raboof/pekko
1
0
Fork 0
Code Issues Pull requests Projects Releases 6 Packages Wiki Activity Actions 10

=cls #17846 Use CRDTs instead of PersistentActor to remember the state of the ShardCoordinator #17871

Browse source
This commit is contained in:
Ostapenko Evgeniy 2015-08-20 13:24:39 +03:00
parent d2f08a3456
commit 6814d08ef1
14 changed files with 654 additions and 360 deletions
Download patch file Download diff file Expand all files Collapse all files

12
akka-docs/rst/scala/cluster-sharding.rst
View file

@ -179,6 +179,18 @@ unused shards due to the round-trip to the coordinator. Rebalancing of shards ma
also add latency. This should be considered when designing the application specific
shard resolution, e.g. to avoid too fine grained shards.
Distributed Data Mode
---------------------
Instead of using ``akka-persistence`` is possible to use ``akka-distributed-data`` module. In such case
state of the ``ShardCoordinator`` will be replicated inside a cluster by the ``akka-distributed-data``
module with the ``WriteMajority`` consistency. This mode could be enabled by setting up
``akka.cluster.sharding.state-store-mode`` as ``ddata``.
It make possible to remove ``akka-persistence`` dependency from a project if this dependency
has not using in user code and ``remember-entities`` is ``off``.
Note that option also could lead to the shards duplication in case of a cluster fragmentation
due to a broken replication between nodes.
Proxy Only Mode
---------------