Prune peers every 5 minutes instead of every cycle of the Controller thread.

This should reduce the amount of time the allKnownPeers lock is held.

src/main/java/org/qortal/controller/Controller.java

@@ -113,6 +113,7 @@ public class Controller extends Thread {
 	private long repositoryBackupTimestamp = startTime; // ms
 	private long repositoryMaintenanceTimestamp = startTime; // ms
 	private long repositoryCheckpointTimestamp = startTime; // ms
+	private long prunePeersTimestamp = startTime; // ms
 	private long ntpCheckTimestamp = startTime; // ms
 	private long deleteExpiredTimestamp = startTime + DELETE_EXPIRED_INTERVAL; // ms
 
@@ -552,6 +553,7 @@ public class Controller extends Thread {
 		final long repositoryBackupInterval = Settings.getInstance().getRepositoryBackupInterval();
 		final long repositoryCheckpointInterval = Settings.getInstance().getRepositoryCheckpointInterval();
 		long repositoryMaintenanceInterval = getRandomRepositoryMaintenanceInterval();
+		final long prunePeersInterval = 5 * 60 * 1000L; // Every 5 minutes
 
 		// Start executor service for trimming or pruning
 		PruneManager.getInstance().start();
@@ -649,10 +651,15 @@ public class Controller extends Thread {
 				}
 
 				// Prune stuck/slow/old peers
-				try {
-					Network.getInstance().prunePeers();
-				} catch (DataException e) {
-					LOGGER.warn(String.format("Repository issue when trying to prune peers: %s", e.getMessage()));
+				if (now >= prunePeersTimestamp + prunePeersInterval) {
+					prunePeersTimestamp = now + prunePeersInterval;
+
+					try {
+						LOGGER.debug("Pruning peers...");
+						Network.getInstance().prunePeers();
+					} catch (DataException e) {
+						LOGGER.warn(String.format("Repository issue when trying to prune peers: %s", e.getMessage()));
+					}
 				}
 
 				// Delete expired transactions