[BACK]Return to iperf3.1 CVS log [TXT] [DIR] Up to [ELWIX - Embedded LightWeight unIX -] / embedaddon / iperf / src

Diff for /embedaddon/iperf/src/iperf3.1 between versions 1.1 and 1.1.1.3

version 1.1, 2016/10/18 13:28:18 version 1.1.1.3, 2023/09/27 11:14:54
Line 1 Line 1
.TH IPERF 1 "October 2015" ESnet "User Manuals".TH IPERF3 1 "September 2022" ESnet "User Manuals"
 .SH NAME  .SH NAME
 iperf3 \- perform network throughput tests  iperf3 \- perform network throughput tests
 .SH SYNOPSIS  .SH SYNOPSIS
Line 6  iperf3 \- perform network throughput tests Line 6  iperf3 \- perform network throughput tests
 .I options  .I options
 .B ]  .B ]
 .br  .br
.B iperf3 -c .B iperf3 -c
 .I server  .I server
 .B [  .B [
 .I options  .I options
 .B ]  .B ]
   
 .SH DESCRIPTION  .SH DESCRIPTION
iperf3 is a tool for performing network throughput measurements.  It can testiperf3 is a tool for performing network throughput measurements.
either TCP or UDP throughput.  To perform an iperf3 test the user mustIt can test TCP, UDP, or SCTP throughput.
establish both a server and a client.To perform an iperf3 test the user must establish both a server and a
client.
 .PP
 The iperf3 executable contains both client and server functionality.
 An iperf3 server can be started using either of the -s or
 --server command-line parameters, for example:
 .IP
 \fCiperf3 -s\fR
 .IP
 \fCiperf3 --server \fR
 .PP
 Note that many iperf3 parameters have both short (-s) and long
 (--server) forms.
 In this section we will generally use the short form of command-line
 flags, unless only the long form of a flag is available.
 .PP
 By default, the iperf3 server listens on TCP port 5201 for connections
 from an iperf3 client.
 A custom port can be specified by using the -p flag, for
 example:
 .IP
 \fCiperf3 -s -p 5002\fR
 .PP
 After the server is started, it will listen for connections from
 iperf3 clients (in other words, the iperf3 program run in client
 mode).
 The client mode can be started using the -c command-line option,
 which also requires a host to which iperf3 should connect.
 The host can by specified by hostname, IPv4 literal, or IPv6 literal:
 .IP
 \fCiperf3 -c iperf3.example.com\fR
 .IP
 \fCiperf3 -c 192.0.2.1\fR
 .IP
 \fCiperf3 -c 2001:db8::1\fR
 .PP
 If the iperf3 server is running on a non-default TCP port, that port
 number needs to be specified on the client as well:
 .IP
 \fCiperf3 -c iperf3.example.com -p 5002\fR
 .PP
 The initial TCP connection is used to exchange test parameters,
 control the start and end of the test, and to exchange test results.
 This is sometimes referred to as the "control connection".
 The actual test data is sent over a separate TCP connection, as a
 separate flow of UDP packets, or as an independent SCTP connection,
 depending on what protocol was specified by the client.
 .PP
 Normally, the test data is sent from the client to the server, and
 measures the upload speed of the client.
 Measuring the download speed from the server can be done by specifying
 the -R flag on the client.
 This causes data to be sent from the server to the client.
 .IP
 \fCiperf3 -c iperf3.example.com -p 5202 -R
 .PP
 Results are displayed on both the client and server.
 There will be at least one line of output per measurement interval (by
 default a measurement interval lasts for one second, but this can be
 changed by the -i option).
 Each line of output includes (at least) the time since the start of
 the test, amount of data transferred during the interval, and the
 average bitrate over that interval.
 Note that the values for each measurement interval are taken from the
 point of view of the endpoint process emitting that output (in other
 words, the output on the client shows the measurement interval data for
 the client.
 .PP
 At the end of the test is a set of statistics that shows (at
 least as much as possible) a summary of the test as seen by both the
 sender and the receiver, with lines tagged accordingly.
 Recall that by default the client is the sender and the server is the
 receiver, although as indicated above, use of the \fC-R\fR flag will
 reverse these roles.
