xref: /freebsd/usr.bin/systat/systat.1 (revision 9768746b)
1.\" Copyright (c) 1985, 1990, 1993
2.\"	The Regents of the University of California.  All rights reserved.
3.\"
4.\" Redistribution and use in source and binary forms, with or without
5.\" modification, are permitted provided that the following conditions
6.\" are met:
7.\" 1. Redistributions of source code must retain the above copyright
8.\"    notice, this list of conditions and the following disclaimer.
9.\" 2. Redistributions in binary form must reproduce the above copyright
10.\"    notice, this list of conditions and the following disclaimer in the
11.\"    documentation and/or other materials provided with the distribution.
12.\" 3. Neither the name of the University nor the names of its contributors
13.\"    may be used to endorse or promote products derived from this software
14.\"    without specific prior written permission.
15.\"
16.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
17.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19.\" ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
20.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26.\" SUCH DAMAGE.
27.\"
28.\"	@(#)systat.1	8.2 (Berkeley) 12/30/93
29.\" $FreeBSD$
30.\"
31.Dd April 1, 2022
32.Dt SYSTAT 1
33.Os
34.Sh NAME
35.Nm systat
36.Nd display system statistics
37.Sh SYNOPSIS
38.Nm
39.Op Fl Ar display
40.Op Ar display-commands
41.Op Ar refresh-interval
42.Sh DESCRIPTION
43The
44.Nm
45utility displays various system statistics in a screen oriented fashion
46using the curses screen display library,
47.Xr ncurses 3 .
48.Pp
49While
50.Nm
51is running the screen is usually divided into two windows (an exception
52is the vmstat display which uses the entire screen).
53The
54upper window depicts the current system load average.
55The
56information displayed in the lower window may vary, depending on
57user commands.
58The last line on the screen is reserved for user
59input and error messages.
60.Pp
61By default
62.Nm
63displays the processes getting the largest percentage of the processor
64in the lower window.
65Other displays show swap space usage, disk I/O statistics (a la
66.Xr iostat 8 ) ,
67virtual memory statistics (a la
68.Xr vmstat 8 ) ,
69TCP/IP statistics,
70and network connections (a la
71.Xr netstat 1 ) .
72.Pp
73Input is interpreted at two different levels.
74A ``global'' command interpreter processes all keyboard input.
75If this command interpreter fails to recognize a command, the
76input line is passed to a per-display command interpreter.
77This
78allows each display to have certain display-specific commands.
79.Pp
80Command line options:
81.Bl -tag -width "refresh_interval"
82.It Fl Ns Ar display
83The
84.Fl
85flag expects
86.Ar display
87to be one of:
88.Ic icmp ,
89.Ic icmp6 ,
90.Ic ifstat ,
91.Ic iolat ,
92.Ic iostat ,
93.Ic ip ,
94.Ic ip6 ,
95.Ic netstat ,
96.Ic pigs ,
97.Ic sctp ,
98.Ic swap ,
99.Ic tcp ,
100.Ic vmstat ,
101or
102.Ic zarc ,
103These displays can also be requested interactively (without the
104.Dq Fl )
105and are described in
106full detail below.
107.It Ar refresh-interval
108The
109.Ar refresh-value
110specifies the screen refresh time interval in seconds.
111Time interval can be fractional.
112.It Ar display-commands
113A list of commands specific to this display.
114These commands can also be entered interactively and are described for
115each display separately below.
116If the command requires arguments, they can be specified as separate
117command line arguments.
118A command line argument
119.Fl -
120will finish display commands.
121For example:
122.Pp
123.Dl Nm Fl ifstat Fl match Ar bge0,em1 Fl pps
124.Pp
125This will display statistics of packets per second for network interfaces
126named as bge0 and em1.
127.Pp
128.Dl Nm Fl iostat Fl numbers Fl - Ar 2.1
129.Pp
130This will display all IO statistics in a numeric format and the information
131will be refreshed each 2.1 seconds.
