Package org.elasticsearch.cluster.routing.allocation

Interface ExistingShardsAllocator

All Known Implementing Classes:

GatewayAllocator

public interface ExistingShardsAllocator

Searches for, and allocates, shards for which there is an existing on-disk copy somewhere in the cluster. The default implementation is GatewayAllocator, but plugins can supply their own implementations too.

Nested Class Summary

Nested Classes

Modifier and Type	Interface	Description
static interface	ExistingShardsAllocator.UnassignedAllocationHandler	Used by allocateUnassigned(org.elasticsearch.cluster.routing.ShardRouting, org.elasticsearch.cluster.routing.allocation.RoutingAllocation, org.elasticsearch.cluster.routing.allocation.ExistingShardsAllocator.UnassignedAllocationHandler) to handle its allocation decisions.

Field Summary

Fields

Modifier and Type	Field	Description
static final Setting<String>	EXISTING_SHARDS_ALLOCATOR_SETTING	Allows plugins to override how we allocate shards that may already exist on disk in the cluster.

Method Summary

Modifier and Type	Method	Description
void	afterPrimariesBeforeReplicas(RoutingAllocation allocation)	Called during a round of allocation after attempting to allocate all the primaries but before any replicas, allowing the allocator to prepare for replica allocation.
void	allocateUnassigned(ShardRouting shardRouting, RoutingAllocation allocation, ExistingShardsAllocator.UnassignedAllocationHandler unassignedAllocationHandler)	Allocate any unassigned shards in the given RoutingAllocation for which this ExistingShardsAllocator is responsible.
void	applyFailedShards(List<FailedShard> failedShards, RoutingAllocation allocation)	Called when the given shards have failed, so that implementations can invalidate caches and clean up any in-flight activity for those shards.
void	applyStartedShards(List<ShardRouting> startedShards, RoutingAllocation allocation)	Called when the given shards have started, so that implementations can invalidate caches and clean up any in-flight activity for those shards.
void	beforeAllocation(RoutingAllocation allocation)	Called before starting a round of allocation, allowing the allocator to invalidate some caches if appropriate.
void	cleanCaches()	Called when this node becomes the elected master and when it stops being the elected master, so that implementations can clean up any in-flight activity from an earlier mastership.
AllocateUnassignedDecision	explainUnassignedShardAllocation(ShardRouting unassignedShard, RoutingAllocation routingAllocation)	Returns an explanation for a single unassigned shard.
int	getNumberOfInFlightFetches()

Field Details

EXISTING_SHARDS_ALLOCATOR_SETTING

static final Setting<String> EXISTING_SHARDS_ALLOCATOR_SETTING

Allows plugins to override how we allocate shards that may already exist on disk in the cluster.

Method Details

beforeAllocation

void beforeAllocation(RoutingAllocation allocation)

Called before starting a round of allocation, allowing the allocator to invalidate some caches if appropriate.

afterPrimariesBeforeReplicas

void afterPrimariesBeforeReplicas(RoutingAllocation allocation)

Called during a round of allocation after attempting to allocate all the primaries but before any replicas, allowing the allocator to prepare for replica allocation.

allocateUnassigned

void allocateUnassigned(ShardRouting shardRouting,
 RoutingAllocation allocation,
 ExistingShardsAllocator.UnassignedAllocationHandler unassignedAllocationHandler)

Allocate any unassigned shards in the given RoutingAllocation for which this ExistingShardsAllocator is responsible.

explainUnassignedShardAllocation

AllocateUnassignedDecision explainUnassignedShardAllocation(ShardRouting unassignedShard,
 RoutingAllocation routingAllocation)

Returns an explanation for a single unassigned shard.

cleanCaches

void cleanCaches()

Called when this node becomes the elected master and when it stops being the elected master, so that implementations can clean up any in-flight activity from an earlier mastership.

applyStartedShards

void applyStartedShards(List<ShardRouting> startedShards,
 RoutingAllocation allocation)

Called when the given shards have started, so that implementations can invalidate caches and clean up any in-flight activity for those shards.

applyFailedShards

void applyFailedShards(List<FailedShard> failedShards,
 RoutingAllocation allocation)

Called when the given shards have failed, so that implementations can invalidate caches and clean up any in-flight activity for those shards.

getNumberOfInFlightFetches

int getNumberOfInFlightFetches()

Returns:

the number of in-flight fetches under this allocator's control.