使用vmstat分析linux主机性能
procps 是一个很强大的工具包，其中包含有free、sysctl 、ps、top 等常用工具，也包含 slabtop、pamp等高级分析工具，还包含了一个重要的工具vmstat 。
一、基本用法
直接使用vmstat会只输出一次当前主机的一些使用情况，后面也可以跟两个数字，每一个数字为多少秒刷新一次结果，后面一个数字代面输出几次。
1$ vmstat 1 5
2procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu------
3 r b swpd free buff cache si so bi bo in cs us sy id wa st
4 3 0 0 44712 110052 623096 0 0 30 28 217 888 13 3 83 1 0
5 0 0 0 44408 110052 623096 0 0 0 0 88 1446 31 4 65 0 0
6 0 0 0 44524 110052 623096 0 0 0 0 84 872 11 2 87 0 0
7 0 0 0 44516 110052 623096 0 0 0 0 149 1429 18 5 77 0 0
8 0 0 0 44524 110052 623096 0 0 0 0 60 431 14 1 85 0 0
如上所示，每秒输出一次，共输出五次系统相关指标的情况。每项指标的意思如下：
- 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.
二、vmstat与CPU load
测试开始前使用vmstat 查看到的主机的总体情况如下；
1# vmstat 1 -w
2procs -----------------------memory---------------------- ---swap-- -----io---- -system-- --------cpu--------
3 r b swpd free buff cache si so bi bo in cs us sy id wa st
4 1 0 0 4205708 2256 3197224 0 0 0 2 9 0 0 0 99 0 0
5 0 0 0 4205732 2256 3197224 0 0 0 0 292 560 0 0 100 0 0
6 0 0 0 4205700 2256 3197224 0 0 0 0 308 577 0 0 100 0 0
7 0 0 0 4205732 2256 3197224 0 0 0 0 294 556 0 0 100 0 0
8 0 0 0 4205732 2256 3197224 0 0 0 4 303 581 0 0 100 0 0
9 0 0 0 4205732 2256 3197224 0 0 0 0 323 606 0 0 100 0 0
从图上可以看到整个CPU都是idle的，在使用dd命令进行如下测试后，会发现CPU的sy项会上去
1# dd if=/dev/urandom of=500MBfile bs=1M count=500
2# vmstat 1 -w
3procs -----------------------memory---------------------- ---swap-- -----io---- -system-- --------cpu--------
4 r b swpd free buff cache si so bi bo in cs us sy id wa st
5 1 0 0 4504584 2256 2897140 0 0 0 0 1525 1144 0 50 50 0 0
6 1 0 0 4493120 2256 2908748 0 0 0 0 1508 1117 0 50 50 0 0
7 1 0 0 4481604 2256 2920360 0 0 0 0 1526 1135 0 50 50 0 0
8 1 0 0 4469112 2256 2932972 0 0 0 0 1523 1132 0 50 50 0 0
9 1 0 0 4457548 2256 2944556 0 0 0 0 1524 1128 0 50 49 0 0
10 1 0 0 4445868 2256 2956252 0 0 0 0 1520 1134 0 50 50 0 0
11 1 0 0 4434344 2256 2967620 0 0 0 36 1494 1081 0 50 50 0 0
12 1 0 0 4422696 2256 2979120 0 0 0 0 1518 1134 0 50 50 0 0
13 1 0 0 4410048 2256 2992040 0 0 0 1 1518 1134 0 50 50 0 0
14 1 0 0 4398500 2256 3003520 0 0 0 0 1515 1123 1 50 50 0 0
15 1 0 0 4387068 2256 3015188 0 0 0 0 1519 1133 0 50 50 0 0
1620 0 0 4375216 2256 3026632 0 0 0 8704 1498 1012 1 53 46 0 0
17 4 0 0 4374288 2256 3027732 0 0 0 106504 306 298 0 100 0 0 0
1815 0 0 4374112 2256 3027920 0 0 0 98312 267 225 0 100 0 0 0
1918 0 0 4374160 2256 3027872 0 0 0 98312 265 214 0 100 0 0 0
20 3 0 0 4367092 2256 3035024 0 0 0 35304 916 478 0 54 46 0 0
上面这个操作其实是分两部分的，前面一部分从/dev/urandom中取随机数会非常耗CPU的system占用，取完在写IO的过程中，会发现io项的block out（即写IO）也会出现增加。
三、vmstat与高IO read
测试前vmstat查看到的主机资源情况如下：
1# vmstat 1 -w
2procs -------------------memory------------------ ---swap-- -----io---- --system-- -----cpu-------
3 r b swpd free buff cache si so bi bo in cs us sy id wa st
4 0 0 137636 246760 187628 1129464 0 0 2 26 7 6 14 4 82 0 0
5 0 0 137636 246760 187628 1129460 0 0 0 0 106 173 0 0 99 0 0
6 0 0 137636 246760 187628 1129460 0 0 0 0 106 203 0 0 100 0 0
7 0 0 137636 244208 187632 1129472 0 0 4 56 477 516 3 3 94 1 0
8 0 0 137636 244208 187632 1129472 0 0 0 0 126 224 0 1 100 0 0
