Class: Flydata::Command::Sync

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

Direct Known Subclasses

Mysql

Defined Under Namespace

Classes: SyncDataEntryError

Constant Summary collapse

INSERT_PROGRESS_INTERVAL =
1000
SERVER_DATA_PROCESSING_TIMEOUT =

seconds

3600
STATUS_PARSING =

for dump.pos file

'PARSING'
STATUS_PARSED =

the value is different from the constant name on purpose for backward compatibility.

'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, 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, #show_purpose_name

Methods included from ExclusiveRunnable

included

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_tablesObject (readonly)

true if full initial sync


36
37
38
# File 'lib/flydata/command/sync.rb', line 36

def ddl_tables
  @ddl_tables
end

#full_initial_syncObject (readonly)

true if full initial sync


36
37
38
# File 'lib/flydata/command/sync.rb', line 36

def full_initial_sync
  @full_initial_sync
end

#full_tablesObject (readonly)

true if full initial sync


36
37
38
# File 'lib/flydata/command/sync.rb', line 36

def full_tables
  @full_tables
end

#input_tablesObject (readonly)

true if full initial sync


36
37
38
# File 'lib/flydata/command/sync.rb', line 36

def input_tables
  @input_tables
end

#new_tablesObject (readonly)

true if full initial sync


36
37
38
# File 'lib/flydata/command/sync.rb', line 36

def new_tables
  @new_tables
end

Class Method Details

.slopObject

Command: flydata sync

- Arguments

50
51
52
53
54
55
56
57
58
59
60
61
# File 'lib/flydata/command/sync.rb', line 50

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_fix_binlogposObject

Command: flydata sync:fix_binlogpos

- Arguments

266
267
268
269
270
# File 'lib/flydata/command/sync.rb', line 266

def self.slop_fix_binlogpos
  Slop.new do
    on 'f', 'force', 'update sent binlog position file forcibly'
  end
end

.slop_flushObject

Command: flydata sync:flush

- Arguments

119
120
121
122
123
124
# File 'lib/flydata/command/sync.rb', line 119

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_ddlObject

Command: flydata sync:generate_table_ddl

- Arguments

240
241
242
243
244
245
246
247
# File 'lib/flydata/command/sync.rb', line 240

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_repairObject

Command: flydata sync:repair

- Arguments

315
316
317
318
319
# File 'lib/flydata/command/sync.rb', line 315

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

.slop_resetObject

Command: flydata sync:reset

- Arguments

146
147
148
149
150
151
# File 'lib/flydata/command/sync.rb', line 146

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(options = {}) ⇒ Object


326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
# File 'lib/flydata/command/sync.rb', line 326

def check(options = {})
  status, pos_mismatch_tables, gap_tables = _check(options)

  if status.include? :OK
    message = "\nNo errors are found.  Sync is clean.\n"
  else
    message = "\nFollowing errors are found.\n"

    if status.include? :STUCK_AT_PROCESS
      message += " - Data is stuck while processing\n"
    end
    if status.include? :STUCK_AT_UPLOAD
      message += " - Data is stuck while uploading\n"
    end
    if status.include? :ABNORMAL_SHUTDOWN
      message += " - Agent was not shut down correctly\n"
    end
    if gap_tables
      message += " - Sync data is missing for the following table(s)\n"
      gap_tables.each do |bt|
        message += "     table:#{bt[:table]}\n"
      end
      message += "\n"
    end
    if pos_mismatch_tables
      message += " - Incorrect table position(s)\n"
      pos_mismatch_tables.each do |bt|
        message += "     table:#{bt[:table]}, agent position:#{bt[:agent_seq] ? bt[:agent_seq] : '(missing)'}, server position:#{bt[:server_seq]}\n"
      end
      message += "\n"
    end
  end
  log_info_stdout message
end

#fix_binlogposObject

Command: flydata sync:fix_binlogpos Set binlog path

- Entry method

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
# File 'lib/flydata/command/sync.rb', line 275

def fix_binlogpos
  de = data_entry
  sync_fm = create_sync_file_manager(de)

  if File.exists?(sync_fm.sent_binlog_path) && !opts.force?
    log_info_stdout("Skip creating sent binlogpos because sent position file is exist already. (#{sync_fm.sent_binlog_path})")
    return
  end

  if Flydata::Command::Sender.new.process_exist?
    log_warn_stderr("flydata is running. flydata process needs to be stopped with 'flydata stop'.")
    return
  end

  binlog_info = sync_fm.load_binlog
  if binlog_info.nil?
    log_info_stdout("Skip creating sent binlogpos because binlog position file is empty or invalid. (#{sync_fm.sent_binlog_path})")
    return
  end

  say("Updating binlog position files...")
  log_info("Updating binlog position files... Original binlog_info:#{binlog_info}")

  # Update binlog.sent.pos file
  #   -1 is because the position in binlog.pos is the next event's position.
  #   on the other hand the position in sent position indicates already processed.
  binlog_info[:pos] -= 1
  log_info("Updating sent position file. #{binlog_info} -> #{sync_fm.sent_binlog_path}")
  sync_fm.save_sent_binlog(binlog_info)

