Using JMX to monitor Cassandra's disk related failures

Question

Cassandra (2.0.7) has two parameters to configure disk failures policies

• My configuration of /etc/cassandra/cassandra.yaml
  • commit_failure_policy = stop_commit
  • disk_failure_policy=best_effort

What are the JMX attributes that designate commit_log / disks failures?

(My goal is to monitor situations when the policies were triggered.)

Answer 1

我不确定是否还有其他有用的MBean，但是在发生文件系统异常时，best_effort策略会设置org.apache.cassandra.db.BlackListedDirectories。{UnvisibleDirectories，UnwritableDirectories}。