Class: Flydata::Command::Sync

Inherits:
Base
  • Object
show all
Includes:
Helpers
Defined in:
lib/flydata/command/sync.rb

Defined Under Namespace

Classes: SyncDataEntryError

Constant Summary collapse

INSERT_PROGRESS_INTERVAL = 
1000
SERVER_DATA_PROCESSING_TIMEOUT =

seconds

600
STATUS_PARSING =

for dump.pos file

'PARSING'
STATUS_WAITING = 
'WAITING'
STATUS_COMPLETE = 
'COMPLETE'

Constants included from Helpers

Helpers::UNIT_PREFIX

Instance Attribute Summary collapse

Attributes inherited from Base

#opts

Class Method Summary collapse

Instance Method Summary collapse

Methods included from Helpers

as_size, development?, env_mode, env_suffix, flydata_api_host_file, flydata_conf_file, flydata_version, format_menu_list, parse_command, retry_on, to_command_class, usage_text

Methods inherited from Base

#ask_input_table_name, #ask_yes_no, #choose_one, #flydata, #initialize, #newline, #register_crontab, #retrieve_data_entries, #separator

Methods included from Flydata::CommandLoggable

#before_logging, #log_error_stderr, #log_info_stdout, #log_warn_stderr

Constructor Details

This class inherits a constructor from Flydata::Command::Base

Instance Attribute Details

#ddl_tables ⇒ Object (readonly)

Returns the value of attribute ddl_tables.



27
28
29
 
# File 'lib/flydata/command/sync.rb', line 27

def ddl_tables
  @ddl_tables
end

#initial_sync(tables, de = nil) ⇒ Object (readonly)

Returns the value of attribute initial_sync.



27
28
29
 
# File 'lib/flydata/command/sync.rb', line 27

def initial_sync
  @initial_sync
end

#new_tables ⇒ Object (readonly)

Returns the value of attribute new_tables.



27
28
29
 
# File 'lib/flydata/command/sync.rb', line 27

def new_tables
  @new_tables
end

Class Method Details

.slop ⇒ Object 



32
33
34
35
36
37
38
39
40
41
42
43
 
# File 'lib/flydata/command/sync.rb', line 32

def self.slop
  Slop.new do
    on 'c', 'skip-cleanup', 'Skip server cleanup'
    on 'f', 'skip-flush', 'Skip server flush'
    on 'y', 'yes', 'Skip command prompt assuming yes to all questions.  Use this for batch operation.'
    on 'd', 'dump-file', 'Dump mysqldump into a file. Use this for debugging after making sure the free space.' # dummy for compatibility
    on 's', 'dump-stream', 'Dump mysqldump stream instead of saving dump file. It might cause timeout error if db size is larger than 10GB.'
    on 'n', 'no-flydata-start', 'Don\'t start the flydata agent after initial sync.'
    #TODO : This option is temp! Should remove soon.
    on 'ff', 'Skip checking query queue and flush'
  end
end

.slop_flush ⇒ Object 



128
129
130
131
132
133
 
# File 'lib/flydata/command/sync.rb', line 128

def self.slop_flush
  Slop.new do
    on 'f', 'skip-flush', 'Skip server flush'
    on 'y', 'yes', 'Skip command prompt assuming yes to all questions.  Use this for batch operation.'
  end
end

.slop_generate_table_ddl ⇒ Object 



280
281
282
283
284
285
286
287
 
# File 'lib/flydata/command/sync.rb', line 280

def self.slop_generate_table_ddl
  Slop.new do
    on 'c', 'ctl-only', 'Only generate FlyData Control definitions'
    on 'y', 'yes', 'Skip command prompt assuming yes to all questions.  Use this for batch operation.'
    on 's', 'skip-primary-key-check', 'Skip primary key check when generating DDL'
    on 'all-tables', 'Generate all table schema'
  end
end

.slop_reset ⇒ Object 



150
151
152
153
154
155
 
# File 'lib/flydata/command/sync.rb', line 150

def self.slop_reset
  Slop.new do
    on 'c', 'client', 'Resets client only.'
    on 'y', 'yes', 'Skip command prompt assuming yes to all questions.  Use this for batch operation.'
  end
end

