Class: WebHDFS::ClientV1

Inherits:
Object
  • Object
show all
Defined in:
lib/webhdfs/client_v1.rb

Direct Known Subclasses

Client

Constant Summary

OPT_TABLE =

This hash table holds command options.

{}
KNOWN_ERRORS =

internal use only

['LeaseExpiredException'].freeze
REDIRECTED_OPERATIONS =
['APPEND', 'CREATE', 'OPEN', 'GETFILECHECKSUM']

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(host = 'localhost', port = 50070, username = nil, doas = nil, proxy_address = nil, proxy_port = nil) ⇒ ClientV1



23
24
25
26
27
28
29
30
31
32
33
34
35
# File 'lib/webhdfs/client_v1.rb', line 23

def initialize(host='localhost', port=50070, username=nil, doas=nil, proxy_address=nil, proxy_port=nil)
  @host = host
  @port = port
  @username = username
  @doas = doas
  @proxy_address = proxy_address
  @proxy_port = proxy_port
  @retry_known_errors = false
  @retry_times = 1
  @retry_interval = 1

  @httpfs_mode = false
end

Instance Attribute Details

#doasObject

Returns the value of attribute doas



14
15
16
# File 'lib/webhdfs/client_v1.rb', line 14

def doas
  @doas
end

#hostObject

Returns the value of attribute host



14
15
16
# File 'lib/webhdfs/client_v1.rb', line 14

def host
  @host
end

#httpfs_modeObject

Returns the value of attribute httpfs_mode



18
19
20
# File 'lib/webhdfs/client_v1.rb', line 18

def httpfs_mode
  @httpfs_mode
end

#open_timeoutObject

Returns the value of attribute open_timeout



16
17
18
# File 'lib/webhdfs/client_v1.rb', line 16

def open_timeout
  @open_timeout
end

#portObject

Returns the value of attribute port



14
15
16
# File 'lib/webhdfs/client_v1.rb', line 14

def port
  @port
end

#proxy_addressObject

Returns the value of attribute proxy_address



14
15
16
# File 'lib/webhdfs/client_v1.rb', line 14

def proxy_address
  @proxy_address
end

#proxy_passObject

Returns the value of attribute proxy_pass



15
16
17
# File 'lib/webhdfs/client_v1.rb', line 15

def proxy_pass
  @proxy_pass
end

#proxy_portObject

Returns the value of attribute proxy_port



14
15
16
# File 'lib/webhdfs/client_v1.rb', line 14

def proxy_port
  @proxy_port
end

#proxy_userObject

Returns the value of attribute proxy_user



15
16
17
# File 'lib/webhdfs/client_v1.rb', line 15

def proxy_user
  @proxy_user
end

#read_timeoutObject

default 30s (in ruby net/http) default 60s (in ruby net/http)



17
18
19
# File 'lib/webhdfs/client_v1.rb', line 17

def read_timeout
  @read_timeout
end

#retry_intervalObject

default false (not to retry) default 1 (ignored when retry_known_errors is false) default 1 ([sec], ignored when retry_known_errors is false)



21
22
23
# File 'lib/webhdfs/client_v1.rb', line 21

def retry_interval
  @retry_interval
end

#retry_known_errorsObject

Returns the value of attribute retry_known_errors



19
20
21
# File 'lib/webhdfs/client_v1.rb', line 19

def retry_known_errors
  @retry_known_errors
end

#retry_timesObject

Returns the value of attribute retry_times



20
21
22
# File 'lib/webhdfs/client_v1.rb', line 20

def retry_times
  @retry_times
end

#usernameObject

Returns the value of attribute username



14
15
16
# File 'lib/webhdfs/client_v1.rb', line 14

def username
  @username
end

Instance Method Details

#api_path(path) ⇒ Object



206
207
208
209
210
211
212
# File 'lib/webhdfs/client_v1.rb', line 206

def api_path(path)
  if path.start_with?('/')
    '/webhdfs/v1' + path
  else
    '/webhdfs/v1/' + path
  end
end

