Linux Performance Measurements using vmstat

From Thomas-Krenn-Wiki
Jump to: navigation, search

vmstat (virtual memory statistics) is a valuable monitoring utility, which also provides information about block IO and CPU activity in addition to memory.

vmstat Basics

vmstat provides a number of values and will typically be called using two numerical parameters.

  • Example: vmstat 1 5
    • 1 -> the values will be re-measured and reported every second
    • 5 -> the values will be reported five times and then the program will stop

The first line of the report will contain the average values since the last time the computer was rebooted. All other lines in the report will represent their respective current values. Vmstat does not need any special user rights. It can run as a normal user.

[user@fedora9 ~]$ vmstat 1 5
procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu------
 r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa st
 3  0      0  44712 110052 623096    0    0    30    28  217  888 13  3 83  1  0
 0  0      0  44408 110052 623096    0    0     0     0   88 1446 31  4 65  0  0
 0  0      0  44524 110052 623096    0    0     0     0   84  872 11  2 87  0  0
 0  0      0  44516 110052 623096    0    0     0     0  149 1429 18  5 77  0  0
 0  0      0  44524 110052 623096    0    0     0     0   60  431 14  1 85  0  0
[user@fedora9 ~]$

Meaning of the individual Values

(Source man vmstat):

  • Procs
    • r: The number of processes waiting for run time.
    • b: The number of processes in uninterruptible sleep.
  • Memory
    • swpd: the amount of virtual memory used.
    • free: the amount of idle memory.
    • buff: the amount of memory used as buffers.
    • cache: the amount of memory used as cache.
    • inact: the amount of inactive memory. (-a option)
    • active: the amount of active memory. (-a option)
  • Swap
    • si: Amount of memory swapped in from disk (/s).
    • so: Amount of memory swapped to disk (/s).
  • IO
    • bi: Blocks received from a block device (blocks/s).
    • bo: Blocks sent to a block device (blocks/s).
  • System
    • in: The number of interrupts per second, including the clock.
    • cs: The number of context switches per second.
  • CPU
    • These are percentages of total CPU time.
    • us: Time spent running non-kernel code. (user time, including nice time)
    • sy: Time spent running kernel code. (system time)
    • id: Time spent idle. Prior to Linux 2.5.41, this includes IO-wait time.
    • wa: Time spent waiting for IO. Prior to Linux 2.5.41, included in idle.
    • st: Time stolen from a virtual machine. Prior to Linux 2.6.11, unknown.

Examples

CPU User Load Example

A standard audio file will be encode as an MP3 file by means of the lame encoder[1] in this example. This process is quite CPU intensive and also demonstrates the execution of vmstat in parallel with a user CPU time of 97%:

[user@fedora8 ~]$ vmstat 1 5
procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu------
 r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa st
 6  1      0 302380  16356 271852    0    0   561   568   80  590 43  7 43  7  0
 1  0      0 300892  16364 273256    0    0     0    52   79  274 97  3  0  0  0
 2  0      0 299544  16364 274532    0    0     0     0   78  372 97  3  0  0  0
 1  0      0 298292  16368 275780    0    0     0     0   53  255 97  3  0  0  0
 1  0      0 296820  16368 277192    0    0     0     0   77  377 97  3  0  0  0
[user@fedora8 ~]$

CPU System Load Example

In this example, a file will be filled with random content using dd.

[user@fedora9 ~]$ dd if=/dev/urandom of=500MBfile bs=1M count=500

For this, /dev/urandom[2] will supply random numbers, which will be generated by the kernel. This will lead to an increased load on the CPU (sy – system time). At the same time, the vmstat executing in parallel will indicate that between 93% and 97% of the CPU time is being used for the execution of kernel code (for the generation of random numbers, in this case).

