b10-transfer 0.1.0__py3-none-any.whl → 0.1.2__py3-none-any.whl

b10_transfer/__init__.py

@@ -1,27 +1,13 @@
-"""B10 Transfer - Lock-free PyTorch compilation cache for Baseten."""
+"""B10 Transfer - Lock-free PyTorch file transfer for Baseten."""
 
-from .core import transfer
-from .torch_cache import load_compile_cache, save_compile_cache, clear_local_cache
-from .async_transfers import (
-    start_transfer_async,
-    get_transfer_status,
-    is_transfer_complete,
-    wait_for_completion,
-    cancel_transfer,
-    list_active_transfers,
-    TransferProgress,
-)
-from .async_torch_cache import (
-    load_compile_cache_async,
-    save_compile_cache_async,
-)
+from .core import load_compile_cache, save_compile_cache, clear_local_cache
 from .utils import CacheError, CacheValidationError
 from .space_monitor import CacheOperationInterrupted
 from .info import get_cache_info, list_available_caches
-from .constants import SaveStatus, LoadStatus, TransferStatus, AsyncTransferStatus
+from .constants import SaveStatus, LoadStatus
 
 # Version
-__version__ = "0.1.0"
+__version__ = "0.1.2"
 
 __all__ = [
     "CacheError",
@@ -29,23 +15,9 @@ __all__ = [
     "CacheOperationInterrupted",
     "SaveStatus",
     "LoadStatus",
-    "TransferStatus",
-    "AsyncTransferStatus",
-    "transfer",
     "load_compile_cache",
     "save_compile_cache",
     "clear_local_cache",
     "get_cache_info",
     "list_available_caches",
-    # Generic async operations
-    "start_transfer_async",
-    "get_transfer_status",
-    "is_transfer_complete",
-    "wait_for_completion",
-    "cancel_transfer",
-    "list_active_transfers",
-    "TransferProgress",
-    # Torch-specific async operations
-    "load_compile_cache_async",
-    "save_compile_cache_async",
 ]

b10_transfer/cleanup.py

@@ -1,4 +1,4 @@
-"""Cooperative cleanup utilities for b10-tcache.
+"""Cooperative cleanup utilities for b10-transfer.
 
 This module provides cooperative cleanup functionality where each pod/replica
 helps maintain the health of shared resources in b10fs by removing stale

b10_transfer/constants.py

@@ -1,4 +1,4 @@
-"""Configuration constants for b10-tcache.
+"""Configuration constants for b10-transfer.
 
 This module defines configuration constants for the PyTorch compilation cache system.
 Some values can be overridden by environment variables, but security caps are enforced
@@ -36,8 +36,7 @@ B10FS_CACHE_DIR = validate_path_security(
     _b10fs_cache_dir, [_REQUIRED_TORCH_CACHE_DIR_PREFIX], "B10FS_CACHE_DIR"
 )
 
-# Validate LOCAL_WORK_DIR - allow /app, /tmp, and /cache paths.
-# This is like a "scratch" directory where you can do work (like compression/archival for example)
+# Validate LOCAL_WORK_DIR - allow /app, /tmp, and /cache paths
 _local_work_dir = os.getenv("LOCAL_WORK_DIR", "/app")
 LOCAL_WORK_DIR = validate_path_security(
     _local_work_dir, ["/app/", "/tmp/", "/cache/"], "LOCAL_WORK_DIR"
@@ -113,7 +112,6 @@ class WorkerStatus(Enum):
     SUCCESS = auto()
     ERROR = auto()
     CANCELLED = auto()
-    FILE_NOT_FOUND = auto()
 
 
 class LoadStatus(Enum):
@@ -131,22 +129,3 @@ class SaveStatus(Enum):
     SUCCESS = auto()
     ERROR = auto()
     SKIPPED = auto()
-
-
-class TransferStatus(Enum):
-    """Status values for generic transfer operations."""
-
-    SUCCESS = auto()
-    ERROR = auto()
-    INTERRUPTED = auto()
-    DOES_NOT_EXIST = auto()
-
-
-class AsyncTransferStatus(Enum):
-    NOT_STARTED = auto()
-    IN_PROGRESS = auto()
-    SUCCESS = auto()
-    ERROR = auto()
-    INTERRUPTED = auto()
-    CANCELLED = auto()
-    DOES_NOT_EXIST = auto()

b10_transfer/core.py

@@ -1,169 +1,394 @@
+import os
 import logging
+import tempfile
+import shutil
 from pathlib import Path
 
