Package com.clickhouse.client

Class ClickHouseNodes

java.lang.Object

com.clickhouse.client.ClickHouseNodes

All Implemented Interfaces:

ClickHouseNodeManager, Serializable, Function<ClickHouseNodeSelector,ClickHouseNode>

Direct Known Subclasses:

ClickHouseCluster

public class ClickHouseNodes
extends Object
implements ClickHouseNodeManager

A generic node manager for managing one or more nodes which may or may not belong to same cluster. It maintains two lists - one for healthy nodes and the other for faulty ones. Behind the scene, there's background thread(s) for discovering nodes and health check. Besides, ClickHouseLoadBalancingPolicy is used to pickup available node and moving node between lists according to its status.

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field	Description
protected AtomicBoolean	checking	Flag for exclusive health check.
protected AtomicReference<ScheduledFuture<?>>	discoveryFuture	Reference holding future of scheduled discovery.
protected LinkedList<ClickHouseNode>	faultyNodes	List of faulty nodes.
protected int	groupSize	Maximum number of nodes can be used for operation at a time.
protected AtomicReference<ScheduledFuture<?>>	healthCheckFuture	Reference holding future of scheduled health check.
protected AtomicInteger	index	Index for retrieving next node.
protected ReentrantReadWriteLock	lock	Lock for read and write nodes and faultyNodes.
protected LinkedList<ClickHouseNode>	nodes	List of healthy nodes.
protected ClickHouseLoadBalancingPolicy	policy	Load balancing policy.
protected ClickHouseNodeSelector	selector	Load balancing tags for filtering out nodes.
protected boolean	singleNode	Flag indicating whether it's single node or not.
protected ClickHouseNode	template	Template node.

Constructor Summary

Constructors

Modifier	Constructor	Description
protected	ClickHouseNodes(Collection<ClickHouseNode> nodes, ClickHouseNode template)	Default constructor.

Method Summary

Modifier and Type	Method	Description
ClickHouseNode	apply(ClickHouseNodeSelector t)
static String	buildCacheKey(String uri, Map<?,?> options)	Build unique key according to the given base URI and options for caching.
void	check()	Checks (faulty) node status.
void	discover()	Discovers nodes in the same cluster by querying against system.clusters table.
boolean	equals(Object obj)
protected ClickHouseNode	get()	Gets next node available.
List<ClickHouseNode>	getFaultyNodes()	Gets a copy of faulty nodes.
List<ClickHouseNode>	getFaultyNodes(ClickHouseNodeSelector selector, int groupSize)	Gets a copy of filtered faulty nodes.
List<ClickHouseNode>	getNodes()	Gets a copy of nodes, which in most cases are in healthy status.
List<ClickHouseNode>	getNodes(ClickHouseNodeSelector selector, int groupSize)	Gets a copy of filtered nodes.
ClickHouseNodeSelector	getNodeSelector()	Gets node selector for filtering out nodes.
ClickHouseLoadBalancingPolicy	getPolicy()	Gets load balancing policy.
ClickHouseNode	getTemplate()
int	hashCode()
boolean	isSingleNode()	Checks whether it's single node or not.
static ClickHouseNodes	of(String endpoints)	Gets or creates list of managed ClickHouseNode for load balancing and fail-over.
static ClickHouseNodes	of(String cacheKey, String endpoints, Map<?,?> options)	Gets or creates list of managed ClickHouseNode for load balancing and fail-over.
static ClickHouseNodes	of(String endpoints, Map<?,?> options)	Gets or creates list of managed ClickHouseNode for load balancing and fail-over.
protected void	queryClusterNodes(Collection<ClickHouseNode> seeds, Collection<ClickHouseNode> allNodes, Collection<ClickHouseNode> newHealthyNodes, Collection<ClickHouseNode> newFaultyNodes, Collection<ClickHouseNode> useless)
Optional<ScheduledFuture<?>>	scheduleDiscovery()	Schedule node discovery task immediately.
Optional<ScheduledFuture<?>>	scheduleHealthCheck()	Schedule node discovery task immediately.
ClickHouseNode	suggestNode(ClickHouseNode server, Throwable failure)	Suggests a different node in order to recover from a failure, which is usually a connection error.
String	toString()
void	update(ClickHouseNode node, ClickHouseNode.Status status)	Updates node status to one of ClickHouseNode.Status.

