public interface AffinityFunction extends Serializable
CacheConfiguration.getAffinity()
method.
Whenever a key is given to cache, it is first passed to a pluggable
AffinityKeyMapper
which may potentially map this key to an alternate
key which should be used for affinity. The key returned from
AffinityKeyMapper.affinityKey(Object)
method is then passed to
partition(Object)
method to find out the partition for the key.
On each topology change, partition-to-node mapping is calculated using
assignPartitions(AffinityFunctionContext)
method, which assigns a collection
of nodes to each partition.
This collection of nodes is used for node affinity. In REPLICATED
cache mode the key will be cached on all returned nodes; generally, all caching nodes
participate in caching every key in replicated mode. In PARTITIONED
mode, only primary and backup nodes are returned with primary node always in the
first position. So if there is 1
backup node, then the returned collection will
have 2
nodes in it - primary
node in first position, and backup
node in second.
For more information about cache affinity and examples refer to AffinityKeyMapper
and
@AffinityKeyMapped
documentation.
Modifier and Type | Method and Description |
---|---|
List<List<ClusterNode>> |
assignPartitions(AffinityFunctionContext affCtx)
Gets affinity nodes for a partition.
|
int |
partition(Object key)
Gets partition number for a given key starting from
0 . |
int |
partitions()
Gets total number of partitions available.
|
void |
removeNode(UUID nodeId)
Removes node from affinity.
|
void |
reset()
Resets cache affinity to its initial state.
|
void reset()
int partitions()
0
inclusively to
N
exclusively without any gaps.int partition(Object key)
0
. Partitioned caches
should make sure that keys are about evenly distributed across all partitions
from 0
to partition count
for best performance.
Note that for fully replicated caches it is possible to segment key sets among different grid node groups. In that case each node group should return a unique partition number. However, unlike partitioned cache, mappings of keys to nodes in replicated caches are constant and a node cannot migrate from one partition to another.
key
- Key to get partition for.List<List<ClusterNode>> assignPartitions(AffinityFunctionContext affCtx)
Note that partitioned affinity must obey the following contract: given that node
N
is primary for some key K
, if any other node(s) leave
grid and no node joins grid, node N
will remain primary for key K
.
affCtx
- Affinity function context. Will provide all required information to calculate
new partition assignments.void removeNode(UUID nodeId)
nodeId
- ID of node to remove.
Follow @ApacheIgnite
Ignite Database and Caching Platform : ver. 2.7.2 Release Date : February 6 2019