Class: Flydata::SyncFileManager

Inherits:
Object
  • Object
show all
Defined in:
lib/flydata/sync_file_manager.rb

Constant Summary collapse

DUMP_DIR = 
ENV['FLYDATA_DUMP'] || File.join(FLYDATA_HOME, 'dump')
BACKUP_DIR = 
ENV['FLYDATA_BACKUP'] || File.join(FLYDATA_HOME, 'backup')
TABLE_POSITIONS_DIR = 
ENV['FLYDATA_TABLE_POSITIONS'] || File.join(FLYDATA_HOME, 'positions')
SYNC_TABLE_POSITIONS = 
0
SOURCE_TABLE_EXT = 
"mysql_table"

Instance Method Summary collapse

Constructor Details

#initialize(data_entry, source = nil) ⇒ SyncFileManager

Returns a new instance of SyncFileManager.



15
16
17
18
19
20
 
# File 'lib/flydata/sync_file_manager.rb', line 15

def initialize(data_entry, source = nil)
  @data_entry = data_entry
  @source = source #for Source dependent objects
  @table_position_files = {} # File objects keyed by table name
  @sync_table_positions_count = SYNC_TABLE_POSITIONS
end

Instance Method Details

#backup_dir ⇒ Object 



441
442
443
 
# File 'lib/flydata/sync_file_manager.rb', line 441

def backup_dir
  BACKUP_DIR
end

#backup_dump_dir ⇒ Object 



431
432
433
434
435
436
437
438
439
 
# File 'lib/flydata/sync_file_manager.rb', line 431

def backup_dump_dir
  backup_dir = BACKUP_DIR.dup
  FileUtils.mkdir_p(backup_dir) unless Dir.exists?(backup_dir)
  dest_dir = File.join(backup_dir, Time.now.strftime("%Y%m%d%H%M%S"))
  FileUtils.mkdir(dest_dir)
  ['info', 'pos', 'stats', SOURCE_TABLE_EXT].each do |ext|
    FileUtils.mv(Dir.glob("#{dump_dir}/*.#{ext}"), dest_dir)
  end
end

#close ⇒ Object 



26
27
28
29
 
# File 'lib/flydata/sync_file_manager.rb', line 26

def close
  @table_position_files.values.each {|f| f.close }
  @table_position_files = {}
end

#delete_dump_file ⇒ Object 



427
428
429
 
# File 'lib/flydata/sync_file_manager.rb', line 427

def delete_dump_file
  FileUtils.rm(dump_file_path) if File.exists?(dump_file_path)
end

#delete_dump_files ⇒ Object 



346
347
348
349
350
351
352
353
354
355
356
357
 
# File 'lib/flydata/sync_file_manager.rb', line 346

def delete_dump_files
  files_to_delete = [
      dump_file_path,
      dump_pos_path,
      source_table_marshal_dump_path,
      sync_info_file,
      stats_path
  ]
  files_to_delete.flatten.each do |file_to_delete|
    FileUtils.rm(file_to_delete) if File.exists?(file_to_delete)
  end
end

#delete_master_position_files ⇒ Object 



359
360
361
362
363
364
365
366
367
368
 
# File 'lib/flydata/sync_file_manager.rb', line 359

def delete_master_position_files
  files_to_delete = [
    source_pos_path,
    sent_source_pos_path,
    lock_pid_file,
  ]
  files_to_delete.flatten.each do |file_to_delete|
    FileUtils.rm(file_to_delete) if File.exists?(file_to_delete)
  end
end

#delete_table_control_files(*tables) ⇒ Object 



322
323
324
325
326
327
328
329
330
331
332
333
 
# File 'lib/flydata/sync_file_manager.rb', line 322

def delete_table_control_files(*tables)
  files_to_delete = [
      table_position_file_paths(*tables),
      table_source_pos_paths(*tables),
      table_source_pos_init_paths(*tables),
      table_rev_file_paths(*tables),
      table_ddl_file_paths(*tables)
  ]
  files_to_delete.flatten.each do |path|
    FileUtils.rm(path) if File.exists?(path)
  end
end