#append(path, body, options = {}) ⇒ Object

curl -i -X POST “



51
52
53
54
55
56
57
58
# File 'lib/webhdfs/client_v1.rb', line 51

def append(path, body, options={})
  if @httpfs_mode
    options = options.merge({'data' => 'true'})
  end
  check_options(options, OPT_TABLE['APPEND'])
  res = operate_requests('POST', path, 'APPEND', options, body)
  res.code == '200'
end

#build_path(path, op, params) ⇒ Object



214
215
216
217
218
219
220
221
222
223
224
225
226
# File 'lib/webhdfs/client_v1.rb', line 214

def build_path(path, op, params)
  opts = if @username and @doas
           {'op' => op, 'user.name' => @username, 'doas' => @doas}
         elsif @username
           {'op' => op, 'user.name' => @username}
         elsif @doas
           {'op' => op, 'doas' => @doas}
         else
           {'op' => op}
         end
  query = URI.encode_www_form(params.merge(opts))
  api_path(path) + '?' + query
end

#check_options(options, optdecl = []) ⇒ Object

def delegation_token(user, options={}) # GETDELEGATIONTOKEN

raise NotImplementedError

end def renew_delegation_token(token, options={}) # RENEWDELEGATIONTOKEN

raise NotImplementedError

end def cancel_delegation_token(token, options={}) # CANCELDELEGATIONTOKEN

raise NotImplementedError

end

Raises:

  • (ArgumentError)


197
198
199
200
# File 'lib/webhdfs/client_v1.rb', line 197

def check_options(options, optdecl=[])
  ex = options.keys.map(&:to_s) - (optdecl || [])
  raise ArgumentError, "no such option: #{ex.join(' ')}" unless ex.empty?
end

#check_success_json(res, attr = nil) ⇒ Object



202
203
204
# File 'lib/webhdfs/client_v1.rb', line 202

def check_success_json(res, attr=nil)
  res.code == '200' and res.content_type == 'application/json' and (attr.nil? or JSON.parse(res.body)[attr])
end

#checksum(path, options = {}) ⇒ Object Also known as: getfilechecksum



124
125
126
127
128
# File 'lib/webhdfs/client_v1.rb', line 124

def checksum(path, options={})
  check_options(options, OPT_TABLE['GETFILECHECKSUM'])
  res = operate_requests('GET', path, 'GETFILECHECKSUM', options)
  check_success_json(res, 'FileChecksum')
end

#content_summary(path, options = {}) ⇒ Object Also known as: getcontentsummary



116
117
118
119
120
# File 'lib/webhdfs/client_v1.rb', line 116

def content_summary(path, options={})
  check_options(options, OPT_TABLE['GETCONTENTSUMMARY'])
  res = operate_requests('GET', path, 'GETCONTENTSUMMARY', options)
  check_success_json(res, 'ContentSummary')
end

#homedir(options = {}) ⇒ Object Also known as: gethomedirectory



132
133
134
135
136
# File 'lib/webhdfs/client_v1.rb', line 132

def homedir(options={})
  check_options(options, OPT_TABLE['GETHOMEDIRECTORY'])
  res = operate_requests('GET', '/', 'GETHOMEDIRECTORY', options)
  check_success_json(res, 'Path')
end

#list(path, options = {}) ⇒ Object Also known as: liststatus



108
109
110
111
112
# File 'lib/webhdfs/client_v1.rb', line 108

def list(path, options={})
  check_options(options, OPT_TABLE['LISTSTATUS'])
  res = operate_requests('GET', path, 'LISTSTATUS', options)
  check_success_json(res, 'FileStatuses')['FileStatus']
end

#mkdir(path, options = {}) ⇒ Object Also known as: mkdirs



72
73
74
75
76
# File 'lib/webhdfs/client_v1.rb', line 72

def mkdir(path, options={})
  check_options(options, OPT_TABLE['MKDIRS'])
  res = operate_requests('PUT', path, 'MKDIRS', options)
  check_success_json(res, 'boolean')
end

