I currently have the current script.
#!/bin/bash
# script.sh
for i in {0..99}; do
script-to-run.sh input/ output/ $i
done
I wish to run it in parallel using xargs. I have tried
script.sh | xargs -P8
But doing the above only executed once at the time. No luck with -n8 as well.
Adding & at the end of the line to be executed in the script for loop would try to run the script 99 times at once. How do I execute the loop only 8 at the time, up to 100 total.
From the xargs man page:
This manual page documents the GNU version of xargs. xargs reads items
from the standard input, delimited by blanks (which can be protected
with double or single quotes or a backslash) or newlines, and executes
the command (default is /bin/echo) one or more times with any initial-
arguments followed by items read from standard input. Blank lines on
the standard input are ignored.
Which means that for your example xargs is waiting and collecting all of the output from your script and then running echo <that output>. Not exactly all that useful nor what you wanted.
The -n argument is how many items from the input to use with each command that gets run (nothing, by itself, about parallelism here).
To do what you want with xargs you would need to do something more like this (untested):
printf %s\\n {0..99} | xargs -n 1 -P 8 script-to-run.sh input/ output/
Which breaks down like this.
printf %s\\n {0..99} - Print one number per-line from 0 to 99.
Run xargs
taking at most one argument per run command line
and run up to eight processes at a time
With GNU Parallel you would do:
parallel script-to-run.sh input/ output/ {} ::: {0..99}
Add in -P8 if you do not want to run one job per CPU core.
Opposite xargs it will do The Right Thing, even if the input contain space, ', or " (not the case here, though). It also makes sure the output from different jobs are not mixed together, so if you use the output you are guaranteed that you will not get half-a-line from two different jobs.
GNU Parallel is a general parallelizer and makes is easy to run jobs in parallel on the same machine or on multiple machines you have ssh access to.
If you have 32 different jobs you want to run on 4 CPUs, a straight forward way to parallelize is to run 8 jobs on each CPU:
GNU Parallel instead spawns a new process when one finishes - keeping the CPUs active and thus saving time:
Installation
If GNU Parallel is not packaged for your distribution, you can do a personal installation, which does not require root access. It can be done in 10 seconds by doing this:
$ (wget -O - pi.dk/3 || lynx -source pi.dk/3 || curl pi.dk/3/ || \
fetch -o - http://pi.dk/3 ) > install.sh
$ sha1sum install.sh | grep 883c667e01eed62f975ad28b6d50e22a
12345678 883c667e 01eed62f 975ad28b 6d50e22a
$ md5sum install.sh | grep cc21b4c943fd03e93ae1ae49e28573c0
cc21b4c9 43fd03e9 3ae1ae49 e28573c0
$ sha512sum install.sh | grep da012ec113b49a54e705f86d51e784ebced224fdf
79945d9d 250b42a4 2067bb00 99da012e c113b49a 54e705f8 6d51e784 ebced224
fdff3f52 ca588d64 e75f6033 61bd543f d631f592 2f87ceb2 ab034149 6df84a35
$ bash install.sh
For other installation options see http://git.savannah.gnu.org/cgit/parallel.git/tree/README
Learn more
See more examples: http://www.gnu.org/software/parallel/man.html
Watch the intro videos: https://www.youtube.com/playlist?list=PL284C9FF2488BC6D1
Walk through the tutorial: http://www.gnu.org/software/parallel/parallel_tutorial.html
Sign up for the email list to get support: https://lists.gnu.org/mailman/listinfo/parallel
You can use this simple 1 line command
seq 1 500 | xargs -n 1 -P 8 script-to-run.sh input/ output/
Here's an example running commands in parallel in conjuction with find:
find -name "*.wav" -print0 | xargs -0 -t -I % -P $(nproc) flac %
-print0 terminates filenames with a null byte rather than a newline so we can use -0 in xargs to prevent filenames with spaces being treated as two seperate arguments.