#delete_table_source_pos(table_name) ⇒ Object 



371
372
373
374
375
376
377
378
 
# File 'lib/flydata/sync_file_manager.rb', line 371

def delete_table_source_pos(table_name)
  file = File.join(table_positions_dir_path, table_name + ".binlog.pos")
  if File.exists?(file)
    FileUtils.rm(file, :force => true)
  else
    puts "#{file} does not exist. Something is wrong. Did you delete the file manually when flydata was running?"
  end
end

#dump_file_path ⇒ Object 



31
32
33
 
# File 'lib/flydata/sync_file_manager.rb', line 31

def dump_file_path
  File.join(dump_dir, @data_entry['name']) + ".dump"
end

#dump_pos_path ⇒ Object

dump pos file for resume



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

def dump_pos_path
  dump_file_path + ".pos"
end

#get_new_table_list(tables, file_type) ⇒ Object 



79
80
81
82
83
84
85
86
 
# File 'lib/flydata/sync_file_manager.rb', line 79

def get_new_table_list(tables, file_type)
  table_positions_dir_path = ENV['FLYDATA_TABLE_POSITIONS'] || File.join(FLYDATA_HOME, 'positions')
  new_tables = []
  tables.each do |table|
    new_tables << table unless File.exists?(File.join(table_positions_dir_path, "#{table}.#{file_type}"))
  end
  new_tables
end

#get_table_position(table_name) ⇒ Object 



256
257
258
259
260
261
 
# File 'lib/flydata/sync_file_manager.rb', line 256

def get_table_position(table_name)
  file = File.join(table_positions_dir_path, table_name + ".pos")
  return nil unless File.exists?(file)

  File.open(file) {|f| f.read}
end

#get_table_source_pos(table_name) ⇒ Object 



396
397
398
399
400
401
 
# File 'lib/flydata/sync_file_manager.rb', line 396

def get_table_source_pos(table_name)
  source_pos_str = get_table_source_raw_pos(table_name)
  return nil unless source_pos_str

  source.source_pos.create_source_pos( source_pos_str )
end

#get_table_source_pos_init(table_name) ⇒ Object 



284
285
286
287
288
289
 
# File 'lib/flydata/sync_file_manager.rb', line 284

def get_table_source_pos_init(table_name)
  file = File.join(table_positions_dir_path, table_name + ".binlog.pos.init")
  return nil unless File.exists?(file)

  source.source_pos.create_source_pos( File.open(file, 'r').readline )
end

#get_table_source_raw_pos(table_name) ⇒ Object

returns String. interface for fluentd



403
404
405
406
407
408
 
# File 'lib/flydata/sync_file_manager.rb', line 403

def get_table_source_raw_pos(table_name) #returns String. interface for fluentd
  file = File.join(table_positions_dir_path, table_name + ".binlog.pos")
  return nil unless File.exists?(file)

  File.open(file, 'r').readline
end

#increment_and_save_table_position(table_name) {|seq| ... } ⇒ Object

Read a sequence number from the table’s position file, increment the number and pass the number to a block. After executing the block, saves the value to the position file.

Yields:

  • (seq) 


212
213
214
215
216
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
 
# File 'lib/flydata/sync_file_manager.rb', line 212

def increment_and_save_table_position(table_name)
  file = File.join(table_positions_dir_path, table_name + ".pos")
  retry_count = 0
  begin
    @table_position_files[table_name] ||= (f = File.open(file, File::RDWR); f.sync = true; f)
  rescue Errno::ENOENT
    raise if retry_count > 0 # Already retried.  Must be a differentfile causing the error
    # File not exist.  Create one with initial value of '0'
    File.open(file, "w") {|f| f.write('0') }
    retry_count += 1
    retry
  end
  f = @table_position_files[table_name]
  seq = f.read
  prev_seq_len = seq.size
  f.rewind
  seq = seq.to_i + 1
  seq = FlydataCore::QueryJob::SYNC_FIRST_SEQ if seq == 1
  new_seq_len = seq.to_s.size