 .PP
 The client can be made to retrieve the server-side output for a given
 test by specifying the --get-server-output flag.
 .PP
 Either the client or the server can produce its output in a JSON
 structure, useful for integration with other programs, by passing it
 the -J flag.
 Because the contents of the JSON structure are only completely known
 after the test has finished, no JSON output will be emitted until the
 end of the test.
 .PP
 iperf3 has a (overly) large set of command-line options that can be
 used to set the parameters of a test.
 They are given in the "GENERAL OPTIONS" section of the manual page
 below, as well as summarized in iperf3's help output, which can be
 viewed by running iperf3 with the -h flag.
 .SH "GENERAL OPTIONS"  .SH "GENERAL OPTIONS"
 .TP  .TP
 .BR -p ", " --port " \fIn\fR"  .BR -p ", " --port " \fIn\fR"
 set server port to listen on/connect to to \fIn\fR (default 5201)  set server port to listen on/connect to to \fIn\fR (default 5201)
 .TP  .TP
 .BR -f ", " --format " "  .BR -f ", " --format " "
[kmKM]   format to report: Kbits, Mbits, KBytes, MBytes[kmgtKMGT]   format to report: Kbits/Mbits/Gbits/Tbits
 .TP  .TP
 .BR -i ", " --interval " \fIn\fR"  .BR -i ", " --interval " \fIn\fR"
pause \fIn\fR seconds between periodic bandwidth reports;pause \fIn\fR seconds between periodic throughput reports;
 default is 1, use 0 to disable  default is 1, use 0 to disable
 .TP  .TP
   .BR -I ", " --pidfile " \fIfile\fR"
   write a file with the process ID, most useful when running as a daemon.
   .TP
 .BR -F ", " --file " \fIname\fR"  .BR -F ", " --file " \fIname\fR"
client-side: read from the file and write to the network, insteadUse a file as the source (on the sender) or sink (on the receiver) of
of using random data;data, rather than just generating random data or throwing it away.
server-side: read from the network and write to the file, insteadThis feature is used for finding whether or not the storage subsystem
of throwing the data awayis the bottleneck for file transfers.
 It does not turn iperf3 into a file transfer tool.
 The length, attributes, and in some cases contents of the received
 file may not match those of the original file.
 .TP  .TP
 .BR -A ", " --affinity " \fIn/n,m\fR"  .BR -A ", " --affinity " \fIn/n,m\fR"
Set the CPU affinity, if possible (Linux and FreeBSD only).Set the CPU affinity, if possible (Linux, FreeBSD, and Windows only).
 On both the client and server you can set the local affinity by using  On both the client and server you can set the local affinity by using
 the \fIn\fR form of this argument (where \fIn\fR is a CPU number).  the \fIn\fR form of this argument (where \fIn\fR is a CPU number).
 In addition, on the client side you can override the server's  In addition, on the client side you can override the server's
 affinity for just that one test, using the \fIn,m\fR form of  affinity for just that one test, using the \fIn,m\fR form of
 argument.  argument.
 Note that when using this feature, a process will only be bound  Note that when using this feature, a process will only be bound
to a single CPU (as opposed to a set containing potentialy multipleto a single CPU (as opposed to a set containing potentially multiple
 CPUs).  CPUs).
 .TP  .TP
.BR -B ", " --bind " \fIhost\fR".BR -B ", " --bind " \fIhost\fR[\fB%\fIdev\fR]"
bind to a specific interfacebind to the specific interface associated with address \fIhost\fR.
 If an optional interface is specified, it is treated as a shortcut
 for \fB--bind-dev \fIdev\fR.
 Note that a percent sign and interface device name are required for IPv6 link-local address literals.
 .TP  .TP
   .BR --bind-dev " \fIdev\fR"
   bind to the specified network interface.
   This option uses SO_BINDTODEVICE, and may require root permissions.
   (Available on Linux and possibly other systems.)
   .TP
 .BR -V ", " --verbose " "  .BR -V ", " --verbose " "
give more detailed output give more detailed output
 .TP  .TP
 .BR -J ", " --json " "  .BR -J ", " --json " "
 output in JSON format  output in JSON format
Line 58  output in JSON format Line 160  output in JSON format
 .BR --logfile " \fIfile\fR"  .BR --logfile " \fIfile\fR"
 send output to a log file.  send output to a log file.
