In vSphere 5.1, one of the major storage enhancements that was part of the new I/O Device Management (IODM) framework was the addition of SMART (Self Monitoring, Analysis And Reporting Technology) data for monitoring FC, FCoE, iSCSI, SAS protocol statistics, this is especially useful for monitoring the health of an SSD device. The SMART data is provided through a SMART daemon which lives inside of ESXi and runs every 30 minutes to gather statistics and diagnostic information from the underlying storage devices and provides the information through the following ESXCLI command:
esxcli storage core device smart get -d [DEVICE]
If you would like to learn more about IODM and SMART, be sure to check out Cormac Hogan's in-depth article here.
The default polling interval for the SMART daemon in vSphere 5.1 was not configurable and 30 minutes was the system default. For most customers, the out of the box configuration should be sufficent. However, for some customers who wish to have greater flexibility in the polling frequency, the default can now be adjusted in vSphere 6.0. The smartd process now includes a new -i option which specifies the polling interval.
[root@mini:~] smartd -h
smartd: option requires an argument -- 'h'
smartd <options>
-i Polling interval (in minutes) for smartd
(default Polling interval is 30 minutes)
If you wish to change the default, you will need to modify the /etc/init.d/smartd init script and include the interval option. One issue that I have found is that changes to the init script do not persist reboots as modifications to these files should not be performed by users. In the case of adjusting the polling interval, we need to add the additional option for smartd startup.
We can still accomplish this by adding the following to /etc/rc.local.d/local.sh make the necessary adjustments and restarting the smartd process:
SMARTD_POLL_INTERVAL=35
/etc/init.d/smartd stop
sed -i "s/^SMARTD_SCHED_PARAM.*/SMARTD_SCHED_PARAM=\"-i ${SMARTD_POLL_INTERVAL} ++group=smartd\"/g" /etc/init.d/smartd
/etc/init.d/smartd start
Note: The -i option is only visible when smartd process is not running
If you wish to see the changes live immediately, then you can run /etc/rc.local.d/local.sh command once or this will automatically happen upon ESXi booting up. If we perform a process look up using "ps", we can see that our smartd is now configured to poll every 35 minutes instead of the default 30.
Not sure how esxcli is formatting output, however the power on hours field is clearly incorrect (overflow?) and pending relocated sectors isn't reported at all...