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

Active active sharding (#29364)

Browse source
This commit is contained in:
Johan Andrén 2020-07-21 12:13:08 +02:00 • committed by Christopher Batey
parent f531d1e57d
commit 9830988566
9 changed files with 754 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

43
akka-docs/src/main/paradox/typed/persistence-active-active.md
View file

@ -241,3 +241,46 @@ When comparing two version vectors `v1` and `v2`:
* `v1`is BEFORE `v2` iff for all i v1(i) <= v2(i) and there exist a j such that v1(j) < v2(j)
* `v1`is AFTER `v2` iff for all i v1(i) >= v2(i) and there exist a j such that v1(j) > v2(j)
* `v1`is CONCURRENT with `v2` otherwise
## Sharded Active Active entities
To simplify what probably are the most common use cases for how you will want to distribute the active active actors there is a minimal API for running multiple instances of @ref[Akka Cluster Sharding](cluster-sharding.md),
each instance holding the entities for a single replica.
The distribution of the replicas can be controlled either through cluster roles or using the @ref[multi datacenter](cluster-dc.md) support in Akka Cluster.
The API consists of bootstrapping logic for starting the sharding instances through @apidoc[ActiveActiveShardingExtension] available from the
`akka-cluster-sharding-typed` module.
Scala
: @@snip [ActiveActiveShardingSpec.scala](/akka-cluster-sharding-typed/src/test/scala/akka/cluster/sharding/typed/ActiveActiveShardingSpec.scala) { #bootstrap }
Java
: @@snip [ActiveActiveShardingTest.java](/akka-cluster-sharding-typed/src/test/java/akka/cluster/sharding/typed/ActiveActiveShardingTest.java) { #bootstrap }
`init` returns an @apidoc[ActiveActiveSharding] instance which gives access to @apidoc[EntityRef]s for each of the replicas for arbitrary routing logic:
Scala
: @@snip [ActiveActiveShardingSpec.scala](/akka-cluster-sharding-typed/src/test/scala/akka/cluster/sharding/typed/ActiveActiveShardingSpec.scala) { #all-entity-refs }
Java
: @@snip [ActiveActiveShardingTest.java](/akka-cluster-sharding-typed/src/test/java/akka/cluster/sharding/typed/ActiveActiveShardingTest.java) { #all-entity-refs }
More advanced routing among the replicas is currently left as an exercise for the reader (or may be covered in a future release [#29281](https://github.com/akka/akka/issues/29281), [#29319](https://github.com/akka/akka/issues/29319)).
## Direct Replication of Events
Normally an event has to be written in the journal and then picked up by the trailing read journal in the other replicas.
As an optimization the active active events can be published across the Akka cluster to the replicas. The read side
query is still needed as delivery is not guaranteed, but can be configured to poll the database less often since most
events will arrive at the replicas through the cluster.
To enable this feature you first need to enable event publishing on the `EventSourcedBehavior` with `withEventPublishing`
(FIXME missing Java API) and then enable direct replication through `withDirectReplication()` on @apidoc[ActiveActiveShardingSettings] (if not using
active active sharding the replication can be run standalone by starting the @apidoc[ActiveActiveShardingDirectReplication] actor).
The "event publishing" feature publishes each event to the local system event bus as a side effect after it has been written,
the `ActiveActiveShardingDirectReplication` actor subscribes to these events and forwards them to the replicas allowing them
to fast forward the stream of events for the origin replica. (With additional potential future support in journals for fast forwarding [#29311](https://github.com/akka/akka/issues/29311)).