  seq_to_write =  seq.to_s
  if new_seq_len < prev_seq_len
    seq_to_write += " " * (prev_seq_len - new_seq_len)
  end
  # logical transaction starts
  yield(seq)
  f.write(seq_to_write)
  if @sync_table_positions_count > 1
    @sync_table_positions_count -= 1
  elsif @sync_table_positions_count == 1
    f.fsync
    @sync_table_positions_count = SYNC_TABLE_POSITIONS
  end
  # logical transaction ends
  f.truncate(new_seq_len) if new_seq_len < prev_seq_len
  f.rewind
end

#increment_table_rev(table_name, base_rev) ⇒ Object 



313
314
315
316
317
318
319
320
 
# File 'lib/flydata/sync_file_manager.rb', line 313

def increment_table_rev(table_name, base_rev)
  file = table_rev_file_path(table_name)
  new_rev = base_rev + 1
  File.open(file, "w") do |f|
    f.write(new_rev)
  end
  new_rev
end

#install_table_source_pos_files(tables) ⇒ Object 



412
413
414
415
416
417
418
419
420
421
422
423
424
425
 
# File 'lib/flydata/sync_file_manager.rb', line 412

def install_table_source_pos_files(tables)
  FileUtils.mkdir_p(table_positions_dir_path) unless Dir.exists?(table_positions_dir_path)
  tables.each do |table_name|
    file_name = table_name + ".binlog.pos"
    src_file = File.join(dump_dir, file_name)
    if ! File.exists?(src_file)
      raise "#{src_file} does not exist. Error!!"
    end
    FileUtils.mv(src_file, table_positions_dir_path)
    # save the position at initial sync.  this is used for repair if
    # necessary.
    FileUtils.cp(File.join(table_positions_dir_path, file_name), File.join(table_positions_dir_path, file_name + ".init"))
  end
end

#load_dump_pos ⇒ Object 



46
47
48
49
50
51
52
53
 
# File 'lib/flydata/sync_file_manager.rb', line 46

def load_dump_pos
  path = dump_pos_path
  return {} unless File.exists?(path)
  content = File.open(path, 'r').readline
  source_table = load_source_table_marshal_dump

  dump_pos_content_to_hash(content).merge( { source_table: source_table} )
end

#load_generated_ddl(tables) ⇒ Object 



55
56
57
58
59
60
61
62
63
64
65
 
# File 'lib/flydata/sync_file_manager.rb', line 55

def load_generated_ddl(tables)
  tables = [ tables ] unless tables.kind_of?(Array)
  paths = table_ddl_file_paths(*tables)
  paths.collect{|path|
    begin
      File.open(path) {|f| f.read }
    rescue Errno::ENOENT
      nil
    end
  }
end

#load_sent_source_pos(file_path = sent_source_pos_path) ⇒ Object 



133
134
135
136
137
138
139
140
141
142
143
 
# File 'lib/flydata/sync_file_manager.rb', line 133

def load_sent_source_pos(file_path = sent_source_pos_path)
  return nil unless File.exists?(file_path)

  source_pos_str = IO.read(file_path).strip
  begin
    source_pos = source.source_pos.create_source_pos(source_pos_str)
  rescue RuntimeError
    return nil
  end
  source_pos
end

#load_source_pos(file_path = source_pos_path) ⇒ Object 



109
110
111
112
113
114
115
116
117
118
119
120
 
# File 'lib/flydata/sync_file_manager.rb', line 109

def load_source_pos(file_path = source_pos_path)
  return nil unless File.exists?(file_path)

  source_pos_str = IO.read(file_path).strip
  begin
    context = source.source_pos
    source_pos = context.create_source_pos(source_pos_str)
  rescue RuntimeError
    return nil
  end
  source_pos
end

