I created a small go program that uses these go bindings to record some commands from the default microphone and act accordingly. It works fine as a standalone binary(both as normal user and root user) but when I tried to convert it into a systemd unit the Capture function in the go bindings failed with error saying connection refused.
The program is failing to capture the microphone input when running as a systemd service. The following is the unit file which is pretty much copy-pasted from here.
[Unit]
Description=Commander service providing voice commands
[Service]
ExecStart=/path/to/binary/binary.sh
[Install]
WantedBy=multi-user.target
binary.sh is a simple shell script that for providing environment variables to the go binary. The script is below
#!/bin/bash
cd /path/to/binary
env LD_LIBRARY_PATH=/usr/local/lib ./binary > stdout 2> stderr
The LD_LIBRARY_PATH is needed for pocketsphinx shared libraries which is being used for recognizing commands from audio.
I think something is wrong with the unit file but I don't know what it is. The whole project can be found here.
Any insight would be helpful. Thanks.
pulseaudio is a daemon which probably has yet to get launched ... update your binary.sh with a
sleep 10
echo about to determine whether pulseaudio is up and running
pulseaudio --check
echo if pulseaudio is running above check silently returns
prior to launching your golang binary
It is required to set XDG_RUNTIME_DIR environmental variable to directory /run/user/<USER_ID>, e.g.
[Service]
Environment="XDG_RUNTIME_DIR=/run/user/1001"
ExecStart=/path/to/binary/binary.sh
Related
I have a server built on bottle that works great when launched from userland. The server appears on port 8088 and appears to be communicating to the outside world, but when I contact the app all I get is the very informative "Critical error while processing request:schema" which is the url of the app.
My systemd file is below:
[Unit]
Description=Survey Service
After=multi-user.target
Conflicts=getty#tty1.service
[Service]
User=ubuntu
Type=simple
Working-directory=/home/ubuntu/survey
ExecStart=/usr/bin/python3 /home/ubuntu/survey/server.py
[Install]
WantedBy=multi-user.target
I've found several articles related to the informative error message, but none related with systemd. As I said, the app runs perfectly when launched as user ubuntu in the project directory with the very simple command "python3 server.py" but seems to be missing... something when systemd tries to launch it.
Systemd reports the process is running and, as I said, I'm able to connect to the app... it just fails in an orderly fashion with this message, and I'm lost as to why. I suspect a permissions problem, but doesn't "user" and "Working-directory" take care of that? All files used by the app are in that directory or directories below it.
Apparently doing it the old fashion way works: set systemd to run a bash script as such:
cat /home/ubuntu/survey/server.sh
#!/bin/bash
cd /home/ubuntu/survey/
python3 server.py
Works just great. So my question now becomes one about systemd: what is the point of "Working-directory" if it does not actually set to that working directory?
I am having a problem with a small application I developed on the BBB running Debian Image 2017-03-19.
I connected a barcode scanner to the usb port and a 2x16 LCD display to the GPIO controlled by BBBioLib.
I developed an application in C to read a barcode label apply to a race tyre, which find a match on an SQLite table and show the racer name on the display.
Application work great but since the all assembly must work stand alone I need to run the program automatically at boot.
I follow all the instruction on creating a bash program and service but I am getting a strange behaviour.
The display after the welcome message hang up and never change but the application work correctly because all the printf to the consolle get logged correctly and once I exit the application I can check them on the log of the service.
If I restart the service manually everything work fine.
This is the bash script
#!/bin/bash
/root/read_barcode
This is the service code
[Unit]
Description=Barcode reader launch
After=syslog.target network.target
[Service]
Type=simple
ExecStart=/usr/bin/barcode.sh
[Install]
WantedBy=multi-user.target
Does anyone can help on solving this problem.
Thanks
Carlo
Run a .service file with sudo systemctl enable YourService.service at this location.
/etc/systemd/system/
Use the enable option for systemd .service files to make your source work on boot or a reboot.
I'd like to run syncthing as a daemon, trying to follow this hint here from the syncthing manual.
I'm running on Fedora 25 and syncthing 0.14.24.
The executable is pointed to via a symlink in /usr/bin/syncthing which can be executed by any user (tested this successfully).
To enable the service, I'm typing (myuser is replaced with my actual username in all of the below):
sudo systemctl enable sycnthing#myuser.service
Which returns:
Failed to lookup unit file state: Invalid argument
I don't understand what the error message means. How could I get to run syncthing as a daemon?
syncthing#myuser.service:
[Unit]
Description=Syncthing - Open Source Continuous File Synchronization for %I
Documentation=man:syncthing(1)
After=network.target
Wants=syncthing-inotify#myuser.service # I also commented this line out; didn't have an effect
[Service]
User=%i
ExecStart=/usr/bin/syncthing -no-browser -no-restart -logflags=0
Restart=on-failure
SuccessExitStatus=3 4
RestartForceExitStatus=3 4
[Install]
WantedBy=multi-user.target
I think myuser should be substituted for your actual username.
Arch wiki has a pretty good article:
System service
Running Syncthing as a system service ensures that it is running at startup even if the user has no active session, it is intended to be used on a server.
Enable and start the syncthing#myuser.service where myuser is the actual name of your user.
Credit: https://wiki.archlinux.org/index.php/Syncthing
I'm trying to create a service that will trigger every time a raspberry pi boots. Currently the service runs a really simple script that sends a POST request to a web service endpoint I control. I can trigger said script manually and that part all works perfectly.
I'm struggling with the next step which is to get that script to run after the pi has finished booting. I also need to be able to get it to run without a user logging in.
CURL Script (algiers-startup.local)
#! /bin/bash
echo "Attempting CURL Request"
curl --data "param1=value1¶m2=value2" http://10.68.159.28:3000/device
Systemd Service
[Unit]
Description=Algiers RaspberryPi Startup
After=network.target
Before=getty#tty1.service
[Service]
Type=oneshot
ExecStart=/usr/local/sbin/algiers-startup.local
TimeoutSec=30
StandardOutput=tty
RemainAfterExit=no
[Install]
WantedBy=multi-user.target
I see no errors or outputs in the console, no hint that anything has happened at all.
I’ll assume your machine is already set up with Systemd. It’s controlled primarily through the systemctl command. I alias it as such since it’s awful to type all the time:
alias sc=systemctl
alias ssc='sudo systemctl'
You just need to “enable” your service to have it run at boot:
sc enable algiers-startup
I’m not sure what distro you’re using, but on Arch and CentOS, you’ll want algiers-startup to live down in /usr/lib/systemd/system/.
You can test your service with sc start algiers-start. journalctl can show you what’s happening.
I have a python script which manages an Erlang daemon. Everything works fine when used through a shell once the system is initialized.
Now, when I included the same script under "/etc/init.d" and with the symlinks properly set in "/etc/rcX.d", the python script still works but my Erlang daemon fails to start and leaves no discernible traces (e.g. crash_dump, dmesg etc.)
I also tried setting the environment variable "HOME" through 'erl -env HOME /root' and still no luck.
Any clues?
To manually run the script the same way the system does, use service daemon start if you have that command, or else try
cd /
env -i LANG="$LANG" PATH="$PATH" TERM="$TERM" /etc/init.d/daemon start
That forces the script to run with a known, minimal environment just like it would at startup.
Thanks for this answer - I was having a devil of a time starting the "Alice" RESTful interface to rabbitmq on startup. The key was using 'env HOME=/root /path/to/alice/startup/script' in my init script.