Cassandra based distributed locking mechanism # 741

- Cassandra tables called lock and lock_priority,
 to manage task execution and synchronization across
 multiple nodes.

CHANGES.md

@@ -6,7 +6,6 @@
 * Create New Repair Type Called "VNODE" - Issue #755
 * Create ReplicaRepairGroup Class for Grouping Replicas and Token Ranges - Issue #721
 * Hot Reload of Nodes List - Issue #699
-* Update nodes when cluster changes, nodes removed or added  #699
 * Investigate Creation of RepairScheduler and ScheduleManager #714
 * Implement ScheduledJobQueue for Prioritized Job Management and Execution - Issue #740
 * Implement RepairGroup Class for Managing and Executing Repair Tasks - Issue #738

application/src/main/java/com/ericsson/bss/cassandra/ecchronos/application/providers/AgentNativeConnectionProvider.java

@@ -94,6 +94,7 @@ public AgentNativeConnectionProvider(
                         .withSslEngineFactory(sslEngineFactory)
                         .withSchemaChangeListener(defaultRepairConfigurationProvider)
                         .withNodeStateListener(defaultRepairConfigurationProvider);
+
         LOG.info("Preparing Agent Connection Config");
         nativeConnectionBuilder = resolveAgentProviderBuilder(nativeConnectionBuilder, agentConnectionConfig);
         LOG.info("Establishing Connection With Nodes");

application/src/main/resources/ecc.yml

@@ -293,12 +293,11 @@ rest_server:
       cache_expiry_time_in_seconds: 30
       ##
       ## Allow to override consistency level for LWT (lightweight transactions). Possible values are:
-      ## "DEFAULT" - Use consistency level based on remoteRouting.
       ## "SERIAL" - Use SERIAL consistency for LWT regardless of remoteRouting.
       ## "LOCAL" - Use LOCAL_SERIAL consistency for LWT regardless of remoteRouting.
       ##
-      ## if you use remoteRouting: false and LOCAL then all locks will be taken locally
-      ## in DC. I.e There's a risk that multiple nodes in different datacenters will be able to lock the
+      ## if you use LOCAL then all locks will be taken locally in DC.
+      ##  I.e There's a risk that multiple nodes in different datacenters will be able to lock the
       ## same nodes causing multiple repairs on the same range/node at the same time.
       ##
-      consistencySerial: "DEFAULT"
+      consistencySerial: "SERIAL"