kotovalexarian-likes-github/capistrano
1
0
Fork 0
mirror of https://github.com/capistrano/capistrano synced 2023-03-27 23:21:18 -04:00
Code Releases Activity

cleanup readme

Browse source
This commit is contained in:
Thomas Kriechbaumer 2015-01-18 14:41:22 +00:00
parent 8c83291c2b
commit 1fe453a278
1 changed files with 31 additions and 268 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

299
README.md
View file

@ -1,16 +1,26 @@
# Capistrano [![Build Status](https://travis-ci.org/capistrano/capistrano.svg?branch=master)](https://travis-ci.org/capistrano/capistrano) [![Code Climate](http://img.shields.io/codeclimate/github/capistrano/capistrano.svg)](https://codeclimate.com/github/capistrano/capistrano) <a href="http://codersclan.net/?repo_id=325&source=small"><img src="http://img.shields.io/badge/get-support-blue.svg"></a>
## Requirements
## Documentation
* Ruby >= 1.9.3 (JRuby and C-Ruby/YARV are supported)
Check out the [online documentation](http://capistranorb.com) of Capistrano 3 hosted via this [repository](https://github.com/capistrano/capistrano.github.io).
## Support
Need help with getting Capistrano up and running? Got a code problem you want to get solved quickly?
Get <a href="http://codersclan.net/?repo_id=325&source=link">Capistrano support on CodersClan.</a>
Get <a href="http://codersclan.net/?repo_id=325&source=link">Capistrano support on CodersClan.</a> <a href="http://codersclan.net/?repo_id=325&source=big"><img src="http://www.codersclan.net/gs_button/?repo_id=325" width="150"></a>
<a href="http://codersclan.net/?repo_id=325&source=big"><img src="http://www.codersclan.net/gs_button/?repo_id=325" width="200"></a>
## Requirements
* Ruby >= 1.9.3 (JRuby and C-Ruby/YARV are supported)
Capistrano support these source code version control systems out of the box:
* Git 1.8 or higher
* Mercurial
* SVN
Binaries for these VCS might be required on the local and/or the remote systems.
## Installation
@ -55,280 +65,26 @@ $ bundle exec cap install STAGES=local,sandbox,qa,production
## Usage
``` sh
# list all available tasks
$ bundle exec cap -T
# deploy to the staging environment
$ bundle exec cap staging deploy
# deploy to the production environment
$ bundle exec cap production deploy
# simulate deploying to the production environment
# does not actually do anything
$ bundle exec cap production deploy --dry-run
# list task dependencies
$ bundle exec cap production deploy --prereqs
# trace through task invocations
$ bundle exec cap production deploy --trace
```
## Tasks
``` ruby
server 'example.com', roles: [:web, :app]
server 'example.org', roles: [:db, :workers]
desc "Report Uptimes"
task :uptime do
on roles(:all) do |host|
execute :any_command, "with args", :here, "and here"
info "Host #{host} (#{host.roles.to_a.join(', ')}):\t#{capture(:uptime)}"
end
end
```
**Note**:
**tl;dr**: `execute(:bundle, :install)` and `execute('bundle install')` don't behave identically!
`execute()` has a subtle behaviour. When calling `within './directory' { execute(:bundle, :install) }` for example, the first argument to `execute()` is a *Stringish* with ***no whitespace***. This allows the command to pass through the [SSHKit::CommandMap](https://github.com/capistrano/sshkit#the-command-map) which enables a number of powerful features.
When the first argument to `execute()` contains whitespace, for example `within './directory' { execute('bundle install') }` (or when using a heredoc), neither Capistrano, nor SSHKit can reliably predict how it should be shell escaped, and thus cannot perform any context, or command mapping, that means that the `within(){}` (as well as `with()`, `as()`, etc) have no effect. There have been a few attempts to resolve this, but we don't consider it a bug although we acknowledge that it might be a little counter intuitive.
