Quantcast
Channel: free – Linuxaria
Viewing all articles
Browse latest Browse all 7

Linux Terminal: Check who uses all your memory with smem

$
0
0

As system administrator, or simple user that uses Linux on its desktop sometime you notice that something it’s eating all the memory of your system.
As first thing be sure to understand how Linux manage memory, I’ve be called too many time by scared users that did a free and were unable to read its output properly, in short, don’t worry if the Linux Kernel it’s using your memory to cache file.

< rant on >

To my “beloved” users:

Be assured that the Kernel developers can do a better job than you (and me) in find a good algorithm to cache file and free that memory area when a process need it, so please don’t ask me to put in cron some job that run something like that :

echo 3 > /proc/sys/vm/drop_caches

After that you’ll have more free memory available on the system, true, but the system will have to re-read all the files from the disk, so in terms of performance this is usually a bad move.

< /rant >

But now let’s take a look at a nice small program that can help us in find which process/users are using, for real, the memory of our systems.




smem is a tool that can give numerous reports on memory usage on Linux systems. Unlike existing tools, smem can report proportional set size (PSS), which is a more meaningful representation of the amount of memory used by libraries and applications in a virtual memory system.

Because large portions of physical memory are typically shared among multiple applications, the standard measure of memory usage known as resident set size (RSS) will significantly overestimate memory usage. The parameter PSS instead measures each application’s “fair share” of each shared area to give a realistic measure.

smem has many features:

  • system overview listing
  • listings by process, mapping, user
  • filtering by process, mapping, or user
  • configurable columns from multiple data sources
  • configurable output units and percentages
  • configurable headers and totals
  • reading live data from /proc
  • reading data snapshots from directory mirrors or compressed tarballs
  • lightweight capture tool for embedded systems
  • built-in chart generation

Installation

If you use Debian, Ubuntu or Mint you can install smem directly from the official repository with the command:

$ sudo apt-get install smem

Arch Linux users can install the package from the AUR repository.

Red Hat / Fedora / CentOS / SuSE users can use the pre-compiled binary that you can download directly from the official website, for your convenience you can copy and paste these commands that will put the command smem in your /usr/local/bin directory:

cd /tmp/
wget http://www.selenic.com/smem/download/smem-1.3.tar.gz
tar xvf smem-1.3.tar.gz
sudo cp /tmp/smem-1.3/smem /usr/local/bin/
sudo chmod +x /usr/local/bin/smem

Basic Usage

smem reports physical memory usage, taking shared memory pages into account. Unshared memory is reported as the USS (Unique Set Size). Shared memory is divided evenly among the processes sharing that memory. The unshared memory (USS) plus a process’s proportion of shared memory is reported as the PSS (Proportional Set Size). The USS and PSS only include physical memory usage. They do not include memory that has been swapped out to disk.

The basic use will show the running processes with the amount of memory used, to see that run the command smem without any option, this will print an output similar to this one:

# smem
  PID User     Command                         Swap      USS      PSS      RSS
....
 4223 linuxaria   /usr/lib/chromium-browser/c        0     9144    11696    29132 
14280 root     python ./smem                      0    11580    11839    13680 
 2180 linuxaria   /usr/bin/python /usr/share/        0    12956    14691    24392 
 2144 linuxaria   python /usr/lib/linuxmint/m        0    20252    22337    35648 
 2121 linuxaria   Thunar --daemon                    0    13316    23122    44404 
 2123 linuxaria   xfdesktop                          0    24004    24924    36032 
13095 linuxaria   /usr/lib/chromium-browser/c        0    27536    30996    53328 
 2577 linuxaria   /usr/bin/python /usr/lib/ub        0    34192    35112    41300 
 2127 linuxaria   /home/riccio/copy/x86_64/Co        0    40268    41081    52248 
10720 linuxaria   /usr/lib/chromium-browser/c        0    38684    42107    63928 
10561 linuxaria   /usr/lib/chromium-browser/c        0    43956    48137    71644 
10579 linuxaria   /usr/lib/chromium-browser/c        0    46276    50290    73484 
 5605 linuxaria   transmission-gtk /tmp/[kick        0    48616    51508    65148 
 8938 linuxaria   /usr/lib/firefox/plugin-con        0    48960    55428    71132 
 2186 linuxaria   /home/linuxaria/.dropbox-dist/        0    55524    56061    64820 
 1722 root     /usr/bin/X :0 -audit 0 -aut        0    52360    73110    96008 
10623 linuxaria   /usr/lib/chromium-browser/c        0   114224   121888   149236 
 4171 linuxaria   chromium-browser --disable-        0   131044   140415   165780 
 7050 linuxaria   /usr/lib/thunderbird/thunde        0   253428   256803   273152 
 2644 linuxaria   /usr/lib/firefox/firefox           0   528420   537423   558992

