While loop in bash script breaks systemd service - linux

I'm on Debian and I have a systemd service that calls a bash script.
The script contains an infinite while loop, as I need it to check something every X seconds infinitely.
The systemd service crashes once it hits the "while true; do" line.
The script runs fine if I execute it manually.
Why doesn't systemd like it? What do I do?
Here are the service and the script. As I've indicated, an echo statement before the "while true; do" prints. The echo statement after the "while true; do" line does not print.
/etc/systemd/system/stream.service:
[Service]
WorkingDirectory=/home/pi/
ExecStart=/home/pi/joi_main.sh
Restart=no
StandardOutput=syslog
StandardError=syslog
SyslogIdentifier=stream_service
User=pi
Group=pi
Environment=NODE_ENV=production
[Install]
WantedBy=multi-user.target
/home/pi/joi_main.sh:
#!/bin/bash -e
today=`/bin/date '+%Y_%m_%d__%H_%M_%S'`
exec 2> "/home/pi/stream_logs/$today.$RANDOM.log"
exec 1>&2
#Wait 120s for system to finish booting
sleep 120
#Initial config
export AUDIODEV=mic_mono
export AUDIODRIVER=alsa
sudo sysctl fs.pipe-max-size=1048576
echo "This line prints"
# Check if video buffer is full every minute. if full, the stream needs to restart
while true; do
echo "This line doesn't"
if grep "100% full" /home/pi/video_buffer_usage.txt; then
echo "Buffer is full!"
# Kill existing processes
pkill -f “raspivid|rec|buffer|ffmpeg”
# Wait 10s
sleep 10
./joi_stream.sh &
fi
sleep 60
done
Journalctl seems completely unhelpful, but here it is. No errors. Why is "session closed"?
Mar 31 02:13:41 raspberrypi sudo[1369]: pi : TTY=unknown ; PWD=/home/pi ; USER=root ; COMMAND=/sbin/sysctl fs.pipe-max-size=1048576
Mar 31 02:13:41 raspberrypi sudo[1369]: pam_unix(sudo:session): session opened for user root by (uid=0)
Mar 31 02:13:41 raspberrypi sudo[1369]: pam_unix(sudo:session): session closed for user root
(Please don't tell me to start yet another systemd service for just this while loop. I want it to be part of this main script because it needs to run after everything else, and if I turn off the main service I don't want the while loop running either, so maintaining two systemd services would only add troube.)

The contents of ./joi_stream.sh were not shared, but here's a problem I see with your systemd solution. It doesn't directly explain your behavior, but may be related:
In your systemd configuration, you redirect both STDOUT and STDERR to syslog, but in your script, you redirect STDERR (file descriptor "2") to a file, and redirect STDOUT (file descriptor "1') to STDERR.
exec 2> "/home/pi/stream_logs/$today.$RANDOM.log"
exec 1>&2
If your ./joi_stream.sh expected your redirection of these file descriptors to another file to work, it may not. If the file is just for logging, I would get rid of these lines and let the systemd journal handle that-- it will tag the logs with your unit you can review your logs specifically:
journalctl -u your-unit-name.service
Also, in systemd, you wouldn't normally put in a sleep to wait until the systemd has booted. Instead, you would use a .timer unit.
The .timer file would instruct to run the main logic every minute, so the "while" loop would not be required. The timer unit would contain directives like:
# Run for the first time 2 minutes after boot
# and every minute after that
OnBootSec=120
OnUnitActiveSec=60
It would be timer unit which is enabled to start on boot. Timer files can be super-simple. Just create a .timer file in /etc/systemd/system and give it the same name as the service file you want it to activate:
[Unit]
Description=Runs my service every minute
[Timer]
# Run for the first time 2 minutes after boot
# and every minute after that
OnBootSec=120
OnUnitActiveSec=60
[Install]
WantedBy=timers.target
To start and test your timer immediately, run:
sudo systemctl start my-service.timer
You can review the status of timers with:
sudo systemctl list-timers
The systemd solution is more robust than the rc.local solution. If your rc.local solution dies for any reason, it will not restart. However, if your script dies will run under systemd, the timer will still run it again a minute later.

FYI, everything works if I call /home/pi/joi_main.sh from /etc/rc.local instead of using a systemd service. I'll use rc.local and kill the service.

