Manual Page Search Parameters
PV(1) User Manuals PV(1)

pv - monitor the progress of data through a pipe

pv [OPTION]... [FILE]...

pv [-pteIrab8kTv] [-A NUM] [-F FORMAT] [-Wgl0fc] [-D SEC] [-s SIZE] [-i SEC] [-m SEC] [-w WIDTH] [-H HEIGHT] [-N NAME] [-o FILE] [-L RATE] [-B BYTES] [-CESYKX] [-Z BYTES] [-U FILE] [-P FILE] [FILE ...]

pv -n [-trbv] [-Wl0] [-D SEC] [-s SIZE] [-i SEC] [-o FILE] [-L RATE] [-B BYTES] [-CESYKX] [-Z BYTES] [-U FILE] [-P FILE] [FILE ...]

pv -q [-v] [-s SIZE] [-o FILE] [-L RATE] [-B BYTES] [-CESYKX] [-Z BYTES] [-U FILE] [-P FILE] [FILE ...]

pv -R PID [-pteIrabT] [-A NUM] [-F FORMAT] [-s SIZE] [-i SEC] [-w WIDTH] [-H HEIGHT] [-N NAME] [-L RATE] [-B BYTES]

pv -d PID:FD [-pteIrab8kTnq] [-F FORMAT] [-i SEC] [-w WIDTH] [-H HEIGHT]

pv -d PID [-pteIrab8kTnq] [-F FORMAT] [-i SEC] [-w WIDTH] [-H HEIGHT]

pv -h

pv -V

pv shows the progress of data through a pipeline by giving information such as time elapsed, percentage completed (with progress bar), current throughput rate, total data transferred, and ETA.

To use it, insert it in a pipeline between two processes, with the appropriate options. Its standard input will be passed through to its standard output and progress will be shown on standard error.

pv will copy each supplied FILE in turn to standard output (- means standard input), or if no FILEs are specified just standard input is copied. This is the same behaviour as cat(1).

A simple example to watch how quickly a file is transferred using nc(1):


pv file | nc -w 1 somewhere.com 3000

A similar example, transferring a file from another process and passing the expected size to pv:


cat file | pv -s 12345 | nc -w 1 somewhere.com 3000

A more complicated example using numeric output to feed into the dialog(1) program for a full-screen progress display:


(tar cf - . \
| pv -n -s $(du -sb . | awk '{print $1}') \
| gzip -9 > out.tgz) 2>&1 \
| dialog --gauge 'Progress' 7 70

Taking an image of a disk, skipping errors:


pv -EE /dev/your/disk/device > disk-image.img

Writing an image back to a disk:


pv disk-image.img > /dev/your/disk/device

Zeroing a disk:


pv < /dev/zero > /dev/your/disk/device

Note that if the input size cannot be calculated, and the output is a block device, then the size of the block device will be used and pv will automatically stop at that size as if -S had been given.

(Linux only): Watching file descriptor 3 opened by another process 1234:


pv -d 1234:3

(Linux only): Watching all file descriptors used by process 1234:


pv -d 1234

pv takes many options, which are divided into display switches, output modifiers, and general options.

If no display switches are specified, pv behaves as if -p, -t, -e, -r, and -b had been given. Otherwise, only those display types that are explicitly switched on will be shown.

