Ubuntu – How to start a Zookeeper daemon after booting under specific user in Ubuntu server 16.04

bashcommand lineserverstartup

I want to start Zookeeper daemon after Ubuntu server 16.04 booting (not after logging) under user named zookeeper. So I changed the file /etc/rc.local like the following:

#!/bin/sh -e
#
# rc.local
#
# This script is executed at the end of each multiuser runlevel.
# Make sure that the script will "exit 0" on success or any other
# value on error.
#
# In order to enable or disable this script just change the execution
# bits.
#
# By default this script does nothing.

echo 'never'; defrag_file_pathname

su -c '$ZOOKEEPER_HOME/bin/zkServer.sh start' zookeeper &

exit 0

, adding the line su -c '$ZOOKEEPER_HOME/bin/zkServer.sh start' zookeeper & before exit 0. But the process is not started after restarting!

What's wrong here?

details: The zookeeper user is in sudo group and has password.

details: When I run command su -c '$ZOOKEEPER_HOME/bin/zkServer.sh start' zookeeper & in terminal, it needs password to run.

Best Answer

Create a .service file in /etc/systemd/system/zoo.service and add the following lines:

[Unit]
Description=Zookeeper Daemon
Wants=syslog.target

[Service]    
Type=forking
WorkingDirectory=/path/to/dir/of/interest
User=zookeeper 
ExecStart=/home/zookeeper_home/bin/zkServer.sh
TimeoutSec=30
Restart=on-failure

[Install]
WantedBy=multi-user.target

Now setup the service:

sudo systemctl start zoo
sudo systemctl enable zoo

Check status:

sudo systemctl status zoo

Please read for further details creating daemons:

https://www.freedesktop.org/software/systemd/man/systemd.unit.html

Related Solutions

Ubuntu – Why is the command in /etc/rc.local not executed during startup

rc.local script exits if any error occurs while executing any of its commands (mention the -e flag in #!/bin/sh -e).

It is possible that some prerequisites are not met when you try to run your commands when rc.local execution takes place, so your command execution fails.

I encountered the same thing while manually setting cpu governor and failing to do so in rc.local. Here's my custom workaround, which uses update-rc.d to make your commands run on startup:

Create a file myscript.sh in directory /etc/init.d with a heading: #!/bin/sh
Put your custom commands as the content
Make it executable: sudo chmod +x /etc/init.d/myscript.sh
Create symlinks for your script for various runlevels: sudo update-rc.d myscript.sh defaults

Also, you could check /etc/network/if-up.d scripts and see if you could trigger your commands when networking starts.

Ubuntu – Start script on system startup (Ubuntu 16.04.1)

The easiest solution is using sudo powers create a file like this in /etc/cron.d/:

SHELL=/bin/sh
PATH=/usr/local/sbin:/usr/local/bin:/sbin:/bin:/usr/sbin:/usr/bin
@reboot   root    mkdir /cronjobs
@reboot   root    sleep 10
@reboot   root    mkdir /cronjobs/demofolder

This avoids the use of a script file altogether and works for all users regardless of their home directory name, ie /home/steve, /home/mary, etc.

Edit - Add `sleep 10`

For whatever reason cron is working too fast, or kernel is working too slow when making directories. An extra line sleep 10 was necessary between the two mkdir lines.

You may not need 10 seconds in between the two make directory commands but 10 works on my system with an SSD.

Edit 2 - Make full directory path in one command

As per comments below a simpler method is to use:

SHELL=/bin/sh
PATH=/usr/local/sbin:/usr/local/bin:/sbin:/bin:/usr/sbin:/usr/bin
@reboot   root    mkdir -p -v /cronjobs/demofolder

-p (long version --parents) tells mkdir to automatically create all directory parent levels if they don't exist.
-v (long version --verbose) tells mkdir to print the names of all directories created.