Instance Method Details

#check ⇒ Object 



256
257
258
259
260
261
262
263
264
265
266
 
# File 'lib/flydata/command/sync.rb', line 256

def check
  de = retrieve_sync_data_entry
  retry_on(RestClient::Exception) do
    status = do_check(de)
    if status['complete']
      nil
    else
      status
    end
  end
end

#flush ⇒ Object 



135
136
137
138
139
140
141
142
143
144
145
146
147
148
 
# File 'lib/flydata/command/sync.rb', line 135

def flush
  begin
    flush_buffer_and_stop
  rescue ServerDataProcessingTimeout => e
    ee = ServerDataProcessingTimeout.new("Delayed Data Processing")
    ee.description = <<EOS
  Data processing is taking more than expected.  Please contact [email protected] to check the system status.

EOS
    ee.set_backtrace e.backtrace
    raise ee
  end
  log_info_stdout("Buffers have been flushed and the sender process has been stopped.")
end

#generate_table_ddl(*tables) ⇒ Object 



289
290
291
292
293
294
 
# File 'lib/flydata/command/sync.rb', line 289

def generate_table_ddl(*tables)
  de = retrieve_sync_data_entry
  dp = flydata.data_port.get
  Flydata::MysqlCompatibilityCheck.new(dp, de['mysql_data_entry_preference']).check
  do_generate_table_ddl(de, tables)
end

#handle_mysql_sync(tables = nil) ⇒ Object 



78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
 
# File 'lib/flydata/command/sync.rb', line 78

def handle_mysql_sync(tables=nil)
  de = retrieve_sync_data_entry

  set_current_tables(de)
  verify_input_tables(tables, de['mysql_data_entry_preference']['tables'])

  unless @new_tables.empty?
    say("We've noticed that these tables have not been synced yet: #{@new_tables.join(", ")}")
    unless @ddl_tables.empty?
      say("  WARNING: We've noticed that at least one of these tables have not had their DDL generated yet.")
      say("    We recommend you run our 'flydata sync:generate_table_ddl > create_table.sql'")
      say("    to generate SQL to run on Redshift to create the correct tables")
      say("    Without running this sql on your Redshift cluster, there may be issues with your data")
    end
    if ask_yes_no("Do you want to run initial sync on all of these tables now?")
      tables = @initial_sync ? [] : @new_tables
      initial_sync(tables, de)
    else
      #If generate_table_ddl has not been run for these tables, warn user
      unless @ddl_tables.empty?
        say("  You can generate DDL SQL for your new tables by running this command")
        say("    $> flydata sync:generate_table_ddl > create_table.sql")
      end
      puts "Without syncing these tables, we cannot start the flydata process"
      raise "Please try again"
    end
  end
end

#reset(*tables) ⇒ Object 



157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
 
# File 'lib/flydata/command/sync.rb', line 157

def reset(*tables)
  msg = tables.empty? ? '' : " for these tables : #{tables.join(" ")}"
  return unless ask_yes_no("This resets the current sync#{msg}.  Are you sure?")
  sender = Flydata::Command::Sender.new
  sender.flush_client_buffer # TODO We should rather delete buffer files
  sender.stop

  de = retrieve_sync_data_entry
  all_tables = de['mysql_data_entry_preference']['tables']
  verify_input_tables(tables, all_tables)

  begin
    wait_for_server_buffer(SERVER_DATA_PROCESSING_TIMEOUT)
  rescue ServerDataProcessingTimeout => e
    ee = ServerDataProcessingTimeout.new("Delayed Data Processing")
    ee.description = <<EOS
  Data processing is taking more than expected.  Please contact [email protected] to check the system status.
  Once checked, you can continue sync reset with the following command

    flydata sync:reset #{tables.empty? ? '' : tables.join(" ")}

EOS
    ee.set_backtrace e.backtrace
    raise ee
  end
  cleanup_sync_server(de, tables) unless opts.client?
  sync_fm = Flydata::FileUtil::SyncFileManager.new(de)
  delete_files = [
    sync_fm.dump_file_path,
    sync_fm.dump_pos_path,
    sync_fm.mysql_table_marshal_dump_path,
    sync_fm.sync_info_file,
    sync_fm.table_position_file_paths(*tables),
    sync_fm.table_binlog_pos_paths(*tables),
    sync_fm.table_binlog_pos_init_paths(*tables),
    sync_fm.table_rev_file_paths(*tables),
    sync_fm.table_ddl_file_paths(*tables)
  ]
  delete_files << sync_fm.binlog_path if tables.empty? or all_tables.empty?
  delete_files.flatten.each do |path|
    FileUtils.rm(path) if File.exists?(path)
  end
  sync_fm.close
  log_info_stdout("Reset completed successfully.")
end

