Package org.apache.hadoop.hdfs.server.datanode.checker

Class ThrottledAsyncChecker<K,V>

java.lang.Object

org.apache.hadoop.hdfs.server.datanode.checker.ThrottledAsyncChecker<K,V>

All Implemented Interfaces:

AsyncChecker<K,V>

@Private
@Unstable
public class ThrottledAsyncChecker<K,V>
extends java.lang.Object
implements AsyncChecker<K,V>

An implementation of AsyncChecker that skips checking recently checked objects. It will enforce at least minMsBetweenChecks milliseconds between two successive checks of any one object. It is assumed that the total number of Checkable objects in the system is small, (not more than a few dozen) since the checker uses O(Checkables) storage and also potentially O(Checkables) threads. minMsBetweenChecks should be configured reasonably by the caller to avoid spinning up too many threads frequently.

Field Summary

Fields

Modifier and Type	Field	Description
static org.slf4j.Logger	LOG

Constructor Summary

Constructors

Constructor	Description
ThrottledAsyncChecker(org.apache.hadoop.util.Timer timer, long minMsBetweenChecks, long diskCheckTimeout, java.util.concurrent.ExecutorService executorService)

Method Summary

Modifier and Type	Method	Description
java.util.Optional<org.apache.hadoop.thirdparty.com.google.common.util.concurrent.ListenableFuture<V>>	schedule(Checkable<K,V> target, K context)	See AsyncChecker.schedule(org.apache.hadoop.hdfs.server.datanode.checker.Checkable<K, V>, K) If the object has been checked recently then the check will be skipped.
void	shutdownAndWait(long timeout, java.util.concurrent.TimeUnit timeUnit)	Cancel all executing checks and wait for them to complete.

Methods inherited from class java.lang.Object

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

Field Detail

LOG

public static final org.slf4j.Logger LOG

Constructor Detail

ThrottledAsyncChecker

public ThrottledAsyncChecker(org.apache.hadoop.util.Timer timer,
                             long minMsBetweenChecks,
                             long diskCheckTimeout,
                             java.util.concurrent.ExecutorService executorService)

Method Detail

schedule

public java.util.Optional<org.apache.hadoop.thirdparty.com.google.common.util.concurrent.ListenableFuture<V>> schedule(Checkable<K,V> target,
                                                                                                                       K context)

See AsyncChecker.schedule(org.apache.hadoop.hdfs.server.datanode.checker.Checkable<K, V>, K) If the object has been checked recently then the check will be skipped. Multiple concurrent checks for the same object will receive the same Future.

Specified by:

schedule in interface AsyncChecker<K,V>

Parameters:

target - object to be checked.

context - the interpretation of the context depends on the target.

Returns:

returns a of ListenableFuture that can be used to retrieve the result of the asynchronous check.

shutdownAndWait

public void shutdownAndWait(long timeout,
                            java.util.concurrent.TimeUnit timeUnit)
                     throws java.lang.InterruptedException

Cancel all executing checks and wait for them to complete. First attempts a graceful cancellation, then cancels forcefully. Waits for the supplied timeout after both attempts. See ExecutorService.awaitTermination(long, java.util.concurrent.TimeUnit) for a description of the parameters.. The results of in-progress checks are not useful during shutdown, so we optimize for faster shutdown by interrupt all actively executing checks.

Specified by:

shutdownAndWait in interface AsyncChecker<K,V>

Throws:

java.lang.InterruptedException