Monitoring Input/Output Performance

To determine where input/output bottlenecks exists on a system the iostat command can be used to display the amount of data moving between the system and a disk or terminal device, on all but the IRIX operating system. The command osview is used in IRIX to duplicate this funtion. It is important to spread the data accesses evenly among the disks. In the example below the load on drive sd1 is 3 to 4 times that of the other disk devices.

#iostat 5
      tty          fd0           sd1           sd2           sd3          cpu
 tin tout Kps tps serv  Kps tps serv  Kps tps serv  Kps tps serv  us sy wt id
   0    5   0   0    0  744  22   76    9   1   54    5   1   75  29 31  4 36
   0   16   0   0    0  352  16   42    0   0    0    0   0    0  40 54  5  0
   0   16   0   0    0  1827  42   91    2   0   24   27   4  107  40 21  3 37
   0   16   0   0    0  1883  45   86    0   0    0    0   0    0  36 19  5 40

Also, the command iostat -D is used to display the read and writes per second for each disk and the disk utilization.

#iostat -D 5
          fd0           sd1           sd2           sd3 
 rps wps util  rps wps util  rps wps util  rps wps util 
   0   0  0.0    8  14 34.7    1   0  1.5    0   0  1.2 
   0   0  0.0   19  11 46.1    1   2  4.8    1   7 11.1 
   0  0  0.0    0   3  3.0    2   0  1.9    0   0  0.0

The Solaris iostat command allows extend information to be displayed.

#iostat -x
                                 extended disk statistics 
disk      r/s  w/s   Kr/s   Kw/s wait actv  svc_t  %w  %b 
fd0       0.0  0.0    0.0    0.0  0.0  0.0    0.0   0   0 
sd1       7.8 14.5  223.9  399.7  0.0  2.0   90.7   0  35 
sd2       0.7  0.4   13.9    4.0  0.0  0.0   37.8   0   1 
sd3       0.4  0.5    2.5    3.8  0.0  0.1   58.1   0   1 
sd6       0.0  0.0    0.0    0.0  0.0  0.0   65.8   0   0 
sd8       0.3  0.2    9.4    9.6  0.0  0.0   38.6   0   1 
sd9       0.7  1.3   12.4   21.3  0.0  0.0   15.2   0   3

The most interesting column is the svc_t. This is the average time, in milliseconds, the disk responds to service requests.

The HP-UX iostat does not follow these conventions. Review the sample output below.

$ iostat 5                                                              

  device    bps     sps    msps  

    c0t6d0      0     0.0     1.0  
    c0t0d0      0     0.0     1.0  
    c0t1d0      0     0.0     1.0

Where device is the disk drive name, bps is the number of kilobytes transferred per second, sps indicates the number of disk seeks per second, and msps is the average time, in milliseconds, for a disk seek.

Please review the man page for more information on iostat command options and display output for the following systems.

Terms used: disk drive.