#load_stats ⇒ Object 



455
456
457
458
 
# File 'lib/flydata/sync_file_manager.rb', line 455

def load_stats
  return nil unless File.exists?(stats_path)
  Hash[*File.read(stats_path).split(/\t/)]
end

#load_sync_info ⇒ Object 



277
278
279
280
281
282
 
# File 'lib/flydata/sync_file_manager.rb', line 277

def load_sync_info
  return nil unless File.exists?(sync_info_file)
  items = File.open(sync_info_file, 'r').readline.split("\t")
  { initial_sync: (items[0] == 'true'),
    tables: items[1].split(" ") }
end

#lock_pid_file ⇒ Object 



263
264
265
 
# File 'lib/flydata/sync_file_manager.rb', line 263

def lock_pid_file
  FLYDATA_LOCK
end

#reset_table_position_files(tables) ⇒ Object

table files



177
178
179
180
181
182
 
# File 'lib/flydata/sync_file_manager.rb', line 177

def reset_table_position_files(tables)
  tables.each do |table_name|
    file = File.join(table_positions_dir_path, table_name + ".pos")
    File.open(file, "w") {|f| f.write('0') }
  end
end

#save_dump_pos(status, table_name, last_pos, source_pos, state = nil, substate = nil) ⇒ Object 



40
41
42
43
44
 
# File 'lib/flydata/sync_file_manager.rb', line 40

def save_dump_pos(status, table_name, last_pos, source_pos, state = nil, substate = nil)
  File.open(dump_pos_path, 'w') do |f|
    f.write(dump_pos_content(status, table_name, last_pos, source_pos, state, substate))
  end
end

#save_generated_ddl(tables, contents = "1") ⇒ Object 



67
68
69
70
71
72
73
74
75
76
77
 
# File 'lib/flydata/sync_file_manager.rb', line 67

def save_generated_ddl(tables, contents = "1")
  tables = [ tables ] unless tables.kind_of?(Array)
  table_positions_dir_path = ENV['FLYDATA_TABLE_POSITIONS'] || File.join(FLYDATA_HOME, 'positions')
  #Create positions if dir does not exist
  unless File.directory?(table_positions_dir_path)
    FileUtils.mkdir_p(table_positions_dir_path)
  end
  tables.each do |tab|
    File.open(File.join(table_positions_dir_path, "#{tab}.generated_ddl"), 'w') {|f| f.write(contents) }
  end
end

#save_record_count_stat(table, record_count) ⇒ Object 



449
450
451
452
453
 
# File 'lib/flydata/sync_file_manager.rb', line 449

def save_record_count_stat(table, record_count)
  stats = load_stats || Hash.new
  stats[table] = stats[table] ? stats[table].to_i + record_count : record_count
  save_stats(stats)
end

#save_sent_source_pos(source_pos) ⇒ Object

sent source pos file (binlog.pos)



127
128
129
130
131
 
# File 'lib/flydata/sync_file_manager.rb', line 127

def save_sent_source_pos(source_pos)
  File.open(sent_source_pos_path, 'w') do |f|
    f.write(source_pos.to_s)
  end
end

#save_source_pos(source_pos) ⇒ Object

master binlog.pos file



102
103
104
105
106
107
 
# File 'lib/flydata/sync_file_manager.rb', line 102

def save_source_pos(source_pos)
  path = source_pos_path
  File.open(path, 'w') do |f|
    f.write(source_pos.to_s)
  end
end

#save_source_table_marshal_dump(source_table) ⇒ Object 



95
96
97
98
99
 
# File 'lib/flydata/sync_file_manager.rb', line 95

def save_source_table_marshal_dump(source_table)
  File.open(source_table_marshal_dump_path, 'w') do |f|
    f.write Marshal.dump(source_table)
  end
end

