Monitoring Disk Access/Usage

[JAW]

What's the best way to monitor the IO of a raid array? I want to see how much the drives are being accessed and if they are bottlenecking.

 

[DougLite]

Start\Run\perfmon.msc if you are on Windows 2000/XP is a good place to start.

 

[unhappy_mage]

"vmstat 3" on Linux reports on this kind of thing (and lots of others). Here's a sample line:

procs -----------memory---------- ---swap-- -----io---- --system-- ----cpu----
 r  b   swpd   free   buff  cache   si   so    bi    bo   in    cs us sy id wa
 2  0      0  11924  42572 114712    0    0     2    24    4    26 86  0 13  0
 2  0      0  11812  42572 114712    0    0     0    11 1012    65 100  0  0  0

The columns relevant to your question are under "io"; bi and bo mean blocks in and out, respectively. Blocks are defined to be one megabyte. So I read 2 blocks and wrote 24, then read 0 and wrote 11. These stats are taken 3 seconds apart, so the throughput was an average of 6MB/s or so writing to disk.

Determining if they're bottlenecking and what's causing it can be difficult to diagnose. Controller? Disks? OS?