1

I'm running debian

iostat -x
Linux 2.6.32-5-amd64 (server56856.uk2net.com)   27/04/13        _x86_64_        (4 CPU)

avg-cpu:  %user   %nice %system %iowait  %steal   %idle
           4.59    0.01    1.08    2.46    0.00   91.85

Device:         rrqm/s   wrqm/s     r/s     w/s   rsec/s   wsec/s avgrq-sz avgqu-sz   await  svctm  %util
sdb               4.63    67.90    2.05   14.91   750.81   654.14    82.80     0.05    2.66   6.91  11.72
sda               4.90    67.90    1.79   14.91   750.10   654.14    84.07     0.06    3.34   7.18  12.00
md0               0.00     0.00    0.12    0.07     1.00     0.57     8.00     0.00    0.00   0.00   0.00
md1               0.00     0.00    2.22   81.44    87.81   650.97     8.83     0.00    0.00   0.00   0.00

Which seems nice and low.

But then:

uptime 15:28:07 up 224 days,  3:53,  1 user,  
load average: 2.46, 2.50, 2.37

I have four cores

Why are the load averages so high?

Juddling
  • 1,065
  • 1
  • 7
  • 11

1 Answers1

0

Your disk subsystem is the bottleneck.

The load average is high if it is higher than the number of CPU cores.

Mircea Vutcovici
  • 16,706
  • 4
  • 52
  • 80
  • I have four cores, so would these load averages not be considered high? – Juddling Apr 27 '13 at 14:47
  • 3
    Depends on what you mean by high. Load avg. is the avg # of running or waiting processes. Since it's below 4, there are not any processes waiting on the CPU (on average). Generally speaking a load average below the number of cores is not considered high. – xofer Apr 27 '13 at 14:50
  • 1
    Load is also more relevent the more interactive / realtime the role of the host is. Its not uncommon for systems performing batch processes to have a load > 50 and still be 'ok'. – Matthew Ife Apr 27 '13 at 17:58