What percentage is considered as a good IOWait?

David Beroff

Well-known member
Registered
Joined
Jun 14, 2016
Messages
1,477
Points
63
As much as I have read about iowait, it is still confusing to me. What percentage of iowait is considered is good or high?
 

VirtuBox

Well-known member
Registered
Joined
May 3, 2016
Messages
1,622
Points
83
As much as I have read about iowait, it is still confusing to me. What percentage of iowait is considered is good or high?
IOWait is time that the processor/processors (cores) are waiting to process/compute a task. If CPU cores are waiting, it's because there is already several disk I/O requests and the disk is not able to handle more requests. So if you have a high IOWait percentage, that mean your CPU can handle more computational work but your not your disk.
If it is constantly high, your disks are a bottleneck and they slow your system down, but it depend on your CPU cores and performance against your disk IOPS.

For example, to process the same program, IOWait percentage will be higher if you have 12 Cores than if you have 6 cores, and it will be lower if your have SSD disks, than if you have classic hard drives.
 

Denis // trabia

Member
Hosting Provider
Registered
Joined
May 29, 2017
Messages
50
Points
6
You can use vmstat command to see the disk performance and not only.
For example running
Code:
vmstat 1 20
will run the test 20 times with a 1 second interval and give you the output.
Code:
$ vmstat 1 20 
procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu----
 r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa
 0  0   3996 168488  57100 1368636    0    0     1     2    0    0  3  1 97  0
 0  0   3996 168604  57100 1368728    0    0     0     0  144  303  1  0 100  0
 0  0   3996 168604  57100 1368984    0    0   256     0  162  464  1  0 99  0
 0  0   3996 168604  57100 1368972    0    0     0     0  239  638  3  0 97  0
 0  0   3996 168604  57100 1368952    0    0     0     0  242  529  1  1 99  0
 0  0   3996 168604  57100 1368952    0    0     0     0  148  430  1  0 99  0
 1  0   3996 168604  57100 1368952    0    0     0     0  222  451  6  0 94  0
 0  0   3996 168604  57100 1368952    0    0     0     0  141  270  2  0 98  0
 0  0   3996 168604  57100 1368952    0    0     0     0  166  450  1  0 99  0
 0  0   3996 168604  57100 1368952    0    0     0     0  133  410  0  0 99  0
 0  0   3996 168604  57100 1368952    0    0     0     0  196  398  0  0 99  0
 0  0   3996 168604  57100 1368952    0    0     0     0  187  510  1  0 100  0
 0  0   3996 168604  57108 1368952    0    0     0    16  263  677  2  0 97  1
 0  0   3996 168604  57108 1368952    0    0     0     0  205  431  1  0 98  0
 0  0   3996 168604  57108 1368964    0    0     0     0  179  467  1  0 98  0
 0  0   3996 168604  57108 1368964    0    0     0     0  169  446  2  0 98  0
 0  0   3996 168604  57108 1368964    0    0     0     0  202  365  2  0 98  0
 0  0   3996 168604  57108 1369208    0    0   256     0  226  458  2  1 96  0
 0  0   3996 168604  57108 1369208    0    0     0    40  202  501  2  0 98  0
 0  0   3996 168604  57108 1369220    0    0     0     0  154  295  2  0 98  0
What you are interested here is the last column (wa) which is the amount of time that the processor spends waiting for IO operations to complete before being able to continue processing tasks. Ideally it should be 0.
More you can read here: https://www.thomas-krenn.com/en/wiki/Linux_Performance_Measurements_using_vmstat
 
Newer Threads
Replies
1
Views
2,052
Replies
9
Views
2,132
Replies
7
Views
4,384
Replies
28
Views
8,545

Latest Hosting OffersNew Reviews

Sponsors

Tag Cloud

You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.

Top