132.El
133.Pp
134Certain characters cause immediate action by
135.Nm .
136These are
137.Bl -tag -width Fl
138.It Ic \&^L
139Refresh the screen.
140.It Ic \&^G
141Print the name of the current ``display'' being shown in
142the lower window and the refresh interval.
143.It Ic \&:
144Move the cursor to the command line and interpret the input
145line typed as a command.
146While entering a command the
147current character erase, word erase, and line kill characters
148may be used.
149.El
150.Pp
151The following commands are interpreted by the ``global''
152command interpreter.
153.Bl -tag -width Fl
154.It Ic help
155Print the names of the available displays on the command line.
156.It Ic load
157Print the load average over the past 1, 5, and 15 minutes
158on the command line.
159.It Ic stop
160Stop refreshing the screen.
161.It Xo
162.Op Ic start
163.Op Ar number
164.Xc
165Start (continue) refreshing the screen.
166If a second, numeric,
167argument is provided it is interpreted as a refresh interval
168(in seconds).
169Supplying only a number will set the refresh interval to this
170value.
171.It Ic quit
172Exit
173.Nm .
174(This may be abbreviated to
175.Ic q . )
176.El
177.Pp
178The available displays are:
179.Bl -tag -width Ic
180.It Ic pigs
181Display, in the lower window, those processes resident in main
182memory and getting the
183largest portion of the processor (the default display).
184When less than 100% of the
185processor is scheduled to user processes, the remaining time
186is accounted to the ``idle'' process.
187.It Ic icmp
188Display, in the lower window, statistics about messages received and
189transmitted by the Internet Control Message Protocol
190.Pq Dq ICMP .
191The left half of the screen displays information about received
192packets, and the right half displays information regarding transmitted
193packets.
194.Pp
195The
196.Ic icmp
197display understands two commands:
198.Ic mode
199and
200.Ic reset .
201The
202.Ic mode
203command is used to select one of four display modes, given as its argument:
204.Bl -tag -width absoluteXX -compact
205.It Ic rate :
206show the rate of change of each value in packets (the default)
207per second
208.It Ic delta :
209show the rate of change of each value in packets per refresh interval
210.It Ic since :
211show the total change of each value since the display was last reset
212.It Ic absolute :
213show the absolute value of each statistic
214.El
215.Pp
216The
217.Ic reset
218command resets the baseline for
219.Ic since
220mode.
221The
222.Ic mode
223command with no argument will display the current mode in the command
224line.
225.It Ic icmp6
226This display is like the
227.Ic icmp
228display,
229but displays statistics for IPv6 ICMP.
230.It Ic ip
231Otherwise identical to the
232.Ic icmp
233display, except that it displays IP and UDP statistics.
234.It Ic ip6
235Like the
236.Ic ip
237display,
238except that it displays IPv6 statistics.
239It does not display UDP statistics.
240.It Ic sctp
241Like
242.Ic icmp ,
243but with SCTP statistics.
244.It Ic tcp
245Like
246.Ic icmp ,
247but with TCP statistics.
248.It Ic iolat
249Display statistics describing the hardware latencies of I/O operations as
250computed by the
251.Va CAM_IOSCHED_DYNAMIC
252option.
253This option must be in the kernel config file of the running kernel for this
254display to work.
255All devices are displayed as there is currently no way to filter them.
256The statistics displayed for the I/O latencies are the percentiles with
257sufficient data during the polling interval to compute.
258If a value cannot be estimated ``-'' is displayed.
259The P50 (also known as the median), P90, P99 and P99.9 values are computed if
260more than 2, 10, 100 or 1000 operations occurred during the polling interval.
261The latency is the hardware latency values, and does not include any software
262queuing time.
263The latencies are estimated based on histogram data computed by the CAM I/O
264scheduler and represent estimates of the actual value that are only good to
265two or three significant digits.
266The display of latency changes based on the scale of the latency to reflect
267the precision of the estimates and to fit on the available screen space.
268All latencies are reported in milliseconds.
