Merge remote-tracking branch 'origin/phase/beta1' into feature/loggingUpdates

Author: Lukasz Gasior

src/main/resources/application.conf

@@ -92,11 +92,11 @@ grothendieck {
 
     peer {
       # Retry delay for failed attempt at connecting to a peer
-      connect-retry-delay = 5 seconds
+      connect-retry-delay = 1 minute
 
       # Maximum number of reconnect attempts after the connection has been initiated.
       # After that, the connection will be dropped until its initiated again (eg. by peer discovery)
-      connect-max-retries = 1
+      connect-max-retries = 2
 
       disconnect-poison-pill-timeout = 5 seconds
 
@@ -111,16 +111,16 @@ grothendieck {
       wait-for-tcp-ack-timeout = 5 seconds
 
       # Maximum block headers in a single response message (as a blockchain host)
-      max-blocks-headers-per-message = 200
+      max-blocks-headers-per-message = 100
 
       # Maximum block bodies in a single response message (as a blockchain host)
-      max-blocks-bodies-per-message = 200
+      max-blocks-bodies-per-message = 100
 
       # Maximum transactions receipts in a single response message (as a blockchain host)
-      max-receipts-per-message = 200
+      max-receipts-per-message = 100
 
       # Maximum MPT components in a single response message (as a blockchain host)
-      max-mpt-components-per-message = 400
+      max-mpt-components-per-message = 200
 
       # Maximum number of peers this node can connect to
       max-peers = 40
@@ -248,7 +248,7 @@ grothendieck {
     peers-scan-interval = 3.seconds
 
     # Duration for blacklisting a peer. Blacklisting reason include: invalid response from peer, response time-out, etc.
-    blacklist-duration = 30.seconds
+    blacklist-duration = 200.seconds
 
     # Retry interval when not having enough peers to start fast-sync
     start-retry-interval = 5.seconds
@@ -257,7 +257,7 @@ grothendieck {
     sync-retry-interval = 5.seconds
 
     # Response time-out from peer during sync. If a peer fails to respond within this limit, it will be blacklisted
-    peer-response-timeout = 10.seconds
+    peer-response-timeout = 3.minutes
 
     # Interval for logging syncing status info
     print-status-interval = 30.seconds
@@ -269,7 +269,7 @@ grothendieck {
     max-concurrent-requests = 50
 
     # Requested number of block headers when syncing from other peers
-    block-headers-per-request = 2048
+    block-headers-per-request = 200
 
     # Requested number of block bodies when syncing from other peers
     block-bodies-per-request = 128
@@ -278,7 +278,7 @@ grothendieck {
     receipts-per-request = 60
 
     # Requested number of MPT nodes when syncing from other peers
-    nodes-per-request = 1000
+    nodes-per-request = 200
 
     # Minimum number of peers required to start fast-sync (by determining the target block)
     min-peers-to-choose-target-block = 2

src/main/scala/io/iohk/ethereum/blockchain/sync/FastSync.scala

@@ -183,6 +183,8 @@ trait FastSync {
           bestBlockHeaderNumber)
       }
 
+    private val heartBeat = scheduler.schedule(syncRetryInterval, syncRetryInterval * 2, self, ProcessSyncing)
+
     def receive: Receive = handlePeerUpdates orElse {
       case EnqueueNodes(hashes) =>
         hashes.foreach {
@@ -226,11 +228,23 @@ trait FastSync {
         cleanupRequestedMaps(ref)
 
       case PrintStatus =>
-        val totalNodesCount = downloadedNodesCount + mptNodesQueue.size + nonMptNodesQueue.size
-        log.info(
-          s"""|Block: ${appStateStorage.getBestBlockNumber()}/${initialSyncState.targetBlock.number}.
-              |Peers: ${assignedHandlers.size}/${handshakedPeers.size} (${blacklistedPeers.size} blacklisted).
-              |State: $downloadedNodesCount/$totalNodesCount known nodes""".stripMargin.replace("\n", " "))
+        printStatus()
+    }
+
+    private def printStatus() = {
+      val totalNodesCount = downloadedNodesCount + mptNodesQueue.size + nonMptNodesQueue.size
+      val formatPeer: (Peer) => String = peer => s"${peer.remoteAddress.getAddress.getHostAddress}:${peer.remoteAddress.getPort}"
+      log.info(
+        s"""|Block: ${appStateStorage.getBestBlockNumber()}/${initialSyncState.targetBlock.number}.
+            |Peers waiting_for_response/connected: ${assignedHandlers.size}/${handshakedPeers.size} (${blacklistedPeers.size} blacklisted).
+            |State: $downloadedNodesCount/$totalNodesCount nodes.
+            |""".stripMargin.replace("\n", " "))
+      log.debug(
+        s"""|Connection status: connected(${assignedHandlers.values.map(formatPeer).toSeq.sorted.mkString(", ")})/
+            |handshaked(${handshakedPeers.keys.map(formatPeer).toSeq.sorted.mkString(", ")})
+            | blacklisted(${blacklistedPeers.map { case (id, _) => id.value }.mkString(", ")})
+            |""".stripMargin.replace("\n", " ")
+      )
     }
 
     private def cleanupRequestedMaps(handler: ActorRef): Unit = {
@@ -291,6 +305,7 @@ trait FastSync {
     }
 
     def cleanup(): Unit = {
+      heartBeat.cancel()
       syncStatePersistCancellable.cancel()
       syncStateStorageActor ! PoisonPill
       fastSyncStateStorage.purge()

src/main/scala/io/iohk/ethereum/network/PeerActor.scala

@@ -201,7 +201,7 @@ class PeerActor[R <: HandshakeResult](
           rlpxConnection.uriOpt.foreach(uri => knownNodesManager ! KnownNodesManager.RemoveKnownNode(uri))
         case _ => // nothing
       }
-      log.debug("Received {}. Closing connection", d)
+      log.debug(s"Received {}. Closing connection with peer ${peerAddress.getHostString}:${peerAddress.getPort}", d)
       context unwatch rlpxConnection.ref
       context stop self
   }

src/test/resources/application.conf

@@ -23,7 +23,7 @@ grothendieck {
     peers-scan-interval = 500.millis
     blacklist-duration = 5.seconds
     start-retry-interval = 500.millis
-    sync-retry-interval = 500.millis
+    sync-retry-interval = 1.second
     peer-response-timeout = 1.second
     print-status-interval = 1.hour
 

src/universal/conf/network.conf

@@ -29,11 +29,11 @@ grothendieck {
 
     peer {
       # Retry delay for failed attempt at connecting to a peer
-      # connect-retry-delay = 20 seconds
+      # connect-retry-delay = 1 minute
 
       # Maximum number of reconnect attempts after the connection has been initiated.
       # After that, the connection will be dropped until its initiated again (eg. by peer discovery)
-      # connect-max-retries = 30
+      # connect-max-retries = 2
 
       # Maximum number of peers this node can connect to
       # max-peers = 40