You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@storm.apache.org by GitBox <gi...@apache.org> on 2020/04/08 18:54:38 UTC

[GitHub] [storm] agresch commented on a change in pull request #3222: STORM-3596 use send assignment status in blacklist scheduling

agresch commented on a change in pull request #3222: STORM-3596 use send assignment status in blacklist scheduling
URL: https://github.com/apache/storm/pull/3222#discussion_r405743357
 
 

 ##########
 File path: storm-server/src/main/java/org/apache/storm/scheduler/blacklist/BlacklistScheduler.java
 ##########
 @@ -129,7 +142,34 @@ public void schedule(Topologies topologies, Cluster cluster) {
         return underlyingScheduler.config();
     }
 
-    private void badSupervisors(Map<String, SupervisorDetails> supervisors) {
+    private Map<String, Integer> getNodeAssignmentFailures() {
 
 Review comment:
   My intent on the rotating map was to drop supervisors over time that have potentially left the cluster.
   
   In my mind, if sending assignments fails every time you talk to a supervisor regardless of how far back, it means you have a suspicious supervisor.  But I'm fine with tracking the exact times of the events to see if they fall within a tolerance window if you prefer.
   
   Instead of tracking a count, I could track a list of failure times and drop the old failures if they're before the tolerance time limit?  Would that do what you desire?
   
   
   
   

----------------------------------------------------------------
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


With regards,
Apache Git Services