Class: Tapsoob::Operation::Pull

Inherits:
Base
  • Object
show all
Defined in:
lib/tapsoob/operation/pull.rb

Instance Attribute Summary

Attributes inherited from Base

#database_url, #dump_path, #opts

Class Method Summary collapse

Instance Method Summary collapse

Methods inherited from Base

#add_completed_table, #apply_table_filter, #can_use_pk_partitioning?, #catch_errors, #completed_tables, #completed_tables_mutex, #data?, #db, #default_chunksize, #exclude_tables, #exiting?, #format_number, #indexes_first?, #initialize, #load_table_order, #log, #parallel?, #parallel_workers, #resuming?, #save_table_order, #schema?, #setup_signal_trap, #store_session, #stream_state, #stream_state=, #table_filter, #table_parallel_workers

Constructor Details

This class inherits a constructor from Tapsoob::Operation::Base

Class Method Details

.factory(db, state) ⇒ Object 



369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
 
# File 'lib/tapsoob/operation/pull.rb', line 369

def self.factory(db, state)
  if defined?(Sequel::MySQL) && Sequel::MySQL.respond_to?(:convert_invalid_date_time=)
    Sequel::MySQL.convert_invalid_date_time = :nil
  end

  if state.has_key?(:klass)
    return eval(state[:klass]).new(db, state)
  end

  if Tapsoob::Utils.single_integer_primary_key(db, state[:table_name].to_sym)
    Tapsoob::DataStream::Keyed.new(db, state)
  else
    Tapsoob::DataStream::Base.new(db, state)
  end
end

Instance Method Details

#fetch_tables_info ⇒ Object 



359
360
361
362
363
364
365
366
367
 
# File 'lib/tapsoob/operation/pull.rb', line 359

def fetch_tables_info
  tables = db.send(:sort_dumped_tables, db.tables, {})

  data = {}
  apply_table_filter(tables).each do |table_name|
    data[table_name] = db[table_name].count
  end
  data
end

#file_prefix ⇒ Object 



9
10
11
 
# File 'lib/tapsoob/operation/pull.rb', line 9

def file_prefix
  "pull"
end

#initialize_dump_directory ⇒ Object 



32
33
34
35
36
37
38
39
40
 
# File 'lib/tapsoob/operation/pull.rb', line 32

def initialize_dump_directory
  %w[data schemas indexes].each do |subdir|
    dir_path = File.join(dump_path, subdir)
    FileUtils.rm_rf(dir_path)
    FileUtils.mkdir_p(dir_path)
  end

  FileUtils.rm_f(File.join(dump_path, "table_order.txt"))
end

#pull_data ⇒ Object 



62
63
64
65
66
67
68
69
70
71
72
73
74
75
 
# File 'lib/tapsoob/operation/pull.rb', line 62

def pull_data
  log.info "Receiving data"

  log.info "#{tables.size} tables, #{format_number(record_count)} records"
  Tapsoob::ProgressEvent.data_start(tables.size, record_count)

  if parallel?
    pull_data_parallel
  else
    pull_data_serial
  end

  Tapsoob::ProgressEvent.data_complete(tables.size, record_count)
end

#pull_data_from_table(stream, progress, total_records = nil) ⇒ Object 



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
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
 
# File 'lib/tapsoob/operation/pull.rb', line 167

def pull_data_from_table(stream, progress, total_records = nil)
  records_processed = 0

  loop do
    if exiting?
      store_session
      exit 0
    end

    row_size = 0
    chunksize = stream.state[:chunksize]

    begin
      chunksize = Tapsoob::Utils.calculate_chunksize(chunksize) do |c|
        stream.state[:chunksize] = c.to_i
        encoded_data, row_size, elapsed_time = nil
        d1 = c.time_delta do
          encoded_data, row_size, elapsed_time = stream.fetch
        end

        data = nil
        d2 = c.time_delta do
          data = {
            :state        => stream.to_hash,
            :checksum     => Tapsoob::Utils.checksum(encoded_data).to_s,
            :encoded_data => encoded_data
          }
        end

