1. Introduction command
vmstat (Virtual Memory Statistics) command is used to report statistics on virtual memory states.
vmstat can not only monitor virtual memory, but also the process of monitoring activities letter, physical memory, paging, disk, and CPU, etc., is the overall situation of the system statistics, the downside is not an in-depth analysis of a process.
vmstat command is a system administrator, are sysstat package, RedHat Linux system can be installed with yum install sysstat directly.
2. Command Format
vmstat [OPTIONS] [DELAY [COUNT]]
DELAY indicates an interval between the report (s). If not specified interval, only a print average statistics since the system started.
COUNT represents the number of reports, if DELAY is specified, do not specify COUNT, represents unlimited reports.
3. Option Description
-a, --active
显示活跃和非活跃内存
-f, --forks
显示从系统启动至今的 fork 数量 。这包括 fork、vfork 和 clone 系统调用，并相当于创建的任务总数。每个进程由一个或多个任务表示，这取决于线程的使用情况
-m, --slabs
显示 slabinfo
-n, --one-header
只在开始时显示一次各字段名称
-s, --stats
显示各种事件计数器和内存统计信息的表
-D, --disk-sum
报告一些有关磁盘活动的汇总统计数据
-d, --disk
显示磁盘相关统计信息
-p, --partition DEVICE
显示指定磁盘分区统计信息
-S, --unit CHAR
使用指定单位显示。CHAR 可取值有 k（1000）、K（1024）、m（1000000） 、M（1048576）。默认单位为 K（1024 Bytes）
-t, --timestamp
将时间戳附加到每行
-w, --wide
宽输出模式，输出大于 80 个字符/行
-h, --help
显示帮助信息并退出
-V, --version
显示版本信息并退出
4. Common examples
(1) show the statistical information system.
vmstat
procs -----------memory---------- ---swap-- -----io---- -system-- ------cpu-----
r b swpd free buff cache si so bi bo in cs us sy id wa st
1 0 0 9120876 554900 5998224 0 0 0 6 1 2 0 0 100 0 0
vmstat command execution result is divided into six parts: procs, memory, swap, io, system, cpu.
Meaning the output contents are as follows:
procs（进程）
r(run)
运行或等待 CPU 时间片的进程数。如果该值长期大于服务器 CPU 的个数，则说明 CPU 资源不足
b(block)
等待 I/O 的进程数量。该数值如果长时间大于 1，则表明系统 I/O 遇到瓶颈
memory（内存，单位 KB）
swpd
虚拟内存（swap 空间）已使用的大小
free
空闲的物理内存的大小
buff
用作缓冲的内存大小，一般存放待写入磁盘的数据
cache
用作缓存的内存大小，一般存放从磁盘中读取的数据
swap（虚拟内存，单位 KB）
si
每秒从交换区写到内存的大小
so
每秒写入交换区的内存大小
io（单位 块/秒）
bi
每秒读取的块数（读磁盘）
bo
每秒写入的块数（写磁盘）
system（系统）
in
每秒 CPU 的中断次数，包括时钟中断
cs
每秒上下文切换数，例如我们调用系统函数，就要进行上下文切换，线程的切换，也要进行上下文切换，这个值越小越好
cpu（以百分比显示）
us
用户进程执行时间
sy
系统进程执行时间
id
空闲时间（包括 IO 等待时间）
wa
等待 IO 时间。wa 的值高时，说明 IO 等待比较严重，这可能由于磁盘大量做随机访问造成的，也有可能是磁盘出现瓶颈
st
被偷走的 CPU 时间
The statistical information (2) three times at intervals of 1s reporting system.
vmstat 1 3
procs -----------memory---------- ---swap-- -----io---- -system-- ------cpu-----
r b swpd free buff cache si so bi bo in cs us sy id wa st
1 0 0 9120312 554900 5999148 0 0 0 6 1 2 0 0 100 0 0
0 0 0 9120048 554900 5999148 0 0 0 0 836 785 0 0 100 0 0
1 0 0 9120064 554900 5999148 0 0 0 52 834 815 0 0 100 0 0
(3) displays active and inactive memory.
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
1 0 0 9124948 2119956 4194464 0 0 0 6 1 2 0 0 100 0 0
When you use the -a option to display active and inactive memory, what is displayed in addition to the increase inact and active, other display content and not content vmstat output options using the same.
Field Description:
inact：非活跃内存大小
active：活跃的内存大小
(4) view the system since its launch has fork how many times.
vmstat -f
5245186 forks
(5) displays statistics about the amount of memory and a variety of system activity.
16165976 K total memory
491596 K used memory
4194836 K active memory
2119968 K inactive memory
9124404 K free memory
554900 K buffer memory
5995076 K swap cache
0 K total swap
0 K used swap
0 K free swap
1660593 non-nice user cpu ticks
26 nice user cpu ticks
1987770 system cpu ticks
1606632174 idle cpu ticks
148581 IO-wait cpu ticks
2 IRQ cpu ticks
9090 softirq cpu ticks
0 stolen cpu ticks
2376421 pages paged in
89871709 pages paged out
0 pages swapped in
0 pages swapped out
1698946367 interrupts
1672804654 CPU context switches
1580713782 boot time
5245774 forks
This information respectively from / proc / meminfo, / proc / stat and / proc / vmstat.
(6) View disk read and write.
vmstat -d
disk- ------------reads------------ ------------writes----------- -----IO------
total merged sectors ms total merged sectors ms cur sec
vda 77168 53491 3302309 349900 3094238 3196966 162673698 7600604 0 1531
vdb 34381 237 1428906 106216 257198 238871 17119528 3120952 0 117
This information mainly from / proc / diskstats.
Output Field Description:
disk 磁盘名
reads
total
读取成功的数据总大小
merged
合并的读请求数
sectors
成功读取的扇区数
ms
读取所花费的毫秒数
Writes
total
写入成功的数据总大小
merged
合并的写请求数
sectors
成功写入的扇区数
ms
写入所花费的毫秒数
IO
cur
当前处于 I/O 等待的进程
s
I/O 花费的秒数
Information slab (7) View system.
Cache Num Total Size Pages
kmalloc-8192 108 128 8192 4
kmalloc-4096 407 488 4096 8
kmalloc-2048 1093 1168 2048 16
...
Since the kernel there will be many small objects that are very frequently destroyed structure, such as i-node, dentry, if every time these objects built on the memory for a page (4KB), but in fact only a few bytes, so would be very wasteful, in order to solve this problem, we introduced a new mechanism to deal with how to allocate a small storage area in the same page, and the slab can be allocated to small objects, so you do not assign a page for each object, thus saving space. Creation of a small kernel object destruction is very frequent, the slab to buffer these small objects, can be reused, reducing the number of memory allocation.
references
[. 1] the vmstat (. 8) Manual
[2] the sysstat official website
[3] blog command linux Park a day (46):. Vmstat command