Methods inherited from class java.lang.Object

clone, finalize, getClass, notify, notifyAll, wait, wait, wait

Methods inherited from interface java.util.function.Function

andThen, compose

Field Detail

checking

protected final AtomicBoolean checking

Flag for exclusive health check.

index

protected final AtomicInteger index

Index for retrieving next node.

lock

protected final ReentrantReadWriteLock lock

Lock for read and write nodes and faultyNodes.

groupSize

protected final int groupSize

Maximum number of nodes can be used for operation at a time.

nodes

protected final LinkedList<ClickHouseNode> nodes

List of healthy nodes.

faultyNodes

protected final LinkedList<ClickHouseNode> faultyNodes

List of faulty nodes.

discoveryFuture

protected final AtomicReference<ScheduledFuture<?>> discoveryFuture

Reference holding future of scheduled discovery.

healthCheckFuture

protected final AtomicReference<ScheduledFuture<?>> healthCheckFuture

Reference holding future of scheduled health check.

policy

protected final ClickHouseLoadBalancingPolicy policy

Load balancing policy.

selector

protected final ClickHouseNodeSelector selector

Load balancing tags for filtering out nodes.

singleNode

protected final boolean singleNode

Flag indicating whether it's single node or not.

template

protected final ClickHouseNode template

Template node.

Constructor Detail

ClickHouseNodes

protected ClickHouseNodes(Collection<ClickHouseNode> nodes,
                          ClickHouseNode template)

Default constructor.

Parameters:

nodes - non-empty list of nodes

template - non-null template node

Method Detail

buildCacheKey

public static String buildCacheKey(String uri,
                                   Map<?,?> options)

Build unique key according to the given base URI and options for caching.

Parameters:

uri - non-null URI

options - options

Returns:

non-empty unique key for caching

of

public static ClickHouseNodes of(String endpoints)

Gets or creates list of managed ClickHouseNode for load balancing and fail-over.

Parameters:

endpoints - non-empty URIs separated by comma

Returns:

non-null list of nodes

of

public static ClickHouseNodes of(String endpoints,
                                 Map<?,?> options)

Gets or creates list of managed ClickHouseNode for load balancing and fail-over.

Parameters:

endpoints - non-empty URIs separated by comma

options - default options

Returns:

non-null list of nodes

of

public static ClickHouseNodes of(String cacheKey,
                                 String endpoints,
                                 Map<?,?> options)

Gets or creates list of managed ClickHouseNode for load balancing and fail-over. Since the list will be cached in a WeakHashMap, as long as you hold strong reference to the cacheKey, same combination of endpoints and options will be always mapped to the exact same list.

Parameters:

cacheKey - non-empty cache key

endpoints - non-empty URIs separated by comma

options - default options

Returns:

non-null list of nodes

queryClusterNodes

protected void queryClusterNodes(Collection<ClickHouseNode> seeds,
                                 Collection<ClickHouseNode> allNodes,
                                 Collection<ClickHouseNode> newHealthyNodes,
                                 Collection<ClickHouseNode> newFaultyNodes,
                                 Collection<ClickHouseNode> useless)

get

protected ClickHouseNode get()

Gets next node available.

Returns:

non-null node

isSingleNode

public boolean isSingleNode()

Checks whether it's single node or not.

Returns:

true if it's single node; false otherwise

apply

public ClickHouseNode apply(ClickHouseNodeSelector t)

Specified by:

apply in interface Function<ClickHouseNodeSelector,ClickHouseNode>

suggestNode

public ClickHouseNode suggestNode(ClickHouseNode server,
                                  Throwable failure)

Description copied from interface: ClickHouseNodeManager

Suggests a different node in order to recover from a failure, which is usually a connection error.

