$1 &
echo $!
is there a different way to launch a command in the background and return the pid immediately?
So when I launch bash run.sh "python worker.py" it will give me the pid of the launched job.
I am using paramiko, a python library which doesn't work with python worker.py &. so I want to create a bash script which will do this for me on the remote server.
Since you're using bash, you can just get the list of background processes from jobs, and instruct it to return the PID via the -l flag. To quote man bash:
jobs [-lnprs] [ jobspec ... ]
jobs -x command [ args ... ]
The first form lists the active jobs. The options have the
following meanings:
-l List process IDs in addition to the normal information.
So in your case, something like
jobs -l | grep 'worker.py' | awk '{print $2}' would probably give you what you want.
Related
I have running python file "cepusender/main.py" (and another python files). How can I restart/kill only main.py file?
Here's a way (there are many):
ps -ef | grep 'cepusender/main.py' | grep -v grep | awk '{print $2}' | xargs kill
ps is the process snapshot command. -e prints every process on the system, and -f prints the full-format listing, which, germanely, includes the command line arguments of each process.
grep prints lines matching a pattern. We first grep for your file, which will match both the python process and the grep process. We then grep -v (invert match) for grep, paring output down to just the python process.
Output now looks like the following:
user 77864 68024 0 13:53 pts/4 00:00:00 python file.py
Next, we use awk to pull out just the second column of the output, which is the process ID or PID.
Finally we use xargs to pass the PID to kill, which asks the python process to shutdown gracefully.
kill is the command to send signals to processes.
You can use kill -9 PID to kill your python process, where 9 is the number for SIGKILL and PID is the python Process ID.
Killing all celery processes involves 'grep'ing on 'ps' command and run kill command on all PID. Greping
on ps command results in showing up self process info of grep command. In order to skip that PID, 'grep
-v grep' command is piplined which executes 'not' condition on 'grep' search key. The 'awk' command is
used to filter only 2nd and 'tr' command translates result of 'awk' command output from rows to columns.
Piplining 'kill' command did not work and so the entire process has been set as command substitution, i.e.,
'$()'. The redirection at the end is not mandatory. Its only to supress the enitre output into background.
kill -9 $(ps aux | grep celery | grep -v grep | awk '{print $2}' | tr '\n' ' ') > /dev/null 2>&1
Apart from the standard unix ways of killing celery processes, Celery also provides API to kill all the workers listening on a particular broker.
To kill using python. You can refer to the docs here or here. Former calls the latter function internally.
app.control.shutdown()
where app is the celery app instance configured with the broker.
Celery command line interface can also be used for the same.
celery -A app_name control shutdown
Is it possible to tag a python program run from the command line?
Context: Said command will be run with nohup in the background, and will be killed and restarted at midnight via cron. My intention is to pipe ps into egrep for said tag, grab the pid, and kill -9 before restarting.
minimal, complete, and verifiable example
Start a python web server:
$ nohup python -m http.server 8888 &
Add a tag to the command. Note that -tag is just my imagination at work.. this is what I want:
$ nohup python -m http.server 8888 & -tag "ced72ca0-cd19-11ea-87d0-0242ac130003"
grep for tag:
$ ps aux | egrep "ced72ca0-cd19-11ea-87d0-0242ac130003"
...grab the pid from this, and kill -9
Because you're saying that you want to kill the processes through isolated cron jobs at nidnight, I guess that the $! based solutions in the linked questions (like (How to get the process ID to kill a nohup process?)) are no option for you.
In order to identity your HTTP server processes, your idea is to 'tag' them with a unique ID so the cron jobs will find them.
What you could do in your specific case is to make use of the fact that the listening TCP sockets are unique on your given machine, and retrieve the associated pid through netstat.
A bash script along the lines of:
#!/bin/bash
port=${1:-"8888"}
IP=${2:-"0.0.0.0"}
pid=`netstat -antp 2>/dev/null | grep -E "^(\S+\s+){3}$IP:$port\s+\S+\s+LISTEN" | sed -E 's/ˆ(\S+\s+){6}([0-9]+).*$/\2/'`
[[ -n "$pid" ]] && kill -TERM $pid
... that you parameterize with IP and port through your cronjob.
You can put code in file with name ced72ca0-cd19-11ea-87d0-0242ac130003,
#!/bin/bash
python -m http.server
set it executable
chmod +x ced72ca0-cd19-11ea-87d0-0242ac130003
run it
nohup ced72ca0-cd19-11ea-87d0-0242ac130003 &
and then you can kill it
pkill ced72ca0-cd19-11ea-87d0-0242ac130003
or even using only beginning of filename
pkill ced
EDIT:
Because new script doesn't get any arguments so you can run it with any argument(s) - ie. some tag/word
nohup ced72ca0-cd19-11ea-87d0-0242ac130003 hello_world &
and then you can kill it using -f
pkill -f hello_world
or even using part of word
pkill -f hello
pkill -f world
This way you can even use normal name for script and add tag
nohup my_script ced72ca0-cd19-11ea-87d0-0242ac130003 &
and kill with -f
pkill -f ced72ca0-cd19-11ea-87d0-0242ac130003
or using only part of word
pkill -f ced
I have a Python script that runs in an infinite loop (it's a server).