[user@fedora9 ~]$ vmstat 1 5
procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu------
 r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa st
 2  0 402944  54000 161912 745324    5   14    54    59  221  867 13  3 82  2  0
 1  0 402944  53232 161916 748396    0    0     0     0   30  213  3 97  0  0  0
 1  0 402944  49752 161920 751452    0    0     0     0   28  290  4 96  0  0  0
 1  0 402944  45804 161924 755564    0    0     0     0   29  188  2 98  0  0  0
 1  0 402944  42568 161936 758608    0    0     0 17456  272  509  7 93  0  0  0
[user@fedora9 ~]$

The time for executing system calls[3][4][5] will be counted as system time (sy).

RAM Bottleneck (swapping) Example

In this example, many applications will be opened (including VirtualBox with a Windows guest system, among others). Almost all of the working memory will be used. Then, one more application (OpenOffice) will be started. The Linux kernel will then swap out several memory pages to the swap file on the hard disk, in order to get more RAM for OpenOffice. Swapping the memory pages to the swap file will be seen in the so (swap out - memory swapped to disk) column as vmstat executes in parallel.

[user@fedora8 ~]$ vmstat 1 5
procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu------
 r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa st
 3  1 244208  10312   1552  62636    4   23    98   249   44  304 28  3 68  1  0
 0  2 244920   6852   1844  67284    0  544  5248   544  236 1655  4  6  0 90  0
 1  2 256556   7468   1892  69356    0 3404  6048  3448  290 2604  5 12  0 83  0
 0  2 263832   8416   1952  71028    0 3788  2792  3788  140 2926 12 14  0 74  0
 0  3 274492   7704   1964  73064    0 4444  2812  5840  295 4201  8 22  0 69  0
[user@fedora8 ~]$

High IO Read Load Example

A large file (such as an ISO file) will be read and written to /dev/null using dd.

[user@fedora9 ~]$ dd if=bigfile.iso of=/dev/null bs=1M

Executed in parallel, vmstat will show the increased IO read load (the bi value).

[user@fedora9 ~]$ vmstat 1 5
procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu------
 r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa st
 3  1 465872  36132  82588 1018364    7   17    70   127  214  838 12  3 82  3  0
 0  1 465872  33796  82620 1021820    0    0 34592     0  357  781  6 10  0 84  0
 0  1 465872  36100  82656 1019660    0    0 34340     0  358  723  5  9  0 86  0
 0  1 465872  35744  82688 1020416    0    0 33312     0  345  892  8 11  0 81  0
 0  1 465872  35716  82572 1020948    0    0 34592     0  358  738  7  8  0 85  0
[user@fedora9 ~]$ 

High IO Write Load Example

In contrast with the previous example, dd will read from /dev/zero and write a file. The oflag=dsync will cause the data to be written immediately to the disk (and not merely stored in the page cache).

[user@fedora9 ~]$ dd if=/dev/zero of=500MBfile bs=1M count=500 oflag=dsync

Executed in parallel, vmstat will show the increased IO write load (the bo value).

[user@fedora9 ~]$ vmstat 1 5
procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu------
 r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa st
 1  1      0  35628  14700 1239164    0    0  1740   652  117  601 11  4 66 20  0
 0  1      0  34852  14896 1239788    0    0     0 23096  300  573  3 16  0 81  0
 0  1      0  32780  15080 1241304    0    0     4 21000  344  526  1 13  0 86  0
 0  1      0  36512  15244 1237256    0    0     0 19952  276  394  1 12  0 87  0
 0  1      0  35688  15412 1237180    0    0     0 18904  285  465  1 13  0 86  0
[user@fedora9 ~]$ 

CPU Waiting for IO Example

In the following example, an updatedb process is already running. The updatedb utility is part of mlocate. It examines the entire file system and accordingly creates the database for the locate command (by means of which file searches can be performed very quickly). Because updatedb reads all of the file names from the entire file system, the CPU must wait to get data from the IO system (the hard disk). For that reason, vmstat running in parallel will display large values for wa (waiting for IO):

