Re: [PR] Stop tracking last compactor check-in for non-existent groups [accumulo]

["ddanielr (via GitHub)" <gi...@apache.org>]

ddanielr commented on code in PR #4403:
URL: https://github.com/apache/accumulo/pull/4403#discussion_r1532500957


##########
server/manager/src/main/java/org/apache/accumulo/manager/compaction/coordinator/CompactionCoordinator.java:
##########
@@ -251,6 +250,18 @@ protected void startIdleCompactionWatcher() {
     ThreadPools.watchNonCriticalScheduledTask(future);
   }
 
+  protected void startGroupCheckThread() {
+    ScheduledFuture<?> future = schedExecutor.scheduleWithFixedDelay(() -> {
+      Set<String> compactorGroups = ExternalCompactionUtil.getCompactorAddrs(ctx).keySet();

Review Comment:
   I looked at `ExternalCompactionUtil.getCompactorAddrs()` and I'm not clear on when the queue is removed from zookeeper based on this comment: 
   https://github.com/apache/accumulo/blob/f3d5fb01d701a6932d37d2f67f52cc0eefa64d50/core/src/main/java/org/apache/accumulo/core/util/compaction/ExternalCompactionUtil.java#L121-L122
   
   Looking at the code, if there are no live compactors for a given queue, the queue name is still added to the returned map with an empty list. 
   https://github.com/apache/accumulo/blob/f3d5fb01d701a6932d37d2f67f52cc0eefa64d50/core/src/main/java/org/apache/accumulo/core/util/compaction/ExternalCompactionUtil.java#L117
   
   If ZK does clean up the queue names quickly then this isn't an issue. but if not, the coordinator could also remove the group if the group's list of compactors is empty.



-- 
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.

To unsubscribe, e-mail: notifications-unsubscribe@accumulo.apache.org

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