Related

systemctl Exec format error when trying to run service

Currently I wanted to run my dedicated server on my vps. When I run system systemctl start csgo.service it gives me error Load: error (Reason: Exec format error) when I run systemctl status csgo.service it gives me /lib/systemd/system/csgo.service:12: Executable path is not absolute: killall -TERM srcds_linux. Below are the service file that I am trying to run, am I making any mistake since it says format error?
[Unit]
Description=CSGO Server
[Service]
Type=simple
User=steam
Group=steam
Restart=on-failure
RestartSec=5
StartLimitInterval=60s
StartLimitBurst=3
ExecStart=/home/steam/steamcmd/csgo/srcds_run -game csgo -console -usercon +game_type 0 +game_mode 1 -tickrate 128 +mapgroup mg_active +map de_dust2 +sv_setsteamaccount GsltKeyHere -net_port_try 1
ExecStop=killall -TERM srcds_linux
[Install]
WantedBy=multi-user.target
My dedicated server files are inside /home/steam/steamcmd/csgo
Quoting the manual on unit files:
Note that shell command lines are not directly supported. If shell command lines are to be used, they need to be passed explicitly to a shell implementation of some kind.
Example: ExecStart=sh -c 'dmesg | tac'
You'll need to either use sh like that or figure out the actual path to your killall executable, e.g.
[Unit]
ExecStop=sh -c 'killall -TERM srcds_linux'
or
[Unit]
ExecStop=/sbin/killall -TERM srcds_linux
As an aside, that's not the best of ExecStop commands; it'll ruthlessly kill all srcds_linux executables, no matter if they're related to this service or not. Having no ExecStop command will have systemd terminate the service by itself:
Note that it is usually not sufficient to specify a command for this setting that only asks the service to terminate (for example, by queuing some form of termination signal for it), but does not wait for it to do so. Since the remaining processes of the services are killed according to KillMode= and KillSignal= as described above immediately after the command exited, this may not result in a clean stop. The specified command should hence be a synchronous operation, not an asynchronous one.

How to debug an upstart script that intermittently fails?

I have a process that I want to start as soon my system is rebooted by whatever means so I was using upstart script for that but sometimes what I am noticing is my process doesn't get started up during hard reboot (plugging off and starting the machine) so I think my upstart script is not getting kicked in after hard reboot. I believe there is no runlevel for Hard Reboot.
I am confuse that why sometimes during reboot it works, but sometimes it doesn't work. And how can I debug this out?
Below is my upstart script:
# sudo start helper
# sudo stop helper
# sudo status helper
start on runlevel [2345]
stop on runlevel [!2345]
chdir /data
respawn
pre-start script
echo "[`date`] Agent Starting" >> /data/agent.log
sleep 30
end script
post-stop script
echo "[`date`] Agent Stopping" >> /data/agent.log
sleep 30
end script
limit core unlimited unlimited
limit nofile 100000 100000
setuid goldy
exec python helper.py
Is there any way to debug this out what's happening? I can easily reproduce this I believe. Any pointers on what I can do here?
Note:
During reboot sometimes I see the logging that I have in pre-start script but sometimes I don't see the logging at all after reboot and that means my upstart script was not triggered. Is there anything I need to change on runlevel to make it work?
I have a VM which is running in a Hypervisor and I am working with Ubuntu.
Your process running nicely, BUT during system startup many things go parallel.
IF mount (which makes available the /data folder) runs later than your pre-start script you will not see the "results" of pre-start script.
I suggest to move sleep 30 earlier (BTW 30 secs seems too looong):
pre-start script
sleep 30 # sleep 10 should be enough
echo "[`date`] Agent Starting" >> /data/agent.log
end script