-t means verbose, makes xargs print every command it's executing, can be useful, remove if not needed.
-I % means replace occurrences of % in the command with arguments read from standard input.
-P $(nproc) means run a maximum of nproc instances of our command in parallel (nproc prints the number of available processing units).
flac % is our command, the -I % from earlier means this will become flac foo.wav
See also: Manual for xargs(1)
Related
I have a bash function that i call in parallel using xargs -P like so
echo ${list} | xargs -n 1 -P 24 -I# bash -l -c 'myAwesomeShellFunction #'
Everything works fine but output is messed up for obvious reasons (no buffering)
Trying to figure out a way to buffer output effectively. I was thinking I could use awk, but I'm not good enough to write such a script and I can't find anything worthwhile on google? Can someone help me write this "output buffer" in sed or awk? Nothing fancy, just accumulate output and spit it out after process terminates. I don't care the order that shell functions execute, just need their output buffered... Something like:
echo ${list} | xargs -n 1 -P 24 -I# bash -l -c 'myAwesomeShellFunction # | sed -u ""'
P.s. I tried to use stdbuf as per
https://unix.stackexchange.com/questions/25372/turn-off-buffering-in-pipe but did not work, i specified buffering on o and e but output still unbuffered:
echo ${list} | xargs -n 1 -P 24 -I# stdbuf -i0 -oL -eL bash -l -c 'myAwesomeShellFunction #'
Here's my first attempt, this only captures first line of output:
$ bash -c "echo stuff;sleep 3; echo more stuff" | awk '{while (( getline line) > 0 )print "got ",$line;}'
$ got stuff
This isn't quite atomic if your output is longer than a page (4kb typically), but for most cases it'll do:
xargs -P 24 bash -c 'for arg; do printf "%s\n" "$(myAwesomeShellFunction "$arg")"; done' _
The magic here is the command substitution: $(...) creates a subshell (a fork()ed-off copy of your shell), runs the code ... in it, and then reads that in to be substituted into the relevant position in the outer script.
Note that we don't need -n 1 (if you're dealing with a large number of arguments -- for a small number it may improve parallelization), since we're iterating over as many arguments as each of your 24 parallel bash instances is passed.
If you want to make it truly atomic, you can do that with a lockfile:
# generate a lockfile, arrange for it to be deleted when this shell exits
lockfile=$(mktemp -t lock.XXXXXX); export lockfile
trap 'rm -f "$lockfile"' 0
xargs -P 24 bash -c '
for arg; do
{
output=$(myAwesomeShellFunction "$arg")
flock -x 99
printf "%s\n" "$output"
} 99>"$lockfile"
done
' _
cat a.txt | xargs -I % echo %
In the example above, xargs takes echo % as the command argument. But in some cases, I need multiple commands to process the argument instead of one. For example:
cat a.txt | xargs -I % {command1; command2; ... }
But xargs doesn't accept this form. One solution I know is that I can define a function to wrap the commands, but I want to avoid that because it is complex. Is there a better solution?
cat a.txt | xargs -d $'\n' sh -c 'for arg do command1 "$arg"; command2 "$arg"; ...; done' _
...or, without a Useless Use Of cat:
<a.txt xargs -d $'\n' sh -c 'for arg do command1 "$arg"; command2 "$arg"; ...; done' _
To explain some of the finer points:
The use of "$arg" instead of % (and the absence of -I in the xargs command line) is for security reasons: Passing data on sh's command-line argument list instead of substituting it into code prevents content that data might contain (such as $(rm -rf ~), to take a particularly malicious example) from being executed as code.
Similarly, the use of -d $'\n' is a GNU extension which causes xargs to treat each line of the input file as a separate data item. Either this or -0 (which expects NULs instead of newlines) is necessary to prevent xargs from trying to apply shell-like (but not quite shell-compatible) parsing to the stream it reads. (If you don't have GNU xargs, you can use tr '\n' '\0' <a.txt | xargs -0 ... to get line-oriented reading without -d).
