[GitHub] [storm] agresch commented on a change in pull request #3378: [STORM-3743] Add new topologies and change TestLargeCluster

[GitBox <gi...@apache.org>]

agresch commented on a change in pull request #3378:
URL: https://github.com/apache/storm/pull/3378#discussion_r578623531



##########
File path: storm-server/src/test/java/org/apache/storm/scheduler/resource/strategies/scheduling/TestLargeCluster.java
##########
@@ -312,57 +320,87 @@ private static void createAndAddOneSupervisor(
     }
 
     /**
-     * Create supervisors.
+     * Create supervisors for a larger cluster configuration.
      *
-     * @param uniformSupervisors true if all supervisors are of the same size, false otherwise.
-     * @return supervisor details indexed by id
+     * @param reducedSupervisorsPerRack number of supervisors to reduce in rack.
+     * @return created supervisors.
      */
-    private static Map<String, SupervisorDetails> createSupervisors(boolean uniformSupervisors) {
-        Map<String, SupervisorDetails> retVal;
-        if (uniformSupervisors) {
-            int numRacks = 16;
-            int numSupersPerRack = 82;
-            int numPorts = 50; // note: scheduling is slower when components with large cpu/mem leave large percent of workerslots unused
-            int rackStart = 0;
-            int superInRackStart = 1;
-            double cpu = 7200; // %percent
-            double mem = 356_000; // MB
-            Map<String, Double> miscResources = new HashMap<>();
-            miscResources.put("network.resource.units", 100.0);
-
-            return TestUtilsForResourceAwareScheduler.genSupervisorsWithRacks(
-                    numRacks, numSupersPerRack, numPorts, rackStart, superInRackStart, cpu, mem, miscResources);
-
+    private static Map<String, SupervisorDetails> createSupervisors(int reducedSupervisorsPerRack) {
+        if (TEST_CLUSTER_NAME.equals(TEST_CLUSTER_02)) {
+            return createSupervisorsForCluster02(reducedSupervisorsPerRack);
         } else {
-            // this non-uniform supervisor distribution closely (but not exactly) mimics a large cluster in use
-            int numSupersPerRack = 82;
-            int numPorts = 50;
-
-            Map<String, SupervisorDetails> retList = new HashMap<>();
+            return createSupervisorsForCluster01(reducedSupervisorsPerRack);
+        }
+    }
 
-            for (int rack = 0 ; rack < 12 ; rack++) {
-                double cpu = 3600; // %percent
-                double mem = 178_000; // MB
-                for (int superInRack = 0; superInRack < numSupersPerRack ; superInRack++) {
-                    createAndAddOneSupervisor(rack, superInRack, cpu - 100 * (superInRack % 2), mem, numPorts, retList);
-                }
+    /**

Review comment:
       Any documentation to allow other naive users like myself grab and anonymize these files and setup a test cluster would be very helpful.  Could be a follow on JIRA.




----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org