joaocenoura / orzaphur Goto Github PK

The backup process must occur after all the others roles backups. On the other hand, the restore process must be the first to occur so the other roles can access the backup repository.
Backup process of git:

• tar.gz home directory
• use ansible fetch to pull the tar file

Restore process of sonar:

• use ansible copy to push the backup file
• untar home directory

Backup process of sonar:

• checkout the 'backup' repository from git server (see #3)
• add:
  • mysql database dump (via mysqldump)
• commit with message "sonar"
• push to git server

Restore process of sonar:

• checkout the 'backup' repository from git server (see #3)
• restore:
  • mysql database dump (via mysql)

Explore git-shell.

Import all the work done so far

Backup process of redmine:

• checkout the 'backup' repository from git server (see #3)
• add:
  • mysql database dump (via mysqldump)
  • uploaded files at /var/lib/redmine (via rsync)
• commit with message "redmine"
• push to git server

Restore process of redmine:

• checkout the 'backup' repository from git server (see #3)
• restore:
  • mysql database dump (via mysql)
  • /var/lib/redmine (via rsync)

Reference: https://wiki.jenkins-ci.org/display/JENKINS/Administering+Jenkins
Note: "Backups can be taken without stopping the server, but when you restore, please do stop the server."
Backup process of jenkins:

• checkout the 'backup' repository from git server (see #3)
• add:
  • jenkins home (via rsync)
    commit with message "jenkins"
• push to git server

Restore process of jenkins:

• checkout the 'backup' repository from git server (see #3)
• stop jenkins
• restore:
  • jenkins home (via rsync)
• start jenkins

Create a new role named "proxy". The main propose, for now, is to cache deb packages in a separated vagrant box. Goal:
Install and configure apt-cacher

Ensure that every role uses variables correctly and when possible. The playbook should be as generic and configurable as possible. This issue should be started after all enhancement issues are closed, in order to spot missing cases. So, check:

• variables: whenever needed, use variables and extend configuration scope
• tags: ensure all tasks have at least one tag

git role should have a 'special' repository for backup proposes. The goal is to create a dedicated repository for other roles places their backup files.
Thus each role (e.g.: redmine, jenkins...) can:

• have a dedicated folder on backup repo
• add/commit/push backup files

(this issue is in a draft stage)

• configure iptables
• allow key based login only
• ...