+import time
+
+from .environment import get_cache_filename
 from .cleanup import cooperative_cleanup_b10fs
 from .utils import (
     timed_fn,
+    critical_section_b10fs_file_lock,
     safe_execute,
+    temp_file_cleanup,
     cache_operation,
+    safe_unlink,
 )
 from .space_monitor import (
     check_sufficient_disk_space,
     CacheSpaceMonitor,
     CacheOperationInterrupted,
-    CacheFileNotFoundError,
     run_monitored_process,
+    worker_process,
 )
 from .constants import (
+    TORCH_CACHE_DIR,
     B10FS_CACHE_DIR,
     LOCAL_WORK_DIR,
+    MAX_CACHE_SIZE_MB,
     REQUIRED_B10FS_SPACE_MB,
     MIN_LOCAL_SPACE_MB,
-    TransferStatus,
+    CACHE_FILE_EXTENSION,
+    CACHE_LATEST_SUFFIX,
+    CACHE_INCOMPLETE_SUFFIX,
+    LoadStatus,
+    SaveStatus,
 )
 
 logger = logging.getLogger(__name__)
 
 
-@timed_fn(logger=logger, name="Generic transfer operation")
-@safe_execute("Transfer failed", TransferStatus.ERROR)
-def transfer(
-    source: Path,
-    dest: Path,
-    callback: callable,
-    *callback_args,
-    monitor_local: bool = True,
-    monitor_b10fs: bool = True,
-    **callback_kwargs,
-) -> TransferStatus:
-    """Generic transfer function with space monitoring and atomic operations.
-
-    The actual transfer logic is provided via callback.
-
-    The function handles:
-    - Cooperative cleanup of stale shared resources
-    - Space monitoring during operations (optional for local and b10fs)
-    - Atomic operations using temp files and rename
-    - Automatic cleanup on interruption or failure
-    - Lock management for b10fs operations
+@timed_fn(logger=logger, name="Loading compile cache")
+@safe_execute("Load failed", False)
+def load_compile_cache() -> LoadStatus:
+    """Load PyTorch compilation cache from b10fs to local torch cache directory.
 
-    Args:
-        source: Source path for the transfer operation
-        dest: Destination path for the transfer operation
-        callback: Function to perform the actual transfer work
-        *callback_args: Positional arguments to pass to callback
-        monitor_local: Whether to monitor local disk space (default: True)
-        monitor_b10fs: Whether to monitor b10fs disk space (default: True)
-        **callback_kwargs: Keyword arguments to pass to callback
+    This function implements a lock-free pattern to safely load cached PyTorch
+    compilation artifacts from the b10fs shared filesystem to the local torch
+    cache directory. It validates b10fs availability, checks for existing cache,
+    and extracts the archive if needed.
+
+    The function monitors local disk space during both the copy from b10fs and
+    extraction phases, interrupting operations if space falls below MIN_LOCAL_SPACE_MB.
 
     Returns:
-        TransferStatus:
-              TransferStatus.SUCCESS if transfer completed successfully
-              TransferStatus.ERROR if transfer failed
-              TransferStatus.INTERRUPTED if transfer was interrupted due to insufficient disk space
+        LoadStatus:
+              LoadStatus.SUCCESS if cache was successfully loaded
+              LoadStatus.SKIPPED if already exists
+              LoadStatus.ERROR if b10fs is unavailable, local disk space is insufficient, or loading failed.
+              LoadStatus.DOES_NOT_EXIST if no cache file was found.
 
     Raises:
-        CacheValidationError: If b10fs is not enabled (caught and returns TransferStatus.ERROR).
+        CacheValidationError: If b10fs is not enabled (caught and returns LoadStatus.ERROR).
         CacheOperationInterrupted: If operations interrupted due to insufficient
-                                  disk space (caught and returns TransferStatus.INTERRUPTED).
-        Exception: Any other errors during transfer (caught and returns TransferStatus.ERROR).
+                                  local disk space (caught and returns LoadStatus.ERROR).
+        Exception: Any other errors during loading (caught and returns LoadStatus.ERROR).
     """
-    with cache_operation("Transfer"):
+    with cache_operation("Load"):
         # Cooperative cleanup of stale shared resources
         cooperative_cleanup_b10fs()
 
         b10fs_dir = Path(B10FS_CACHE_DIR)
+        torch_dir = Path(TORCH_CACHE_DIR)
         work_dir = Path(LOCAL_WORK_DIR)
 