 .TP  .TP
   .BR --forceflush " "
   force flushing output at every interval.
   Used to avoid buffering when sending output to pipe.
   .TP
   .BR --timestamps "[\fB=\fIformat\fR]"
   prepend a timestamp at the start of each output line.
   By default, timestamps have the format emitted by
   .BR ctime ( 1 ).
   Optionally, \fC=\fR followed by
   a format specification can be passed to customize the
   timestamps, see
   .BR strftime ( 3 ).
   If this optional format is given, the \fC=\fR must immediately
   follow the \fB--timestamps\fR option with no whitespace intervening.
   .TP
   .BR --rcv-timeout " \fI#\fR"
   set idle timeout for receiving data during active tests. The receiver
   will halt a test if no data is received from the sender for this
   number of ms (default to 12000 ms, or 2 minutes).
   .TP
   .BR --snd-timeout " \fI#\fR"
   set timeout for unacknowledged TCP data (on both test and control
   connections) This option can be used to force a faster test timeout
   in case of a network partition during a test. The required
   parameter is specified in ms, and defaults to the system settings.
   This functionality depends on the TCP_USER_TIMEOUT socket option, and
   will not work on systems that do not support it.
   .TP
 .BR -d ", " --debug " "  .BR -d ", " --debug " "
 emit debugging output.  emit debugging output.
 Primarily (perhaps exclusively) of use to developers.  Primarily (perhaps exclusively) of use to developers.
Line 76  run in server mode Line 206  run in server mode
 .BR -D ", " --daemon " "  .BR -D ", " --daemon " "
 run the server in background as a daemon  run the server in background as a daemon
 .TP  .TP
 .BR -I ", " --pidfile " \fIfile\fR"  
 write a file with the process ID, most useful when running as a daemon.  
 .TP  
 .BR -1 ", " --one-off  .BR -1 ", " --one-off
handle one client connection, then exit.handle one client connection, then exit.  If an idle time is set, the
server will exit after that amount of time with no connection.
 .TP
 .BR --idle-timeout " \fIn\fR"
 restart the server after \fIn\fR seconds in case it gets stuck.  In
 one-off mode, this is the number of seconds the server will wait
 before exiting.
 .TP
 .BR --server-bitrate-limit " \fIn\fR[KMGT]"
 set a limit on the server side, which will cause a test to abort if
 the client specifies a test of more than \fIn\fR bits per second, or
 if the average data sent or received by the client (including all data
 streams) is greater than \fIn\fR bits per second.  The default limit
 is zero, which implies no limit.  The interval over which to average
 the data rate is 5 seconds by default, but can be specified by adding
 a '/' and a number to the bitrate specifier.
 .TP
 .BR --rsa-private-key-path " \fIfile\fR"
 path to the RSA private key (not password-protected) used to decrypt
 authentication credentials from the client (if built with OpenSSL
 support).
 .TP
 .BR --authorized-users-path " \fIfile\fR"
 path to the configuration file containing authorized users credentials to run
 iperf tests (if built with OpenSSL support).
 The file is a comma separated list of usernames and password hashes;
 more information on the structure of the file can be found in the
 EXAMPLES section.
 .TP
 .BR --time-skew-threshold second " \fIseconds\fR"
 time skew threshold (in seconds) between the server and client
 during the authentication process.
 .SH "CLIENT SPECIFIC OPTIONS"  .SH "CLIENT SPECIFIC OPTIONS"
 .TP  .TP
.BR -c ", " --client " \fIhost\fR".BR -c ", " --client " \fIhost\fR[\fB%\fIdev\fR]"
run in client mode, connecting to the specified serverrun in client mode, connecting to the specified server.
 By default, a test consists of sending data from the client to the
 server, unless the \-R flag is specified.
 If an optional interface is specified, it is treated as a shortcut
 for \fB--bind-dev \fIdev\fR.
 Note that a percent sign and interface device name are required for IPv6 link-local address literals.
