj["key"] for obj in data["objects"]]
+            assert len(first_page_keys) == 2
+            assert data["is_truncated"] is True
+            
+            # Get second page
+            token = data["next_continuation_token"]
+            resp = client.get(f"/ui/buckets/test-bucket/objects?max_keys=2&continuation_token={token}")
+            assert resp.status_code == 200
+            data = resp.get_json()
+            
+            second_page_keys = [obj["key"] for obj in data["objects"]]
+            assert len(second_page_keys) == 2
+            
+            # No overlap between pages
+            assert set(first_page_keys).isdisjoint(set(second_page_keys))
+    
+    def test_objects_api_prefix_filter(self, tmp_path):
+        """Objects API should support prefix filtering."""
+        app = _make_app(tmp_path)
+        storage = app.extensions["object_storage"]
+        storage.create_bucket("test-bucket")
+        
+        # Create objects with different prefixes
+        storage.put_object("test-bucket", "logs/access.log", BytesIO(b"log"))
+        storage.put_object("test-bucket", "logs/error.log", BytesIO(b"log"))
+        storage.put_object("test-bucket", "data/file.txt", BytesIO(b"data"))
+        
+        with app.test_client() as client:
+            client.post("/ui/login", data={"access_key": "test", "secret_key": "secret"}, follow_redirects=True)
+            
+            # Filter by prefix
+            resp = client.get("/ui/buckets/test-bucket/objects?prefix=logs/")
+            assert resp.status_code == 200
+            data = resp.get_json()
+            
+            keys = [obj["key"] for obj in data["objects"]]
+            assert all(k.startswith("logs/") for k in keys)
+            assert len(keys) == 2
+    
+    def test_objects_api_requires_authentication(self, tmp_path):
+        """Objects API should require login."""
+        app = _make_app(tmp_path)
+        storage = app.extensions["object_storage"]
+        storage.create_bucket("test-bucket")
+        
+        with app.test_client() as client:
+            # Don't login
+            resp = client.get("/ui/buckets/test-bucket/objects")
+            # Should redirect to login
+            assert resp.status_code == 302
+            assert "/ui/login" in resp.headers.get("Location", "")
+    
+    def test_objects_api_returns_object_metadata(self, tmp_path):
+        """Objects API should return complete object metadata."""
+        app = _make_app(tmp_path)
+        storage = app.extensions["object_storage"]
+        storage.create_bucket("test-bucket")
+        storage.put_object("test-bucket", "test.txt", BytesIO(b"test content"))
+        
+        with app.test_client() as client:
+            client.post("/ui/login", data={"access_key": "test", "secret_key": "secret"}, follow_redirects=True)
+            
+            resp = client.get("/ui/buckets/test-bucket/objects")
+            assert resp.status_code == 200
+            data = resp.get_json()
+            
+            assert len(data["objects"]) == 1
+            obj = data["objects"][0]
+
+            # Check all expected fields
+            assert obj["key"] == "test.txt"
+            assert obj["size"] == 12  # len("test content")
+            assert "last_modified" in obj
+            assert "last_modified_display" in obj
+            assert "etag" in obj
+
+            # URLs are now returned as templates (not per-object) for performance
+            assert "url_templates" in data
+            templates = data["url_templates"]
+            assert "preview" in templates
+            assert "download" in templates
+            assert "delete" in templates
+            assert "KEY_PLACEHOLDER" in templates["preview"]
+    
+    def test_bucket_detail_page_loads_without_objects(self, tmp_path):
+        """Bucket detail page should load even with many objects."""
+        app = _make_app(tmp_path)
+        storage = app.extensions["object_storage"]
+        storage.create_bucket("test-bucket")
+        
+        # Create many objects
+        for i in range(100):
+            storage.put_object("test-bucket", f"file{i:03d}.txt", BytesIO(b"x"))
+        
+        with app.test_client() as client:
+            client.post("/ui/login", data={"access_key": "test", "secret_key": "secret"}, follow_redirects=True)
+            
+            # The page should load quickly (objects loaded via JS)
+            resp = client.get("/ui/buckets/test-bucket")
+            assert resp.status_code == 200
+            
+            html = resp.data.decode("utf-8")
+            # Should have the JavaScript loading infrastructure
+            assert "loadObjects" in html or "objectsApiUrl" in html

tests/test_ui_policy.py

@@ -70,8 +70,12 @@ def test_ui_bucket_policy_enforcement_toggle(tmp_path: Path, enforce: bool):
         assert b"Access denied by bucket policy" in response.data
     else:
         assert response.status_code == 200
-        assert b"vid.mp4" in response.data
         assert b"Access denied by bucket policy" not in response.data
+        # Objects are now loaded via async API - check the objects endpoint
+        objects_response = client.get("/ui/buckets/testbucket/objects")
+        assert objects_response.status_code == 200
+        data = objects_response.get_json()
+        assert any(obj["key"] == "vid.mp4" for obj in data["objects"])
 
 
 def test_ui_bucket_policy_disabled_by_default(tmp_path: Path):
@@ -109,5 +113,9 @@ def test_ui_bucket_policy_disabled_by_default(tmp_path: Path):
     client.post("/ui/login", data={"access_key": "test", "secret_key": "secret"}, follow_redirects=True)
     response = client.get("/ui/buckets/testbucket", follow_redirects=True)
     assert response.status_code == 200
-    assert b"vid.mp4" in response.data
     assert b"Access denied by bucket policy" not in response.data
+    # Objects are now loaded via async API - check the objects endpoint
+    objects_response = client.get("/ui/buckets/testbucket/objects")
+    assert objects_response.status_code == 200
+    data = objects_response.get_json()
+    assert any(obj["key"] == "vid.mp4" for obj in data["objects"])