Comment 3 for bug 1296012

Revision history for this message
Nick Lock (nick-lock) wrote :

This is related to the bug in kernel 3.11 described here:

https://lkml.org/lkml/2013/9/7/126

As the contents of /sys/devices/system/cpu/cpu1/cpufreq/stats/time_in_state are incorrect (and never update), firstly there is a strong chance that the plugin will return values below $MINHZ - resulting in NaN being recorded.

Setting $MINHZ to 0 manually in the plugin code then returns 0 values on the graph, as the type is set as DERIVE... as the values returned from time_in_state never change, the derive is always zero. ;)

A kernel upgrade will probably fix the problem - the above page says that the patch was submitted for 3.12-rc1. You could either wait until 3.12 is released for Saucy, or upgrade to Tahr which has 3.13(?)