 .TP  .TP
 .BR --sctp  .BR --sctp
 use SCTP rather than TCP (FreeBSD and Linux)  use SCTP rather than TCP (FreeBSD and Linux)
Line 93  use SCTP rather than TCP (FreeBSD and Linux) Line 255  use SCTP rather than TCP (FreeBSD and Linux)
 .BR -u ", " --udp  .BR -u ", " --udp
 use UDP rather than TCP  use UDP rather than TCP
 .TP  .TP
.BR -b ", " --bandwidth " \fIn\fR[KM]".BR --connect-timeout " \fIn\fR"
set target bandwidth to \fIn\fR bits/sec (default 1 Mbit/sec for UDP, unlimited for TCP).set timeout for establishing the initial control connection to the
If there are multiple streams (\-P flag), the bandwidth limit is appliedserver, in milliseconds.
 The default behavior is the operating system's timeout for TCP
 connection establishment.
 Providing a shorter value may speed up detection of a down iperf3
 server.
 .TP
 .BR -b ", " --bitrate " \fIn\fR[KMGT]"
 set target bitrate to \fIn\fR bits/sec (default 1 Mbit/sec for UDP,
 unlimited for TCP/SCTP).
 If there are multiple streams (\-P flag), the throughput limit is applied
 separately to each stream.  separately to each stream.
You can also add a '/' and a number to the bandwidth specifier.You can also add a '/' and a number to the bitrate specifier.
 This is called "burst mode".  This is called "burst mode".
 It will send the given number of packets without pausing, even if that  It will send the given number of packets without pausing, even if that
temporarily exceeds the specified bandwidth limit.temporarily exceeds the specified throughput limit.
Setting the target bandwidth to 0 will disable bandwidth limitsSetting the target bitrate to 0 will disable bitrate limits
 (particularly useful for UDP tests).  (particularly useful for UDP tests).
On platforms supporting the \fCSO_MAX_PACING_RATE\fR socket optionThis throughput limit is implemented internally inside iperf3, and is
(currently only Linux), fair-queueing socket-level pacing, implemented inavailable on all platforms.
the kernel, will be used.Compare with the \--fq-rate flag.
On other platforms, iperf3 will implement its own rate control.This option replaces the \--bandwidth flag, which is now deprecated
 but (at least for now) still accepted.
 .TP  .TP
   .BR --pacing-timer " \fIn\fR[KMGT]"
   set pacing timer interval in microseconds (default 1000 microseconds,
   or 1 ms).
   This controls iperf3's internal pacing timer for the \-b/\--bitrate
   option.
   The timer fires at the interval set by this parameter.
   Smaller values of the pacing timer parameter smooth out the traffic
   emitted by iperf3, but potentially at the cost of performance due to
   more frequent timer processing.
   .TP
   .BR --fq-rate " \fIn\fR[KMGT]"
   Set a rate to be used with fair-queueing based socket-level pacing,
   in bits per second.
   This pacing (if specified) will be in addition to any pacing due to
   iperf3's internal throughput pacing (\-b/\--bitrate flag), and both can be
   specified for the same test.
   Only available on platforms supporting the
   \fCSO_MAX_PACING_RATE\fR socket option (currently only Linux).
   The default is no fair-queueing based pacing.
   .TP
 .BR --no-fq-socket-pacing  .BR --no-fq-socket-pacing
disable the use of fair-queueing based socket-level pacing with the -bThis option is deprecated and will be removed.
option, and rely on iperf3's internal rate control.It is equivalent to specifying --fq-rate=0.
 .TP  .TP
 .BR -t ", " --time " \fIn\fR"  .BR -t ", " --time " \fIn\fR"
 time in seconds to transmit for (default 10 secs)  time in seconds to transmit for (default 10 secs)
 .TP  .TP
.BR -n ", " --bytes " \fIn\fR[KM]".BR -n ", " --bytes " \fIn\fR[KMGT]"
 number of bytes to transmit (instead of \-t)  number of bytes to transmit (instead of \-t)
 .TP  .TP
.BR -k ", " --blockcount " \fIn\fR[KM]".BR -k ", " --blockcount " \fIn\fR[KMGT]"
 number of blocks (packets) to transmit (instead of \-t or \-n)  number of blocks (packets) to transmit (instead of \-t or \-n)
 .TP  .TP
.BR -l ", " --length " \fIn\fR[KM]".BR -l ", " --length " \fIn\fR[KMGT]"
length of buffer to read or write (default 128 KB for TCP, 8KB for UDP)length of buffer to read or write.  For TCP tests, the default value
 is 128KB.