The _ is a placeholder for $0, such that other data values added by xargs become $1 and onward, which happens to be the default set of values a for loop iterates over.
You can use
cat file.txt | xargs -i sh -c 'command {} | command2 {} && command3 {}'
{} = variable for each line on the text file
With GNU Parallel you can do:
cat a.txt | parallel 'command1 {}; command2 {}; ...; '
For security reasons it is recommended you use your package manager to
install. But if you cannot do that then you can use this 10 seconds
installation.
The 10 seconds installation will try to do a full installation; if
that fails, a personal installation; if that fails, a minimal
installation.
$ (wget -O - pi.dk/3 || lynx -source pi.dk/3 || curl pi.dk/3/ || \
fetch -o - http://pi.dk/3 ) > install.sh
$ sha1sum install.sh | grep 883c667e01eed62f975ad28b6d50e22a
12345678 883c667e 01eed62f 975ad28b 6d50e22a
$ md5sum install.sh | grep cc21b4c943fd03e93ae1ae49e28573c0
cc21b4c9 43fd03e9 3ae1ae49 e28573c0
$ sha512sum install.sh | grep da012ec113b49a54e705f86d51e784ebced224fdf
79945d9d 250b42a4 2067bb00 99da012e c113b49a 54e705f8 6d51e784 ebced224
fdff3f52 ca588d64 e75f6033 61bd543f d631f592 2f87ceb2 ab034149 6df84a35
$ bash install.sh
I prefer style which allows dry run mode (without | sh) :
cat a.txt | xargs -I % echo "command1; command2; ... " | sh
Works with pipes too:
cat a.txt | xargs -I % echo "echo % | cat " | sh
This is just another approach without xargs nor cat:
while read stuff; do
command1 "$stuff"
command2 "$stuff"
...
done < a.txt
This seems to be the safest version.
tr '[\n]' '[\0]' < a.txt | xargs -r0 /bin/bash -c 'command1 "$#"; command2 "$#";' ''
(-0 can be removed and the tr replaced with a redirect (or the file can be replaced with a null separated file instead). It is mainly in there since I mainly use xargs with find with -print0 output) (This might also be relevant on xargs versions without the -0 extension)
It is safe, since args will pass the parameters to the shell as an array when executing it. The shell (at least bash) would then pass them as an unaltered array to the other processes when all are obtained using ["$#"][1]
If you use ...| xargs -r0 -I{} bash -c 'f="{}"; command "$f";' '', the assignment will fail if the string contains double quotes. This is true for every variant using -i or -I. (Due to it being replaced into a string, you can always inject commands by inserting unexpected characters (like quotes, backticks or dollar signs) into the input data)
If the commands can only take one parameter at a time:
tr '[\n]' '[\0]' < a.txt | xargs -r0 -n1 /bin/bash -c 'command1 "$#"; command2 "$#";' ''
Or with somewhat less processes:
tr '[\n]' '[\0]' < a.txt | xargs -r0 /bin/bash -c 'for f in "$#"; do command1 "$f"; command2 "$f"; done;' ''
If you have GNU xargs or another with the -P extension and you want to run 32 processes in parallel, each with not more than 10 parameters for each command:
tr '[\n]' '[\0]' < a.txt | xargs -r0 -n10 -P32 /bin/bash -c 'command1 "$#"; command2 "$#";' ''
This should be robust against any special characters in the input. (If the input is null separated.) The tr version will get some invalid input if some of the lines contain newlines, but that is unavoidable with a newline separated file.
The blank first parameter for bash -c is due to this: (From the bash man page) (Thanks #clacke)
-c If the -c option is present, then commands are read from the first non-option argument com‐
mand_string. If there are arguments after the command_string, the first argument is assigned to $0
and any remaining arguments are assigned to the positional parameters. The assignment to $0 sets
the name of the shell, which is used in warning and error messages.