-        # Determine which paths to monitor based on source/dest
-        local_path = None
-        b10fs_path = None
-
-        if str(source).startswith(str(b10fs_dir)) or str(dest).startswith(
-            str(b10fs_dir)
-        ):
-            b10fs_path = b10fs_dir
-
-        if (
-            str(source).startswith(str(work_dir))
-            or str(dest).startswith(str(work_dir))
-            or not str(source).startswith(str(b10fs_dir))
-            or not str(dest).startswith(str(b10fs_dir))
-        ):
-            local_path = work_dir
-
-        # Initial disk space checks
-        if monitor_local and local_path:
-            check_sufficient_disk_space(
-                local_path, MIN_LOCAL_SPACE_MB, "local transfer operations"
-            )
-            logger.debug(
-                f"Initial local space check passed: {MIN_LOCAL_SPACE_MB:.1f}MB required"
-            )
-
-        if monitor_b10fs and b10fs_path:
-            check_sufficient_disk_space(
-                b10fs_path, REQUIRED_B10FS_SPACE_MB, "b10fs transfer operations"
-            )
-            logger.debug(
-                f"Initial b10fs space check passed: {REQUIRED_B10FS_SPACE_MB:.1f}MB required"
-            )
-
-        # Determine primary space monitor (prioritize b10fs if both are monitored)
-        primary_monitor = None
-        if monitor_b10fs and b10fs_path:
-            primary_monitor = CacheSpaceMonitor(REQUIRED_B10FS_SPACE_MB, b10fs_path)
-        elif monitor_local and local_path:
-            primary_monitor = CacheSpaceMonitor(MIN_LOCAL_SPACE_MB, local_path)
-
-        if primary_monitor is None:
-            # No monitoring requested, execute callback directly
-            logger.info(f"Starting transfer (no monitoring): {source} -> {dest}")
-            try:
-                callback(source, dest, *callback_args, **callback_kwargs)
-                logger.info("Transfer complete")
-                return TransferStatus.SUCCESS
-            except (FileNotFoundError, CacheFileNotFoundError) as e:
-                logger.info(f"Transfer failed - file not found: {e}")
-                return TransferStatus.DOES_NOT_EXIST
-
-        # Start the primary space monitor
-        primary_monitor.start()
+        cache_filename = get_cache_filename()
+        cache_file = (
+            b10fs_dir / f"{cache_filename}{CACHE_LATEST_SUFFIX}{CACHE_FILE_EXTENSION}"
+        )
+        logger.debug(f"Looking for cache file: {cache_file}")
+
+        if not cache_file.exists():
+            logger.info("No cache file found in b10fs")
+            return LoadStatus.DOES_NOT_EXIST
+
+        # Skip if already loaded
+        if torch_dir.exists() and any(torch_dir.iterdir()):
+            logger.info("Torch cache already loaded, skipping extraction")
+            return LoadStatus.SKIPPED
+
+        # Initial disk space check for local operations
+        check_sufficient_disk_space(
+            work_dir, MIN_LOCAL_SPACE_MB, "cache load operations"
+        )
+        logger.debug(
+            f"Initial space check passed: {MIN_LOCAL_SPACE_MB:.1f}MB required on local machine"
+        )
+
+        # Start background space monitoring for local disk
+        space_monitor = CacheSpaceMonitor(MIN_LOCAL_SPACE_MB, work_dir)
+        space_monitor.start()
+
+        # Create temp local copy
+        with tempfile.NamedTemporaryFile(
+            suffix=CACHE_FILE_EXTENSION, dir=work_dir, delete=False
+        ) as f:
+            temp_path = Path(f.name)
+        logger.debug(f"Created temporary file for cache: {temp_path}")
 
         try:
-            # Execute the callback using monitored process for continuous space monitoring
-            logger.info(f"Starting monitored transfer: {source} -> {dest}")
+            with temp_file_cleanup(temp_path):
+                # Phase 1: Copy from b10fs to local temp file in separate process
+                @critical_section_b10fs_file_lock("copy_out")
+                def _monitored_copy_from_b10fs():
+                    logger.info(
+                        f"Starting copy from b10fs: {cache_file} -> {temp_path}"
+                    )
+                    run_monitored_process(
+                        _cache_copy_from_b10fs_worker,
+                        (str(cache_file), str(temp_path)),
+                        space_monitor,
+                        "b10fs to local copy",
+                    )
+
+                _monitored_copy_from_b10fs()
 