269When color is enabled
270.Bl -bullet
271.It
272Values below the medium latency threshold are displayed in green.
273.It
274Values between the minimum latency and high latency thresholds are displayed
275in magenta.
276.It
277Values above the high latency thresholds are displayed in red.
278.Pp
279When color is disabled, the default foreground and background colors are always
280used.
281.Pp
282The following commands are specific to the
283.Ic iolat
284display; the minimum unambiguous prefix may be supplied.
285.Pp
286.Bl -tag -width Fl -compact
287.It Cm color
288Toggle the use of color in the display.
289The default is on.
290.It Cm hi=XXX
291Set the high latency threshold to XXX milliseconds.
292.It Cm med=XXX
293Set the medium latency threshold to XXX milliseconds.
294.It Cm read
295Toggle the display of statistics about read operations.
296The default is on.
297.It Cm write
298Toggle the display of statistics about write operations.
299The default is on.
300.It Cm trim
301Toggle the display of statistics about trim operations.
302The default is on.
303.El
304.El
305.It Ic iostat
306Display, in the lower window, statistics about processor use
307and disk throughput.
308Statistics on processor use appear as
309bar graphs of the amount of time executing in user mode (``user''),
310in user mode running low priority processes (``nice''), in
311system mode (``system''), in interrupt mode (``interrupt''),
312and idle (``idle'').
313Statistics
314on disk throughput show, for each drive, megabytes per second,
315average number of disk transactions per second, and
316average kilobytes of data per transaction.
317This information may be
318displayed as bar graphs or as rows of numbers which scroll downward.
319Bar
320graphs are shown by default.
321.Pp
322The following commands are specific to the
323.Ic iostat
324display; the minimum unambiguous prefix may be supplied.
325.Pp
326.Bl -tag -width Fl -compact
327.It Cm numbers
328Show the disk I/O statistics in numeric form.
329Values are
330displayed in numeric columns which scroll downward.
331.It Cm bars
332Show the disk I/O statistics in bar graph form (default).
333.It Cm kbpt
334Toggle the display of kilobytes per transaction.
335(the default is to
336not display kilobytes per transaction).
337.El
338.It Ic swap
339Show information about swap space usage on all the
340swap areas compiled into the kernel and processes that are swapped out
341as well as a summary of disk activity.
342.Pp
343The swap areas are displayed first with their name, sizes and
344usage percentage.
345The
346.Ar Used
347column indicates the total blocks used so far;
348the graph shows the percentage of space in use on each partition.
349If there are more than one swap partition in use,
350a total line is also shown.
351Areas known to the kernel, but not in use are shown as not available.
352.Pp
353Below the swap space statistics,
354processes are listed in order of higher swap area usage.
355Pid, username, a part of command line, the total use of swap space
356in bytes, the size of process, as well as per-process swap usage percentage and
357per-system swap space percentage are shown per process.
358.Pp
359At the bottom left is the disk usage display.
360It reports the number of
361kilobytes per transaction, transactions per second, megabytes
362per second and the percentage of the time the disk was busy averaged
363over the refresh period of the display (by default, five seconds).
364The system keeps statistics on most every storage device.
365In general, up
366to seven devices are displayed.
367The devices displayed by default are the
368first devices in the kernel's device list.
369See
370.Xr devstat 3
371and
372.Xr devstat 9
373for details on the devstat system.
374.It Ic vmstat
375Take over the entire display and show a (rather crowded) compendium
376of statistics related to virtual memory usage, process scheduling,
377device interrupts, system name translation caching, disk I/O etc.
378.Pp
379The upper left quadrant of the screen shows the number
380of users logged in and the load average over the last one, five,
381and fifteen minute intervals.
382Below this line are statistics on memory utilization.
383The first row of the table reports memory usage only among
384active processes, that is processes that have run in the previous
385twenty seconds.
386The second row reports on memory usage of all processes.
387The first column reports on the number of kilobytes in physical pages
388claimed by processes.