Turn the progress bar on. If any inputs are not files, or are unreadable, and no size was explicitly given (with the -s modifier), the progress bar cannot indicate how close to completion the transfer is, so it will just move left and right to indicate that data is moving - or, with --gauge, the bar will indicate the current rate as a percentage of the maximum rate seen so far.
Turn the timer on. This will display the total elapsed time that pv has been running for.
Turn the ETA timer on. This will attempt to guess, based on current transfer rates and the total data size, how long it will be before completion. The countdown is prefixed with "ETA". This option will have no effect if the total data size cannot be determined.
Turn the ETA timer on, but display the estimated local time at which the transfer will finish, instead of the amount of time remaining. When the estimated time is more than 6 hours in the future, the date is shown as well. The time is prefixed with "FIN" for finish time. As with --eta, this option will have no effect if the total data size cannot be determined.
Turn the rate counter on. This will display the current rate of data transfer. The rate is shown in square brackets [].
Turn the average rate counter on. This will display the current average rate of data transfer (default: last 30s, see -m). The average rate is shown in brackets ().
Turn the total byte counter on. This will display the total amount of data transferred so far.
-8, --bits
Display the total bits instead of the total bytes. The output suffix will be "b" instead of "B".
Display and interpret suffixes as multiples of 1000 rather than the default of 1024. Note that this only takes effect on options after this one, so for consistency, specify this option first.
Turn on the transfer buffer percentage display. This will show the percentage of the transfer buffer in use - but see the caveat under %T in the FORMATTING section below. Implies -C. The transfer buffer percentage is shown in curly brackets {}.
Show the last NUM bytes written - but see the caveat under %nA in the FORMATTING section below. Implies -C.
Ignore the options -p, -t, -e, -r, -a, -b, -T, and -A, and instead use the format string FORMAT to determine the output format. See the FORMATTING section below.
At the end of the transfer, write an additional line showing the transfer rate minimum, maximum, mean, and standard deviation. The values are always in bytes per second (or bits, with -8).
Numeric output. Instead of giving a visual indication of progress, pv will give an integer percentage, one per line, on standard error, suitable for piping (via convoluted redirection) into dialog(1). Note that -f is not required if -n is being used.
If --numeric is in use, then adding --bytes will cause the number of bytes processed so far to be output instead of a percentage; if --line-mode is also in use as well as --bytes and --numeric, then instead of bytes or a percentage, the number of lines so far is output. If --rate is added, then the transfer rate is also output (if --bytes is in use as well, the rate comes after the byte/line count). If --timer is also added, then each output line is prefixed with the elapsed time so far, as a decimal number of seconds.
No output. Useful if the -L option is being used on its own to just limit the transfer rate of a pipe.

Wait until the first byte has been transferred before showing any progress information or calculating any ETAs. Useful if the program you are piping to or from requires extra information before it starts, eg piping data into gpg(1) or mcrypt(1) which require a passphrase before data can be processed.
Wait until SEC seconds have passed before showing any progress information, for example in a script where you only want to show a progress bar if it starts taking a long time. Note that this can be a decimal such as 0.5.
Assume the total amount of data to be transferred is SIZE bytes when calculating percentages and ETAs. The same suffixes of "k", "m" etc can be used as with -L.
If SIZE starts with @, the size of file whose name follows the @ will be used.
Note that --size has no effect if used with -d PID to watch all file descriptors of a process, but will work with -d PID:FD.
If the progress bar is shown but the size is not known, then instead of moving the bar left and right to show progress, show the current transfer rate as a percentage of the maximum rate seen so far.
Instead of counting bytes, count lines (newline characters). The progress bar will only move when a new line is found, and the value passed to the -s option will be interpreted as a line count.
If this option is used without -s, the "total size" (in this case, total line count) is calculated by reading through all input files once before transfer starts. If any inputs are pipes or non-regular files, or are unreadable, the total size will not be calculated.
-0, --null
Count lines as terminated with a null byte instead of with a newline. This option implies --line-mode.
Wait SEC seconds between updates. The default is to update every second. Note that this can be a decimal such as 0.1.
Compute current average rate over a SEC seconds window for average rate and ETA calculations (default 30 seconds).
Assume the terminal is WIDTH characters wide, instead of trying to work it out (or assuming 80 if it cannot be guessed). If this option is used, the output width will not be adjusted if the width of the terminal changes while the transfer is running.
Assume the terminal is HEIGHT rows high, instead of trying to work it out (or assuming 25 if it cannot be guessed). If this option is used, the output height will not be adjusted if the height of the terminal changes while the transfer is running.
Prefix the output information with NAME. Useful in conjunction with -c if you have a complicated pipeline and you want to be able to tell different parts of it apart.
Force output. Normally, pv will not output any visual display if standard error is not a terminal. This option forces it to do so.
Use cursor positioning escape sequences instead of just using carriage returns. This is useful in conjunction with -N (name) if you are using multiple pv invocations in a single, long, pipeline.

