bbeaudreault commented on a change in pull request #3660:
URL: https://github.com/apache/hbase/pull/3660#discussion_r717643028
##########
File path:
hbase-server/src/main/java/org/apache/hadoop/hbase/rsgroup/RSGroupInfoManagerImpl.java
##########
@@ -1192,6 +1192,10 @@ public BalanceResponse balanceRSGroup(String groupName,
BalanceRequest request)
return responseBuilder.build();
}
+ if (request.isReloadConfigs()) {
Review comment:
Thanks for the review @Apache9. I could do this, but I didn't do it
because currently we need to implement the clone and reload logic twice: here
and in HMaster.java. Would you prefer that I implement it twice?
##########
File path:
hbase-balancer/src/main/java/org/apache/hadoop/hbase/master/balancer/BaseLoadBalancer.java
##########
@@ -617,4 +617,14 @@ private void roundRobinAssignment(BalancerClusterState
cluster, List<RegionInfo>
public void onConfigurationChange(Configuration conf) {
loadConf(conf);
}
+
+ @Override
+ public void reloadConfiguration() {
+ LOG.info("Reloading balancer configs");
+ // We clone and reload the clone because the original conf may be used in
other
+ // non-balancer contexts. We only want to reload configs for the balancer.
+ Configuration conf = new Configuration(getConf());
Review comment:
The reason I did this is I felt like it was even more confusing to
developers if a Balancer-specific reloadConfigs option caused the entire
HMaster to reload all of its configs. I think the expectation for such a
feature would be that it only reloads the balancer-related configs.
One thing I could do is rename isReloadConfigs to isReloadMasterConfigs or
something. This would make it clearer that the operation will reload everything.
What do you think of the trade off?
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
