Tag Archives: Linux

My Play Framework Systemd script

Ubuntu deprecated Upstart so I had to turn to Systemd for my app controls in Ubuntu 18.04. In this script I set 2 environment variables (HOME and LANG), change directory to the app directory and starts the Play Framework application

# Myapp systemd script
#
# Location:/lib/systemd/system/myapp.service
#
# Useful commands:
#
# Start Myapp: 		systemctl start myapp.service
# Stop Myapp:		systemctl stop myapp.service
# Restart Myapp:	systemctl restart myapp.service
# Show status:		systemctl status myapp.service
# Enable start on boot:	systemctl enable myapp.service
# Disable start on boot:systemctl disable myapp.service
#
# List all services running: systemctl
# Check config: systemd-analyze verify myapp.service
#
####################################################################################

[Unit]
Description=Job that runs my app daemon

[Service]
Type=forking
Environment=HOME=/opt/tankmin/app
Environment=LANG=en_US.UTF-8
ExecStartPre=/bin/bash -c 'cd /opt/myapp/app'
ExecStart=/bin/bash -c 'bin/myapp -J-Xms256M -J-Xmx768m -J-server -Dhttp.port=80 -Dconfig.file=conf/application.conf -Dlogger.file=conf/application-logger.xml'

[Install]
WantedBy=multi-user.target

The arguments for the Play service are what I normally use for AWS. You might need other settings

Tested on Ubuntu 18.04 and Play Framework 2.3

My Play Framework logrotate configuration

Here is another script I need to keep handy for the future. My logrotate script for Play Framework applications. The important thing here is the parameter “copytruncate”. I tried many times without it and the result was always the same: after the logrotate was done no more logging was made by the application. “copytruncate” solved this

/opt/myapp/logs/*.log {
  missingok
  notifempty
  size 50M
  copytruncate
  create 0644 root root
  rotate 9999
}

How to run logrotate manually:

logrotate --force /etc/logrotate.d/myapp

Tested on Ubuntu 14.04 and Play Framework 2.3

A small MySQL backup script using mysqldump and creates one file per database

This is the script I use for all my Amazon RDS instances local backups (called from an EC2). It gives me one db backup file for every database in the RDS instance, and at the end it makes sure that I don’t have any files older then 30 days. I put it here now so I don’t loose it again 🙂

#!/bin/bash
 
DATE=$(date +"%Y%m%d_%H%M")
DATABASES=`mysql -hsomedbname-someregion.rds.amazonaws.com -udbadmin -pmypassword -e "show databases;"`
 
for db in $DATABASES; do
        echo $db
        mysqldump --databases $db --single-transaction -hsomedbname-someregion.rds.amazonaws.com -udbadmin -pmypassword | gzip > /opt/backup/mysql.backup.$db.$DATE.sql.gz
done
find /opt/backup/* -mtime +30 -exec rm {} \;

I set this on a two hour cron to make sure that I have fresh backups of my production databases

0 */2 * * * /opt/scripts/mysql_db_backup.sh > /var/log/mysql_db_backup.log