Server Fault is a question and answer site for system and network administrators. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Using Linux mdadm, I've setup a six disk raid10,f2 array. I'm using 7200RPM SATA drives. The biggest load I am worried about is recording four streams of live video (worst case of 20Mbits each, final file size ends up being 10GB in the worst case), while watching another previously-saved file (similar characteristics).

I've been seeing glitches in the recordings that appear to be dropped frames. I suspect my the RAID array isn't keeping up, but I'd like to know more about what specifically is going on. Is it seeking too much? Is there some random wait? Is there a specific device in the array that's taking longer than the others, etc.

What sort of tools should I be using, what information should I be gathering, and how do I interpret it? Ideally, I'd like to have something gather data for several days then I could go back and look at it, but if that's not possible more live-snapshot type tools will work too.

share|improve this question
up vote 1 down vote accepted

For recording use sar - in many distributions part of the sysstat package.

Another tool, that is part of sysstat is iostat.

Let`s assume you got a raid1 (md0) composed of sdd and sde. To watch the disk-io:

iostat -d -m 1 md0 sdd sde.

This is how I did watch a multipathed SAN software raid1 today to see if the traffic gets balanced across all active paths.

share|improve this answer
What doe the various stats mean? What do they tell me about what is going wrong? – kbyrd Oct 18 '12 at 21:56
iostat with the above options will show transactions per second and tranfer-rates in MB/s. That way you will see how traffic is distributed across the physical disks. sar can trace almost everything - including your recording process. – Nils Oct 19 '12 at 19:53
Thanks this is a good start. I'd still like to know how to interpret things like iowait times and queue sizes, but this will get me started. After a few days of sar recording and a quick look, I suspect I have a couple of drives in the array writing much slower than the others. – kbyrd Oct 20 '12 at 16:15
@kbyrd another thing that might go wrong is that the data does not get streamed to the disks, but gets synced (blocking your system since you have got many outstanding writes). Try a sync-mount and have a look again. – Nils Oct 20 '12 at 20:58

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.