 In the case of UDP, iperf3 tries to dynamically determine a reasonable
 sending size based on the path MTU; if that cannot be determined it
 uses 1460 bytes as a sending size.
 For SCTP tests, the default size is 64KB.
 .TP  .TP
 .BR --cport " \fIport\fR"  .BR --cport " \fIport\fR"
 bind data streams to a specific client port (for TCP and UDP only,  bind data streams to a specific client port (for TCP and UDP only,
 default is to use an ephemeral port)  default is to use an ephemeral port)
 .TP  .TP
 .BR -P ", " --parallel " \fIn\fR"  .BR -P ", " --parallel " \fIn\fR"
number of parallel client streams to runnumber of parallel client streams to run. Note that iperf3 is single threaded, so if you are CPU bound, this will not yield higher throughput.
 .TP  .TP
 .BR -R ", " --reverse  .BR -R ", " --reverse
run in reverse mode (server sends, client receives)reverse the direction of a test, so that the server sends data to the
 client
 .TP  .TP
.BR -w ", " --window " \fIn\fR[KM]".BR --bidir
window size / socket buffer size (this gets sent to the server and used on that side too)test in both directions (normal and reverse), with both the client and
 server sending and receiving data simultaneously
 .TP  .TP
   .BR -w ", " --window " \fIn\fR[KMGT]"
   set socket buffer size / window size.
   This value gets sent to the server and used on that side too; on both
   sides this option sets both the sending and receiving socket buffer sizes.
   This option can be used to set (indirectly) the maximum TCP window size.
   Note that on Linux systems, the effective maximum window size is approximately
   double what is specified by this option (this behavior is not a bug in iperf3
   but a "feature" of the Linux kernel, as documented by tcp(7) and socket(7)).
   .TP
 .BR -M ", " --set-mss " \fIn\fR"  .BR -M ", " --set-mss " \fIn\fR"
 set TCP/SCTP maximum segment size (MTU - 40 bytes)  set TCP/SCTP maximum segment size (MTU - 40 bytes)
 .TP  .TP
Line 150  only use IPv4 Line 358  only use IPv4
 only use IPv6  only use IPv6
 .TP  .TP
 .BR -S ", " --tos " \fIn\fR"  .BR -S ", " --tos " \fIn\fR"
set the IP 'type of service'set the IP type of service. The usual prefixes for octal and hex can be used,
 i.e. 52, 064 and 0x34 all specify the same value.
 .TP  .TP
   .BR "--dscp " \fIdscp\fR
   set the IP DSCP bits.  Both numeric and symbolic values are accepted. Numeric
   values can be specified in decimal, octal and hex (see --tos above).
   .TP
 .BR -L ", " --flowlabel " \fIn\fR"  .BR -L ", " --flowlabel " \fIn\fR"
 set the IPv6 flow label (currently only supported on Linux)  set the IPv6 flow label (currently only supported on Linux)
 .TP  .TP
Line 178  Use a "zero copy" method of sending data, such as send Line 391  Use a "zero copy" method of sending data, such as send
 instead of the usual write(2).  instead of the usual write(2).
 .TP  .TP
 .BR -O ", " --omit " \fIn\fR"  .BR -O ", " --omit " \fIn\fR"
Omit the first n seconds of the test, to skip past the TCP slow-startPerform pre-test for N seconds and omit the pre-test statistics, to skip past the TCP slow-start
 period.  period.
 .TP  .TP
 .BR -T ", " --title " \fIstr\fR"  .BR -T ", " --title " \fIstr\fR"
 Prefix every output line with this string.  Prefix every output line with this string.
 .TP  .TP
   .BR --extra-data " \fIstr\fR"
   Specify an extra data string field to be included in JSON output.