[user@fedora9 ~]$ vmstat 1 5
procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu------
 r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa st
 2  1 403256 602848  17836 400356    5   15    50    50  207  861 13  3 83  1  0
 1  0 403256 601568  18892 400496    0    0  1048   364  337 1903  5  7  0 88  0
 0  1 403256 600816  19640 400568    0    0   748     0  259 1142  6  4  0 90  0
 0  1 403256 600300  20116 400800    0    0   476     0  196  630  8  5  0 87  0
 0  1 403256 599328  20792 400792    0    0   676     0  278 1401  7  5  0 88  0
[user@fedora9 ~]$

Additional vmstat Options

vmstat --help

[user@fedora9 ~]$ vmstat --help
usage: vmstat [-V] [-n] [delay [count]]
              -V prints version.
              -n causes the headers not to be reprinted regularly.
              -a print inactive/active page stats.
              -d prints disk statistics
              -D prints disk table
              -p prints disk partition statistics
              -s prints vm table
              -m prints slabinfo
              -S unit size
              delay is the delay between updates in seconds. 
              unit size k:1000 K:1024 m:1000000 M:1048576 (default is K)
              count is the number of updates.
[user@fedora9 ~]$

vmstat

[user@fedora9 ~]$ vmstat
procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu------
 r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa st
 2  0  14960  38016   6584 1069284    0    1   506    81  727 1373 12  4 81  3  0
[user@fedora9 ~]$

vmstat -V

[user@fedora9 ~]$ vmstat -V
procps version 3.2.7
[user@fedora9 ~]$

vmstat -a

[user@fedora9 ~]$ vmstat -a
procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu------
 r  b   swpd   free  inact active   si   so    bi    bo   in   cs us sy id wa st
 3  0  14960  38024 988284 461704    0    1   506    81  726 1372 12  4 81  3  0
[user@fedora9 ~]$

vmstat -d

[user@fedora9 ~]$ vmstat -d
disk- ------------reads------------ ------------writes----------- -----IO------
       total merged sectors      ms  total merged sectors      ms    cur    sec
ram0       0      0       0       0      0      0       0       0      0      0
ram1       0      0       0       0      0      0       0       0      0      0
ram2       0      0       0       0      0      0       0       0      0      0
ram3       0      0       0       0      0      0       0       0      0      0
ram4       0      0       0       0      0      0       0       0      0      0
ram5       0      0       0       0      0      0       0       0      0      0
ram6       0      0       0       0      0      0       0       0      0      0
ram7       0      0       0       0      0      0       0       0      0      0
ram8       0      0       0       0      0      0       0       0      0      0
ram9       0      0       0       0      0      0       0       0      0      0
ram10      0      0       0       0      0      0       0       0      0      0
ram11      0      0       0       0      0      0       0       0      0      0
ram12      0      0       0       0      0      0       0       0      0      0
ram13      0      0       0       0      0      0       0       0      0      0
ram14      0      0       0       0      0      0       0       0      0      0
ram15      0      0       0       0      0      0       0       0      0      0
sda   136909  31536 13893867 1197609  58190 219323 2233264 7688807      0    677
sda1   35703   6048 1326394  511477   6728  16136  182984  419232      0    222
sda2      85   1489    2935     653    141   3603   29952    5254      0      1
sda3  101111  23961 12564154  685330  51321 199584 2020328 7264321      0    512
sr0        0      0       0       0      0      0       0       0      0      0
fd0        0      0       0       0      0      0       0       0      0      0
[user@fedora9 ~]

vmstat -D

[user@fedora9 ~]$ vmstat -D
           22 disks 
            0 partitions 
       273820 total reads
        63034 merged reads
     27787446 read sectors
      2395193 milli reading
       116450 writes
       438666 merged writes
      4467248 written sectors
     15377932 milli writing
            0 inprogress IO
         1412 milli spent IO

vmstat -p

vmstat -p will not work under Fedora: https://bugzilla.redhat.com/show_bug.cgi?id=485246. The following report comes from an Ubuntu 9.10 system.

