Module org.elasticsearch.server

Package org.elasticsearch.cluster

Class AckedClusterStateUpdateTask

java.lang.Object

org.elasticsearch.cluster.ClusterStateUpdateTask

org.elasticsearch.cluster.AckedClusterStateUpdateTask

All Implemented Interfaces:

ClusterStateAckListener, ClusterStateTaskListener

Direct Known Subclasses:

RepositoriesService.RegisterRepositoryTask, RepositoriesService.UnregisterRepositoryTask

public abstract class AckedClusterStateUpdateTask
extends ClusterStateUpdateTask
implements ClusterStateAckListener

An extension interface to ClusterStateUpdateTask that allows the caller to be notified after the master has computed, published, accepted, committed, and applied the cluster state update AND only after the rest of the nodes (or a specified subset) have also accepted and applied the cluster state update.

Constructor Summary

Constructors

Modifier	Constructor	Description
protected	AckedClusterStateUpdateTask(AcknowledgedRequest<?> request, ActionListener<? extends AcknowledgedResponse> listener)
protected	AckedClusterStateUpdateTask(Priority priority, AcknowledgedRequest<?> request, ActionListener<? extends AcknowledgedResponse> listener)
protected	AckedClusterStateUpdateTask(Priority priority, TimeValue masterNodeTimeout, TimeValue ackTimeout, ActionListener<? extends AcknowledgedResponse> listener)
protected	AckedClusterStateUpdateTask(TimeValue masterNodeTimeout, TimeValue ackTimeout, ActionListener<? extends AcknowledgedResponse> listener)

Method Summary

Modifier and Type	Method	Description
final TimeValue	ackTimeout()
boolean	mustAck(DiscoveryNode discoveryNode)	Called to determine which nodes the acknowledgement is expected from
protected AcknowledgedResponse	newResponse(boolean acknowledged)
void	onAckFailure(Exception e)	Called after all the selected nodes have acknowledged the cluster state update request but at least one of them failed.
void	onAckTimeout()	Called if the acknowledgement timeout defined by ClusterStateAckListener.ackTimeout() expires while still waiting for an acknowledgement from one or more of the selected nodes.
void	onAllNodesAcked()	Called once all the selected nodes have acknowledged the cluster state update request.
void	onFailure(Exception e)	A callback for when task execution fails.

Methods inherited from class org.elasticsearch.cluster.ClusterStateUpdateTask

clusterStateProcessed, execute, priority, timeout

Methods inherited from class java.lang.Object

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

Constructor Details

AckedClusterStateUpdateTask

protected AckedClusterStateUpdateTask(AcknowledgedRequest<?> request,
 ActionListener<? extends AcknowledgedResponse> listener)

AckedClusterStateUpdateTask

protected AckedClusterStateUpdateTask(TimeValue masterNodeTimeout,
 TimeValue ackTimeout,
 ActionListener<? extends AcknowledgedResponse> listener)

AckedClusterStateUpdateTask

protected AckedClusterStateUpdateTask(Priority priority,
 AcknowledgedRequest<?> request,
 ActionListener<? extends AcknowledgedResponse> listener)

AckedClusterStateUpdateTask

protected AckedClusterStateUpdateTask(Priority priority,
 TimeValue masterNodeTimeout,
 TimeValue ackTimeout,
 ActionListener<? extends AcknowledgedResponse> listener)

Method Details

mustAck

public boolean mustAck(DiscoveryNode discoveryNode)

Called to determine which nodes the acknowledgement is expected from

Specified by:

mustAck in interface ClusterStateAckListener

Parameters:

discoveryNode - a node

Returns:

true if the node is expected to send ack back, false otherwise

onAllNodesAcked

public void onAllNodesAcked()

Description copied from interface: ClusterStateAckListener

Called once all the selected nodes have acknowledged the cluster state update request. Must be very lightweight execution, since it is executed on the cluster service thread.

Specified by:

onAllNodesAcked in interface ClusterStateAckListener

onAckFailure

public void onAckFailure(Exception e)

Description copied from interface: ClusterStateAckListener

Called after all the selected nodes have acknowledged the cluster state update request but at least one of them failed. Must be very lightweight execution, since it is executed on the cluster service thread.

Specified by:

onAckFailure in interface ClusterStateAckListener

Parameters:

e - exception representing the failure.

newResponse

protected AcknowledgedResponse newResponse(boolean acknowledged)

onAckTimeout

public void onAckTimeout()

Description copied from interface: ClusterStateAckListener

Called if the acknowledgement timeout defined by ClusterStateAckListener.ackTimeout() expires while still waiting for an acknowledgement from one or more of the selected nodes.

Specified by:

onAckTimeout in interface ClusterStateAckListener

onFailure

public void onFailure(Exception e)

Description copied from interface: ClusterStateTaskListener

A callback for when task execution fails. May receive a NotMasterException if this node stopped being the master before this task was executed or a ProcessClusterEventTimeoutException if the task timed out before it was executed. If the task fails during execution then this method receives the corresponding exception. If the task executes successfully but the resulting cluster state publication fails then this method receives a FailedToCommitClusterStateException. If publication fails then a new master is elected and the update might or might not take effect, depending on whether or not the newly-elected master accepted the published state that failed to be committed.

Use MasterService.isPublishFailureException(java.lang.Exception) to detect the "expected" master failure cases if needed.

Implementations of this callback must not throw exceptions: an exception thrown here is logged by the master service at ERROR level and otherwise ignored, except in tests where it raises an AssertionError. If log-and-ignore is the right behaviour then implementations must do so themselves, typically using a more specific logger and at a less dramatic log level.

Specified by:

onFailure in interface ClusterStateTaskListener

ackTimeout

public final TimeValue ackTimeout()

Specified by:

ackTimeout in interface ClusterStateAckListener

Returns:

acknowledgement timeout, i.e. the maximum time interval to wait for a full set of acknowledgements. This time interval is measured from the start of the publication (which is after computing the new cluster state and serializing it as a transport message). If the cluster state is committed (i.e. a quorum of master-eligible nodes have accepted the new state) and then the timeout elapses then the corresponding listener is completed via ClusterStateAckListener.onAckTimeout(). Although the time interval is measured from the start of the publication, it does not have any effect until the cluster state is committed:

• If the cluster state update fails before committing then the failure is always reported via ClusterStateAckListener.onAckFailure(Exception) rather than ClusterStateAckListener.onAckTimeout(), and this may therefore happen some time after the timeout period elapses.
• If the cluster state update is eventually committed, but takes longer than ackTimeout to do so, then the corresponding listener will be completed via ClusterStateAckListener.onAckTimeout() when it is committed, and this may therefore happen some time after the timeout period elapses.

A timeout of TimeValue.MINUS_ONE means that the master should wait indefinitely for acknowledgements.

A timeout of TimeValue.ZERO means that the master will complete this listener (via ClusterStateAckListener.onAckTimeout()) as soon as the state is committed, before any nodes have applied the new state.