4BSD/usr/man/man1/vmstat.1

.TH VMSTAT 1 4/22/80
.de s1
.if n .sp
.if t .sp .1i
..
.de t1
.if n .ta 5n
.if t .ta 1i
..
.UC 4
.SH NAME
vmstat \- report virtual memory statistics
.SH SYNOPSIS
.B vmstat
[
.B \-fs
] [
.B \-i
]
[ interval [ count ] ]
.SH DESCRIPTION
.I Vmstat
delves into the system and normally reports certain statistics kept about
process, virtual memory, disk, trap and cpu activity.
If given a
.B \-f
argument, it instead reports on the number of
.I forks
and
.I vforks
since system startup and the number of pages of virtual memory involved in each
kind of fork.
If given a
.B \-s
argument, it instead prints the contents
of the
.I sum
structure, giving the total number of several kinds of paging related
events which have occurred since boot.
.PP
If none of these options are given,
.I vmstat
will report in a (usually) iterative fashion on the virtual memory
activity in the system.
In this case, the optional
.I interval
argument causes
.I vmstat
to report once each
.I interval
seconds;
``vmstat 5'' will print what the system is doing every five seconds;
this is a good choice of printing interval since this is how often
the statistics are sampled in the system.
If a
.I count
is given, the statistics are repeated
.I count
times.
There are two formats; fields for the older format which is forced
by the
.B \-i
option are tersely described later.  The default format fields are:
.PP
Procs: information about numbers of processes in various states.
.s1
.t1
.nf
R	in run queue
B	blocked for resources (i/o, paging, etc.)
W	runnable or short sleeper (< 20 secs) but swapped
.fi
.s1
Memory: information about the usage of virtual and real memory.
Virtual pages are considered active if they belong to processes which
are running or have run in the last 20 seconds.
A ``page'' here is 1024 bytes.
.s1
.t1
.nf
AVM	active virtual pages
FRE	size of the free list
.fi
.s1
Page: information about page faults and paging activity.
These are averaged each five seconds, and given in units per second.
.s1
.t1
.nf
RE	page reclaims (simulating reference bits)
PI	page-in events
PO	page-out events
FR	pages freed per second
DE	anticipated short term memory shortfall
SR	scan rate: pageout daemon rpm
.fi
.s1
Swap: information about the activity of the swap daemon (process 0).
.s1
.t1
.nf
I	process swap ins in last 5 seconds
O	process swap outs in last 5 seconds
.fi
.s1
Disk: operations per second (this field is system dependent).
Typically paging will be split across several of the available drives.
.s1
.t1
.nf
D0	disk 0; on csvax: /usr file system
D1	disk 1; on csvax: on csvax: /va, /tmp and some paging
D2	disk 2; on csvax: /, /vb, /3bsd, /3bsd/usr and some paging
D3	disk 3; on csvax: /arch, /4bsd, /4bsd/usr and some paging
.fi
.br
.ne 8
.s1
Faults: trap/interrupt rate averages per second over last 5 seconds.
.s1
.t1
.nf
IN	(non clock) device interrupts per second
SY	system calls per second
PD	pseudo-dma interrupts on DZ
CS	cpu context switch rate (switches/sec)
.fi
.s1
Cpu: breakdown of percentage usage of CPU time
.s1
.nf
US	user time for normal and low priority processes
SY	system time
ID	cpu idle
.fi
.s1
For the older format, process states are broken down into
RU (runnable), DW (disk i/o wait), PW (page wait), SL (sleeping in core),
and SW (swapped out while recently active).  For memory usage, the
percentage of memory used by sharable pages is printed as TX.
Fault information is not printed, but the percentage of CPU time given
to low priority jobs is given as NI.
.SH FILES
/dev/kmem, /vmunix
.SH SEE ALSO
The sections starting with ``Interpreting system activity'' in
.I "Setting up the Fourth Berkeley Software Tape"
by W. Joy, O. Babaoglu, and K. Sklower
.SH AUTHORS
William Joy and Ozalp Babaoglu
.SH BUGS