I see a lot of people are using the following command:
nohup ./sendData.sh >logs/send.log 2>&1 &
Since the nohup has already been used, do I still need to add & on the trail?
nohup tells the terminal to not hang up, even when the parent terminal is closed. & tells the terminal to send the process to the background. So
command &
will send the command to the background. You can then use the shell to invoke other commands. If the terminal is exited, the jobs are killed.
nohup command
will not send the command to the background, but when you close the terminal, the process will live on until it is finished or killed.
nohup command &
will send the command to the background and the process will not be killed even when you close the terminal. Usually when using nohup, one also wants the process to be in the background, that's why you see the combination of nohup and & more often than not.
Related
I have a bash script:
node web/dist/web/src/app.js & node api/dist/api/src/app.js &
$SHELL
It successfully starts both my node servers. However:
I do not receive any output (from console.log etc) in my terminal window
If I cancel by (Ctrl +C) the processes are not exited, so then I annoyingly have to manually do a taskkill /F /PID etc afterwards.
Is there anyway around this?
The reason you can't stop your background jobs with Ctrl+C is because signals (SIGINT in this case) are received only by the foreground process.
When your foreground process (the non-interactive main script) exits, its children processes become orphans which are immediately adopted by the init process. To kill them, you need their PIDs. (When you run a background process in an interactive shell, it will receive the SIGHUP, and probably exit, when shell exits.)
The solution in your case is to make your script wait for its children, using the shell built-in wait command. wait will ensure your script receives the SIGINT, which you can then handle (with trap) and kill the background jobs (with kill 0):
#!/bin/bash
trap 'kill 0' EXIT
node app1.js &
node app2.js &
wait
By setting trap on EXIT (special pseudo-signal in bash), you'll ensure background processes will terminate whenever your main script exits (either by Ctrl+C/SIGINT, or by any other signal like SIGTERM, SIGHUP, SIGKILL). The kill 0 command kills all processes in the current process group.
Regarding the output -- on Linux, background processes will inherit the standard output/error from shell (if not redirected), and continue to write to your TTY/terminal. If that's not working on Windows, I'm not sure why not.
However, even if your background processes somehow lost their way to your TTY, you can, as a workaround, append to a log file:
node app1.js >>/path/to/file.log 2>&1 &
node app2.js >>/path/to/file.log 2>&1 &
and then tail -f that log file, either in this, or some other terminal:
tail -f /path/to/file.log
I want to run a command silently via ssh and exit the shell, but the program should continue running.
I tried screen and nohup, but apparently with those it executes 3 processes instead of 1:
user:/bin/bash ./[script]
root: sudo [commandInTheScript]
root: [commandInTheScript]
What am I doing wrong?
P.S.: The thing is that I want to run this command with the Workflow app (iOS), but the app waits until the command is finished, so it freezes 'forever'
To run your process back ground, at end of the command you have to use &.
In your case, you have to run without session since you are planning to exit from ssh after execute the command, so you need nohup
nohup <command> &
nohup < command > &
This makes your command runs on background and shows its PID
How did you use nohup?
Eg.
nohup ruby server.rb &
Ampersand (&) is necessary to let command run in the background.
I'm running some script on remote server using ssh. The task is downloading images to remote server. I'm wondering will the script keep running after I log out the ssh session? Why? Could anyone explain in detail?
If you want the script keep running after logout you need to detach it from the terminal and run it in the background:
nohup ./script.sh &
If you close the terminal where you launched a process in, the process will receive SIGHUP and unless it handles it this means the process will get terminated. HUP means hang up, like in a phone call.
The nohup command can be used to start a process and prevent it from SIGHUP signals getting send to it. An alternative would be to use the bash builtin disown, which does basically the same:
./script.sh &
disown %1
Note that the 1 represents the job id. If you running multiple processes in the background you need to specify the correct job id.
I'm using reptyr to bring a background process to a new terminal.
The description says reptyr should recognize ctrl+z, but it doesn't.
I'm wondering how I can set the process I just brought to the terminal, back to the background.
What I had in mind was:
reptyr <processid>
(ctrl+z)
bg
disown
on the other console you can run:
nohup ./reptyr $PID_OF_PREVOUS_REPTYL_COMMAND &> /dev/null &
$PID_OF_PREVOUS_REPTYL_COMMAND you can retrieve from ps fx command
I have a PHP script with infinite loop. I need this script running forever. So, I run
php /path/to/script.php > /dev/null &
And it works in background in my current user's security context. But when I close terminal window (log off), of course, CentOS Linux kills my program.
I see two guesses: run from a different user in background or make a daemon. I need help in each situation.
Thanks a lot!
nohup is your friend.
nohup command &
I think the general solution to that is nohup:
nohup is a POSIX command to ignore the HUP (hangup) signal, enabling the command to keep running after the user who issues the command has logged out. The HUP (hangup) signal is by convention the way a terminal warns depending processes of logout.
nohup is most often used to run commands in the background as daemons. Output that would normally go to the terminal goes to a file called nohup.out if it has not already been redirected. This command is very helpful when there is a need to run numerous batch jobs which are inter-dependent.
nohup is your friend.
You could:
Install screen and run the command from there. screen is a persistent terminal session that you can leave running.
Write an init/upstart (whatever you use) script so it loads on boot
Use the pear lib system_daemon
Use cron if batch work fits the scenario better (just remember to check for running instances before you launch another, iff concurrency is an issue)
Edit: or as everybody else and their brother has just said, nohup
Using command
nohup your_command &
For example
nohup phantomjs highcharts-convert.js -host 127.0.0.1 -port 3003 &
here "phantomjs highcharts-convert.js -host 127.0.0.1 -port 3003" was my command