#save_ssl_ca(ssl_ca_content, path = ssl_ca_path) ⇒ Object 



157
158
159
160
161
 
# File 'lib/flydata/sync_file_manager.rb', line 157

def save_ssl_ca(ssl_ca_content, path = ssl_ca_path)
  File.open(path, 'w') do |f|
    f.write(ssl_ca_content)
  end
end

#save_ssl_cipher(ssl_cipher_content, path = ssl_cipher_path) ⇒ Object 



170
171
172
173
174
 
# File 'lib/flydata/sync_file_manager.rb', line 170

def save_ssl_cipher(ssl_cipher_content, path = ssl_cipher_path)
  File.open(path, 'w') do |f|
    f.write(ssl_cipher_content)
  end
end

#save_sync_info(initial_sync, tables) ⇒ Object 



271
272
273
274
275
 
# File 'lib/flydata/sync_file_manager.rb', line 271

def save_sync_info(initial_sync, tables)
  File.open(sync_info_file, "w") do |f|
    f.write([initial_sync, tables.join(" ")].join("\t"))
  end
end

#save_table_position(table_name, seq) ⇒ Object 



250
251
252
253
254
 
# File 'lib/flydata/sync_file_manager.rb', line 250

def save_table_position(table_name, seq)
  file = File.join(table_positions_dir_path, table_name + ".pos")

  File.open(file, "w") {|f| f.write(seq)}
end

#save_table_source_pos(tables, source_pos, options = {}) ⇒ Object 



380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
 
# File 'lib/flydata/sync_file_manager.rb', line 380

def save_table_source_pos(tables, source_pos, options = {})
  dest_dir = case options[:destination]
             when :positions; table_positions_dir_path
             when :dump; dump_dir
             else dump_dir
             end

  tables = [ tables ] unless tables.kind_of?(Array)
  tables.each do |table_name|
    file = File.join(dest_dir, table_name + ".binlog.pos")
    File.open(file, "w") do |f|
      f.write(source_pos.to_s)
    end
  end
end

#sent_source_pos_path(master_source_pos_path = source_pos_path) ⇒ Object 



145
146
147
148
 
# File 'lib/flydata/sync_file_manager.rb', line 145

def sent_source_pos_path(master_source_pos_path = source_pos_path)
  validate_master_source_pos_path(master_source_pos_path)
  "#{master_source_pos_path[0..-5]}.sent.pos"
end

#source ⇒ Object 



22
23
24
 
# File 'lib/flydata/sync_file_manager.rb', line 22

def source
  @source
end

#source_pos_path ⇒ Object 



122
123
124
 
# File 'lib/flydata/sync_file_manager.rb', line 122

def source_pos_path
  File.join(FLYDATA_HOME, @data_entry['name'] + ".binlog.pos")
end

#source_table_marshal_dump_path ⇒ Object

SourceTable marshal file



91
92
93
 
# File 'lib/flydata/sync_file_manager.rb', line 91

def source_table_marshal_dump_path
  dump_file_path + ".#{SOURCE_TABLE_EXT}"
end

#ssl_ca_path(master_source_pos_path = source_pos_path) ⇒ Object

ssl_ca file path



151
152
153
154
155
 
# File 'lib/flydata/sync_file_manager.rb', line 151

def ssl_ca_path(master_source_pos_path = source_pos_path)
  validate_master_source_pos_path(master_source_pos_path)
  # <data-entry-name>.ssl_ca.pem
  "#{master_source_pos_path[0..-12]}.ssl_ca.pem"
end

#ssl_cipher_path(master_source_pos_path = source_pos_path) ⇒ Object

ssl_cipher file path



164
165
166
167
168
 
# File 'lib/flydata/sync_file_manager.rb', line 164

def ssl_cipher_path(master_source_pos_path = source_pos_path)
  validate_master_source_pos_path(master_source_pos_path)
  # <data-entry-name>.ssl_cipher
  "#{master_source_pos_path[0..-12]}.ssl_cipher"
end