One thing I do is to add to .bashrc/.profile this function:
function each() {
while read line; do
for f in "$#"; do
$f $line
done
done
}
then you can do things like
... | each command1 command2 "command3 has spaces"
which is less verbose than xargs or -exec. You could also modify the function to insert the value from the read at an arbitrary location in the commands to each, if you needed that behavior also.
Another possible solution that works for me is something like -
cat a.txt | xargs bash -c 'command1 $#; command2 $#' bash
Note the 'bash' at the end - I assume it is passed as argv[0] to bash. Without it in this syntax the first parameter to each command is lost. It may be any word.
Example:
cat a.txt | xargs -n 5 bash -c 'echo -n `date +%Y%m%d-%H%M%S:` ; echo " data: " $#; echo "data again: " $#' bash
My current BKM for this is
... | xargs -n1 -I % perl -e 'system("echo 1 %"); system("echo 2 %");'
It is unfortunate that this uses perl, which is less likely to be installed than bash; but it handles more input that the accepted answer. (I welcome a ubiquitous version that does not rely on perl.)
#KeithThompson's suggestion of
... | xargs -I % sh -c 'command1; command2; ...'
is great - unless you have the shell comment character # in your input, in which case part of the first command and all of the second command will be truncated.
Hashes # can be quite common, if the input is derived from a filesystem listing, such as ls or find, and your editor creates temporary files with # in their name.
Example of the problem:
$ bash 1366 $> /bin/ls | cat
#Makefile#
#README#
Makefile
README
Oops, here is the problem:
$ bash 1367 $> ls | xargs -n1 -I % sh -i -c 'echo 1 %; echo 2 %'
1
1
1
1 Makefile
2 Makefile
1 README
2 README
Ahh, that's better:
$ bash 1368 $> ls | xargs -n1 -I % perl -e 'system("echo 1 %"); system("echo 2 %");'
1 #Makefile#
2 #Makefile#
1 #README#
2 #README#
1 Makefile
2 Makefile
1 README
2 README
$ bash 1369 $>
Try this:
git config --global alias.all '!f() { find . -d -name ".git" | sed s/\\/\.git//g | xargs -P10 -I{} git --git-dir={}/.git --work-tree={} $1; }; f'
It runs ten threads in parallel and does what ever git command you want to all repos in the folder structure. No matter if the repo is one or n levels deep.
E.g: git all pull
I have good idea to solve the problem.
Only write a comman mcmd, then you can do
find . -type f | xargs -i mcmd echo {} ## cat {} #pipe sed -n '1,3p'
The mcmd content as follows:
echo $* | sed -e 's/##/\n/g' -e 's/#pipe/|/g' | csh
I found this script on the net, I don't know to work in bash too much is too weird but..
Here's my script:
CONTOR=0
for i in `cat targets`
do
CONTOR=`ps aux | grep -c php`
while [ $CONTOR -ge 250 ];do
CONTOR=`ps aux | grep -c php`
sleep 0.1
done
if [ $CONTOR -le 250 ]; then
php b $i > /dev/null &
fi
done
My targets are urls, and the b php file is a crawler which save some links into a file. The problem is max numbers of threads is 50-60 and that's because the crawler finish very fast and that bash script code doesn't have time to open my all 250 threads. It's any chance to do something to open all threads (250) ? It is possible to run more than one thread per ps -aux process? Right know seems he open 1 thread after execute ps -aux.
First: Bash has no multithreading support whatsoever. foo & starts a separate process, not a thread.
Second: launching ps to check for children is both prone to false positives (treating unrelated invocations of php as if they were jobs in the current process) and extremely inefficient if done in a loop (since every invocation involves a fork()/exec()/wait() cycle).
Thus, don't do it that way: Use a release of GNU xargs with -P, or (if you must) GNU parallel.
