Module: CloudCrowd

Defined in:
lib/cloud-crowd.rb,
lib/cloud_crowd/node.rb,
lib/cloud_crowd/models.rb,
lib/cloud_crowd/server.rb,
lib/cloud_crowd/worker.rb,
lib/cloud_crowd/action.rb,
lib/cloud_crowd/helpers.rb,
lib/cloud_crowd/inflector.rb,
lib/cloud_crowd/exceptions.rb,
lib/cloud_crowd/models/job.rb,
lib/cloud_crowd/asset_store.rb,
lib/cloud_crowd/command_line.rb,
lib/cloud_crowd/models/work_unit.rb,
lib/cloud_crowd/helpers/resources.rb,
lib/cloud_crowd/models/node_record.rb,
lib/cloud_crowd/asset_store/s3_store.rb,
lib/cloud_crowd/helpers/authorization.rb,
lib/cloud_crowd/asset_store/cloudfiles_store.rb,
lib/cloud_crowd/asset_store/filesystem_store.rb

Defined Under Namespace

Modules: Helpers, Inflector, ModelStatus Classes: Action, AssetStore, CommandLine, Error, Job, Node, NodeRecord, Server, WorkUnit, Worker

Constant Summary

VERSION =

Keep this version in sync with the gemspec.

'0.7.2'
SCHEMA_VERSION =

Increment the schema version when there's a backwards incompatible change.

4
ROOT =

Root directory of the CloudCrowd gem.

File.expand_path(File.dirname(__FILE__) + '/..')
LOG_PATH =

Default folder to log daemonized servers and nodes into.

'log'
PID_PATH =

Default folder to contain the pids of daemonized servers and nodes.

'tmp/pids'
MIN_RETRIES =

Minimum number of attempts per work unit.

1
PROCESSING =

A Job is processing if its WorkUnits are in the queue to be handled by nodes.

1
SUCCEEDED =

A Job has succeeded if all of its WorkUnits have finished successfully.

2
FAILED =

A Job has failed if even a single one of its WorkUnits has failed (they may be attempted multiple times on failure, however).

3
SPLITTING =

A Job is splitting if it's in the process of dividing its inputs up into multiple WorkUnits.

4
MERGING =

A Job is merging if it's busy collecting all of its successful WorkUnits back together into the final result.

5
COMPLETE =

A Job is considered to be complete if it succeeded or if it failed.

[SUCCEEDED, FAILED]
INCOMPLETE =

A Job is considered incomplete if it's being processed, split up or merged.

[PROCESSING, SPLITTING, MERGING]
DISPLAY_STATUS_MAP =

Mapping of statuses to their display strings.

['unknown', 'processing', 'succeeded', 'failed', 'splitting', 'merging']

Class Attribute Summary (collapse)

Class Method Summary (collapse)

Class Attribute Details

+ (Object) config (readonly)

Returns the value of attribute config



95
96
97
# File 'lib/cloud-crowd.rb', line 95

def config
  @config
end

+ (Object) identity

Returns the value of attribute identity



96
97
98
# File 'lib/cloud-crowd.rb', line 96

def identity
  @identity
end

Class Method Details

+ (Object) action_paths

Retrieve the list of every installed Action for this node or server.



193
194
195
196
197
198
# File 'lib/cloud-crowd.rb', line 193

def action_paths
  default_actions   = config[:disable_default_actions] ? [] : Dir["#{ROOT}/actions/*.rb"]
  installed_actions = Dir["#{@config_path}/actions/*.rb"]
  custom_actions    = CloudCrowd.config[:actions_path] ? Dir["#{CloudCrowd.config[:actions_path]}/*.rb"] : []
  default_actions + installed_actions + custom_actions
end

+ (Object) actions

CloudCrowd::Actions are requested dynamically by name. Access them through this actions property, which behaves like a hash. At load time, we load all installed Actions and CloudCrowd's default Actions into it. If you wish to have certain nodes be specialized to only handle certain Actions, then install only those into the actions directory.



178
179
180
181
182
183
184
185
186
187
188
189
190
# File 'lib/cloud-crowd.rb', line 178

def actions
  return @actions if @actions
  @actions = action_paths.inject({}) do |memo, path|
    path = Pathname.new(path)
    require path.relative? ? path.basename : path
    name = path.basename('.*').to_s
    memo[name] = Module.const_get( Inflector.camelize( name ) )
    memo
  end
rescue NameError => e
  adjusted_message = "One of your actions failed to load. Please ensure that the name of your action class can be deduced from the name of the file. ex: 'word_count.rb' => 'WordCount'\n#{e.message}"
  raise NameError.new(adjusted_message, e.name)
end

+ (Object) central_server

Get a reference to the central server, including authentication if configured.