How to redirect output of systemd service to a file

I am trying to redirect output of a systemd service to a file but it doesn't seem to work:
[Unit]
Description=customprocess
After=network.target
[Service]
Type=forking
ExecStart=/usr/local/bin/binary1 agent -config-dir /etc/sample.d/server
StandardOutput=/var/log1.log
StandardError=/var/log2.log
Restart=always
[Install]
WantedBy=multi-user.target
Please correct my approach.
I think there's a more elegant way to solve the problem: send the stdout/stderr to syslog with an identifier and instruct your syslog manager to split its output by program name.
Use the following properties in your systemd service unit file:
StandardOutput=syslog
StandardError=syslog
SyslogIdentifier=<your program identifier> # without any quote
Then, assuming your distribution is using rsyslog to manage syslogs, create a file in /etc/rsyslog.d/<new_file>.conf with the following content:
if $programname == '<your program identifier>' then /path/to/log/file.log
& stop
Now make the log file writable by syslog:
# ls -alth /var/log/syslog
-rw-r----- 1 syslog adm 439K Mar 5 19:35 /var/log/syslog
# chown syslog:adm /path/to/log/file.log
Restart rsyslog (sudo systemctl restart rsyslog) and enjoy! Your program stdout/stderr will still be available through journalctl (sudo journalctl -u <your program identifier>) but they will also be available in your file of choice.
Source via archive.org
If you have a newer distro with a newer systemd (systemd version 236 or newer), you can set the values of StandardOutput or StandardError to file:YOUR_ABSPATH_FILENAME.
Long story:
In newer versions of systemd there is a relatively new option (the github request is from 2016 ish and the enhancement is merged/closed 2017 ish) where you can set the values of StandardOutput or StandardError to file:YOUR_ABSPATH_FILENAME. The file:path option is documented in the most recent systemd.exec man page.
This new feature is relatively new and so is not available for older distros like centos-7 (or any centos before that).
I would suggest adding stdout and stderr file in systemd service file itself.
Referring : https://www.freedesktop.org/software/systemd/man/systemd.exec.html#StandardOutput=
As you have configured it should not like:
StandardOutput=/home/user/log1.log
StandardError=/home/user/log2.log
It should be:
StandardOutput=file:/home/user/log1.log
StandardError=file:/home/user/log2.log
This works when you don't want to restart the service again and again.
This will create a new file and does not append to the existing file.
Use Instead:
StandardOutput=append:/home/user/log1.log
StandardError=append:/home/user/log2.log
NOTE: Make sure you create the directory already. I guess it does not support to create a directory.
You possibly get this error:
Failed to parse output specifier, ignoring: /var/log1.log
From the systemd.exec(5) man page:
StandardOutput=
Controls where file descriptor 1 (STDOUT) of the executed processes is connected to. Takes one of inherit, null, tty, journal, syslog, kmsg, journal+console, syslog+console, kmsg+console or socket.
The systemd.exec(5) man page explains other options related to logging. See also the systemd.service(5) and systemd.unit(5) man pages.
Or maybe you can try things like this (all on one line):
ExecStart=/bin/sh -c '/usr/local/bin/binary1 agent -config-dir /etc/sample.d/server 2>&1 > /var/log.log'
If for a some reason can't use rsyslog, this will do:
ExecStart=/bin/bash -ce "exec /usr/local/bin/binary1 agent -config-dir /etc/sample.d/server >> /var/log/agent.log 2>&1"
Short answer:
StandardOutput=file:/var/log1.log
StandardError=file:/var/log2.log
If you don't want the files to be cleared every time the service is run, use append instead:
StandardOutput=append:/var/log1.log
StandardError=append:/var/log2.log
We are using Centos7, spring boot application with systemd. I was running java as below. and setting StandardOutput to file was not working for me.
ExecStart=/bin/java -jar xxx.jar -Xmx512-Xms32M
Below workaround solution working without setting StandardOutput. running java through sh as below.
ExecStart=/bin/sh -c 'exec /bin/java -jar xxx.jar -Xmx512M -Xms32M >> /data/logs/xxx.log 2>&1'
Assume logs are already put to stdout/stderr, and have systemd unit's log in /var/log/syslog
journalctl -u unitxxx.service
Jun 30 13:51:46 host unitxxx[1437]: time="2018-06-30T11:51:46Z" level=info msg="127.0.0.1
Jun 30 15:02:15 host unitxxx[1437]: time="2018-06-30T13:02:15Z" level=info msg="127.0.0.1
Jun 30 15:33:02 host unitxxx[1437]: time="2018-06-30T13:33:02Z" level=info msg="127.0.0.1
Jun 30 15:56:31 host unitxxx[1437]: time="2018-06-30T13:56:31Z" level=info msg="127.0.0.1
Config rsyslog (System Logging Service)
# Create directory for log file
mkdir /var/log/unitxxx
# Then add config file /etc/rsyslog.d/unitxxx.conf
if $programname == 'unitxxx' then /var/log/unitxxx/unitxxx.log
& stop
Restart rsyslog
systemctl restart rsyslog.service
In my case 2>&1(stdout and stderr file descriptor symbol) had to be placed correctly,then log redirection worked as I expected
[Unit]
Description=events-server
[Service]
User=manjunath
Type=simple
ExecStart=/bin/bash -c '/opt/events-server/bin/start.sh my-conf 2>&1 >> /var/log/events-server/events.log'
[Install]
WantedBy=multi-user.target
Make your service file call a shell script instead of running the app directly. This way you have extra control. For example, you can make output files like those in /var/log/
Make a shell script like /opt/myapp/myapp.sh
#!/bin/sh
/usr/sbin/logrotate --force /opt/myapp/myapp.conf --state /opt/myapp/state.tmp
logger "[myapp] Run" # send a marker to syslog
myapp > /opt/myapp/myapp.log 2>&1 &
And your service file myapp.service contains:
...
[Service]
Type=forking
ExecStart=/bin/sh -c /opt/myapp/myapp.sh
...
A sample of log config file /opt/myapp/myapp.conf
/opt/myapp/myapp.log {
daily
rotate 20
missingok
compress
}
Then you will get myapp.log, and zipped myapp.log.1.gz ... for each time the service was started, and previous zipped.