## Before / After
Where calling on the same task name, executed in order of inclusion
``` ruby
# call an existing task
before :starting, :ensure_user
after :finishing, :notify
# or define in block
before :starting, :ensure_user do
#
end
after :finishing, :notify do
#
end
```
If it makes sense for your use case (often, that means *generating a file*)
the Rake prerequisite mechanism can be used:
``` ruby
desc "Create Important File"
file 'important.txt' do |t|
sh "touch #{t.name}"
end
desc "Upload Important File"
task :upload => 'important.txt' do |t|
on roles(:all) do
upload!(t.prerequisites.first, '/tmp')
end
end
```
The final way to call out to other tasks is to simply `invoke()` them:
``` ruby
namespace :example do
task :one do
on roles(:all) { info "One" }
end
task :two do
invoke "example:one"
on roles(:all) { info "Two" }
end
end
```
This method is widely used.
## Getting User Input
``` ruby
desc "Ask about breakfast"
task :breakfast do
ask(:breakfast, "pancakes")
on roles(:all) do |h|
execute "echo \"$(whoami) wants #{fetch(:breakfast)} for breakfast!\""
end
end
```
Perfect, who needs telephones.
When using `ask` to get user input, you can pass `echo: false` to prevent the input from being displayed:
```ruby
ask(:database_password, "default", echo: false)
```
## Using password authentication
Password authentication can be done via `set` and `ask` in your deploy environment file (e.g.: config/deploy/production.rb)
```ruby
set :password, ask('Server password', nil)
server 'server.domain.com', user: 'ssh_user_name', port: 22, password: fetch(:password), roles: %w{web app db}
```
## Running local tasks
Local tasks can be run by replacing `on` with `run_locally`
``` ruby
desc 'Notify service of deployment'
task :notify do
run_locally do
with rails_env: :development do
rake 'service:notify'
end
end
end
```
Of course, you can always just use standard ruby syntax to run things locally
``` ruby
desc 'Notify service of deployment'
task :notify do
%x('RAILS_ENV=development bundle exec rake "service:notify"')
end
```
Alternatively you could use the rake syntax
``` ruby
desc "Notify service of deployment"
task :notify do
sh 'RAILS_ENV=development bundle exec rake "service:notify"'
end
```
## Console
**Note:** Here be dragons. The console is very immature, but it's much more
cleanly architected than previous incarnations and it'll only get better from
here on in.
Execute arbitrary remote commands, to use this simply add
`require 'capistrano/console'` which will add the necessary tasks to your
environment:
``` sh
$ bundle exec cap staging console
```
Then, after setting up the server connections, this is how that might look:
``` sh
$ bundle exec cap production console
capistrano console - enter command to execute on production
production> uptime
INFO [94db8027] Running /usr/bin/env uptime on leehambley@example.com:22
DEBUG [94db8027] Command: /usr/bin/env uptime
DEBUG [94db8027] 17:11:17 up 50 days, 22:31, 1 user, load average: 0.02, 0.02, 0.05
INFO [94db8027] Finished in 0.435 seconds command successful.
production> who
INFO [9ce34809] Running /usr/bin/env who on leehambley@example.com:22
DEBUG [9ce34809] Command: /usr/bin/env who
DEBUG [9ce34809] leehambley pts/0 2013-06-13 17:11 (port-11262.pppoe.wtnet.de)
INFO [9ce34809] Finished in 0.420 seconds command successful.
```
## VCS Requirements
Capistano requires modern versions of Git, Mercurial and Subversion
respectively (if you are using that particular VCS). Git requirement is at
least version 1.8.x. Mercurial and Subversion, any modern version should
suffice.
## A word about PTYs
There is a configuration option which asks the backend driver to ask the remote host
to assign the connection a *pty*. A *pty* is a pseudo-terminal, which in effect means
*tell the backend that this is an __interactive__ session*. This is normally a bad idea.
Most of the differences are best explained by [this page](https://github.com/sstephenson/rbenv/wiki/Unix-shell-initialization) from the author of *rbenv*.
**When Capistrano makes a connection it is a *non-login*, *non-interactive* shell.
This was not an accident!**
It's often used as a band aid to cure issues related to RVM and rbenv not loading login
and shell initialisation scripts. In these scenarios RVM and rbenv are the tools at fault,
or at least they are being used incorrectly.
Whilst, especially in the case of language runtimes (Ruby, Node, Python and friends in
particular) there is a temptation to run multiple versions in parallel on a single server
and to switch between them using environmental variables, this is an anti-pattern, and
symptomatic of bad design (e.g. you're testing a second version of Ruby in production because
your company lacks the infrastructure to test this in a staging environment).