135
136
137
# File 'lib/cloud-crowd.rb', line 135

def central_server
  @central_server ||= RestClient::Resource.new(CloudCrowd.config[:central_server], CloudCrowd.client_options)
end

+ (Object) client_options

The standard RestClient options for the central server talking to nodes, as well as the other way around. There's a timeout of 5 seconds to open a connection, and a timeout of 30 to finish reading it.



154
155
156
157
158
159
160
161
162
163
164
165
# File 'lib/cloud-crowd.rb', line 154

def client_options
  return @client_options if @client_options
  @client_options = {
    :timeout => (self.server? ? config[:node_timeout] : config[:server_timeout]) || 30,
    :open_timeout => config[:open_timeout] || 5
  }
  if CloudCrowd.config[:http_authentication]
    @client_options[:user]      = CloudCrowd.config[:login]
    @client_options[:password]  = CloudCrowd.config[:password]
  end
  @client_options
end

+ (Object) configure(config_path)

Configure CloudCrowd by passing in the path to config.yml.



99
100
101
102
103
104
105
106
107
# File 'lib/cloud-crowd.rb', line 99

def configure(config_path)
  @config_path = File.expand_path(File.dirname(config_path))
  @config = YAML.load(ERB.new(File.read(config_path)).result)
  @config[:work_unit_retries] ||= MIN_RETRIES
  if @config[:actions_path]
    path = Pathname.new( @config[:actions_path] ).realpath
    $LOAD_PATH.unshift( path ) unless $LOAD_PATH.include?( path )
  end
end

+ (Object) configure_database(config_path, validate_schema = true)

Configure the CloudCrowd central database (and connect to it), by passing in a path to database.yml. The file should use the standard ActiveRecord connection format.



112
113
114
115
116
117
118
119
120
121
# File 'lib/cloud-crowd.rb', line 112

def configure_database(config_path, validate_schema=true)
  configuration = YAML.load(ERB.new(File.read(config_path)).result)
  ActiveRecord::Base.establish_connection(configuration)
  if validate_schema
    version = ActiveRecord::Base.connection.select_values('select max(version) from schema_migrations').first.to_i
    return true if version == SCHEMA_VERSION
    puts "Your database schema is out of date. Please use `crowd load_schema` to update it. This will wipe all the tables, so make sure that your jobs have a chance to finish first.\nexiting..."
    exit
  end
end

+ (Object) defer

Starts a new thread with a ActiveRecord connection_pool and yields for peforming work inside the blocks



125
126
127
128
129
130
131
# File 'lib/cloud-crowd.rb', line 125

def defer
  Thread.new do
    ActiveRecord::Base.connection_pool.with_connection do
      yield
    end
  end
end

+ (Object) display_status(status)

Return the displayable status name of an internal CloudCrowd status number. (See the above constants).



169
170
171
# File 'lib/cloud-crowd.rb', line 169

def display_status(status)
  DISPLAY_STATUS_MAP[status] || 'unknown'
end

+ (Object) log_path(log_file = nil)

The path that daemonized servers and nodes will log to.



140
141
142
143
# File 'lib/cloud-crowd.rb', line 140

def log_path(log_file=nil)
  @log_path ||= config[:log_path] || LOG_PATH
  log_file ? File.join(@log_path, log_file) : @log_path
end

+ (Boolean) node?

Or is it a node?

Returns:

  • (Boolean)


207
208
209
# File 'lib/cloud-crowd.rb', line 207

def node?
  @identity == :node
end

+ (Object) pid_path(pid_file = nil)

The path in which daemonized servers and nodes will store their pids.



146
147
148
149
# File 'lib/cloud-crowd.rb', line 146

def pid_path(pid_file=nil)
  @pid_path ||= config[:pid_path] || PID_PATH
  pid_file ? File.join(@pid_path, pid_file) : @pid_path
end

+ (Boolean) server?

Is this CloudCrowd instance a server? Useful for avoiding loading unneeded code from actions.

Returns:

  • (Boolean)


202
203
204
# File 'lib/cloud-crowd.rb', line 202

def server?
  @identity == :server
end