        stream.fetch_data_from_database(data) do |rows|
          next if rows == {}

          # Update progress bar by 1 chunk
          progress.inc(1) if progress

          # Track records processed for progress events
          if rows[:data]
            records_processed += rows[:data].size
            Tapsoob::ProgressEvent.table_progress(stream.table_name, records_processed, total_records) if total_records
          end

          if dump_path.nil?
            puts JSON.generate(rows)
          else
            Tapsoob::Utils.export_rows(dump_path, stream.table_name, rows)
          end
        end
        log.debug "row size: #{row_size}"
        stream.error = false
        self.stream_state = stream.to_hash

        c.idle_secs = (d1 + d2)

        elapsed_time
      end
    rescue Tapsoob::CorruptedData => e
      log.info "Corrupted Data Received #{e.message}, retrying..."
      stream.error = true
      next
    end

    break if stream.complete?
  end

  progress.finish if progress
  add_completed_table(stream.table_name)
  self.stream_state = {}

  # Emit final table complete event
  Tapsoob::ProgressEvent.table_complete(stream.table_name, records_processed)
end

#pull_data_from_table_parallel(table_name, row_count, num_workers, parent_progress = nil) ⇒ Object 



239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
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
327
328
329
330
331
332
333
334
335
336
337
338
339
340
 
# File 'lib/tapsoob/operation/pull.rb', line 239

def pull_data_from_table_parallel(table_name, row_count, num_workers, parent_progress = nil)
  # Mutex for coordinating file writes and progress tracking
  write_mutex = Mutex.new
  records_processed = 0

  begin
    # Determine partitioning strategy
    use_pk_partitioning = can_use_pk_partitioning?(table_name)

    if use_pk_partitioning
      # PK-based partitioning for efficient range queries
      ranges = Tapsoob::DataStream::Keyed.calculate_pk_ranges(db, table_name, num_workers)
      log.debug "Table #{table_name}: using PK-based partitioning with #{ranges.size} ranges"
    else
      # Interleaved chunking for tables without integer PK
      log.debug "Table #{table_name}: using interleaved chunking with #{num_workers} workers"
    end

    # Progress tracking - create ONE shared progress bar for the entire table
    estimated_chunks = [(row_count.to_f / default_chunksize).ceil, 1].max
    shared_progress = parent_progress ? parent_progress.create_bar(table_name.to_s, estimated_chunks) : nil

    workers = (0...num_workers).map do |worker_id|
      Thread.new do
        # Create worker-specific stream
        if use_pk_partitioning
          min_pk, max_pk = ranges[worker_id]
          stream = Tapsoob::DataStream::KeyedPartition.new(db, {
            :table_name => table_name,
            :chunksize => default_chunksize,
            :partition_range => [min_pk, max_pk]
          }, { :debug => opts[:debug] })
        else
          stream = Tapsoob::DataStream::Interleaved.new(db, {
            :table_name => table_name,
            :chunksize => default_chunksize,
            :worker_id => worker_id,
            :num_workers => num_workers
          }, { :debug => opts[:debug] })
        end

        # Process data chunks
        loop do
          break if exiting? || stream.complete?

          begin
            encoded_data, row_size, elapsed_time = stream.fetch

            # Skip processing empty results
            if row_size.positive?
              data = {
                :state => stream.to_hash,
                :checksum => Tapsoob::Utils.checksum(encoded_data).to_s,
                :encoded_data => encoded_data
              }

              stream.fetch_data_from_database(data) do |rows|
                next if rows == {}

                # Thread-safe file write and progress tracking
                write_mutex.synchronize do
                  if dump_path.nil?
                    puts JSON.generate(rows)
                  else
                    Tapsoob::Utils.export_rows(dump_path, stream.table_name, rows)
                  end