-            # Try direct callback with run_monitored_process first
-            try:
+                # Phase 2: Extract archive in separate process
+                logger.info(f"Starting extraction: {temp_path} -> {torch_dir}")
                 run_monitored_process(
-                    callback,
-                    (source, dest, *callback_args),
-                    primary_monitor,
-                    "transfer callback",
+                    _cache_extract_worker,
+                    (str(temp_path), str(torch_dir)),
+                    space_monitor,
+                    "archive extraction",
+                    cleanup_func=lambda: _cleanup_torch_dir(torch_dir),
                 )
-                logger.info("Transfer complete (monitored)")
-                return TransferStatus.SUCCESS
 
-            except (TypeError, AttributeError, ImportError, OSError) as e:
-                # Callback not pickleable or other serialization issue
-                logger.warning(
-                    f"Callback not suitable for process isolation, running without monitoring: {e}"
+            logger.info("Cache load complete")
+            return LoadStatus.SUCCESS
+
+        except CacheOperationInterrupted as e:
+            logger.warning(f"Cache load interrupted: {e}")
+            return LoadStatus.ERROR
+
+        finally:
+            space_monitor.stop()
+
+
+"""
+FIXME(SRAY):
+What about the case in @b10-transfer/ where a single pod finishes an inference request,
+and then the client calls save_compile_cache. And while we are creating the local archive,
+another inference call on the same pod is kicked off, which then modifies the torch cache.
+How would this be handled? Maybe just accept that the cache will be recompiled/overwritten?
+Otherwise you'd need application level coordination to ensure that the cache is not modified
+while we are creating the archive, but this doesn't really seem like a good idea in terms of adoption.
+
+FIXME(SR):
+More things to consider:
+- [possible] What if b10fs dies *during* an op? right now we check for b10fs availability in the beginning of the op... Add some constants instead of just False for load().
+- [possible, and really bad if it happens] potential memory exhaustion during compression if the cache is super super large. very very edge case. higher compression levels also have high memory usage.
+"""
+
+
+@timed_fn(logger=logger, name="Saving compile cache")
+@safe_execute("Save failed", False)
+def save_compile_cache() -> SaveStatus:
+    """Save local PyTorch compilation cache to b10fs using atomic journal pattern.
+
+    This function creates an archive of the local torch cache directory and
+    atomically saves it to b10fs using a journal pattern (write to temp file,
+    then rename). This ensures concurrent saves don't corrupt each other.
+
+    The function validates b10fs availability, checks if cache already exists
+    (early exit), performs initial space checks using pre-calculated requirements
+    for concurrent saves, starts background space monitoring, then runs compression
+    and copy operations in separate worker processes that can be terminated if disk
+    space becomes insufficient, finally performing an atomic rename to the final cache file.
+
+    Returns:
+        SaveStatus:
+              SaveStatus.SUCCESS if cache was successfully saved or already exists
+              SaveStatus.ERROR if b10fs is unavailable, insufficient disk space caused interruption,
+                no cache exists to save, or saving failed.
+              SaveStatus.SKIPPED if no cache exists to save or cache already exists in b10fs
+
+    Raises:
+        CacheValidationError: If b10fs is not enabled (caught and returns SaveStatus.ERROR).
+        CacheOperationInterrupted: If operations interrupted due to insufficient
+                                  disk space (caught and returns SaveStatus.ERROR).
+        ArchiveError: If archive creation fails (caught and returns SaveStatus.ERROR).
+        Exception: Any other errors during saving (caught and returns SaveStatus.ERROR).
+    """
+    with cache_operation("Save"):
+        # Cooperative cleanup of stale shared resources
+        cooperative_cleanup_b10fs()
+
+        b10fs_dir = Path(B10FS_CACHE_DIR)
+        torch_dir = Path(TORCH_CACHE_DIR)
+        work_dir = Path(LOCAL_WORK_DIR)
+
+        # Check if anything to save
+        if not torch_dir.exists() or not any(torch_dir.iterdir()):
+            logger.info("No torch cache to save")
+            return SaveStatus.SKIPPED
+
+        cache_filename = get_cache_filename()
+        final_file = (
+            b10fs_dir / f"{cache_filename}{CACHE_LATEST_SUFFIX}{CACHE_FILE_EXTENSION}"
+        )
+
+        # Check for existing cache first (early exit)
+        if final_file.exists():
+            logger.info("Cache already exists in b10fs, skipping save")
+            return SaveStatus.SKIPPED
+
+        # Initial disk space checks using calculated space requirements
+        check_sufficient_disk_space(
+            work_dir, MAX_CACHE_SIZE_MB, "local temp file creation"
+        )
+        check_sufficient_disk_space(
+            b10fs_dir, REQUIRED_B10FS_SPACE_MB, "cache save to b10fs"
+        )
+        logger.debug(
+            f"Initial space checks passed: {MAX_CACHE_SIZE_MB:.1f}MB local, {REQUIRED_B10FS_SPACE_MB:.1f}MB b10fs"
+        )
+
+        temp_file = (
+            b10fs_dir
+            / f"{cache_filename}{CACHE_INCOMPLETE_SUFFIX}{CACHE_FILE_EXTENSION}"
+        )
+
+        # Start background space monitoring
+        space_monitor = CacheSpaceMonitor(REQUIRED_B10FS_SPACE_MB, b10fs_dir)
+        space_monitor.start()
+
+        with tempfile.NamedTemporaryFile(
+            suffix=CACHE_FILE_EXTENSION, dir=work_dir, delete=False
+        ) as f:
+            local_temp = Path(f.name)
+        logger.debug(f"Created local temp file for archive: {local_temp}")
+
+        try:
+            with temp_file_cleanup(local_temp):
+                # Phase 1: Compression in separate process
+                logger.info(f"Starting compression: {torch_dir} -> {local_temp}")
+                run_monitored_process(
+                    _cache_compression_worker,
+                    (str(torch_dir), str(local_temp), MAX_CACHE_SIZE_MB),
+                    space_monitor,
+                    "compression",
                 )
 
