I need to execute a command with a timeout in OpenWRT, but it seems that the command timeout is not installed by default neither can be installed using opkg. I know that I can do a work around (using command &; sleep $DELAY; kill $!), but I wish to do this more properly without getting the risk of kill trying to kill a process in case the command finished before the timeout.
Yes you can install timeout on openWRT
$ opkg update
$ opkg install coreutils-timeout
$ timeout 2 sleep 10
This has been tested with AA, pretty sure that would also work with BB.
In short: it is not possible. I have to do it using sleep && kill.
timeout is a shell command so it executes in a subshell
timeout 6 sleep 20 will work if executed in direct shell terminal but same command won't work if initiated from a shell script.
So to run timeout in a shell script , use like this
out="$(timeout 6 sleep 20)"
OR
echo "$(timeout 10 sleep 20)"
this will run your timeout and your command in one subshell
Related
I am trying to open appium, avd, and ride to run a test in Jeninks. I have the following in execute windows batch command:
C:
cd C:\Users\Gebruiker
appium
adb start-server
cd C:\Users\Gebruiker\AppData\Local\Android\Sdk\emulator
emulator -avd Pixel_3_API_23
ride.py
D:
cd D:\RIDE - testproject\Avans-One
robot -d Open_settings_connect_wifi AvansOne
It keeps loading after opening Appium, the first step. I have tried seperating the steps but it still doesnt work. It completes the first step and then just keeps loading. How do i fix this?
Most probably, this is happening because the command used to start the app is not returning while the app is running. So you will have to use a command that allows you to run the application in the background.
Check with something like below.
START /B title program
e.g.:
C:
cd C:\Users\Gebruiker
START /B "" appium
You can read more about the background process here
The field Exec command of execute shell over ssh, I only put command sudo su in there.
But when I build the project, it always stop at the command sudo su till timeout.
The image show what I get in console output.
Welcome to StackOverflow.
sudo
sudo is used for custom installations, system files modification, etc
Interactive shells (human in front of monitor)
When you exec some command preceded by sudo, this shell will ask you for the password of some privileged user. This ask event waits for user response:
Background executions
Some process, crons, scripts or applications like jenkins needs to execute sudo commands. But in this mode , there is no way to enter the password. This is the reason of your timeout
Solutions
disable requiretty
https://gist.github.com/jrichardsz/1adaaa07885b45d497b519431701a943
Security risk of this approach:
https://unix.stackexchange.com/a/65789/188975
use a pseudo TTY in Jenkins
https://unix.stackexchange.com/a/373843/188975
You can try using the expect directive.
I have a server written in Erlang, compiled with Rebar, and I make a release with Relx. Starts nicely with
/root/rel/share3/bin/share3 start
The next step is to start when the server boots.
I have tried different approaches, the last one is using the /etc/init.d/skeleton where I changed the following
NAME=share3
DAEMON=/root/rel/share3/bin/share3
DAEMON_ARGS="$1"
After that, I run update-rc.d, but I have not gotten it too work. (Ubuntu 14.04)
The service runs until the machine reboots, and I need to login and start it again.
For Windows, it is really elegant, since it can create the Windows service.
Ubuntu uses upstart as init system, so you could try something like that:
description "Start my awesome service"
start on runlevel [2345]
stop on runlevel [!2345]
respawn
exec /root/rel/share3/bin/share3
You have to place this script in /etc/init/ directory with '.conf' extension like '/etc/init/share3.coinf'. To start it invoke sudo start share3.
At last, I solved it!
I have told to relx to place the result at /home/mattias/rel. The script from relx is /home/mattias/rel/share3/bin/share3
Replace the row
SCRIPT_DIR="$(dirname "$0")"
by (you need to fix the path /home/mattias/rel)
HOME=/home/mattias
export HOME
SCRIPT_DIR="/home/mattias/rel/share3/bin"
Copy the file to /etc/init.d/share3 using
sudo cp ~/rel/share3/bin/share3 /etc/init.d/
Test that it works using
/etc/init.d/share3 start
and
/etc/init.d/share3 stop
In order to make it start at boot, install sysv-rc-conf
sudo apt-get install sysv-rc-conf
Enable boot at start using
sudo sysv-rc-conf share3 on
and disable
sudo sysv-rc-conf share3 off
Alternatives are welcome.
I'm currently struggling with executing a simple command which I know works when I run it manually when logged in as either root or non-root user:
god -c path/to/app/queue_worker.god
I'm trying to run this when the server starts (I'm running Ubuntu 12.04), and I've investigated adding it to /etc/rc.local just to see if it runs. I know I can add it to /etc/init.d and then use update-rc.d but as far as I understand it's basically the same thing.
My question is how I run this command after everything has booted up as clean as possible without any fuzz.
Im probably missing something in the lifecycle of how everything's initialized, but then I gladly encourage some education! Are there alternative ways or places of putting this command?
Thanks!
You could write a bash script to determine when Apache has started and then set it to run as a cron job at a set interval...
if [ "$(pidof apache)" ]
then
# process was found
else
# process not found
fi
Of course then you'll have a useless cron job running all the time, and you'll have to somehow flip a switch once it's run once so it doesn't run again.. This should give you an idea to start from..
Struggling with cron jobs. Ubuntu 11.10 on the server.
Until recently had whenever cron jobs running successfully several times a day; then due to another problem I had to remove RVM from the server and go back to ruby 1.9.3 installed without RVM (I'm sure this is something to do with it)
There is no .rvmrc file in my app
Now, the cron jobs are somehow failing as I can see from syslog:
Jun 30 08:03:01 ip-10-251-30-96 CRON[18706]: (ubuntu) CMD (/bin/bash -l -c 'cd /var/www/my_app/app/releases/201300629090954 && script/rails runner -e production '\''User.remind_non_confirmed_users'\''')
Jun 30 08:03:01 ip-10-251-30-96 CRON[18705]: (CRON) error (grandchild #18706 failed with exit status 127)
Jun 30 08:03:01 ip-10-251-30-96 CRON[18705]: (CRON) info (No MTA installed, discarding output)
If I run that command manually (with env - /bin/bash -l -c '...' ) it runs fine..
I'm going to add "set :output, 'tmp/whenever.log'" to whenever to see what is going on, but I suspect it is an issue with the ruby version / path or something.
Any idea how I could diagnose / fix this properly??
this is my cron/whenever job:
3 8 * * * /bin/bash -l -c 'cd /var/www/my_app/app/releases/20130629090954 && script/rails runner -e production '\''User.remind_non_confirmed_users'\'''
many thanks
To help diagnose what's going on, I usually capture the cron output into a separate log file. There's probably an error that's just not being recorded anywhere.
#hourly bash -lc 'cd /path/to/app; RAILS_ENV=production bundle exec rake remind_non_confirmed_users' >> /path/to/app/log/tasks.log
Also, I prefer creating rake tasks for cron jobs as opposed to runners. A little easier to invoke via the command line than runners, for me at least.
I'm still not sure what was going on, running Whatever with 'set :output' should have created log files, but it didn't, yet the jobs are still failing (and write permissions were there for the log files).
I got so fed up I redeveloped the solution without using script/runner, in stead have cron just call a URL that then takes care of matters as a delayed job. For our particular situation this has a number of additional benefits, though I know it is not ideal for many.
thanks for the suggestions