Package org.apache.polaris.service.task

Class FileCleanupTaskHandler

java.lang.Object

org.apache.polaris.service.task.FileCleanupTaskHandler

All Implemented Interfaces:

TaskHandler

Direct Known Subclasses:

BatchFileCleanupTaskHandler, ManifestFileCleanupTaskHandler

public abstract class FileCleanupTaskHandler
extends Object
implements TaskHandler

Abstract base class for handling file cleanup tasks within Apache Polaris.

This class is for performing asynchronous file deletions with retry logic.

Subclasses must implement canHandleTask(TaskEntity) and handleTask(TaskEntity, CallContext) to define task-specific handling logic.

Field Summary

Fields

Modifier and Type	Field	Description
final ExecutorService	executorService
static final int	FILE_DELETION_RETRY_MILLIS
final TaskFileIOSupplier	fileIOSupplier
static final int	MAX_ATTEMPTS

Constructor Summary

Constructors

Constructor	Description
FileCleanupTaskHandler(TaskFileIOSupplier fileIOSupplier, ExecutorService executorService)

Method Summary

Modifier and Type	Method	Description
abstract boolean	canHandleTask(org.apache.polaris.core.entity.TaskEntity task)
abstract void	handleTask(org.apache.polaris.core.entity.TaskEntity task, org.apache.polaris.core.context.CallContext callContext)	Handles the provided task.
CompletableFuture<Void>	tryDelete(org.apache.iceberg.catalog.TableIdentifier tableId, org.apache.iceberg.io.FileIO fileIO, Iterable<String> files, String type, Boolean isConcurrent, Throwable e, int attempt)	Attempts to delete multiple files in a batch operation with retry logic.
CompletableFuture<Void>	tryDelete(org.apache.iceberg.catalog.TableIdentifier tableId, org.apache.iceberg.io.FileIO fileIO, String baseFile, String file, Throwable e, int attempt)	Attempts to delete a single file with retry logic.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Details

MAX_ATTEMPTS

public static final int MAX_ATTEMPTS

See Also:

• Constant Field Values

FILE_DELETION_RETRY_MILLIS

public static final int FILE_DELETION_RETRY_MILLIS

See Also:

• Constant Field Values

fileIOSupplier

public final TaskFileIOSupplier fileIOSupplier

executorService

public final ExecutorService executorService

Constructor Details

FileCleanupTaskHandler

public FileCleanupTaskHandler(TaskFileIOSupplier fileIOSupplier,
 ExecutorService executorService)

Method Details

canHandleTask

public abstract boolean canHandleTask(org.apache.polaris.core.entity.TaskEntity task)

Specified by:

canHandleTask in interface TaskHandler

handleTask

public abstract void handleTask(org.apache.polaris.core.entity.TaskEntity task,
 org.apache.polaris.core.context.CallContext callContext)

Handles the provided task. A normal return indicates success and the task entity will be dropped; throwing indicates failure and triggers the task retry path.

Specified by:

handleTask in interface TaskHandler

tryDelete

public CompletableFuture<Void> tryDelete(org.apache.iceberg.catalog.TableIdentifier tableId,
 org.apache.iceberg.io.FileIO fileIO,
 String baseFile,
 String file,
 Throwable e,
 int attempt)

Attempts to delete a single file with retry logic. If the file does not exist, it logs a message and does not retry. If an error occurs, it retries up to MAX_ATTEMPTS times before failing.

Parameters:

tableId - The identifier of the table associated with the file.

fileIO - The FileIO instance used for file operations.

baseFile - An optional base file associated with the file being deleted (can be null).

file - The path of the file to be deleted.

e - The exception from the previous attempt, if any.

attempt - The current retry attempt count.

Returns:

A CompletableFuture representing the asynchronous deletion operation.

tryDelete

public CompletableFuture<Void> tryDelete(org.apache.iceberg.catalog.TableIdentifier tableId,
 org.apache.iceberg.io.FileIO fileIO,
 Iterable<String> files,
 String type,
 Boolean isConcurrent,
 Throwable e,
 int attempt)

Attempts to delete multiple files in a batch operation with retry logic. If an error occurs, it retries up to MAX_ATTEMPTS times before failing.

Parameters:

tableId - The identifier of the table associated with the files.

fileIO - The FileIO instance used for file operations.

files - The list of file paths to be deleted.

type - The type of files being deleted (e.g., data files, metadata files).

isConcurrent - Whether the deletion should be performed concurrently.

e - The exception from the previous attempt, if any.

attempt - The current retry attempt count.

Returns:

A CompletableFuture representing the asynchronous batch deletion operation.