#operate_requests(method, path, op, params = {}, payload = nil) ⇒ Object



229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
# File 'lib/webhdfs/client_v1.rb', line 229

def operate_requests(method, path, op, params={}, payload=nil)
  if not @httpfs_mode and REDIRECTED_OPERATIONS.include?(op)
    res = request(@host, @port, method, path, op, params, nil)
    unless res.is_a?(Net::HTTPRedirection) and res['location']
      msg = "NameNode returns non-redirection (or without location header), code:#{res.code}, body:#{res.body}."
      raise WebHDFS::RequestFailedError, msg
    end
    uri = URI.parse(res['location'])
    rpath = if uri.query
              uri.path + '?' + uri.query
            else
              uri.path
            end
    request(uri.host, uri.port, method, rpath, nil, {}, payload, {'Content-Type' => 'application/octet-stream'})
  else
    if @httpfs_mode and not payload.nil?
      request(@host, @port, method, path, op, params, payload, {'Content-Type' => 'application/octet-stream'})
    else
      request(@host, @port, method, path, op, params, payload)
    end
  end
end

#rename(path, dest, options = {}) ⇒ Object

curl -i -X PUT “<HOST>:<PORT>/webhdfs/v1/<PATH>?op=RENAME&destination=<PATH>”



81
82
83
84
85
86
87
88
# File 'lib/webhdfs/client_v1.rb', line 81

def rename(path, dest, options={})
  check_options(options, OPT_TABLE['RENAME'])
  unless dest.start_with?('/')
    dest = '/' + dest
  end
  res = operate_requests('PUT', path, 'RENAME', options.merge({'destination' => dest}))
  check_success_json(res, 'boolean')
end

#request(host, port, method, path, op = nil, params = {}, payload = nil, header = nil, retries = 0) ⇒ Object

IllegalArgumentException 400 Bad Request UnsupportedOperationException 400 Bad Request SecurityException 401 Unauthorized IOException 403 Forbidden FileNotFoundException 404 Not Found RumtimeException 500 Internal Server Error



258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
# File 'lib/webhdfs/client_v1.rb', line 258

def request(host, port, method, path, op=nil, params={}, payload=nil, header=nil, retries=0)
  conn = Net::HTTP.new(host, port, @proxy_address, @proxy_port)
  conn.proxy_user = @proxy_user if @proxy_user
  conn.proxy_pass = @proxy_pass if @proxy_pass
  conn.open_timeout = @open_timeout if @open_timeout
  conn.read_timeout = @read_timeout if @read_timeout

  request_path = if op
                   build_path(path, op, params)
                 else
                   path
                 end

  res = nil
  if !payload.nil? and payload.is_a?(IO)
    req = Net::HTTPGenericRequest.new(method,(payload ? true : false),true,request_path,header)
    raise WebHDFS::IOError, 'Error reading given IO data source' unless payload.respond_to? :read and payload.respond_to? :size
    raise WebHDFS::ClientError, 'Error accepting given IO resource as data payload, Not valid in methods other than PUT and POST' unless (method == 'PUT' or method == 'POST')

    req.body_stream = payload
    req.content_length = payload.size
    res = conn.request(req)
  else
    res = conn.send_request(method, request_path, payload, header)
  end

  case res
  when Net::HTTPSuccess
    res
  when Net::HTTPRedirection
    res
  else
    message = if res.body and not res.body.empty?
                res.body.gsub(/\n/, '')
              else
                'Response body is empty...'
              end

    if @retry_known_errors && retries < @retry_times
      detail = nil
      if message =~ /^\{"RemoteException":\{/
        begin
          detail = JSON.parse(message)
        rescue
          # ignore broken json response body
        end
      end
      if detail && detail['RemoteException'] && KNOWN_ERRORS.include?(detail['RemoteException']['exception'])
        sleep @retry_interval if @retry_interval > 0
        return request(host, port, method, path, op, params, payload, header, retries+1)
      end
    end

    case res.code
    when '400'
      raise WebHDFS::ClientError, message
    when '401'
      raise WebHDFS::SecurityError, message
    when '403'
      raise WebHDFS::IOError, message
    when '404'
      raise WebHDFS::FileNotFoundError, message
    when '500'
      raise WebHDFS::ServerError, message
    else
      raise WebHDFS::RequestFailedError, "response code:#{res.code}, message:#{message}"
    end
  end
end