389The second column reports the number of kilobytes in physical pages that
390are devoted to read only text pages.
391The third and fourth columns report the same two figures for
392virtual pages, that is the number of kilobytes in pages that would be
393needed if all processes had all of their pages.
394Finally the last column shows the number of kilobytes in physical pages
395on the free list.
396.Pp
397Below the memory display is a list of the
398average number of threads (over the last refresh interval)
399that are runnable (`r'), in page wait (`p'),
400in disk wait other than paging (`d'),
401sleeping (`s'), and swapped out but desiring to run (`w').
402The row also shows the average number of context switches
403(`Csw'), traps (`Trp'; includes page faults), system calls (`Sys'),
404interrupts (`Int'), network software interrupts (`Sof'), and page
405faults (`Flt').
406.Pp
407Below the process queue length listing is a numerical listing and
408a bar graph showing the amount of
409system (shown as `='), interrupt (shown as `+'), user (shown as `>'),
410nice (shown as `-'), and idle time (shown as ` ').
411.Pp
412Below the process display are statistics on name translations.
413It lists the number of names translated in the previous interval,
414the number and percentage of the translations that were
415handled by the system wide name translation cache, and
416the number and percentage of the translations that were
417handled by the per process name translation cache.
418.Pp
419To the right of the name translations display are lines showing
420the number of dirty buffers in the buffer cache (`dtbuf'),
421desired maximum size of vnode cache (`desvn'),
422number of vnodes actually allocated (`numvn'),
423and
424number of allocated vnodes that are free (`frevn').
425.Pp
426At the bottom left is the disk usage display.
427It reports the number of
428kilobytes per transaction, transactions per second, megabytes
429per second and the percentage of the time the disk was busy averaged
430over the refresh period of the display (by default, five seconds).
431The system keeps statistics on most every storage device.
432In general, up
433to seven devices are displayed.
434The devices displayed by default are the
435first devices in the kernel's device list.
436See
437.Xr devstat 3
438and
439.Xr devstat 9
440for details on the devstat system.
441.Pp
442Under the date in the upper right hand quadrant are statistics
443on paging and swapping activity.
444The first two columns report the average number of pages
445brought in and out per second over the last refresh interval
446due to page faults and the paging daemon.
447The third and fourth columns report the average number of pages
448brought in and out per second over the last refresh interval
449due to swap requests initiated by the scheduler.
450The first row of the display shows the average
451number of disk transfers per second over the last refresh interval;
452the second row of the display shows the average
453number of pages transferred per second over the last refresh interval.
454.Pp
455Below the paging statistics is a column of lines regarding the virtual
456memory system.
457The first few lines describe,
458in units (except as noted below)
459of pages per second averaged over the sampling interval,
460pages copied on write (`cow'),
461pages zero filled on demand (`zfod'),
462pages optimally zero filled on demand (`ozfod'),
463the ratio of the (average) ozfod / zfod as a percentage (`%ozfod'),
464pages freed by the page daemon (`daefr'),
465pages freed by exiting processes (`prcfr'),
466total pages freed (`totfr'),
467pages reactivated from the free list (`react'),
468the average number of
469times per second that the page daemon was awakened (`pdwak'),
470pages analyzed by the page daemon (`pdpgs'),
471and
472in-transit blocking page faults (`intrn').
473Note that the units are special for `%ozfod' and `pdwak'.
474The next few lines describe,
475as amounts of memory in kilobytes,
476pages wired down (`wire'),
477active pages (`act'),
478inactive pages (`inact'),
479dirty pages queued for laundering (`laund'),
480and
481free pages (`free').
482Note that the values displayed are the current transient ones;
483they are not averages.
484.Pp
485At the bottom of this column is a line showing the
486amount of virtual memory, in kilobytes, mapped into the buffer cache (`buf').
487This statistic is not useful.
488It exists only as a placeholder for the corresponding useful statistic
489(the amount of real memory used to cache disks).
