rememberingEntities with ddata mode, #22154

* one Replicator per configured role
* log LMDB directory at startup
* clarify the imporantce of the LMDB directory
* use more than one key to support many entities

akka-distributed-data/src/main/resources/reference.conf

@@ -72,6 +72,12 @@ akka.cluster.distributed-data {
       #    and its remote port.
       # 2. Otherwise the path is used as is, as a relative or absolute path to
       #    a directory.
+      #
+      # When running in production you may want to configure this to a specific
+      # path (alt 2), since the default directory contains the remote port of the
+      # actor system to make the name unique. If using a dynamically assigned 
+      # port (0) it will be different each time and the previously stored data 
+      # will not be loaded.
       dir = "ddata"
       
       # Size in bytes of the memory mapped file.

akka-distributed-data/src/main/scala/akka/cluster/ddata/DurableStore.scala

@@ -115,6 +115,7 @@ final class LmdbDurableStore(config: Config) extends Actor with ActorLogging {
       case path ⇒
         new File(path)
     }
+    log.info("Using durable data in LMDB directory [{}]", dir.getCanonicalPath)
     dir.mkdirs()
     Env.create()
       .setMapSize(mapSize)

akka-distributed-data/src/test/scala/akka/cluster/ddata/protobuf/ReplicatedDataSerializerSpec.scala

@@ -50,10 +50,11 @@ class ReplicatedDataSerializerSpec extends TestKit(ActorSystem(
     newBlob should equal(oldBlob)
   }
 
-  def checkSerialization(obj: AnyRef): Unit = {
+  def checkSerialization(obj: AnyRef): Int = {
     val blob = serializer.toBinary(obj)
     val ref = serializer.fromBinary(blob, serializer.manifest(obj))
     ref should be(obj)
+    blob.length
   }
 
   def checkSameContent(a: AnyRef, b: AnyRef): Unit = {
@@ -101,6 +102,31 @@ class ReplicatedDataSerializerSpec extends TestKit(ActorSystem(
       checkSameContent(s3.merge(s4), s4.merge(s3))
     }
 
+    "serialize large GSet" in {
+      val largeSet = (10000 until 20000).foldLeft(GSet.empty[String]) {
+        case (acc, n) ⇒ acc.add(n.toString)
+      }
+      val numberOfBytes = checkSerialization(largeSet)
+      info(s"size of GSet with ${largeSet.size} elements: $numberOfBytes bytes")
+      numberOfBytes should be <= (80000)
+    }
+
+    "serialize large ORSet" in {
+      val largeSet = (10000 until 20000).foldLeft(ORSet.empty[String]) {
+        case (acc, n) ⇒
+          val address = (n % 3) match {
+            case 0 ⇒ address1
+            case 1 ⇒ address2
+            case 2 ⇒ address3
+          }
+          acc.add(address, n.toString)
+      }
+      val numberOfBytes = checkSerialization(largeSet)
+      // note that ORSet is compressed, and therefore smaller than GSet
+      info(s"size of ORSet with ${largeSet.size} elements: $numberOfBytes bytes")
+      numberOfBytes should be <= (50000)
+    }
+
     "serialize Flag" in {
       checkSerialization(Flag())
       checkSerialization(Flag().switchOn)