RAID iostat showing 0.00 for wait's and %util

Question

I have a EC2 server running ubuntu 11.10 with 6 raid devices, 5 with 8 EBS drives each as part of a RAID 10 setup, the other one with 4 EBS drives as part of a RAID 0 setup.

I am finding that even though each of the individual EBS drive devices are showing correct iostat's that the md devices are showing 0.00 for avgqu-sz, await, r_await, w_await, svctm and %util. The other stats for the md devices (rrqm/s, wrqm/s, r/s, w/s, rkB/s, wkB/s, avgrq-sz) all seem to be correct.

Any ideas on how I might be able to get stats for the missing columns?

Answer 1

The zeros are actually the correct result (in a way) considering the values of the virtual layer itself. Hopefully, your virtual layers are fast enough that they are always zero. For example, avgqu-sz is number of requests in queue, but it passes the request to the disk almost immediately.