So in this example (my desktop) the process that is eating more RAM is Firefox and at the second place Thunderbird.

An useful option it’s the -u that shows the sum of memory used per user:

#smem -u
User     Count     Swap      USS      PSS      RSS 
daemon       1        0      196      197      368 
rtkit        1        0      300      312     1396 
nobody       1        0      428      442     1600 
mdm          2        0      600      695     1808 
avahi        2        0      568      806     2420 
syslog       1        0     1032     1043     1884 
messagebus     1        0     1124     1207     2048 
root        40        0    96568   123002   203372 
linuxaria      83        0  1539864  1640378  2222556

It’s possible to add a -p to this command (and also with all the others) to show percentages values:

#smem -u -p
User     Count     Swap      USS      PSS      RSS 
daemon       1    0.00%    0.00%    0.00%    0.00% 
rtkit        1    0.00%    0.00%    0.00%    0.02% 
nobody       1    0.00%    0.01%    0.01%    0.02% 
mdm          2    0.00%    0.01%    0.01%    0.02% 
avahi        2    0.00%    0.01%    0.01%    0.03% 
syslog       1    0.00%    0.01%    0.01%    0.02% 
messagebus     1    0.00%    0.01%    0.01%    0.03% 
root        40    0.00%    1.18%    1.47%    2.41% 
linuxaria      83    0.00%   19.29%   20.48%   27.58%

Or if you prefer a system wide memory usage view, you can use -w

#smem -w -p
Area                           Used      Cache   Noncache 
firmware/hardware             0.00%      0.00%      0.00% 
kernel image                  0.00%      0.00%      0.00% 
kernel dynamic memory        64.55%     62.42%      2.13% 
userspace memory             22.12%      3.00%     19.12% 
free memory                  13.33%     13.33%      0.00%

With the -w I suggest to use also the option -R REALMEM, this value it’s the amount of physical RAM. This lets smem detect the amount of memory used by firmware/hardware in the systemwide (-w) output. If provided, it will also be used as the total memory size to base percentages on.

#smem -R 8G -w -p
Area                           Used      Cache   Noncache 
firmware/hardware             2.83%      0.00%      2.83% 
kernel image                  0.00%      0.00%      0.00% 
kernel dynamic memory        63.02%     60.95%      2.07% 
userspace memory             21.60%      2.91%     18.68% 
free memory                  12.56%     12.56%      0.00%


Making graphs of the linux memory with smem

With smem (and having the matplotlib library) it’s possible to generate bar and pie chart with just some more options:

Show a bar graph with the columns “pss and uss”:

#smem --bar pid -c "pss uss"

smem

Creates a pie chart of the processes that start with ‘k’ on a running system:

#smem -P '^k' --pie=name

smem_pie

Conclusions

Smem it’s an easy to use tool that can give with ease some useful information when searching for memory hoggers, it’s also simple to get a graph of them, so you’ll be able to talk with your users and persuade them that after all their process are not so small and well done, memory wise.

Flattr this!


Viewing all articles
Browse latest Browse all 7

Trending Articles