I want to use GNU parallel to execute some commands in parallel, but I don't know how to send the argument to my bash scripts.
My bash_script.sh:
scp $1.zip [email protected]:~/
scp $1.zip [email protected]:~/
scp $1.zip [email protected]:~/
scp $1.zip [email protected]:~/
scp $1.zip [email protected]:~/
I can send the argument to the bash_script and execute it in sequence.
bash bash_script.sh argument
but how to do it in parallel?
parallel -a bash_script argument
The solution that worked for me is writing the commands to a text file (e.g: commands.txt) and the executing parallel:
parallel -j 2 < commands.txt
parallel can execute individual processes in parallel, but a script is usually intended to be run as a single process.
You could do
parallel scp "$1" xxx@{}1.xyz.com: ::: {1..5}
or if you really wanted to split the script file into individual lines
sed "s#\$1#$1#g" bash_script.sh | parallel
If the script was just static text, you could simply have parallel read it line by line; but parallel doesn't have a value for $1 in that context, so we substitute it in, and then piped the substituted lines for parallel execution.
Tangentially, notice proper variable quoting and don't put a .sh extension on a Bash script. (Don't put any extension on any script, basically.)
Maybe something like this:
for server in server1 server2 server 3
do
scp "$1" $server: &
done
The & causes the job to run in the background. If I'm not understanding your requirements, please clarify.
bash_script.sh
parallel scp "$1" xxx@{}.com: ::: {1..5}
Usage:
bash bash_script.sh argument
~/ is redundant because scp by default copies to the user's home directory.)