1VMSTAT(8) System Administration VMSTAT(8)
2
6 vmstat - Report virtual memory statistics
7
9 vmstat [options] [delay [count]]
10
12 vmstat reports information about processes, memory, paging, block IO,
13 traps, disks and cpu activity.
14 The first report produced gives averages since the last reboot. Addi‐
16 tional reports give information on a sampling period of length delay.
17 The process and memory reports are instantaneous in either case.
18
20 delay The delay between updates in seconds. If no delay is specified,
21 only one report is printed with the average values since boot.
22 count Number of updates. In absence of count, when delay is defined,
24 default is infinite.
25 -a, --active
27 Display active and inactive memory, given a 2.5.41 kernel or
28 better.
29 -f, --forks
31 The -f switch displays the number of forks since boot. This
32 includes the fork, vfork, and clone system calls, and is equiva‐
33 lent to the total number of tasks created. Each process is rep‐
34 resented by one or more tasks, depending on thread usage. This
35 display does not repeat.
36 -m, --slabs
38 Displays slabinfo.
39 -n, --one-header
41 Display the header only once rather than periodically.
42 -s, --stats
44 Displays a table of various event counters and memory statis‐
45 tics. This display does not repeat.
46 -d, --disk
48 Report disk statistics (2.5.70 or above required).
49 -D, --disk-sum
51 Report some summary statistics about disk activity.
52 -p, --partition device
54 Detailed statistics about partition (2.5.70 or above required).
55 -S, --unit character
57 Switches outputs between 1000 (k), 1024 (K), 1000000 (m), or
58 1048576 (M) bytes. Note this does not change the swap (si/so)
59 or block (bi/bo) fields.
60 -t, --timestamp
62 Append timestamp to each line
63 -w, --wide
65 Wide output mode (useful for systems with higher amount of mem‐
66 ory, where the default output mode suffers from unwanted column
67 breakage). The output is wider than 80 characters per line.
68 -V, --version
70 Display version information and exit.
71 -h, --help
73 Display help and exit.
74
76 Procs
77 r: The number of runnable processes (running or waiting for run time).
78 b: The number of processes in uninterruptible sleep.
79 Memory
81 swpd: the amount of virtual memory used.
82 free: the amount of idle memory.
83 buff: the amount of memory used as buffers.
84 cache: the amount of memory used as cache.
85 inact: the amount of inactive memory. (-a option)
86 active: the amount of active memory. (-a option)
87 Swap
89 si: Amount of memory swapped in from disk (/s).
90 so: Amount of memory swapped to disk (/s).
91 IO
93 bi: Blocks received from a block device (blocks/s).
94 bo: Blocks sent to a block device (blocks/s).
95 System
97 in: The number of interrupts per second, including the clock.
98 cs: The number of context switches per second.
99 CPU
101 These are percentages of total CPU time.
102 us: Time spent running non-kernel code. (user time, including nice time)
103 sy: Time spent running kernel code. (system time)
104 id: Time spent idle. Prior to Linux 2.5.41, this includes IO-wait time.
105 wa: Time spent waiting for IO. Prior to Linux 2.5.41, included in idle.
106 st: Time stolen from a virtual machine. Prior to Linux 2.6.11, unknown.
107
109 Reads
110 total: Total reads completed successfully
111 merged: grouped reads (resulting in one I/O)
112 sectors: Sectors read successfully
113 ms: milliseconds spent reading
114 Writes
116 total: Total writes completed successfully
117 merged: grouped writes (resulting in one I/O)
118 sectors: Sectors written successfully
119 ms: milliseconds spent writing
120 IO
122 cur: I/O in progress
123 s: seconds spent for I/O
124
126 reads: Total number of reads issued to this partition
127 read sectors: Total read sectors for partition
128 writes : Total number of writes issued to this partition
129 requested writes: Total number of write requests made for partition
130
132 cache: Cache name
133 num: Number of currently active objects
134 total: Total number of available objects
135 size: Size of each object
136 pages: Number of pages with at least one active object
137
139 vmstat does not require special permissions.
140 These reports are intended to help identify system bottlenecks. Linux
142 vmstat does not count itself as a running process.
143 All linux blocks are currently 1024 bytes. Old kernels may report
145 blocks as 512 bytes, 2048 bytes, or 4096 bytes.
146 Since procps 3.1.9, vmstat lets you choose units (k, K, m, M). Default
148 is K (1024 bytes) in the default mode.
149 vmstat uses slabinfo 1.1
151
153 /proc/meminfo
154 /proc/stat
155 /proc/*/stat
156
158 free(1), iostat(1), mpstat(1), ps(1), sar(1), top(1)
159
161 Does not tabulate the block io per device or count the number of system
162 calls.
163
165 Written by Henry Ware ⟨al172@yfn.ysu.edu⟩.
166 Fabian Frédérick ⟨ffrederick@users.sourceforge.net⟩ (diskstat, slab,
167 partitions...)
168
170 Please send bug reports to ⟨procps@freelists.org⟩
171procps-ng September 2011 VMSTAT(8)