Merge pull request #143 from lerna-stack/use-akka-actor-testkit-typed-config-instead

Ignore `LogReplicationDuringSnapshotSyncSpec` due to unstable

Author: negokaz

src/multi-jvm/scala/lerna/akka/entityreplication/typed/LogReplicationDuringSnapshotSyncSpec.scala

@@ -14,6 +14,7 @@ import akka.util.Timeout
 import com.typesafe.config.{ Config, ConfigFactory }
 import lerna.akka.entityreplication.util.AtLeastOnceComplete
 import lerna.akka.entityreplication.{ STMultiNodeSerializable, STMultiNodeSpec }
+import org.scalatest.Ignore
 
 import scala.concurrent.Future
 import scala.concurrent.duration._
@@ -30,7 +31,9 @@ object LogReplicationDuringSnapshotSyncSpecConfig extends MultiNodeConfig {
 
   testTransport(true)
 
-  private implicit val testKitSettings: TestKitSettings = TestKitSettings(ConfigFactory.load())
+  private implicit val testKitSettings: TestKitSettings = TestKitSettings(
+    ConfigFactory.load().getConfig("akka.actor.testkit.typed"),
+  )
 
   private val testConfig: Config =
     ConfigFactory.parseString {
@@ -45,11 +48,7 @@ object LogReplicationDuringSnapshotSyncSpecConfig extends MultiNodeConfig {
         raft-actor-auto-start.number-of-actors = 1
         
         sharding {
-          // Sharding will be available as possible quick.
-          retry-interval = 500ms
-          waiting-for-state-timeout = 500ms
-          updating-state-timeout = 500ms
-          
+          // drop old messages actively 
           buffer-size = 50
         }
         compaction {
@@ -61,8 +60,16 @@ object LogReplicationDuringSnapshotSyncSpecConfig extends MultiNodeConfig {
       // This spec requires longer timeouts
       akka.actor.testkit.typed.filter-leeway = 30s
       akka.testconductor.barrier-timeout = 5m
-      // This spec don't use the eventsourced feature
-      lerna.akka.entityreplication.raft.eventsourced.commit-log-store.retry.attempts = 0
+      
+      akka.cluster.sharding {
+        // Sharding will be available as possible quick.
+        retry-interval = 500ms
+        waiting-for-state-timeout = 500ms
+        updating-state-timeout = 500ms
+        distributed-data.majority-min-cap = 2
+        coordinator-state.write-majority-plus = 0
+        coordinator-state.read-majority-plus = 0
+      }
       """
     }
 
@@ -126,12 +133,20 @@ object LogReplicationDuringSnapshotSyncSpecConfig extends MultiNodeConfig {
   })
 }
 
-class LogReplicationDuringSnapshotSyncSpecMultiJvmController extends LogReplicationDuringSnapshotSyncSpec
-class LogReplicationDuringSnapshotSyncSpecMultiJvmNode1      extends LogReplicationDuringSnapshotSyncSpec
-class LogReplicationDuringSnapshotSyncSpecMultiJvmNode2      extends LogReplicationDuringSnapshotSyncSpec
-class LogReplicationDuringSnapshotSyncSpecMultiJvmNode3      extends LogReplicationDuringSnapshotSyncSpec
-class LogReplicationDuringSnapshotSyncSpecMultiJvmNode4      extends LogReplicationDuringSnapshotSyncSpec
-
+// This test is ignored due to that stabilizing this test in a CI environment is difficult.
+// To enable this test, remove all @Ignore below:
+@Ignore class LogReplicationDuringSnapshotSyncSpecMultiJvmController extends LogReplicationDuringSnapshotSyncSpec
+@Ignore class LogReplicationDuringSnapshotSyncSpecMultiJvmNode1      extends LogReplicationDuringSnapshotSyncSpec
+@Ignore class LogReplicationDuringSnapshotSyncSpecMultiJvmNode2      extends LogReplicationDuringSnapshotSyncSpec
+@Ignore class LogReplicationDuringSnapshotSyncSpecMultiJvmNode3      extends LogReplicationDuringSnapshotSyncSpec
+@Ignore class LogReplicationDuringSnapshotSyncSpecMultiJvmNode4      extends LogReplicationDuringSnapshotSyncSpec
+
+/**
+  * This test doesn't verify specific features but reproduces a specific fault.
+  *
+  * This test verifies that the committed events don't disappear even if new events are produced by entities
+  * while only the leader and the member synchronizing the snapshot exist.
+  */
 class LogReplicationDuringSnapshotSyncSpec
     extends MultiNodeSpec(LogReplicationDuringSnapshotSyncSpecConfig)
     with STMultiNodeSpec {
@@ -200,7 +215,7 @@ class LogReplicationDuringSnapshotSyncSpec
   "LogReplicationDuringSnapshotSyncSpec" should {
 
     "elect node1 as a leader" in {
-      newCluster(controller, node1, node2, node3)
+      newCluster(node1, node2, node3)
       runOn(node1) {
         expectNewLeaderElected {
           clusterReplication.init(Register(typedSystem))