Arch Linux / systemd - prevent any kind of shutdown/rebboot

I'm running Arch-based Manjaro Linux and wrote myself a little update program, that starts every 7 hours and runs completely in the background. This update program is started by systemd.
What I wanna know is: How can I prevent any system shutdown/reboot during the time this program runs no matter if the user just wants to turn it off or any program wants to do so.
The best would be, if any shutdown/reboot action wouldn't be cancelled but delayed instead, so when the update program has finished its run, the shutdown/reboot continues.
My systemd parts are:
uupgrades.timer
[Unit]
Description=UU Upgrades Timer
[Timer]
OnBootSec=23min
OnUnitActiveSec=7h
Unit=uupgrades.target
[Install]
WantedBy=basic.target
uupgrades.target
[Unit]
Description=UU Upgrades Timer Target
StopWhenUnneeded=yes
and in the folder uupgrades.target.wants
uupgrades.service
[Unit]
Description=UU Update Program
[Service]
Nice=19
IOSchedulingClass=2
IOSchedulingPriority=7
ExecStart=/usr/bin/uupgrades
How can I achieve this?
If a user with sufficient permissions to reboot the server or manipulate processes wants to stop or reboot the machine you cant stop them. That's just how linux works. You should set up permissions and accounts such that no other users have root permissions or permissions sufficient to manipulate the process or user that the process is running as.
When I want to block myself from rebooting or shutdown, I alias my usual shutdown and reboot aliases to beep;beep;beep;.
In multiuser environments you could move the reboot, shutdown etc. binaries and move them back, when shutdown should be allowed again.
You could also temporarily move an executable shell script outputting information about the postponed shutdown possibility in place of the corresponding binaries. This script could set a flag, if a shutdown was requested.
Q&D example script:
#!/usr/bin/env bash
echo "preventing reboot"
BACKUPBINARY_REBOOT=$(mktemp);
mv /bin/reboot $BACKUPBINARY_REBOOT;
FLAGFILE=$(mktemp);
echo '#!/usr/bin/env bash' > /bin/reboot;
echo '# original reboot binary was moved to'"$BACKUPBINARY_REBOOT" >> /bin/reboot;
echo 'echo request-reboot > '"$FLAGFILE" >> /bin/reboot;
echo 'echo reboot is prevented, your request will trigger later' >> /bin/reboot;
chmod 666 "$FLAGFILE";
chmod +x /bin/reboot;
echo "postponed reboot - press enter to allow it again and make up for requested reboot";
read;
mv "$BACKUPBINARY_REBOOT" /bin/reboot;
if grep -q request-reboot "$FLAGFILE"; then
rm $FLAGFILE;
/bin/reboot;
fi
You can add another systemd service at /usr/lib/systemd/system-shutdown/ which will be run at shutdown, and have it check if your update script is running, and if so, cancel or delay the shutdown.

Can upstart expect/respawn be used on processes that fork more than twice?

I am using upstart to start/stop/automatically restart daemons. One of the daemons forks 4 times. The upstart cookbook states that it only supports forking twice. Is there a workaround?
How it fails
If I try to use expect daemon or expect fork, upstart uses the pid of the second fork. When I try to stop the job, nobody responds to upstarts SIGKILL signal and it hangs until you exhaust the pid space and loop back around. It gets worse if you add respawn. Upstart thinks the job died and immediately starts another one.
Bug acknowledged by upstream
A bug has been entered for upstart. The solutions presented are stick with the old sysvinit, rewrite your daemon, or wait for a re-write. RHEL is close to 2 years behind the latest upstart package, so by the time the rewrite is released and we get updated the wait will probably be 4 years. The daemon is written by a subcontractor of a subcontractor of a contractor so it will not be fixed any time soon either.
I came up with an ugly hack to make this work. It works for my application on my system. YMMV.
start the application in the pre-start section
in the script section run a script that runs as long as the application runs. The pid of this script is what upstart will track.
in the post-stop section kill the application
example
env DAEMON=/usr/bin/forky-application
pre-start script
su -s /bin/sh -c "$DAEMON" joeuseraccount
end script
script
sleepWhileAppIsUp(){
while pidof $1 >/dev/null; do
sleep 1
done
}
sleepWhileAppIsUp $DAEMON
end script
post-stop script
if pidof $DAEMON;
then
kill `pidof $DAEMON`
#pkill $DAEMON # post-stop process (19300) terminated with status 1
fi
end script
a similar approach could be taken with pid files.

Resources