taklwu commented on a change in pull request #4181:
URL: https://github.com/apache/hbase/pull/4181#discussion_r823234036
##########
File path:
hbase-server/src/main/java/org/apache/hadoop/hbase/master/HMaster.java
##########
@@ -389,7 +391,7 @@
// the key is table name, the value is the number of compactions in that
table.
private Map<TableName, AtomicInteger> mobCompactionStates =
Maps.newConcurrentMap();
- MasterCoprocessorHost cpHost;
+ volatile MasterCoprocessorHost cpHost;
Review comment:
and even if we can measure the impact, this may be still required
because at the moment of the some getter is using the coprocessor for reading
or calculating their methods, this loaded CP(s) could be changed or removed.
I noticed `RackManager` in `BaseLoadBalancer`/`LoadBalancer` is also having
this penalty and if we don't change the configuration, we should not go into
the `onConfigurationChange` hooked called by `update_all_config`.
So, IMO the benefit is better than restarting the JVM .
--
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]
