In order to calculate the volumes I/O metrics, a Machine Agent must use the disk file system command or df as well as the cat /proc/diskstats command in Linux. Then the Machine Agent matches the mount point and file system names of both of the command outputs.
Example:
If the output of df is:
Linux-VirtualBox:~$ df
Filesystem 1K-blocks Used Available Use% Mounted on
udev 368708 4 368704 1% /dev
tmpfs 75884 948 74936 2% /run
/dev/sda1 11479264 6694412 4178696 62% /
and the output of proc/diskstats is:
Linux-VirtualBox:~$ cat /proc/diskstats
8 0 sda 45542 1694 1800102 44732 173309 171383 3502920 37456 0 43884 82128
8 1 sda1 42589 483 1766554 43576 161862 141753 3175080 33336 0 41848 76856
As /dev/sda1 is mounted on / (root), the partition name will be dev-sda1 and the disk stats for sda1 is mapped to it.
So now on the Controller UI, the I/O metric for dev-sda1 will be displayed as the same I/O metric which displays in the SIM dashboard for / (root) volume.
In what circumstance will a Machine Agent not report the I/O metric for volumes?
If the volumes mounted on volume groups, the Machine Agent will not report the I/O metrics for volume.
Linux-VirtualBox:~$:/home/adm/cosg# df
Filesystem 1048576-blocks Used Available Capacity Mounted on
/dev/mapper/vg00-lv_root 1984 1624 258 87% /
/dev/mapper/vg00-lv_tmp 496 51 420 11% /tmp
/dev/mapper/vg00-lv_home 93 7 82 8% /home