#run(*tables) ⇒ Object 



45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
 
# File 'lib/flydata/command/sync.rb', line 45

def run(*tables)
  sender = Flydata::Command::Sender.new
  if (sender.process_exist?)
    if tables.empty?
      # full sync
      log_warn_stderr("FlyData Agent is already running.  If you'd like to restart FlyData Sync from scratch, run 'flydata sync:reset' first.")
    else
      # per-table sync
      log_warn_stderr("Flydata Agent is already running.  If you'd like to Sync the table(s), run 'flydata sync:flush' first.")
    end
    exit 1
  end

  handle_mysql_sync(tables)

  unless opts.no_flydata_start?
    log_info_stdout("Starting FlyData Agent...")
    Flydata::Command::Sender.new.start(quiet: true)
    log_info_stdout("  -> Done")
  end

  dashboard_url = "#{flydata.flydata_api_host}/dashboard"
  redshift_console_url = "#{flydata.flydata_api_host}/redshift_clusters/query/new"
  last_message = ALL_DONE_MESSAGE_TEMPLATE % [redshift_console_url, dashboard_url]
  log_info_stdout(last_message)
end

#skip ⇒ Object

skip initial sync



269
270
271
272
273
274
275
276
277
278
 
# File 'lib/flydata/command/sync.rb', line 269

def skip
  de = retrieve_sync_data_entry
  sync_fm = Flydata::FileUtil::SyncFileManager.new(de)
  binlog_path = sync_fm.binlog_path
  sync_fm.close
  `touch #{binlog_path}`
  log_info_stdout("Created an empty binlog position file.")
  log_info_stdout("-> #{binlog_path}")
  log_info_stdout("Run 'flydata start' to start continuous sync.")
end

#try_mysql_sync ⇒ Object 



72
73
74
75
76
 
# File 'lib/flydata/command/sync.rb', line 72

def try_mysql_sync
  handle_mysql_sync()
rescue SyncDataEntryError
  return
end

#wait_for_server_buffer(timeout = 0) ⇒ Object 



203
204
205
206
207
208
209
210
211
212
213
214
215
 
# File 'lib/flydata/command/sync.rb', line 203

def wait_for_server_buffer(timeout = 0)
  start_time = Time.now
  log_info_stdout("Waiting for the server buffer to get empty.")
  prev_message =nil
  while (status = check) && (status['state'] == 'processing')
    prev_message = status['message']
    if timeout > 0 && Time.now - start_time > timeout
      raise ServerDataProcessingTimeout.new
    end
    print_progress(status)
    sleep 10
  end
end

#wait_for_server_data_processing(timeout = 0) ⇒ Object 



217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
 
# File 'lib/flydata/command/sync.rb', line 217

def wait_for_server_data_processing(timeout = 0)
  state = :PROCESS
  start_time = Time.now
  log_info_stdout("Uploading data to Redshift...")
  sleep 10
  status = nil
  prev_message =nil
  while (status = check)
    if state == :PROCESS && status['state'] == 'uploading'
      log_info_stdout("  -> Done")
      state = :UPLOAD
      log_info_stdout("Finishing data upload...")
    end

    #TODO This is based on a temporary option
    if state == :UPLOAD && opts.ff?
      log_info_stdout("Skip checking for pending uploads")
      break
    end

    if status['message'] != prev_message
      # making some progress.  Reset timer
      start_time = Time.now
    end
    prev_message = status['message']
    if timeout > 0 && Time.now - start_time > timeout
      raise ServerDataProcessingTimeout.new
    end
    print_progress(status)
    sleep 10
  end
  if (state == :PROCESS)
    # :UPLOAD state was skipped due to no data
    log_info_stdout("  -> Done")
    log_info_stdout("Finishing data upload...")
  end
  log_info_stdout("  -> Done")
end