I have a server with a high average load when it's not doing much work see below:

I added an additional processor yesterday but it has not made any difference to the load averages.

I am using fail2ban and wonder if that would cause this problem?

free -h shows:

total used free shared buffers cached
Mem: 1.1G 964M 150M 14M 147M 477M
-/+ buffers/cache: 339M 776M
Swap: 2.0G 6.3M 2.0G

Using ps awwlx --sort=vsz. This shows processes sorted by virtual sizes I have included the bigest ones below:

1 1000 1178 1 20 0 639220 53452 futex_ Sl ? 2:25 ./insync-headless start
4 104 1058 1 20 0 878608 77380 poll_s Ssl ? 1:11 /usr/sbin/mysqld
5 0 1354 1 20 0 1167388 14868 poll_s Sl ? 1:01 /usr/bin/python /usr/bin/fail2ban-server -b -s /var/run/fail2ban/fail2ban.sock -p /var/run/fail2ban/fail2ban.pid

Operating system Ubuntu Linux 14.04.1

Kernel and CPU Linux 3.13.0-76-generic on x86_64

Processor information Intel(R) Xeon(R) CPU E5-2650 v2 @ 2.60GHz, 2 cores

Running processes 244

CPU load averages 2.00 (1 min) 2.02 (5 mins) 2.05 (15 mins)

CPU usage 0% user, 0% kernel, 0% IO, 100% idle Real memory 385.79 MB used, 1.09 GB total

Virtual memory 12.76 MB used, 2 GB total

Local disk space 8.41 GB used, 49.08 GB total

iostat: Linux 3.13.0-76-generic (server) 30/03/17 _x86_64_ (2 CPU)

avg-cpu: %user %nice %system %iowait %steal %idle 0.88 0.02 0.80 0.21 0.00 98.08

Device: tps kB_read/s kB_wrtn/s kB_read kB_wrtn sda 5.33 6.54 57.48 1020105 8969285 sdb 0.01 0.02 0.10 3812 16060 dm-0 6.90 6.52 57.48 1017981 8969264



  • Without a full list of processes, it's hard to say, but with your CPUs 100% idle it's obviously not that your CPUs are busy. The load average is a (moving average of) the number of processes that are ready to run. This often means needing CPU time, but can also mean they're waiting for I/O to complete for example (like reading from disk). Mar 30, 2017 at 17:25
  • I have added iostat results to my question - which may help
    – RentUrApp
    Mar 30, 2017 at 17:58
  • @andrew-mccarthy process list
    – RentUrApp
    Mar 30, 2017 at 19:10

1 Answer 1


If there is potential bruteforce on ssh, apache etc, fail2ban will become aggresive and consume lots of cpu time.

The easiest way to see, which one is causing high load is by top command.

To me, this load seems perfectly ok due amount of cores.

I would chech IO of HDDs to check does they are bottleneck.

Check the wa column in top as well. If it is high, then you have a problem.

  • See my updated question which now includes iostat results. Should I add some more cores to the server?
    – RentUrApp
    Mar 30, 2017 at 17:59
  • @RentUrApp Like Andrew McCarthy mentioned, your CPU's are totally idle. I am not sure what is actually running on the server, but it seems fine. You'll need to make a conclusion after some time of monitoring do you need hardware upgrade or not, or resolve potential bottleneck.You need to see what is expected load for the server in the normal circumvents, and then compare that result with monitoring tools.
    – fugitive
    Mar 30, 2017 at 19:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.