490The most important component of the latter (the amount of real memory
491used by the vm system to cache disks) is not available,
492but can be guessed from the `inact' amount under some system loads.
493.Pp
494Running down the right hand side of the display is a breakdown
495of the interrupts being handled by the system.
496At the top of the list is the total interrupts per second
497over the time interval.
498The rest of the column breaks down the total on a device
499by device basis.
500Only devices that have interrupted at least once since boot time are shown.
501.Pp
502The following commands are specific to the
503.Ic vmstat
504display; the minimum unambiguous prefix may be supplied.
505.Pp
506.Bl -tag -width Ar -compact
507.It Cm boot
508Display cumulative statistics since the system was booted.
509.It Cm run
510Display statistics as a running total from the point this
511command is given.
512.It Cm time
513Display statistics averaged over the refresh interval (the default).
514.It Cm zero
515Reset running statistics to zero.
516.El
517.It Ic zarc
518display arc cache usage and hit/miss statistics.
519.It Ic netstat
520Display, in the lower window, network connections.
521By default,
522network servers awaiting requests are not displayed.
523Each address
524is displayed in the format ``host.port'', with each shown symbolically,
525when possible.
526It is possible to have addresses displayed numerically,
527limit the display to a set of ports, hosts, and/or protocols
528(the minimum unambiguous prefix may be supplied):
529.Pp
530.Bl -tag -width Ar -compact
531.It Cm all
532Toggle the displaying of server processes awaiting requests (this
533is the equivalent of the
534.Fl a
535flag to
536.Xr netstat 1 ) .
537.It Cm numbers
538Display network addresses numerically.
539.It Cm names
540Display network addresses symbolically.
541.It Cm proto Ar protocol
542Display only network connections using the indicated
543.Ar protocol .
544Supported protocols are ``tcp'', ``udp'', and ``all''.
545.It Cm ignore Op Ar items
546Do not display information about connections associated with
547the specified hosts or ports.
548Hosts and ports may be specified
549by name (``vangogh'', ``ftp''), or numerically.
550Host addresses
551use the Internet dot notation (``128.32.0.9'').
552Multiple items
553may be specified with a single command by separating them with
554spaces.
555.It Cm display Op Ar items
556Display information about the connections associated with the
557specified hosts or ports.
558As for
559.Ar ignore ,
560.Op Ar items
561may be names or numbers.
562.It Cm show Op Ar ports\&|hosts
563Show, on the command line, the currently selected protocols,
564hosts, and ports.
565Hosts and ports which are being ignored
566are prefixed with a `!'.
567If
568.Ar ports
569or
570.Ar hosts
571is supplied as an argument to
572.Cm show ,
573then only the requested information will be displayed.
574.It Cm reset
575Reset the port, host, and protocol matching mechanisms to the default
576(any protocol, port, or host).
577.El
578.It Ic ifstat
579Display the network traffic going through active interfaces on the
580system.
581Idle interfaces will not be displayed until they receive some
582traffic.
583.Pp
584For each interface being displayed, the current, peak and total
585statistics are displayed for incoming and outgoing traffic.
586By default,
587the
588.Ic ifstat
589display will automatically scale the units being used so that they are
590in a human-readable format.
591The scaling units used for the current and
592peak
593traffic columns can be altered by the
594.Ic scale
595command.
596.Bl -tag -width ".Cm scale Op Ar units"
597.It Cm scale Op Ar units
598Modify the scale used to display the current and peak traffic over all
599interfaces.
600The following units are recognised: kbit, kbyte, mbit,
601mbyte, gbit, gbyte and auto.
602.It Cm pps
603Show statistics in packets per second instead of bytes/bits per second.
604A subsequent call of
605.Ic pps
606switches this mode off.
607.It Cm match Op Ar patterns
608Display only interfaces that match pattern provided as an argument.
609Patterns should be in shell syntax separated by whitespaces or commas.
610If this command is called without arguments then all interfaces are displayed.
611For example:
612.Pp
613.Dl match em0, bge1
614.Pp
615This will display em0 and bge1 interfaces.
