jiajunwang commented on a change in pull request #365: Fix RoutingTableProvider
statePropagationLatency metric reporting bug
URL: https://github.com/apache/helix/pull/365#discussion_r308554643
##########
File path:
helix-core/src/main/java/org/apache/helix/spectator/RoutingTableProvider.java
##########
@@ -558,23 +558,28 @@ protected void refresh(Map<String, Map<String,
Map<String, CurrentState>>> curre
private void resetRoutingTableAndNotify(long startTime, RoutingTable
newRoutingTable) {
_routingTableRef.set(newRoutingTable);
- logger.info("Refresh the RoutingTable for cluster {}, takes {} ms.",
- (_helixManager != null ? _helixManager.getClusterName() : null),
+ String clusterName = _helixManager != null ?
_helixManager.getClusterName() : null;
+ logger.info("Refresh the RoutingTable for cluster {}, takes {} ms.",
clusterName,
(System.currentTimeMillis() - startTime));
- notifyRoutingTableChange();
+
+ notifyRoutingTableChange(_helixManager != null ?
_helixManager.getClusterName() : null);
// Update timestamp for last refresh
if (_isPeriodicRefreshEnabled) {
_lastRefreshTimestamp = System.currentTimeMillis();
}
}
- private void notifyRoutingTableChange() {
- for (Map.Entry<RoutingTableChangeListener, ListenerContext> entry :
_routingTableChangeListenerMap
- .entrySet()) {
- entry.getKey().onRoutingTableChange(new
RoutingTableSnapshot(_routingTableRef.get()),
- entry.getValue().getContext());
+ private void notifyRoutingTableChange(String clusterName) {
+ // This call back is called in the main event queue of
RoutingTableProvider. We add log to record time spent
+ // here. Potentially, we should call this callback in a separate thread if
this is a bottleneck.
+ long startTime = System.currentTimeMillis();
+ for (Map.Entry<RoutingTableChangeListener, ListenerContext> entry :
_routingTableChangeListenerMap.entrySet()) {
+ entry.getKey()
+ .onRoutingTableChange(new
RoutingTableSnapshot(_routingTableRef.get()), entry.getValue().getContext());
Review comment:
Might not be valid, but the constructing of RoutingTableSnapshot might take
some time. Shall we exclude this outside the startTime - endTime?
I have no strong preference here.
----------------------------------------------------------------
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:
[email protected]
With regards,
Apache Git Services
