This will only make failure detector calculations less accurate, so if you're not having nodes flapping UP/DOWN or being incorrectly marked as down, this shouldn't be a big problem. You'll probably want to tune cassandra.fd_max_interval_ms for your cluster environment but I'm not really familiar with these settings so tuning suggestions are welcome.
If you upgraded from a previous version, this started being logged on debug.log only on 2.2 so there`s a chance it has been like this for a while but was never noticed. See CASSANDRA-11388, CASSANDRA-7307 and CASSANDRA-4375 for more background information. See CASSANDRA-11388, CASSANDRA-7307 and CASSANDRA-4375 for more background information. 2016-03-21 14:08 GMT-03:00 Herbert Fischer <herbert.fisc...@crossengage.io>: > Hi, > > I'm using Cassandra 2.2.4 and I we are getting ~4k messages per hour, per > node. > > It doesn't look good to me. Is it normal? If not, any idea on what might > be wrong? > > Thanks > > Herbert >
