I am creating a playbook that takes a variable ver at runtime and then runs tasks based on that variable. If that ver is defined, it takes the version number and installs a version according to that. If the ver is not defined, it defaults to a default version.
My playbook for now looks like:
- name: Download Maven from remote repository.
unarchive:
src: "{{ baseurl }}/{{ versions[ver] }}"
dest: /usr/local/
remote_src: yes
become: yes
when: ver is defined
- name: Rename maven folder.
command: mv /usr/local/apache-maven-{{ver}} /usr/local/maven-{{ver}}
become: yes
when: ver is defined
- name: Create symbolic link to maven folder.
file:
src: "/usr/local/maven-{{ver}}"
dest: "/usr/local/maven"
state: link
become: yes
when: ver is defined
- name: Download Default Maven from remote repository.
unarchive:
src: "{{ baseurl }}/{{ versions[default_version] }}"
dest: /usr/local/
remote_src: yes
list_files: yes
become: yes
when: ver is not defined
- name: Rename maven folder.
command: mv /usr/local/apache-maven-{{default_version}} /usr/local/maven-{{default_version}}
become: yes
when: ver is not defined
- name: Create symbolic link to maven folder.
file:
src: "/usr/local/maven-{{default_version}}"
dest: "/usr/local/maven"
state: link
become: yes
when: ver is not defined
As you can see, based on whether a variable is defined or not, I have to write whole new tasks for it.
The above works but I think it's not optimized as I'm writing too much code to get this to work. One idea that I have is to create have main.yaml invoke another task file when ver is defined and a separate task file when ver is not defined.
Is there another way to optimize this? I have many other tasks that deal with version number.
