如何解读iostat？

Question

I track a lot of parameters on my Server and the only thing I can't realy put in perspective is the IOstat. It is a MySQL Server, is this a good result, or should I worry?

root:/var/lib/mysql# iostat -xc
Linux 2.6.28-11-server ()     07/25/2009      _x86_64_        (8 CPU)

avg-cpu:  %user   %nice %system %iowait  %steal   %idle
           3.66    0.19    0.45    1.04    0.00   94.69

Device:         rrqm/s   wrqm/s     r/s     w/s   rsec/s   wsec/s avgrq-sz avgqu-sz   await  svctm  %util
sda               2.55   871.36    1.46   27.67   392.40  7200.45   260.64     1.02   34.85   2.48   7.22
sda1              0.18     0.61    0.03    0.01     3.60     4.98   215.91     0.01  185.95  19.25   0.08
sda2              0.01     0.00    0.00    0.00     1.03     0.02   919.32     0.00   21.36   6.94   0.00
sda3              2.36   870.75    1.43   27.66   387.76  7195.46   260.68     1.01   34.65   2.48   7.21
sdb               2.37   871.36    1.63   27.67   392.69  7200.45   259.12     0.65   22.07   2.51   7.35
sdb1              0.17     0.61    0.04    0.01     3.59     4.98   187.33     0.01  110.67  12.54   0.06
sdb2              0.00     0.00    0.00    0.00     1.03     0.02   256.48     0.00    2.36   1.50   0.00
sdb3              2.19   870.75    1.60   27.66   388.06  7195.46   259.23     0.64   21.93   2.51   7.34
md0               0.00     0.00    0.38    0.62     3.06     4.96     8.00     0.00    0.00   0.00   0.00
md1               0.00     0.00    0.00    0.00     0.00     0.02     8.36     0.00    0.00   0.00   0.00
md2               0.00     0.00    2.01  898.28    62.49  7186.28     8.05     0.00    0.00   0.00   0.00

Also what war options for decreasing read / write activity?

• delay_______key_______writes
• memory based Tables
• less indicies

The write load is quite high on the tables.

Answer 1

If anyone worrying about the disk IO bottlenecks, please have a check with the following command. iostat

If this tool is not installed then,

apt-get install sysstat

on Debian based servers.

yum install sysstat

on Redhat/CentOS based servers.

Then,

iostat -x -d sda -here "sda" denotes your HDD

Output:

root@forum.innovationframes.com:~# iostat -x -d sda
Linux 2.6.32-24-server (forum.innovationframes.com)    10/01/2011      _x86_64_        (1 CPU)

Dev:  rrqm/s  wrqm/s   r/s   w/s  rsec/s  wsec/s avgrq-sz avgqu-sz   await  svctm  **%util**

sda   0.01    0.04    0.06   0.03  1.34     0.51    21.77     0.00    5.23  0.30   **0.00**

Note: If Util shows more than 75-80% then you should keep an eye on your HDD.