   .TP
 .BR -C ", " --congestion " \fIalgo\fR"  .BR -C ", " --congestion " \fIalgo\fR"
 Set the congestion control algorithm (Linux and FreeBSD only).  An  Set the congestion control algorithm (Linux and FreeBSD only).  An
 older  older
Line 198  JSON format, otherwise it will be in human-readable fo Line 414  JSON format, otherwise it will be in human-readable fo
 If the client is run with \fB--json\fR, the server output is included  If the client is run with \fB--json\fR, the server output is included
 in a JSON object; otherwise it is appended at the bottom of the  in a JSON object; otherwise it is appended at the bottom of the
 human-readable output.  human-readable output.
   .TP
   .BR --udp-counters-64bit
   Use 64-bit counters in UDP test packets.
   The use of this option can help prevent counter overflows during long
   or high-bitrate UDP tests.  Both client and server need to be running
   at least version 3.1 for this option to work.  It may become the
   default behavior at some point in the future.
   .TP
   .BR --repeating-payload
   Use repeating pattern in payload, instead of random bytes.
   The same payload is used in iperf2 (ASCII '0..9' repeating).
   It might help to test and reveal problems in networking gear with hardware
   compression (including some WiFi access points), where iperf2 and iperf3
   perform differently, just based on payload entropy.
   .TP
   .BR --dont-fragment
   Set the IPv4 Don't Fragment (DF) bit on outgoing packets.
   Only applicable to tests doing UDP over IPv4.
   .TP
   .BR --username " \fIusername\fR"
   username to use for authentication to the iperf server (if built with
   OpenSSL support).
   The password will be prompted for interactively when the test is run.  Note,
   the password to use can also be specified via the IPERF3_PASSWORD environment
   variable. If this variable is present, the password prompt will be skipped.
   .TP
   .BR --rsa-public-key-path " \fIfile\fR"
   path to the RSA public key used to encrypt authentication credentials
   (if built with OpenSSL support)
   
   .SH EXAMPLES
   .SS "Authentication - RSA Keypair"
   The authentication feature of iperf3 requires an RSA public keypair.
   The public key is used to encrypt the authentication token containing the
   user credentials, while the private key is used to decrypt the authentication token.
   The private key must be in PEM format and additionally must not have a
   password set.
   The public key must be in PEM format and use SubjectPrefixKeyInfo encoding.
   An example of a set of UNIX/Linux commands using OpenSSL
   to generate a correctly-formed keypair follows:
   .sp 1
   .in +.5i
   > openssl genrsa -des3 -out private.pem 2048
   .sp 0
   > openssl rsa -in private.pem -outform PEM -pubout -out public.pem
   .sp 0
   > openssl rsa -in private.pem -out private_not_protected.pem -outform PEM
   .in -.5i
   .sp 1
   After these commands, the public key will be contained in the file
   public.pem and the private key will be contained in the file
   private_not_protected.pem.
   .SS "Authentication - Authorized users configuration file"
   A simple plaintext file must be provided to the iperf3 server in order to specify
   the authorized user credentials.
   The file is a simple list of comma-separated pairs of a username and a
   corresponding password hash.
   The password hash is a SHA256 hash of the string "{$user}$password".
   The file can also contain commented lines (starting with the \fC#\fR
   character).
   An example of commands to generate the password hash on a UNIX/Linux system
   is given below:
   .sp 1
   .in +.5i
   > S_USER=mario S_PASSWD=rossi
   .sp 0
   > echo -n "{$S_USER}$S_PASSWD" | sha256sum | awk '{ print $1 }'
   .in -.5i
   .sp 1
   An example of a password file (with an entry corresponding to the
   above username and password) is given below:
   .sp 0
   .in +.5i
   > cat credentials.csv
   .sp 0
   # file format: username,sha256
   .sp 0
   mario,bf7a49a846d44b454a5d11e7acfaf13d138bbe0b7483aa3e050879700572709b
   .in -.5i
   .sp 1
   
 .SH AUTHORS  .SH AUTHORS
 A list of the contributors to iperf3 can be found within the  A list of the contributors to iperf3 can be found within the
 documentation located at  documentation located at
\fChttp://software.es.net/iperf/dev.html#authors\fR.\fChttps://software.es.net/iperf/dev.html#authors\fR.
   
 .SH "SEE ALSO"  .SH "SEE ALSO"
 libiperf(3),  libiperf(3),
http://software.es.net/iperfhttps://software.es.net/iperf
Removed from v.1.1  
changed lines
  Added in v.1.1.1.3

FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>