-                # Fallback to direct execution without process isolation
-                callback(source, dest, *callback_args, **callback_kwargs)
-                logger.info("Transfer complete (unmonitored)")
-                return TransferStatus.SUCCESS
+                b10fs_dir.mkdir(parents=True, exist_ok=True)
 
-        except (FileNotFoundError, CacheFileNotFoundError) as e:
-            logger.info(f"Transfer failed - file not found: {e}")
-            return TransferStatus.DOES_NOT_EXIST
+                # Phase 2: Copy to b10fs in separate process
+                @critical_section_b10fs_file_lock("copy_in")
+                def _monitored_copy_to_b10fs():
+                    logger.info(f"Starting copy to b10fs: {local_temp} -> {temp_file}")
+                    run_monitored_process(
+                        _cache_copy_worker,
+                        (str(local_temp), str(temp_file)),
+                        space_monitor,
+                        "b10fs copy",
+                        cleanup_func=lambda: safe_unlink(
+                            temp_file, f"Failed to cleanup interrupted copy {temp_file}"
+                        ),
+                    )
+
+                _monitored_copy_to_b10fs()
+
+                # Phase 3: Atomic rename (fast, don't interrupt)
+                logger.info(
+                    f"Renaming temp file to final cache file: {temp_file} -> {final_file}"
+                )
+                temp_file.rename(final_file)
+
+            logger.info("Cache save complete")
+            return SaveStatus.SUCCESS
 
         except CacheOperationInterrupted as e:
-            logger.warning(f"Transfer interrupted: {e}")
-            return TransferStatus.INTERRUPTED
+            logger.warning(f"Cache save interrupted: {e}")
+            return SaveStatus.ERROR
 