Specified by:

suggestNode in interface ClickHouseNodeManager

Parameters:

server - node related to the failure(e.g. the node couldn't be connected)

failure - recoverable failure

Returns:

non-null node which may or may not be same as the given one

update

public void update(ClickHouseNode node,
                   ClickHouseNode.Status status)

Description copied from interface: ClickHouseNodeManager

Updates node status to one of ClickHouseNode.Status. It simply delegates the call to getPolicy().update(node, status) in a thread-safe manner.

Specified by:

update in interface ClickHouseNodeManager

Parameters:

node - non-null node to update

status - non-null status of the node

check

public void check()

Checks (faulty) node status.

discover

public void discover()

Discovers nodes in the same cluster by querying against system.clusters table.

getTemplate

public ClickHouseNode getTemplate()

getNodes

public final List<ClickHouseNode> getNodes()

Description copied from interface: ClickHouseNodeManager

Gets a copy of nodes, which in most cases are in healthy status. However, this really depends on how ClickHouseLoadBalancingPolicy manages node status. In first-alive policy, it's acutally a full list regardless node status.

Specified by:

getNodes in interface ClickHouseNodeManager

Returns:

non-null nodes

getNodes

public List<ClickHouseNode> getNodes(ClickHouseNodeSelector selector,
                                     int groupSize)

Description copied from interface: ClickHouseNodeManager

Gets a copy of filtered nodes.

Specified by:

getNodes in interface ClickHouseNodeManager

Parameters:

selector - node selector for filtering out nodes, null means no filter

groupSize - maximum number of nodes to get, zero or negative value means all

Returns:

non-null nodes

getFaultyNodes

public final List<ClickHouseNode> getFaultyNodes()

Description copied from interface: ClickHouseNodeManager

Gets a copy of faulty nodes.

Specified by:

getFaultyNodes in interface ClickHouseNodeManager

Returns:

non-null faulty nodes

getFaultyNodes

public List<ClickHouseNode> getFaultyNodes(ClickHouseNodeSelector selector,
                                           int groupSize)

Description copied from interface: ClickHouseNodeManager

Gets a copy of filtered faulty nodes.

Specified by:

getFaultyNodes in interface ClickHouseNodeManager

Parameters:

selector - node selector for filtering out nodes, null means no filter

groupSize - maximum number of nodes to get, zero or negative value means all

Returns:

non-null faulty nodes

getPolicy

public ClickHouseLoadBalancingPolicy getPolicy()

Description copied from interface: ClickHouseNodeManager

Gets load balancing policy.

Specified by:

getPolicy in interface ClickHouseNodeManager

Returns:

non-null load balancing policy

getNodeSelector

public ClickHouseNodeSelector getNodeSelector()

Description copied from interface: ClickHouseNodeManager

Gets node selector for filtering out nodes.

Specified by:

getNodeSelector in interface ClickHouseNodeManager

Returns:

non-null node selector

scheduleDiscovery

public Optional<ScheduledFuture<?>> scheduleDiscovery()

Description copied from interface: ClickHouseNodeManager

Schedule node discovery task immediately. Nothing will happen when task scheduler does not exist(e.g. getPolicy().getScheduler() returns null) or there's a task running for node discovery.

Specified by:

scheduleDiscovery in interface ClickHouseNodeManager

Returns:

optional future for retrieving the running task

scheduleHealthCheck

public Optional<ScheduledFuture<?>> scheduleHealthCheck()

Description copied from interface: ClickHouseNodeManager

Schedule node discovery task immediately. Nothing will happen when task scheduler does not exist(e.g. getPolicy().getScheduler() returns null) or there's a task running for health check.

Specified by:

scheduleHealthCheck in interface ClickHouseNodeManager

Returns:

optional future for retrieving the running task

hashCode

public int hashCode()

Overrides:

hashCode in class Object

equals

public boolean equals(Object obj)

Overrides:

equals in class Object

toString

public String toString()

Overrides:

toString in class Object