Claire is an open source CLI cluster management tool that works over SSH. It provides a fast and clear interface to the state of your servers.

To install Claire clone this repository in your personal folder and type:

$ cd claire && python setup.py install

To use Claire simply type in your shell:

$ claire stats

To run Claire as a daemon, you need to have super user privileges:

$ sudo claire stats --daemon

To stop the daemon:

$ sudo claire stats --stop-daemon

Claire uses by default the config file in /etc/claire/config. To use another file as config, type:

$ claire stats --conf <path of the config file>

The config file is an ini file structured in this way:

[default]
pidfile=<pid file location> DEFAULT=/etc/claire/claire.pid
interval=<check interval (only for daemon mode)> DEFAULT=300

[reporter]
name=<reporter name>
<reporter configuration key>=<reporter configuration value>

[loggers]
...

[handlers]
...

[formatters]
...

The config file contains also the logging configuration, in the standard python logging config format.

An example of the config file can be found in the config_example folder.

The reporter is the object in charge of reporting the errors. It inherits from the class Reporter, defined in reporter.py, which methods report_server and report_process are called when an error occours reaching a server or checking for a process.

Claire uses by default the hosts file in /etc/claire/hosts. To use another file as hosts, type:

$ claire stats --hosts <path of the hosts file>

The hosts file is an ini file structured in this way:

[group_name]
host_name
check=<check if process with this name is active>
instances=<desired number of running instances of the process>
failure=<run this command in case of failure and retry>

An example of the hosts file can be found in the config_example folder.

To only make Claire check for a single hosts group:

$ claire stats --group <group name>

• Make server checks parallel
• Check service status