#stats_path ⇒ Object 



445
446
447
 
# File 'lib/flydata/sync_file_manager.rb', line 445

def stats_path
  File.join(dump_dir, @data_entry['name']) + ".stats"
end

#sync_info_file ⇒ Object 



267
268
269
 
# File 'lib/flydata/sync_file_manager.rb', line 267

def sync_info_file
  File.join(dump_dir, "sync.info")
end

#table_ddl_file_paths(*tables) ⇒ Object 



193
194
195
196
 
# File 'lib/flydata/sync_file_manager.rb', line 193

def table_ddl_file_paths(*tables)
  tables.empty? ? Dir.glob(File.join(table_positions_dir_path, '*.generated_ddl')) :
    tables.map{|table| File.join(table_positions_dir_path, table + '.generated_ddl')}
end

#table_position_file_paths(*tables) ⇒ Object 



188
189
190
191
 
# File 'lib/flydata/sync_file_manager.rb', line 188

def table_position_file_paths(*tables)
  tables.empty? ? Dir.glob(File.join(table_positions_dir_path, '*.pos')) :
    tables.map{|table| File.join(table_positions_dir_path, table + '.pos')}
end

#table_positions_dir_path ⇒ Object 



184
185
186
 
# File 'lib/flydata/sync_file_manager.rb', line 184

def table_positions_dir_path
  TABLE_POSITIONS_DIR
end

#table_rev(table_name) ⇒ Object 



300
301
302
303
304
305
306
307
308
309
310
311
 
# File 'lib/flydata/sync_file_manager.rb', line 300

def table_rev(table_name)
  file = table_rev_file_path(table_name)
  return 1 unless File.exists?(file) #default revision is 1
  File.open(file, "r+") do |f|
    seq = f.read
    if seq.empty?
      return 1
    else
      return seq.to_i
    end
  end
end

#table_rev_file_path(table_name) ⇒ Object 



291
292
293
 
# File 'lib/flydata/sync_file_manager.rb', line 291

def table_rev_file_path(table_name)
  File.join(table_positions_dir_path, table_name + ".rev")
end

#table_rev_file_paths(*tables) ⇒ Object 



295
296
297
298
 
# File 'lib/flydata/sync_file_manager.rb', line 295

def table_rev_file_paths(*tables)
  tables.empty? ? Dir.glob(File.join(table_positions_dir_path, "*.rev")) :
    tables.map{|table| table_rev_file_path(table)}
end

#table_source_pos_init_paths(*tables) ⇒ Object 



203
204
205
206
 
# File 'lib/flydata/sync_file_manager.rb', line 203

def table_source_pos_init_paths(*tables)
  tables.empty? ? Dir.glob(File.join(table_positions_dir_path, '*.binlog.pos.init')) :
    tables.map{|table| File.join(table_positions_dir_path, table + '.binlog.pos.init')}
end

#table_source_pos_paths(*tables) ⇒ Object 



198
199
200
201
 
# File 'lib/flydata/sync_file_manager.rb', line 198

def table_source_pos_paths(*tables)
  tables.empty? ? Dir.glob(File.join(table_positions_dir_path, '*.binlog.pos')) :
    tables.map{|table| File.join(table_positions_dir_path, table + '.binlog.pos')}
end

#tables_from_positions_dir ⇒ Object 



335
336
337
338
339
340
341
342
343
344
 
# File 'lib/flydata/sync_file_manager.rb', line 335

def tables_from_positions_dir
  all_table_control_files = Dir.glob(File.join(table_positions_dir_path, '*.{pos,generated_ddl,init,rev}'))
  tables = Set.new
  all_table_control_files.each do |control_file|
    file_name = File.basename(control_file)
    file_name = file_name.slice(0...(file_name.index('.')))
    tables << file_name
  end
  tables.to_a
end