I want to write an AppleScript that will start this script if it isn't started yet, and otherwise force-quit and restart it. This will make it easy for me to make changes to the server code while programming.
Currently I only know how to start it: do shell script "python server.py"
Note that AppleScript's do shell script starts the shell (/bin/sh) in the root directory (/) by default, so you should specify an explicit path to server.py
In the following examples I'll assume directory path ~/srv.
Here's the shell command:
pid=$(pgrep -fx 'python .*/server\.py'); [ "$pid" ] && kill -9 $pid; python ~/srv/server.py
As an AppleScript statement, wrapped in do shell script - note the \-escaped inner " and \ chars.:
do shell script "pid=$(pgrep -fx 'python .*/server\\.py'); [ \"$pid\" ] && kill -9 $pid; python ~/srv/server.py"
pgrep -fx '^python .*/server\.py$' uses pgrep to find your running command by regex against the full command line (-f), requiring a full match (-x), and returns the PID (process ID), if any.
Note that I've used a more abstract regex to underscore the fact that pgrep (always) treats its search term as a regular expression.
To specify the full launch command line as the regex, use python ~/srv/server\.py - note the \-escaping of . for full robustness.
[ "$pid" ] && kill -9 $pid kills the process, if a PID was found ([ "$pid" ] is short for [ -n "$pid" ] and evaluates to true only if $pid is nonempty); -9 sends signal SIGKILL, which forcefully terminates the process.
python ~/srv/server.py then (re)starts your server.
On the shell, if you do ps aux | grep python\ server.py | head -n1, you'll get the ID of the process running server.py. You can then use kill -9 to kill that process and restart it:
kill -9 `ps aux | grep python\ server.py | head -n1 | python -c 'import sys; print(sys.stdin.read().split()[1])'`
That'll kill it. Al you have to do now is to restart it:
python server.py
You can combine the two with &&:
kill -9 `ps aux | grep python\ server.py | head -n1 | python -c 'import sys; print(sys.stdin.read().split()[1])'` && python server.py
Of course, you already know how to put that in a do shell script!
I run a bash script with which start a python script to run in background
#!/bin/bash
python test.py &
So how i can i kill the script with bash script also?
I used the following command to kill but output no process found
killall $(ps aux | grep test.py | grep -v grep | awk '{ print $1 }')
I try to check the running processes by ps aux | less and found that the running script having command of python test.py
Please assist, thank you!
Use pkill command as
pkill -f test.py
(or) a more fool-proof way using pgrep to search for the actual process-id
kill $(pgrep -f 'python test.py')
Or if more than one instance of the running program is identified and all of them needs to be killed, use killall(1) on Linux and BSD
killall test.py
You can use the ! to get the PID of the last command.
I would suggest something similar to the following, that also check if the process you want to run is already running:
#!/bin/bash
if [[ ! -e /tmp/test.py.pid ]]; then # Check if the file already exists
python test.py & #+and if so do not run another process.
echo $! > /tmp/test.py.pid
else
echo -n "ERROR: The process is already running with pid "
cat /tmp/test.py.pid
echo
fi
Then, when you want to kill it:
#!/bin/bash
if [[ -e /tmp/test.py.pid ]]; then # If the file do not exists, then the
kill `cat /tmp/test.py.pid` #+the process is not running. Useless
rm /tmp/test.py.pid #+trying to kill it.
else
echo "test.py is not running"
fi
Of course if the killing must take place some time after the command has been launched, you can put everything in the same script:
#!/bin/bash
python test.py & # This does not check if the command
echo $! > /tmp/test.py.pid #+has already been executed. But,
#+would have problems if more than 1
sleep(<number_of_seconds_to_wait>) #+have been started since the pid file would.
#+be overwritten.
if [[ -e /tmp/test.py.pid ]]; then
kill `cat /tmp/test.py.pid`
else
echo "test.py is not running"
fi
If you want to be able to run more command with the same name simultaneously and be able to kill them selectively, a small edit of the script is needed. Tell me, I will try to help you!
With something like this you are sure you are killing what you want to kill. Commands like pkill or grepping the ps aux can be risky.
ps -ef | grep python
it will return the "pid" then kill the process by
sudo kill -9 pid
eg output of ps command:
user 13035 4729 0 13:44 pts/10 00:00:00 python (here 13035 is pid)
With the use of bashisms.
#!/bin/bash
python test.py &
kill $!
$! is the PID of the last process started in background. You can also save it in another variable if you start multiple scripts in the background.
killall python3
will interrupt any and all python3 scripts running.