본문 : http://www.ischo.net -- 조인상 //시스템 엔지니어

+++++++++++++++++++++++++++++++++++++++++++++++++++++++




Input Output statistics (  iostat )

iostat  reports terminal and disk  I/O  activity and  CPU utilization.  The first line of output is for the  time period  since boot  &  each subsequent line is for  the  prior  interval . Kernel maintains  a number of counters to keep track of  the  values.

iostat's activity class options default  to  tdc  (terminal,  disk, and CPU). If any other option/s are specified,  this  default is completely overridden i.e.  iostat -d will report only statistics about the disks.



syntax:

Basic synctax is iostat  <options>  interval  count

option - let you specify the device for which information is needed like disk , cpu or terminal. (-d , -c , -t  or -tdc ) .  x options gives the extended statistics .

interval -  is time period in seconds between two samples . iostat  4  will give data at each 4 seconds interval.

count  - is the  number of times the data is needed .  iostat 4 5  will give data at 4 seconds interval  5 times


 









Example

$ iostat -xtc 5 2
                  extended disk statistics    tty      cpu
    disk r/s  w/s Kr/s Kw/s wait actv svc_t  %w  %b  tin tout us sy wt id
    sd0  2.6 3.0 20.7 22.7 0.1  0.2  59.2  6  19  0  84  3  85 11 0
    sd1  4.2 1.0 33.5  8.0 0.0  0.2  47.2  2  23
    sd2  0.0 0.0  0.0  0.0 0.0  0.0  0.0  0  0
    sd3  10.2 1.6 51.4 12.8 0.1  0.3  31.2  3  31


    The fields have the following meanings:

    disk  name of the disk
    r/s    reads per second
    w/s    writes per second
    Kr/s  kilobytes read per second
    Kw/s  kilobytes written per second
    wait  average number of transactions waiting for service (Q length)
    actv  average number of transactions  actively 
          being serviced (removed  from  the
      queue but not yet
          completed)
    %w    percent of time there are transactions  waiting
          for service (queue non-empty)
    %b    percent of time the disk is busy  (transactions
            in progress)




Results and Solutions:

The values to look from the iostat output  are:

Reads/writes  per second (r/s , w/s)
Percentage busy (%b)
Service time (svc_t)
If a disk shows consistently high reads/writes along with , the percentage busy (%b) of the disks is greater than 5 percent, and the average service time  (svc_t) is greater than 30 milliseconds, then  one of the following action needs to be taken

1.)Tune the application to use disk i/o more efficiently  by modifying the disk queries and using available cache facilities of application servers .

2.) Spread the file system of the disk on to two or more disk  using disk striping feature of volume manager /disksuite  etc.

3.) Increase the system parameter values for  inode cache  , ufs_ninode ,  which is  Number of inodes to be held in memory. Inodes are cached globally (for UFS), not on a per-file system basis 

4.) Move the file system to another faster disk /controller  or replace existing disk/controller to a faster one.