I am trying to figure out how sysstat is reporting the utilization of the device mapper logical volume, lower than the individual disks that make up the logical volume. I keep trying to do the math but I cant make sense of it.
This is the description of that column in the man page:
%util
Percentage of elapsed time during which I/O requests were issued to the device
(bandwidth utilization for the device). Device saturation occurs when this
value is close to 100%.
I have a logical volume (dm-2) which contains xvdb, xvdc, xvdd.
if the %util is the percentage of time during I/O requests were issued, id expect that the combination of the disks that make up dm-2 would be equivalent to the value reported by dm-2.
If we add it up:
xvdb - 32.36
xvdc - 51.31
xvdd - 48.30
the value is 131.97
but dm-2 is reporting a util of 91.02.
It does not make sense to me how the logical volume in LVM can report a different utilization than the physical volumes/disks that make it.
When I add up wr_sec/s, or even tps, the results are much closer to what I would expect. but the Util I am stumped on the difference.
12:00:01 AM DEV tps rd_sec/s wr_sec/s avgrq-sz avgqu-sz await svctm %util
...
07:56:01 AM xvda 3.01 0.00 26.40 8.76 0.01 1.85 0.58 0.17
07:56:01 AM xvdb 329.24 0.00 4680.19 14.22 3.71 11.26 0.98 32.36
07:56:01 AM xvdc 370.90 0.00 5396.45 14.55 7.92 21.39 1.38 51.31
07:56:01 AM xvdd 329.44 0.14 4392.93 13.33 7.41 22.49 1.47 48.30
07:56:01 AM dm-0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
07:56:01 AM dm-1 0.20 0.00 1.62 8.00 0.00 9.83 2.00 0.04
07:56:01 AM dm-2 953.43 0.14 14461.30 15.17 19.46 20.43 0.95 91.02
07:56:01 AM dm-3 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
07:56:01 AM dm-4 1.10 0.00 8.93 8.12 0.00 2.57 1.05 0.12
07:56:01 AM dm-5 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
07:56:01 AM dm-6 1.81 0.00 15.84 8.75 0.00 0.54 0.10 0.02
07:56:01 AM dm-7 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Anyone have the explanation of this difference?
This is not just one by the way, there are MANY examples of this.