An Ansible role to deploy a CouchDB 2.0 on CentOS 6.x
This repo contains an Ansible role to install CouchDB 2.0 on CentOS 6.x machines. It's been tested with Ansible 1.9.4 and CentOS 6.7/6/8.
The role roughly follows the instructions from the CouchDB 2.0 Release Testing Plan (see https://docs.google.com/document/d/1BtndYr-0KDQTqBSLVdJoR_8C5ObYjT1RBo_Qyh5ykdQ/edit#)
The role is fairly self explanatory. It installs SpiderMonkey, Erlang (from Erlang Solutions), Node.js 4.x and then proceeds to build CouchDB 2.0 from the tar file.
The process may take a little while, but at the end, you will have a running CouchDB 2.0 installed as a Linux service (via a sysv script).
If you deploy on multiple nodes at once, you can also use the role to configure those nodes in a cluster, simply set the variable configure_cluster to True (by default it is set to False).
The role will then connect all the machines which are in the couches inventory group as per the steps described in https://github.com/apache/couchdb-setup (the instructions on that page are not 100% correct, some of the URLs have typos in them).
This can be done in your inventory file, or via a command-line parameter, like so :
ansible-playbook -i inventory -e "configure_cluster=True" couchdb2.yaml
In Cluster mode, the role automatically creates an admin account (user: admin, password: password).
This role is intended to be for experimental use, but feel free to ping me with comments, or even better, submit pull requests.
At present, the playbook disables the firewall, so use at your own risk. In the future, I might be working on an updated version which uses Ansible to poke the correct holes in IPTables.