Monitoring disk I/O on hadoop cluster with ganglia

Question

I deployed hadoop 2 (HDP2), and would like to get the disk i/o metrics of my slave nodes in ganglia. I didn't find any relevant metric so far.

What metric do you suggest to use or to add to ganglia ?

Thanks

Answer 1

cloudera-manager can help monitor the cluster status including Disk I/O

Answer 2

You can also use Ganglia Python Module Disk-stat provided here

Answer 3

As stated by @vishnu: Ganglia Python Module Disk-stat provided here

You can also use gmetric scripts that you would run from cron from here

Finally, you can use collectl and export disk metrics to ganglia, see here