1,403 bytes added
, 14:54, 27 June 2016
This is an article in the [[Performance tuning]] series.
You usually get asked to figure out why load is high on a system that lacks any sort of systematic baseline graphs and statistics like you would have with a professionally managed environment. If you're looking to install performance monitoring systems proactively, we'll cover that separately in articles like [[wp:Munin (software)]]. In Linux Journal, Kyle Rankin introduces some methods for diagnosing [https://www.linuxjournal.com/magazine/hack-and-linux-troubleshooting-part-i-high-load High Load] problems on your linux server. He goes through typical usage of commands like <code>top</code> and <code>iotop</code>.
{{highlight|text=Performance issues should be addressed on a system wide basis in addition to the specific requirements of each application or service.}}
== General ==
Some of the commands you will find useful for high load issues
<source lang="bash">
# look at any hardware or unusual issues first
dmesg
# look at the partitions you have
mount
# see total stats for the machine
uptime
# look at processes sorted by virtual size
ps awwlx --sort=vsz
# get three cycles of info from iostat
iostat -d -x 5 3
# show one iteration of iotop
iotop -bto --iter=1
</source>
== Applications ==
Troubleshooting High Load issues by configuring specific services or applications
See
* [[PHP Accelerator]]
* [[CiviCRM/debugging]]