Assuming your targets file is newline-delimited, and has no special quoting or characters, this could be as simple as:
xargs -d $'\n' -n 1 -P 250 php b <targets
...or, for pure POSIX shells:
xargs -d "
" -n 1 -P 250 php b <targets
With GNU Parallel it looks like this (choose the style you like best):
cat targets | parallel -P 250 php b
parallel -a targets -P 250 php b
parallel -P 250 php b :::: targets
There is no risk of false positives if there are other php processes running. And unlike xargs there is no risk if the file targets contain space, " or '.
My problem can be summed up by making this simple command works :
nice -n 10 "ls|xargs -I% echo \"%\""
Which fails :
nice: ls|xargs -I% echo "%": No such file or directory
Removing the quotes makes it works, but my point is to wrap multiple quoted commands into one to do something more complex like :
ftphost="192.168.1.1"
dirinputtopush="/tmp/archivedir/"
ftpoutputdir="mydir/"
nice -n 19 ls $dirinputtopush | xargs -I% "lftp $ftphost -e \"mirror -R $dirinputtopush% $ftpoutputdirrecent ;quit\"; sleep 10"
Try using nice -n 10 bash -c 'your; commands | or_complex pipelines' as command. This way bash is the binary and the string after -c contains a sequence interpreted by bash so it can contain pipelines, loops etc. Watch out for proper quoting. You need to do it this way because nice expects a binary, not expressions interpreted by the shell. In contrast, shell builtins such as time (but not /usr/bin/time which is a separate binary) will accept shell expressions as the command to execute. They can because they're built into the shell. nice is not, so it requires a binary to execute.
Children inherit nice value:
nice -n 10 bash -c 'ls | xargs -I% echo %'
Nice each command separately:
nice -n 10 ls | nice -n 10 xargs -I% echo %
I'm trying to use xargs in a shell script to run parallel instances of a function I've defined in the same script. The function times the fetching of a page, and so it's important that the pages are actually fetched concurrently in parallel processes, and not in background processes (if my understanding of this is wrong and there's negligible difference between the two, just let me know).
The function is:
function time_a_url ()
{
oneurltime=$($time_command -p wget -p $1 -O /dev/null 2>&1 1>/dev/null | grep real | cut -d" " -f2)
echo "Fetching $1 took $oneurltime seconds."
}
How does one do this with an xargs pipe in a form that can take number of times to run time_a_url in parallel as an argument? And yes, I know about GNU parallel, I just don't have the privilege to install software where I'm writing this.
Here's a demo of how you might be able to get your function to work:
$ f() { echo "[$#]"; }
$ export -f f
$ echo -e "b 1\nc 2\nd 3 4" | xargs -P 0 -n 1 -I{} bash -c f\ \{\}
[b 1]
[d 3 4]
[c 2]
The keys to making this work are to export the function so the bash that xargs spawns will see it and to escape the space between the function name and the escaped braces. You should be able to adapt this to work in your situation. You'll need to adjust the arguments for -P and -n (or remove them) to suit your needs.
You can probably get rid of the grep and cut. If you're using the Bash builtin time, you can specify an output format using the TIMEFORMAT variable. If you're using GNU /usr/bin/time, you can use the --format argument. Either of these will allow you to drop the -p also.
You can replace this part of your wget command: 2>&1 1>/dev/null with -q. In any case, you have those reversed. The correct order would be >/dev/null 2>&1.
On Mac OS X:
xargs: max. processes must be >0 (for: xargs -P [>0])
f() { echo "[$#]"; }
export -f f
echo -e "b 1\nc 2\nd 3 4" | sed 's/ /\\ /g' | xargs -P 10 -n 1 -I{} bash -c f\ \{\}
echo -e "b 1\nc 2\nd 3 4" | xargs -P 10 -I '{}' bash -c 'f "$#"' arg0 '{}'
If you install GNU Parallel on another system, you will see the functionality is in a single file (called parallel).
You should be able to simply copy that file to your own ~/bin.