616.Pp
617.Dl match em*, bge*, lo0
618.Pp
619This will display all
620.Ic em
621interfaces, all
622.Ic bge
623interfaces and the loopback interface.
624.El
625.El
626.Pp
627Commands to switch between displays may be abbreviated to the
628minimum unambiguous prefix; for example, ``io'' for ``iostat''.
629Certain information may be discarded when the screen size is
630insufficient for display.
631For example, on a machine with 10
632drives the
633.Ic iostat
634bar graph displays only 3 drives on a 24 line terminal.
635When
636a bar graph would overflow the allotted screen space it is
637truncated and the actual value is printed ``over top'' of the bar.
638.Pp
639The following commands are common to each display which shows
640information about disk drives.
641These commands are used to
642select a set of drives to report on, should your system have
643more drives configured than can normally be displayed on the
644screen.
645.Pp
646.Bl -tag -width Ar -compact
647.It Cm ignore Op Ar drives
648Do not display information about the drives indicated.
649Multiple
650drives may be specified, separated by spaces.
651.It Cm display Op Ar drives
652Display information about the drives indicated.
653Multiple drives
654may be specified, separated by spaces.
655.It Cm only Op Ar drives
656Display only the specified drives.
657Multiple drives may be specified,
658separated by spaces.
659.It Cm drives
660Display a list of available devices.
661.It Cm match Xo
662.Ar type , Ns Ar if , Ns Ar pass
663.Op | Ar ...
664.Xc
665Display devices matching the given pattern.
666The basic matching
667expressions are the same as those used in
668.Xr iostat 8
669with one difference.
670Instead of specifying multiple
671.Fl t
672arguments which are then ORed together, the user instead specifies multiple
673matching expressions joined by the pipe
674.Pq Ql \&|
675character.
676The comma
677separated arguments within each matching expression are ANDed together, and
678then the pipe separated matching expressions are ORed together.
679Any
680device matching the combined expression will be displayed, if there is room
681to display it.
682For example:
683.Pp
684.Dl match da,scsi | cd,ide
685.Pp
686This will display all SCSI Direct Access devices and all IDE CDROM devices.
687.Pp
688.Dl match da | sa | cd,pass
689.Pp
690This will display all Direct Access devices, all Sequential Access devices,
691and all passthrough devices that provide access to CDROM drives.
692.El
693.Sh FILES
694.Bl -tag -width /boot/kernel/kernel -compact
695.It Pa /boot/kernel/kernel
696For the namelist.
697.It Pa /dev/kmem
698For information in main memory.
699.It Pa /etc/hosts
700For host names.
701.It Pa /etc/networks
702For network names.
703.It Pa /etc/services
704For port names.
705.El
706.Sh SEE ALSO
707.Xr netstat 1 ,
708.Xr kvm 3 ,
709.Xr icmp 4 ,
710.Xr icmp6 4 ,
711.Xr ip 4 ,
712.Xr ip6 4 ,
713.Xr tcp 4 ,
714.Xr udp 4 ,
715.Xr gstat 8 ,
716.Xr iostat 8 ,
717.Xr vmstat 8
718.Sh HISTORY
719The
720.Nm
721program appeared in
722.Bx 4.3 .
723The
724.Ic icmp ,
725.Ic ip ,
726and
727.Ic tcp
728displays appeared in
729.Fx 3.0 ;
730the notion of having different display modes for the
731ICMP, IP, TCP, and UDP statistics was stolen from the
732.Fl C
733option to
734.Xr netstat 1
735in Silicon Graphics' IRIX system.
736.Sh BUGS
737Certain displays presume a minimum of 80 characters per line.
738Ifstat does not detect new interfaces.
739The
740.Ic vmstat
741display looks out of place because it is (it was added in as
742a separate display rather than created as a new program).
743The
744.Ic iolat
745command does not implement the common device commands including
746filtering, as it does not use the
747.Xr devstat 3
748mechanism to obtain its statistics.
749