1VMSTAT(8) Linux Administrator's Manual VMSTAT(8)
2
6 vmstat - Report virtual memory statistics
7
9 vmstat [-a] [-n] [-t] [-S unit] [delay [ count]]
10 vmstat [-s] [-n] [-S unit]
11 vmstat [-m] [-n] [delay [ count]]
12 vmstat [-d] [-n] [delay [ count]]
13 vmstat [-p disk partition] [-n] [delay [ count]]
14 vmstat [-f]
15 vmstat [-V]
16
18 vmstat reports information about processes, memory, paging, block IO,
19 traps, and cpu activity.
20 The first report produced gives averages since the last reboot. Addi‐
22 tional reports give information on a sampling period of length delay.
23 The process and memory reports are instantaneous in either case.
24 Options
27 The -a switch displays active/inactive memory, given a 2.5.41 kernel or
28 better.
29 The -f switch displays the number of forks since boot. This includes
31 the fork, vfork, and clone system calls, and is equivalent to the total
32 number of tasks created. Each process is represented by one or more
33 tasks, depending on thread usage. This display does not repeat.
34 The -t switch adds timestamp to the output.
36 The -m switch displays slabinfo.
38 The -n switch causes the header to be displayed only once rather than
40 periodically.
41 The -s switch displays a table of various event counters and memory
43 statistics. This display does not repeat.
44 delay is the delay between updates in seconds. If no delay is speci‐
46 fied, only one report is printed with the average values since boot.
47 count is the number of updates. If no count is specified and delay is
49 defined, count defaults to infinity.
50 The -d reports disk statistics (2.5.70 or above required)
52 The -w enlarges field width for big memory sizes
54 The -p followed by some partition name for detailed statistics (2.5.70
56 or above required)
57 The -S followed by k or K or m or M switches outputs between 1000,
59 1024, 1000000, or 1048576 bytes
60 The -V switch results in displaying version information.
62
64 Procs
65 r: The number of processes waiting for run time.
66 b: The number of processes in uninterruptible sleep.
67 Memory
69 swpd: the amount of virtual memory used.
70 free: the amount of idle memory.
71 buff: the amount of memory used as buffers.
72 cache: the amount of memory used as cache.
73 inact: the amount of inactive memory. (-a option)
74 active: the amount of active memory. (-a option)
75 Swap
77 si: Amount of memory swapped in from disk (/s).
78 so: Amount of memory swapped to disk (/s).
79 IO
81 bi: Blocks received from a block device (blocks/s).
82 bo: Blocks sent to a block device (blocks/s).
83 System
85 in: The number of interrupts per second, including the clock.
86 cs: The number of context switches per second.
87 CPU
89 These are percentages of total CPU time.
90 us: Time spent running non-kernel code. (user time, including nice time)
91 sy: Time spent running kernel code. (system time)
92 id: Time spent idle. Prior to Linux 2.5.41, this includes IO-wait time.
93 wa: Time spent waiting for IO. Prior to Linux 2.5.41, included in idle.
94 st: Time stolen from a virtual machine. Prior to Linux 2.6.11, unknown.
95
98 Reads
99 total: Total reads completed successfully
100 merged: grouped reads (resulting in one I/O)
101 sectors: Sectors read successfully
102 ms: milliseconds spent reading
103 Writes
105 total: Total writes completed successfully
106 merged: grouped writes (resulting in one I/O)
107 sectors: Sectors written successfully
108 ms: milliseconds spent writing
109 IO
111 cur: I/O in progress
112 s: seconds spent for I/O
113
116 reads: Total number of reads issued to this partition
117 read sectors: Total read sectors for partition
118 writes : Total number of writes issued to this partition
119 requested writes: Total number of write requests made for partition
120
124 cache: Cache name
125 num: Number of currently active objects
126 total: Total number of available objects
127 size: Size of each object
128 pages: Number of pages with at least one active object
129 totpages: Total number of allocated pages
130 pslab: Number of pages per slab
131
134 vmstat does not require special permissions.
135 These reports are intended to help identify system bottlenecks. Linux
137 vmstat does not count itself as a running process.
138 All linux blocks are currently 1024 bytes. Old kernels may report
140 blocks as 512 bytes, 2048 bytes, or 4096 bytes.
141 Since procps 3.1.9, vmstat lets you choose units (k, K, m, M) default
143 is K (1024 bytes) in the default mode
144 vmstat uses slabinfo 1.1 FIXME
146
148 /proc/meminfo
149 /proc/stat
150 /proc/*/stat
151
154 iostat(1), sar(1), mpstat(1), ps(1), top(1), free(1)
155
157 Does not tabulate the block io per device or count the number of system
158 calls.
159
161 Written by Henry Ware <al172@yfn.ysu.edu>.
162 Fabian Frédérick <ffrederick@users.sourceforge.net> (diskstat, slab, partitions...)
163Throatwobbler Ginkgo Labs 27 July 1994 VMSTAT(8)