Write data to FILE rather than standard output. If the file already exists, it will be truncated.
Limit the transfer to a maximum of RATE bytes per second. A suffix of "K", "M", "G", or "T" can be added to denote kibibytes (*1024), mebibytes, and so on. If --si was also passed, suffixes will denote kilobytes (*1000), megabytes, etc. Note the caveat about the positioning of --si .
Use a transfer buffer size of BYTES bytes. A suffix of "K", "M", "G", or "T" can be added to denote kibibytes (*1024), mebibytes, and so on. The default buffer size is the block size of the input file's filesystem multiplied by 32 (512KiB max), or 400KiB if the block size cannot be determined. This can be useful on platforms like MacOS which perform better in pipelines with specific buffer sizes such as 1024. Implies -C.
Never use splice(2), even if it would normally be possible. The splice(2) system call is a more efficient way of transferring data from or to a pipe than regular read(2) and write(2), but means that the transfer buffer may not be used. This prevents -A and -T from working, cannot work with -X, and makes -B redundant, so using -A, -T, -X, or -B automatically switches on -C. Switching on -C results in a small loss of transfer efficiency. (This option has no effect on systems where splice(2) is unavailable).
Ignore read errors by attempting to skip past the offending sections. The corresponding parts of the output will be null bytes. At first only a few bytes will be skipped, but if there are many errors in a row then the skips will move up to chunks of 512. This is intended to be similar to dd conv=sync,noerror but has not been as thoroughly tested.
Specify -E twice to only report a read error once per file, instead of reporting each byte range skipped.
When ignoring read errors with -E, instead of trying to adaptively skip by reading small amounts and skipping progressively larger sections until a read succeeds, move to the next file block of BYTES bytes as soon as an error occurs. There may still be some shorter skips where the block being skipped coincides with the end of the transfer buffer.
This option can only be used with -E and is intended for use when reading from a block device, such as -E -Z 4K to skip in 4 kibibyte blocks. This will speed up reads from faulty media, at the expense of potentially losing more data.
If a size was specified with -s, stop transferring data once that many bytes have been written, instead of continuing to the end of input.
After every write operation, synchronise the buffer caches to disk - see fdatasync(2). This has no effect when the output is a pipe. Using -Y may improve the accuracy of the progress bar when writing to a slow disk.
Set the O_DIRECT flag on all inputs and outputs, if it is available. This will minimise the effect of caches, at the cost of performance. Due to memory alignment requirements, it also may cause read or write failures with an error of "Invalid argument", especially if reading and writing files across a variety of filesystems in a single pv call. Use this option with caution.
Instead of transferring input data to standard output, discard it. This is equivalent to redirecting standard output to /dev/null, except that write(2) is never called. Implies -C.
Instead of passing data through immediately, do it in two stages - first read all input and write it to FILE, and then once the input is exhausted, read all of FILE and write it to the output. FILE remains in place afterwards, unless it is "-", in which case pv creates a temporary file for this purpose, and automatically removes it afterwards.
This can be useful if you have a pipeline which generates data (your input) quickly but you don't know the size, and you wish to pass it to some slower process, once all of the input has been generated and you know its size, so you can see its progress. Note that when doing this with relatively small amounts of data, --no-splice may be preferable so that pipe buffering doesn't affect the progress display.
Instead of transferring data, watch file descriptor FD of process PID, and show its progress. The pv process will exit when FD either changes to a different file, changes read/write mode, or is closed; other data transfer modifiers - and remote control - may not be used with this option.
If only a PID is specified, then that process will be watched, and all regular files and block devices it opens will be shown with a progress bar. The pv process will exit when process PID exits.
If PID is an instance of pv that is already running, -R PID will cause that instance to act as though it had been given this instance's command line instead. For example, if pv -L 123K is running with process ID 9876, then running pv -R 9876 -L 321K will cause it to start using a rate limit of 321KiB instead of 123KiB. Note that some options cannot be changed while running, such as -c, -l, -f, -D, -E, and -S.

Save the process ID of pv in FILE. The file will be replaced if it already exists, and will be removed when pv exits. While pv is running, it will contain a single number - the process ID of pv - followed by a newline.
Print a usage message on standard output and exit successfully.
Print version information on standard output and exit successfully.