## Configuration
The following variables are settable:
| Variable Name | Description | Notes |
|:---------------------:|----------------------------------------------------------------------|-----------------------------------------------------------------|
| `:repo_url` | The URL of your scm repository (git, hg, svn) | file://, https://, ssh://, or svn+ssh:// are all supported |
| `:repo_tree` | The subtree of the scm repository to deploy (git, hg) | Only implemented for git and hg repos. Extract just this tree |
| `:branch` | The branch you wish to deploy | This only has meaning for git and hg repos, to specify the branch of an svn repo, set `:repo_url` to the branch location. |
| `:scm` | The source control system used | `:git`, `:hg`, `:svn` are currently supported |
| `:tmp_dir` | The (optional) temp directory that will be used (default: /tmp) | if you have a shared web host, this setting may need to be set (i.e. /home/user/tmp/capistrano). |
__Support removed__ for following variables:
| Variable Name | Description | Notes |
|:---------------------:|---------------------------------------------------------------------|-----------------------------------------------------------------|
| `:copy_exclude` | The (optional) array of files and/or folders excluded from deploy | Replaced by Git's native `.gitattributes`, see [#515](https://github.com/capistrano/capistrano/issues/515) for more info. |
## Host and Role Filtering
Capistrano enables the declaration of servers and roles, each of which may have properties
associated with them. Tasks are then able to use these definitions in two distinct ways:
* To determine _configurations_: typically by using the `roles()`, `release_roles()` and
`primary()` methods. Typically these are used outside the scope of the `on()` method.
* To _interact_ with remote hosts using the `on()` method
An example of the two would be to create a `/etc/krb5.conf` file containing the list of
available KDC's by using the list of servers returned by `roles(:kdc)` and then uploading
it to all client machines using `on(roles(:all)) do upload!(file) end`
A problem with this arises when _filters_ are used. Filters are designed to limit the
actual set of hosts that are used to a subset of those in the overall stage, but how
should that apply in the above case?
If the filter applies to both the _interaction_ and _configuration_ aspects, any configuration
files deployed will not be the same as those on the hosts excluded by the filters. This is
almost certainly not what is wanted, the filters should apply only to the _interactions_
ensuring that any configuration files deployed will be identical across the stage.
Another type of filtering is done by defining properties on servers and selecting on that
basis. An example of that is the 'no_release' property and it's use in the
`release_roles()` method. To distinguish these two types of filtering we name them:
* On-Filtering
Specified in the following ways:
* Via environment variables HOSTS and ROLES
* Via command line options `--hosts` and `--roles`
* Via the `:filter` variable set in a stage file
* Property-Filtering
These are specified by options passed to the `roles()` method (and implicitly in methods
like `release_roles()` and `primary()`)
To increase the utility of On-Filters they can use regular expressions:
* If the host name in a filter doesn't match `/^[-A-Za-z0-9.]+$/` (the set of valid characters
for a DNS name) then it's assumed to be a regular expression.
* Since role names are Ruby symbols they can legitimately contain any characters. To allow multiple
of them to be specified on one line we use the comma. To use a regexp for a role filter begin
and end the string with '/'. These may not contain a comma.
When filters are specified using comma separated lists, the final filter is the _union_ of
all of the components. However when multiple filters are declared the result is the
_intersection_.
## SSHKit
[SSHKit](https://github.com/leehambley/sshkit) is the driver for SSH
connections behind the scenes in Capistrano. Depending on how deep you dig, you
might run into interfaces that come directly from SSHKit (the configuration is
a good example).
## Testing
Capistrano has two test suites: an RSpec suite and a Cucumber suite. The
@ -370,6 +126,13 @@ tty](https://github.com/capistrano/stats/blob/77c9993d3ee604520712261aa2a70c90f3
allow us to differentiate between many members of the same team deploying the
same project vs. many people on many projects.
## SSHKit
[SSHKit](https://github.com/leehambley/sshkit) is the driver for SSH
connections behind the scenes in Capistrano. Depending on how deep you dig, you
might run into interfaces that come directly from SSHKit (the configuration is
a good example).
## License
MIT License (MIT)