[Ganglia-developers] diskstat bandwidth calculation

[Christopher H. <chr...@un...>]

Dear Ganglia developers

first, thanks for your work on this great project!

I currently stuck in an issue with monitoring disk access. I'm using the 
diskstat python module and tried to get out how much bytes actually were 
read/written from/to the disk. I noticed the calculation for 
bytes_per_second in the code [1] what leads to something like

"reads_merged_delta * 512_Bytes_per_sector / 15s_delta_time"

My question: is the value in reads_merged spanning over only one sector 
or could one read be from multiple sectors?
If multiple sectors were read with for 1 read_merged, should the # of 
read sectors also considered in this calculation?

I'm not sure a) if I understand the /proc/diskstat info wrong or b) if 
there is a bug in the diskstat module.

Can you help me to figure that out?

Thank you all!

Best
Christopher

[1] 
https://github.com/ganglia/monitor-core/blob/master/gmond/python_modules/disk/diskstat.py#L304