         finally:
-            # Stop space monitor
-            primary_monitor.stop()
+            space_monitor.stop()
+
+
+@safe_execute("Clear failed", False)
+def clear_local_cache() -> bool:
+    """Clear the local PyTorch compilation cache directory.
+
+    This function removes the entire local torch cache directory and all its
+    contents. This is useful for cleaning up disk space or forcing recompilation.
+
+    Returns:
+        bool: True if cache was successfully cleared or didn't exist, False if
+              clearing failed due to permissions or other filesystem errors.
+
+    Raises:
+        Exception: Any errors during directory removal (caught and returns False).
+    """
+    torch_dir = Path(TORCH_CACHE_DIR)
+    if not torch_dir.exists():
+        return True
+    shutil.rmtree(torch_dir)
+    return True
+
+
+@worker_process("Compression was cancelled before starting")
+def _cache_compression_worker(
+    torch_dir_str: str, local_temp_str: str, max_size_mb: int
+) -> None:
+    """Worker process that handles cache compression.
+
+    This function runs in a separate process to compress the torch cache directory
+    into an archive. It can be terminated externally if disk space becomes insufficient.
+
+    Args:
+        torch_dir_str: String path to the torch cache directory to compress.
+        local_temp_str: String path where the compressed archive will be created.
+        max_size_mb: Maximum allowed archive size in megabytes.
+    """
+    torch_dir = Path(torch_dir_str)
+    local_temp = Path(local_temp_str)
+
+    # Import here to avoid issues with multiprocessing
+    from .archive import create_archive
+
+    create_archive(torch_dir, local_temp, max_size_mb)
+
+
+@worker_process("Copy was cancelled before starting")
+def _cache_copy_worker(source_path_str: str, dest_path_str: str) -> None:
+    """Worker process that handles file copy to b10fs.
+
+    This function runs in a separate process to copy the compressed cache file
+    to the b10fs filesystem. It can be terminated externally if disk space becomes insufficient.
+
+    Args:
+        source_path_str: String path to the source file to copy.
+        dest_path_str: String path where the file will be copied.
+    """
+    source_path = Path(source_path_str)
+    dest_path = Path(dest_path_str)
+
+    shutil.copy2(source_path, dest_path)
+
+
+@worker_process("Copy from b10fs was cancelled before starting")
+def _cache_copy_from_b10fs_worker(source_path_str: str, dest_path_str: str) -> None:
+    """Worker process that handles file copy from b10fs to local machine.
+
+    This function runs in a separate process to copy the cache file from b10fs
+    to the local filesystem. It can be terminated externally if local disk space becomes insufficient.
+
+    Args:
+        source_path_str: String path to the source file in b10fs to copy.
+        dest_path_str: String path where the file will be copied locally.
+    """
+    source_path = Path(source_path_str)
+    dest_path = Path(dest_path_str)
+
+    shutil.copy2(source_path, dest_path)
+
+
+def _cleanup_torch_dir(torch_dir: Path) -> None:
+    """Helper function to safely cleanup torch directory during interrupted extraction."""
+    try:
+        if torch_dir.exists():
+            shutil.rmtree(torch_dir)
+            logger.debug(f"Cleaned up torch directory: {torch_dir}")
+    except Exception as e:
+        logger.error(f"Failed to cleanup torch directory {torch_dir}: {e}")
+
+
+@worker_process("Extraction was cancelled before starting")
+def _cache_extract_worker(archive_path_str: str, dest_dir_str: str) -> None:
+    """Worker process that handles archive extraction.
+
+    This function runs in a separate process to extract the cache archive to
+    the torch cache directory. It can be terminated externally if local disk space becomes insufficient.
+
+    Args:
+        archive_path_str: String path to the archive file to extract.
+        dest_dir_str: String path to the directory where archive will be extracted.
+    """
+    archive_path = Path(archive_path_str)
+    dest_dir = Path(dest_dir_str)
+
+    # Import here to avoid issues with multiprocessing
+    from .archive import extract_archive
+
+    extract_archive(archive_path, dest_dir)

b10_transfer/space_monitor.py

@@ -1,4 +1,4 @@
-"""Space monitoring utilities for b10-tcache.
+"""Space monitoring utilities for b10-transfer.
 
 This module provides disk space monitoring functionality to prevent cache operations
 from exhausting available disk space and causing system instability.
@@ -24,12 +24,6 @@ class CacheOperationInterrupted(Exception):
     pass
 
 
-class CacheFileNotFoundError(Exception):
-    """Raised when a cache file is not found during transfer operations."""
-
-    pass
-
-
 def worker_process(cancelled_message: str):
     """Decorator for worker process functions to handle common try/catch/result_queue pattern.
 
@@ -66,8 +60,6 @@ def worker_process(cancelled_message: str):
                 # If we get here, the function completed successfully
                 result_queue.put((WorkerStatus.SUCCESS.value, None))
 
-            except FileNotFoundError as e:
-                result_queue.put((WorkerStatus.FILE_NOT_FOUND.value, str(e)))
             except Exception as e:
                 result_queue.put((WorkerStatus.ERROR.value, str(e)))
 
@@ -294,11 +286,6 @@ def run_monitored_process(
                 if cleanup_func:
                     cleanup_func()
                 raise CacheOperationInterrupted(error_msg)
-            elif status == WorkerStatus.FILE_NOT_FOUND.value:
-                logger.info(
-                    f"{operation_name} worker failed - file not found: {error_msg}"
-                )
-                raise CacheFileNotFoundError(error_msg)
             # status == WorkerStatus.SUCCESS.value - continue normally
 
         logger.debug(f"{operation_name} completed successfully")