If the -F option is given, then the output format is determined by the given format string. Within that string, the following sequences can be used:

%p
Progress bar. Expands to fill the remaining space. Should only be specified once. Equivalent to -p.
%t
Elapsed time. Equivalent to -t.
%e
ETA as time remaining. Equivalent to -e.
%I
ETA as local time at which the transfer will finish. Equivalent to -I.
%r
Current data transfer rate. Equivalent to -r.
%a
Average data transfer rate. Equivalent to -a.
%b
Bytes transferred so far (or lines if -l was specified). Equivalent to -b. If --bits was specified, %b shows the bits transferred so far, not bytes.
%T
Percentage of the transfer buffer in use. Equivalent to -T. Shows "{----}" if the transfer is being done with splice(2), since splicing to or from pipes does not use the buffer.
%nA
Show the last n bytes written (e.g. %16A for the last 16 bytes). Shows only dots if the transfer is being done with splice(2), since splicing to or from pipes does not use the buffer.
%N
Name prefix given by -N. Padded to 9 characters with spaces, and suffixed with :.
%%
A single %.

The format string equivalent of turning on all display switches is `%N %b %T %t %r %a %p %e %I'.

Some suggested common switch combinations:

Show a progress bar, elapsed time, estimated completion time, byte counter, average rate, and current rate.
Show a progress bar, elapsed time, estimated completion time, line counter, and average rate, counting lines instead of bytes.
Show the amount transferred, elapsed time, current rate, and a gauge showing the current rate as a percentage of the maximum rate seen - useful in a pipeline where the total size is unknown. (If the size is known, these options will show the percentage completion instead of the rate gauge).
Show only the elapsed time - useful as a simple timer, e.g. sleep 10m | pv -t.
The default behaviour: progress bar, elapsed time, estimated completion time, current rate, and byte counter.

On MacOS, it may be useful to specify -B 1024 in a pipeline, as this may improve performance.

An exit status of 1 indicates a problem with the -R or -P options.

Any other exit status is a bitmask of the following:

2
One or more files could not be accessed, stat(2)ed, or opened.
4
An input file was the same as the output file.
8
Internal error with closing a file or moving to the next file.
16
There was an error while transferring data from one or more input files.
32
A signal was caught that caused an early exit.
64
Memory allocation failed.

A zero exit status indicates no problems.

The following environment variables may affect pv:

The current user's home directory. This may be used by the remote control mechanism (the --remote option) to exchange messages between pv instances: if the /run/user/UID/ directory does not exist (where UID is the current user ID), then $HOME/.pv/ will be used instead.
The directory to create per-tty lock files for the terminal when using the --cursor option. If TMPDIR is set to a non-empty value, it is the directory under which lock files are created. Otherwise, if TMP is set, then it is used; and if neither are set, then /tmp is used.

Written by Andrew Wood, with patches submitted by various other people. Please see the package's ACKNOWLEDGEMENTS file for a complete list of contributors.

The following problems are known to exist in pv:

  • In some versions of bash(1) and zsh(1), the construct <(pv filename) will not output any progress to the terminal when run from an interactive shell, due to the subprocess being run in a separate process group from the one that owns the terminal. In these cases, use --force.
  • If pv is used in a pipeline in zsh version 5.8, and the last command in the pipeline is based on shell builtins, zsh takes control of the terminal away from pv, preventing progress from being displayed. For example, this will produce no progress bar:

pv InputFile | { while read -r line; do sleep 0.1; done; }

To work around this, put the last commands of the pipeline in normal brackets to force the use of a subshell:

pv InputFile | ( while read -r line; do sleep 0.1; done; )

Refer to issue #105 for full details.
  • The -c option does not work properly on Cygwin without cygserver running, if started near the bottom of the screen (IPC is needed to handle the terminal scrolling). To fix this, start cygserver before using pv -c.
  • The -R option requires that either /run/user/<uid>/ or $HOME/ can be written to, for inter-process communication.

If you find any other problems, please report them.

Please report any bugs to pv@ivarch.com.

Alternatively, use the issue tracker linked from the pv home page.

cat(1), dialog(1), splice(2), open(2) (for O_DIRECT)

Copyright © 2002-2008, 2010, 2012-2015, 2017, 2021, 2023-2024 Andrew Wood.

License GPLv3+: GNU GPL version 3 or later <https://www.gnu.org/licenses/gpl-3.0.html>.

This is free software: you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law.

October 2024 Linux