user@ubuntu-9-10:~$ vmstat -p /dev/sda9
sda9          reads   read sectors  writes    requested writes
               23420     411365      24464     530801
user@ubuntu-9-10:~$

vmstat -s

[user@fedora9 ~]$ vmstat -s
      1553972  total memory
      1516180  used memory
       461892  active memory
       988304  inactive memory
        37792  free memory
         6644  buffer memory
      1069388  swap cache
      1052248  total swap
        14960  used swap
      1037288  free swap
       161467 non-nice user cpu ticks
         7586 nice user cpu ticks
        46310 system cpu ticks
      1108919 idle cpu ticks
        46832 IO-wait cpu ticks
         2694 IRQ cpu ticks
         2452 softirq cpu ticks
            0 stolen cpu ticks
      6947021 pages paged in
      1116896 pages paged out
          183 pages swapped in
         3744 pages swapped out
      9985406 interrupts
     18852586 CPU context switches
   1239004323 boot time
        15072 forks
[user@fedora9 ~]$

vmstat -m

[user@fedora9 ~]$ vmstat -m
Cache                       Num  Total   Size  Pages
fuse_request                 11     11    368     11
fuse_inode                    9      9    448      9
rpc_inode_cache               8      8    512      8
nf_conntrack_expect           0      0    168     24
nf_conntrack                 26     80    248     16
dm_uevent                     0      0   2464      3
UDPv6                        22     22    704     11
TCPv6                         6      6   1344      6
kmalloc_dma-512               8      8    512      8
sgpool-128                   12     12   2048      4
scsi_io_context               0      0    104     39
ext3_inode_cache           6822   8360    496      8
ext3_xattr                   85     85     48     85
journal_handle              170    170     24    170
journal_head                 76    219     56     73
revoke_record               256    256     16    256
flow_cache                    0      0     80     51
bsg_cmd                       0      0    288     14
mqueue_inode_cache            7      7    576      7
isofs_inode_cache             0      0    376     10
hugetlbfs_inode_cache        11     11    344     11
dquot                         0      0    128     32
shmem_inode_cache          1058   1071    448      9
xfrm_dst_cache                0      0    320     12
UDP                          19     21    576      7
TCP                          17     24   1216      6
blkdev_queue                 21     21   1080      7
biovec-256                    2      2   3072      2
biovec-128                    5      5   1536      5
biovec-64                     7     10    768      5
sock_inode_cache            619    650    384     10
file_lock_cache              39     39    104     39
Acpi-Operand               2935   2958     40    102
Acpi-Namespace             1700   1700     24    170
Cache                       Num  Total   Size  Pages
taskstats                    25     26    312     13
proc_inode_cache            233    242    360     11
sigqueue                     28     28    144     28
radix_tree_node            7888   8606    296     13
bdev_cache                   24     24    512      8
inode_cache                 370    462    344     11
dentry                     6592  15390    136     30
names_cache                   2      2   4096      2
avc_node                     73     73     56     73
selinux_inode_security     9888  10030     48     85
idr_layer_cache             627    644    144     28
buffer_head                2308   2688     64     64
mm_struct                   659    693    448      9
vm_area_struct            11110  11592     88     46
files_cache                 115    130    384     10
sighand_cache               141    150   1344      6
task_struct                 246    248   3696      2
anon_vma                   4778   5120     16    256
kmalloc-4096                 95    112   4096      8
kmalloc-2048                272    304   2048     16
kmalloc-1024                518    524   1024      4
kmalloc-512                 764    888    512      8
kmalloc-256                 198    208    256     16
kmalloc-128                 629    832    128     32
kmalloc-64                 4322   5568     64     64
kmalloc-32                 1554   1664     32    128
kmalloc-16                 2644   3584     16    256
kmalloc-8                  3561   3584      8    512
kmalloc-192                6349   6930    192     21
kmalloc-96                  885   1176     96     42
[user@fedora9 ~]$ 

References

  1. http://lame.sourceforge.net/
  2. http://en.wikipedia.org/wiki//dev/random
  3. http://en.wikipedia.org/wiki/System_call
  4. http://www.softpanorama.org/Internals/unix_system_calls.shtml
  5. http://www.ibm.com/developerworks/linux/library/l-system-calls/

Related articles

Using rdiff-backup under Linux
Wake On LAN under Linux
Linux I/O Performance Tests using dd