Monitoring Memory Usage in Linux Machines

For any one running a linux desktop/server , the amount of free memory available on the system  is one of the critical parameters that deterimine the system response. We will look at  some  commands which give you information about memory usage 


 free
   The free command provides the quickest way to see how much memory is being used on  your system. It shows the total amount of RAM   and swap space, along
with the amount currently being used.

Here are some examples of the free command:

$ free
             total       used       free     shared    buffers     cached
Mem:       2051656    1333508     718148          0     157452     461856
-/+ buffers/cache:     714200    1337456
Swap:      4416664          0    4416664

  free -m List memory usage in megabytes

  $ free -m
             total       used       free     shared    buffers     cached
Mem:          2003       1302        701          0        153        451
-/+ buffers/cache:        697       1306
Swap:         4313          0       4313

 free -b List memory usage in blocks

 sunil@Jaunty:~$ free -b
             total       used       free     shared    buffers     cached
Mem:    2100895744 1366093824  734801920          0  161419264  473022464
-/+ buffers/cache:  731652096 1369243648
Swap:   4522663936          0 4522663936

 

  free -mt List memory usage with totals displayed (Swap + Mem)

sunil@Jaunty:~$ free -mt
             total       used       free     shared    buffers     cached
Mem:          2003       1302        700          0        154        451
-/+ buffers/cache:        697       1305
Swap:         4313          0       4313
Total:        6316       1302       5013

     To avoid wasting RAM and speed up applications, Linux uses as much otherwise unused RAM as possible for the disc cache. For that reason, the first line of output from free that often shows little free RAM can be misleading. Please  pay closer attention to the second line of output, which shows the amount of RAM actually available for applications. That amount is 1305 MB in the above  example which shows
-/+ buffers/cache:        697       1305

 top 

The top command provides a means of watching the currently running processes, along with their memory usage.

$ top

 top - 10:53:00 up 29 min,  2 users,  load average: 0.42, 0.46, 0.45
Tasks: 199 total,   3 running, 196 sleeping,   0 stopped,   0 zombie
Cpu(s):  5.9%us,  2.6%sy,  0.0%ni, 91.5%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
Mem:   2051656k total,  1335604k used,   716052k free,   158124k buffers
Swap:  4416664k total,        0k used,  4416664k free,   462016k cached

To exit top, press q. Like the output for free, top shows total of memory usage for
RAM  and swap space. However, because top is screen oriented and
provides ongoing monitoring, you can watch memory usage change every three seconds (by default). 

   With top running, press Shift+m and the running processes will be
displayed in memory-use order (so you can watch which processes are consuming themost memory). See the screen shot above.
    The most useful column to analyze a process’ memory usage is RES, which shows the process’ actual physical RAM usage, also known as resident size.
The MEM column is based on this resident size. You can look at the man page of top for more information.

 vmstat

 This command gives you virtual memory statistics.

sunil@Jaunty:~$ vmstat
procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu----
 r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa
 1  0      0 682412 158804 489856    0    0    69    11  203  679  6  2 88  3

You can  view  memory use over a given time period. The following out put gives you vmstat every 5 seconds.

sunil@Jaunty:~$ vmstat 5
procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu----
 r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa
 0  0      0 682856 159000 489892    0    0    67    11  204  680  6  2 89  3
 1  0      0 682832 159004 489888    0    0     1     1  792 1969  4  2 94  0
 0  0      0 682684 159012 489892    0    0     0     2  752 1249  4  2 94  0
 0  0      0 682792 159012 489892    0    0     0     0  836 1839  4  2 94  0

The out put is printed every 5 seconds. To exit vmstat, press Ctrl+c. T$ vmstat -S m Display output in 1000k megabytes. You can try loading several applications and watch how the vmstat output changes .
   The kernel itself, however, has its own memory cache to keep track of its resources, called the kernel slab. You can use the vmstat command to display kernel slab memory cache statistics (from /proc/slabinfo) as follows:
sunil@Jaunty:~$ vmstat  -m
Cache                       Num  Total   Size  Pages
RAWv6                        23     23    704     23
TCPv6                        96     96   1344     24
flow_cache                    0      0     80     51
kcopyd_job                    0      0    264     31
dm_uevent                     0      0   2464     13
kmalloc_dma-512              16     16    512     16
mqueue_inode_cache           28     28    576     28
fuse_request                 62     63    376     21
       --------  ---       --------
kmalloc-16                 3772   4096     16    256
kmalloc-8                  6093   6144      8    512
kmalloc-192                9739  10038    192     21
kmalloc-96                  971   1008     96     42
          --- ----- ----
...
The slab memory cache information shows each cache name, the number of objects active for that cache type, the total number of objects available for that cache type, the   size of the cache (in bytes), and the number of pages for each cache. 

slabtop
  You can also use slabtop , similar to top to display the kernel slab info.

  

The slabtop output updates every three seconds. By default, slab caches are sorted by the number of objects (first column) in each cache. By pressing c you can sort by cache size instead.