Fabric
Fabric is a python package that helps to automate tasks. It is flexible and simple to use, it will help you to define a repeatable installation / configuration process.The tasks defined in your fabric's script, often call fabfile.py, can be run on any host that is accessible via ssh.
Installation
This is one of the few package that I install globally because I use it all the time and not always in the context of a virtualenv.
1 | pip install Fabric |
Usage
Once Fabric is installed you can use the command called "fab" to run the tasks defined in the "fabfile.py". The tasks can be chained and like this
1 | fab machine environment my-task |
In the example above machine, environment, my-task are 3 tasks defined in your "fabfile.py". I used the first 2 tasks to set some environment variable that are used by the third task "my-task" to actually do something useful.
You can also pass argument to you task
1 | fab my-task:arg1="my first argument" |
Writing a task
Writing a fabric task is as simple as writing a python method, Fabric comes with a set of small wrapper that will help you to get things done quickly : local, run, cd, ..
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 | from uuid import uuid1 from os import path from fabric.api import * def localhost(): env.hosts = ['localhost'] def example(): env.hosts = ['example.com'] def project_settings(): env.db_user = '<database Username>' env.webapps = '/opt/webapps/' env.backups = '/opt/backups/' env.www = '/var/www/' env.project_name = 'example-project' env.test_settings = 'project.conf.dev.settings' env.static_cache = path.join('project_cms','themes','example', 'static','CACHE') env.pip_requirements = 'requirements.pip' def prod(): env.site = 'example.com' env.db_name = env.site env.settings = 'project_cms.conf.prod.settings' project_settings() def dev(): env.site = 'dev.example.com' env.db_name = env.site env.settings = 'project_cms.conf.dev.settings' project_settings() def backup_db(identifier=None, db_name=None, db_user=None, site=None, backups=None): """ Backup db """ db_name = db_name or env.db_name db_user = db_user or env.db_user site = site or env.site backups = backups or env.backups backup_file = "%s-dump-%s.sql" % (identifier or uuid1(), db_name) backup_site = path.join(backups, site) with cd(backup_site): run("pg_dump -U %s %s > %s" % (db_user, db_name, backup_file)) def backup_media(identifier=None,backups=None, www=None, site=None): """ Back up the media directory """ www = www or env.www site = site or env.site backups = backups or env.backups backup_file = path.join(backups,site, "%s-media.tar.bz2" % (identifier or uuid1())) with cd(path.join(www,site)): run("tar -cjf %s media/" %(backup_file)) |
The example above shows how fabric can be used, as you can read each task aim to automate a particular activity. You should have a task for all your installation / configuration / deployment / backup / restore. This will enable you to write some macro task like :
1 | fab example prod install |
or
1 | fab example prod delete |
The above tasks can be used like this :
1 | fab example dev backup_db |
This will execute a backup of the db on example.com via ssh. I usually have a complete automation of all the activity done on my servers. Do not be afraid of having a large fabfiles.py, this is just python so you can organize it in modules just like you would do with the rest of your code.
Conclusion
This article just scratch the tip of the iceberg and only cover 20 % of the Fabric package however they are the one I use 80 % of the time. I hope that you find this article useful and I would be glad to hear from you if you can improve this article.