Fabric task runner and helper. Execute tasks via web browser.
oozappa is 大雑把.
Install from Cheese Shop (pypi).
$ pip install oozappa
If you use Xcode 5.1(above) and failed with clang: error: unknown argument: '-mno-fused-madd', export flags before install pycrypt (before pip install -r requirements.txt).
$ export CPPFLAGS=-Qunused-arguments
$ export CFLAGS=-Qunused-arguments
Environment is a category that has fibfile directory and vars.py .
A problem about fabric with large project is, too many fabric tasks and complicated task orders.
So you should separate fabfile into domain category. That is Environment.
Job is like a normal fabric execution command.
eg. fab task1 task2 is a job.
Jobset is oozappa operation unit.
Jobset can contain multiple job, even extend over environments.
see sample oozappa project(sample/ops).
.
├── common
│ ├── __init__.py
│ ├── files
│ ├── functions
│ │ ├── common_multiple_fabric_environment.py
│ ├── templates
│ │ └── sample_a.txt
│ └── vars.py
├── construction
│ ├── fabfile
│ │ ├── __init__.py
│ ├── templates
│ └── vars.py
├── deployment
│ ├── fabfile
│ │ ├── __init__.py
│ ├── templates
│ └── vars.py
├── production
│ ├── fabfile
│ │ ├── __init__.py
│ ├── templates
│ └── vars.py
└── staging
├── fabfile
│ ├── __init__.py
├── templates
│ └── sample_a.txt
└── vars.py
common is reserved directory. construction and others are environment directory. These names are just example.
common and each environment's vars.py might have oozappa.config.OozappaSetting instance named settings. OozappaSetting is dict like object. common.vars.setting is updated by executed environment's vars.setting, so you can set base configuration to common.vars.setting and environment's one overwrite it.
Check printsetting task on staging environment.
You can run fabric task within environment directory as usual.
$ cd sample/ops/staging
$ fab printsetting
{'instance_type': 't1.micro', 'domain': 'localhost', 'sample_template_vars': {'sample_a': {'key_a_2': "a's 2 value from common.vars", 'key_a_1': "a's 1 value from stging.vars"}}, 'email': 'mtsuyuki at gmail.com'}
See common/vars.py and staging/vars.py .
Same as vars, oozappa.fabrictools.upload_template search template. upload_template is almost same as fabric.contrib.files.upload_template . oozappa's upload_template doesn't accept use_jinja, because oozappa's upload_template pass use_jinja=True to fabric.contrib.files.upload_template.
Jinja2 has inheritance template system and search template from multipul paths. fabric's upload_template accept only one template_dir string not list. fabric doesn't assume multiplu environment, so it's reasonable. Because of this, oozappa's upload_template search template path is limited only one template_dir that found filename. It mean that you can't store child template and parent template separately.
Call oozappa.config.procure_common_functions() and add commons/functions directory to sys.path for convinient to using on multiple fabric environment.
Change directory to outside environment directory.
$ cd ..
$ ls
common production staging
$ gunicorn -t 3600 -w 4 -k flask_sockets.worker oozappa.webui:app
Running oozappa:app creates /tmp/oozappa.sqlite .
Open your web browser and browse http://localhost:8000/ .
rapid execution function removed(version 0.9.0).
Modify Run fabric in raw's left hand side input to staging and click run_tasks. You can see what tasks exists.
Then input ls ps to right hand side input and click run_tasks .
That's it.
- Click environment button via top menu.
- Add new Environment
- name: constructiton
- sort_order: 1
- execute_path: constructiton
- Add 3 more.
- Click environment you created
- Create new Job.
- Click task from Possible tasks in order
- Click jobset button via top menu.
- Click jobs you'd like to execute once.
- Click navigation button or jobset button via top menu.
- Click jobset you'd like to execute.
- Click run jobset button.
- Running log displays Running log.
- Reload page when jobset done. or Go to top(via navigation button)
- You see Execute Logs and show raw log when you click success (or fail).
Change directory your own oozappa.
$ mkdir devops
$ cd devops
Then run zappa command.
$ zappa init
Create common environment here? [y/N] : y
Sqlite database stored path. [/tmp/oozappa/data.sqlite] :
Log files stored path. [/tmp/oozappa] :
Create directory or exit? "/tmp/oozappa" [y/N] : y
created common directory. db/log file path and flask secret key are in common/vars.py.
attention:: Default stored path is not for production use. You should input your own file/directory path. Or data/results disapear when you reboot your machine or server.
Open common/vars.py and change settings.
- OOZAPPA_DB : sqlite's data store path.
- OOZAPPA_LOG_BASEDIR : Jobset execute log store directory path.
Run oozappa.create_environment with environment name(s).
$ zappa create_environment --names construction deployment
2014-04-20 16:43:26,543 INFO create environment : construction
2014-04-20 16:43:26,544 INFO create environment : deployment
Then you can write fabfile normally and execute via oozappa.
$ zappa list
----------------------------------------
|id:001|CLI_ONLY:False|title:デプローイのテスト１|
|id:002|CLI_ONLY:True|title:デプロイプリント|
----------------------------------------
$ zappa run_jobset --jobset 1
Only jobset which CLI_ONLY flag is False can run using webui.
You can set CLI_ONLY flag to True with zappa command.
$ zappa manage
<manage:/Users/makoto/oozappa/oozappa>
commands
-> q: quit
-> c: Modify Jobset's CLI Flag.
-> l: List Jobsets
-> n: next
c
----------------------------------------
|id:001|CLI_ONLY:False|title:デプローイのテスト１|
|id:002|CLI_ONLY:True|title:デプロイプリント|
----------------------------------------
Which Jobset id?2
----------------------------------------
|id:001|CLI_ONLY:False|title:デプローイのテスト１|
|id:002|CLI_ONLY:False|title:デプロイプリント|
----------------------------------------
commands
-> q: quit
-> c: Modify Jobset's CLI Flag.
-> l: List Jobsets
-> n: next
q