How to measure memory usage in Linux
Article by me, first published on Openlogic.com
Whether you are a system administrator or a developer, sometimes you need to consider the use of memory in GNU/Linux processes and programs. Memory is a critical resource, and limited memory plus processes that use a lot of RAM can cause a situation where the kernel goes out of memory (OOM). In this state Linux activates an OOM killer kernel process that attempts to recover the system by terminating one or more low-priority processes. Which processes the system kills is unpredictable, so though the OOM killer may keep the server from going down, it can cause problems in the delivery of services that should stay running.
In this article we’ll look at three utilities that report information about the memory used on a GNU/Linux system. Each has strengths and weaknesses, with accuracy being their Achilles’ heel. I’ll use CentOS 6.4 as my demo system, but these programs are available on any Linux distribution.
ps displays information about active processes, with a number of custom fields that you can decide to show or not. For the purposes of this article I’ll focus on how to display information about memory usage. ps shows the percentage of memory that is used by each process or task running on the system, so you can easily identify memory-hogging processes.
Running ps aux shows every process on the system. Typical output looks something like this:
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
root 1 0.0 0.0 19228 1488 ? Ss 18:59 0:01 /sbin/init
root 2 0.0 0.0 0 0 ? S 18:59 0:00 [kthreadd]
root 3 0.0 0.0 0 0 ? S 18:59 0:00 [migration/0]
root 742 0.0 0.0 0 0 ? S 19:00 0:00 [ext4-dio-unwrit]
root 776 0.0 0.0 0 0 ? S 19:00 0:00 [kauditd]
root 785 0.0 0.0 0 0 ? S 19:00 0:00 [flush-253:0]
root 939 0.0 0.0 27636 808
Read more at Linux Aria
Comments are closed.