9 0 0 137636 244208 187632 1129472 0 0 0 0 160 347 0 0 99 0 0
使用dd命令读取一个大文件，并将其内容写入/dev/null
1# dd if=361way.com_redhat_dvd.iso of=/dev/null bs=1M
2# vmstat 1 -w
3procs -------------------memory------------------ ---swap-- -----io---- --system-- -----cpu-------
4 r b swpd free buff cache si so bi bo in cs us sy id wa st
5 0 1 137764 148004 184180 1243028 0 0 195840 0 1793 2767 0 16 47 36 0
6 1 0 137764 138208 184180 1255772 0 0 204672 0 1992 2906 0 18 48 34 0
7 1 0 137764 149368 184180 1248132 0 0 206208 4 1932 2882 0 18 48 35 0
8 0 1 137764 129156 184180 1271188 0 0 211200 32 1891 3096 1 17 48 35 0
9 0 1 137764 117748 184180 1285796 0 0 210176 0 1947 2945 1 17 48 34 0
10 0 1 137764 133000 184180 1273800 0 0 207616 0 2062 3064 0 18 47 34 0
11 1 0 137764 125932 184180 1283836 0 0 216064 0 2089 3194 0 17 48 35 0
12 0 1 137764 129156 184180 1283624 0 0 224768 0 2201 3417 0 18 47 35 0
四、CPU 与CPU wait
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):
1$ vmstat 1 5
2procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu------
3 r b swpd free buff cache si so bi bo in cs us sy id wa st
4 2 1 403256 602848 17836 400356 5 15 50 50 207 861 13 3 83 1 0
5 1 0 403256 601568 18892 400496 0 0 1048 364 337 1903 5 7 0 88 0
6 0 1 403256 600816 19640 400568 0 0 748 0 259 1142 6 4 0 90 0
7 0 1 403256 600300 20116 400800 0 0 476 0 196 630 8 5 0 87 0
8 0 1 403256 599328 20792 400792 0 0 676 0 278 1401 7 5 0 88 0
五、vmstat其他
vmstat帮助信息：
1$ vmstat --help
2usage: vmstat [-V] [-n] [delay [count]]
3 -V prints version.
4 -n causes the headers not to be reprinted regularly.
5 -a print inactive/active page stats.
6 -d prints disk statistics
7 -D prints disk table
8 -p prints disk partition statistics
9 -s prints vm table
10 -m prints slabinfo
11 -S unit size
12 delay is the delay between updates in seconds.
13 unit size k:1000 K:1024 m:1000000 M:1048576 (default is K)
14 count is the number of updates
vmstat -a结果
1$ vmstat -a
2procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu------
3 r b swpd free inact active si so bi bo in cs us sy id wa st
4 3 0 14960 38024 988284 461704 0 1 506 81 726 1372 12 4 81 3 0
vmstat -d结果
1$ vmstat -d
2disk- ------------reads------------ ------------writes----------- -----IO------
3 total merged sectors ms total merged sectors ms cur sec
4ram0 0 0 0 0 0 0 0 0 0 0
5ram1 0 0 0 0 0 0 0 0 0 0
6ram2 0 0 0 0 0 0 0 0 0 0
7ram3 0 0 0 0 0 0 0 0 0 0
8ram4 0 0 0 0 0 0 0 0 0 0
9ram5 0 0 0 0 0 0 0 0 0 0
10ram6 0 0 0 0 0 0 0 0 0 0
11ram7 0 0 0 0 0 0 0 0 0 0
12ram8 0 0 0 0 0 0 0 0 0 0
13ram9 0 0 0 0 0 0 0 0 0 0
14ram10 0 0 0 0 0 0 0 0 0 0
15ram11 0 0 0 0 0 0 0 0 0 0
16ram12 0 0 0 0 0 0 0 0 0 0
17ram13 0 0 0 0 0 0 0 0 0 0
18ram14 0 0 0 0 0 0 0 0 0 0
19ram15 0 0 0 0 0 0 0 0 0 0
20sda 136909 31536 13893867 1197609 58190 219323 2233264 7688807 0 677
21sda1 35703 6048 1326394 511477 6728 16136 182984 419232 0 222
22sda2 85 1489 2935 653 141 3603 29952 5254 0 1
23sda3 101111 23961 12564154 685330 51321 199584 2020328 7264321 0 512
vmstat -D
1$ vmstat -D
2 22 disks
3 0 partitions
4 273820 total reads
5 63034 merged reads
6 27787446 read sectors