  # Update binlog.pos file to start from head of the current binlog file
  new_binlog_info = binlog_info.dup.tap{|h| h[:pos] = 4}  # 4 is the first position of binlog file.
  log_info("Updating original position file. #{new_binlog_info} -> #{sync_fm.binlog_path}")
  sync_fm.save_binlog(new_binlog_info)
  log_info_stdout("Done!")
end

#flush(*tables) ⇒ Object

Command: flydata sync:flush

- Entry method

128
129
130
131
132
133
134
135
136
137
138
139
140
141
# File 'lib/flydata/command/sync.rb', line 128

def flush(*tables)
  begin
    flush_buffer_and_stop(tables, skip_flush: opts.skip_flush?)
  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

Command: flydata sync:generate_table_ddl

- Entry method

251
252
253
254
255
256
257
258
259
260
261
# File 'lib/flydata/command/sync.rb', line 251

def generate_table_ddl(*tables)
  # Compatibility check
  de = data_entry
  dp = flydata.data_port.get
  Flydata::MysqlCompatibilityCheck.new(dp, de['mysql_data_entry_preference']).check

  # Set instance variables
  set_current_tables(tables, include_all_tables: true)

  do_generate_table_ddl(de)
end

#repairObject


321
322
323
# File 'lib/flydata/command/sync.rb', line 321

def repair
  _repair
end

#reset(*tables) ⇒ Object

Command: flydata sync:reset

- Entry method

155
156
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
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
# File 'lib/flydata/command/sync.rb', line 155

def reset(*tables)
  show_purpose_name

  # Set instance variables
  sync_resumed = set_current_tables(tables, resume: true)

  message = ''
  if sync_resumed && !tables.empty?
    log_info_stdout <<EOS
Initial sync is in progress.  In this case, you can only reset the initial sync.  To reset specific table(s), please resume and complete the initial sync by running the 'flydata start' command first.
If you'd like to reset the initial sync in progress, run the 'flydata reset' command with no arguments.
EOS
    return
  end

  # Flush client buffer
  msg_tables = @input_tables.empty? ? '' : " for these tables : #{@input_tables.join(" ")}"
  msg_sync_type = sync_resumed ? "the current initial sync" : "the current sync"
  return unless ask_yes_no("This resets #{msg_sync_type}#{msg_tables}.  Are you sure?")
  sender = Flydata::Command::Sender.new
  sender.flush_client_buffer # TODO We should rather delete buffer files
  sender.stop

  begin
    wait_for_server_buffer(timeout: SERVER_DATA_PROCESSING_TIMEOUT, tables: target_tables_for_api)
  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 tables on server
  de = data_entry
  cleanup_sync_server(de, @input_tables) unless opts.client?

  # Delete local files
  sync_fm = create_sync_file_manager(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.stats_path,
    sync_fm.table_position_file_paths(*@input_tables),
    sync_fm.table_binlog_pos_paths(*@input_tables),
    sync_fm.table_binlog_pos_init_paths(*@input_tables),
    sync_fm.table_rev_file_paths(*@input_tables),
    sync_fm.table_ddl_file_paths(*@input_tables)
  ]
  new_tables_after_reset = @unsynced_tables + @input_tables
  if @input_tables.empty? or @full_tables.empty? or @full_tables.all?{|ft| new_tables_after_reset.include?(ft)}
    delete_files << sync_fm.binlog_path
    delete_files << sync_fm.sent_binlog_path
  end
  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

Command: flydata sync

- Entry method

65
66
67
68
69
70
71
72
73
74
75
76
77
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 65

def run(*tables)
  # Process check
  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

  fluentd_started = false

  quiet_option = false

  start_fluentd = Proc.new do |binlog_pos|
    # Start continuous sync by starting fluentd process
    unless opts.no_flydata_start?
      log_info_stdout("Starting FlyData Agent...") unless quiet_option
      Flydata::Command::Sender.new.start(quiet: true)
      log_info_stdout("  -> Done") unless quiet_option
    end
    fluentd_started = true
  end

  quiet_option = true
  # Start initial sync with check
  handle_mysql_sync(tables, binlog_ready_callback: start_fluentd)
  quiet_option = false

  start_fluentd.call unless fluentd_started

  # Show message
  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

#skipObject

Depricated Command: flydata sync:skip skip initial sync


226
227
228
229
230
231
232
233
234
235
# File 'lib/flydata/command/sync.rb', line 226

def skip
  de = data_entry
  sync_fm = create_sync_file_manager(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(options) ⇒ Object

Public method

- Called from Sender#start/restart

110
111
112
113
114
115
# File 'lib/flydata/command/sync.rb', line 110

def try_mysql_sync(options)
  # Start initial sync
  handle_mysql_sync(nil, options)
rescue SyncDataEntryError
  return
end