ConfigUpdatedEvent every second during graceful failover

java
connections

#1

I am noticing during graceful failover of a node in my cluster, my client gets a ConfigUpdatedEvent with the same exact message every 1 second until the rebalance is ready to be run.

Couchbase Version: 4.1.0-5005 Enterprise Edition (build-5005) on CentOS release 6.7
couchbase-core-io-1.2.9.jar
couchbase-java-client-2.2.8.jar
rxjava-1.0.17.jar

//hit “failover” -> “graceful” against “10.17.4.159”

2016.07.12 14:54:12:361 EDT | Info | CommonPersistency CouchbaseEvent: ConfigUpdatedEvent{bucketNames=[bwecl], clusterNodes=[/10.17.4.155, /10.17.4.159, /10.17.4.158]}

[…] repeat the previous every 1 second until…

2016.07.12 14:56:34:784 EDT | Info       | CommonPersistency
	CouchbaseEvent: NodeDisconnectedEvent{host=10.17.4.159/10.17.4.159}

2016.07.12 14:56:34:785 EDT | Info       | CommonPersistency

	CouchbaseEvent: ConfigUpdatedEvent{bucketNames=[bwecl], clusterNodes=[/10.17.4.155, 10.17.4.158/10.17.4.158]}

#2

Currently we don’t do any sophisticated filtering here and send this message every time we actually update it internally, even if its not part of the properties of this event itself (the actual internal config of course has many more properties) - so I’d say this is nothing unexpected.

What are you doing downstream with that message?


#3

OK, I see. It is not a problem but did not know if it was expected. Since we’re looking at integrating this with our system monitoring, downstream we will check the cluster nodes in the message and if the list is different from what we have cached against the cluster, we update the cached list and lower any alarms for any disconnected nodes that are now gone, like I suppose could happen during a failover-rebalance.