7 2395193 milli reading
8 116450 writes
9 438666 merged writes
10 4467248 written sectors
11 15377932 milli writing
12 0 inprogress IO
13 1412 milli spent IO
vmstat -p
1vmstat -p /dev/sda9
2sda9 reads read sectors writes requested writes
3 23420 411365 24464 530801
vmstat -s
1$ vmstat -s
2 1553972 total memory
3 1516180 used memory
4 461892 active memory
5 988304 inactive memory
6 37792 free memory
7 6644 buffer memory
8 1069388 swap cache
9 1052248 total swap
10 14960 used swap
11 1037288 free swap
12 161467 non-nice user cpu ticks
13 7586 nice user cpu ticks
14 46310 system cpu ticks
15 1108919 idle cpu ticks
16 46832 IO-wait cpu ticks
17 2694 IRQ cpu ticks
18 2452 softirq cpu ticks
19 0 stolen cpu ticks
20 6947021 pages paged in
21 1116896 pages paged out
22 183 pages swapped in
23 3744 pages swapped out
24 9985406 interrupts
25 18852586 CPU context switches
26 1239004323 boot time
27 15072 forks
vmstat -m
1$ vmstat -m
2Cache Num Total Size Pages
3fuse_request 11 11 368 11
4fuse_inode 9 9 448 9
5rpc_inode_cache 8 8 512 8
6nf_conntrack_expect 0 0 168 24
7nf_conntrack 26 80 248 16
8dm_uevent 0 0 2464 3
9UDPv6 22 22 704 11
10TCPv6 6 6 1344 6
11kmalloc_dma-512 8 8 512 8
12sgpool-128 12 12 2048 4
13scsi_io_context 0 0 104 39
14ext3_inode_cache 6822 8360 496 8
15ext3_xattr 85 85 48 85
16journal_handle 170 170 24 170
17journal_head 76 219 56 73
18revoke_record 256 256 16 256
19flow_cache 0 0 80 51
20bsg_cmd 0 0 288 14
21mqueue_inode_cache 7 7 576 7
22isofs_inode_cache 0 0 376 10
23hugetlbfs_inode_cache 11 11 344 11
24dquot 0 0 128 32
25shmem_inode_cache 1058 1071 448 9
26xfrm_dst_cache 0 0 320 12
27UDP 19 21 576 7
28TCP 17 24 1216 6
29blkdev_queue 21 21 1080 7
30biovec-256 2 2 3072 2
31biovec-128 5 5 1536 5
32biovec-64 7 10 768 5
33sock_inode_cache 619 650 384 10
34file_lock_cache 39 39 104 39
35Acpi-Operand 2935 2958 40 102
36Acpi-Namespace 1700 1700 24 170
37Cache Num Total Size Pages
38taskstats 25 26 312 13
39proc_inode_cache 233 242 360 11
40sigqueue 28 28 144 28
41radix_tree_node 7888 8606 296 13
42bdev_cache 24 24 512 8
43inode_cache 370 462 344 11
44dentry 6592 15390 136 30
45names_cache 2 2 4096 2
46avc_node 73 73 56 73
47selinux_inode_security 9888 10030 48 85
48idr_layer_cache 627 644 144 28
49buffer_head 2308 2688 64 64
50mm_struct 659 693 448 9
51vm_area_struct 11110 11592 88 46
52files_cache 115 130 384 10
53sighand_cache 141 150 1344 6
54task_struct 246 248 3696 2
55anon_vma 4778 5120 16 256
56kmalloc-4096 95 112 4096 8
57kmalloc-2048 272 304 2048 16
58kmalloc-1024 518 524 1024 4
59kmalloc-512 764 888 512 8
60kmalloc-256 198 208 256 16
61kmalloc-128 629 832 128 32
62kmalloc-64 4322 5568 64 64
63kmalloc-32 1554 1664 32 128
64kmalloc-16 2644 3584 16 256
65kmalloc-8 3561 3584 8 512
66kmalloc-192 6349 6930 192 21
67kmalloc-96 885 1176 96 42
捐赠本站(Donate)
如您感觉文章有用，可扫码捐赠本站！(If the article useful, you can scan the QR code to donate))
- Author: shisekong
- Link: https://blog.361way.com/vmstat/5137.html
- License: This work is under a 知识共享署名-非商业性使用-禁止演绎 4.0 国际许可协议. Kindly fulfill the requirements of the aforementioned License when adapting or creating a derivative of this work.