                  # Track records for progress events
                  if rows[:data]
                    records_processed += rows[:data].size
                    Tapsoob::ProgressEvent.table_progress(table_name, records_processed, row_count)
                  end
                end

                shared_progress.inc(1) if shared_progress
              end
            end

            # Check completion AFTER processing data to avoid losing the last chunk
            break if stream.complete?

          rescue Tapsoob::CorruptedData => e
            log.info "Worker #{worker_id}: Corrupted Data Received #{e.message}, retrying..."
            next
          rescue StandardError => e
            log.error "Worker #{worker_id} error: #{e.message}"
            log.error e.backtrace.join("\n")
            raise
          end
        end
      end
    end

    workers.each(&:join)
    shared_progress.finish if shared_progress

    add_completed_table(table_name)
  ensure
    # Always emit table_complete event, even if there was an error
    Tapsoob::ProgressEvent.table_complete(table_name, records_processed)
  end
end

#pull_data_parallel ⇒ Object 



99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
 
# File 'lib/tapsoob/operation/pull.rb', line 99

def pull_data_parallel
  log.info "Using #{parallel_workers} parallel workers for table-level parallelization"

  # Reserve space for both table-level and intra-table workers
  # With 4 table workers and potentially 8 intra-table workers per table,
  # we could have many concurrent progress bars. Show up to 8 at once.
  max_visible_bars = 8
  multi_progress = opts[:progress] ? Tapsoob::Progress::MultiBar.new(max_visible_bars) : nil
  table_queue = Queue.new
  tables.each { |table_name, count| table_queue << [table_name, count] }

  workers = (1..parallel_workers).map do
    Thread.new do
      loop do
        break if table_queue.empty?

        table_name, count = table_queue.pop(true) rescue break

        # Check if this table should use intra-table parallelization
        table_workers = table_parallel_workers(table_name, count)

        if table_workers > 1
          # Large table - use intra-table parallelization
          info_msg = "Table #{table_name}: using #{table_workers} workers for #{format_number(count)} records"
          if multi_progress
            multi_progress.set_info(info_msg)
          else
            log.info info_msg
          end

          Tapsoob::ProgressEvent.table_start(table_name, count, workers: table_workers)
          # Run intra-table parallelization, passing parent progress bar
          pull_data_from_table_parallel(table_name, count, table_workers, multi_progress)
        else
          # Small table - use single-threaded processing
          Tapsoob::ProgressEvent.table_start(table_name, count)
          stream = Tapsoob::DataStream::Base.factory(db, {
            :chunksize  => default_chunksize,
            :table_name => table_name
          }, { :debug => opts[:debug] })

          estimated_chunks = [(count.to_f / default_chunksize).ceil, 1].max
          progress = multi_progress ? multi_progress.create_bar(table_name.to_s, estimated_chunks) : nil

          pull_data_from_table(stream, progress, count)
        end
      end
    end
  end

  workers.each(&:join)
  multi_progress.stop if multi_progress
end

#pull_data_serial ⇒ Object 



77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
 
# File 'lib/tapsoob/operation/pull.rb', line 77

def pull_data_serial
  tables.each do |table_name, count|
    # Auto-detect if we should use intra-table parallelization
    table_workers = table_parallel_workers(table_name, count)

    if table_workers > 1
      log.info "Table #{table_name}: using #{table_workers} workers for #{format_number(count)} records"
      Tapsoob::ProgressEvent.table_start(table_name, count, workers: table_workers)
      pull_data_from_table_parallel(table_name, count, table_workers)
    else
      Tapsoob::ProgressEvent.table_start(table_name, count)
      stream = Tapsoob::DataStream::Base.factory(db, {
        :chunksize  => default_chunksize,
        :table_name => table_name
      }, { :debug => opts[:debug] })
      estimated_chunks = [(count.to_f / default_chunksize).ceil, 1].max
      progress = (opts[:progress] ? Tapsoob::Progress::Bar.new(table_name.to_s, estimated_chunks) : nil)
      pull_data_from_table(stream, progress, count)
    end
  end
end

#pull_indexes ⇒ Object 



385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
 
# File 'lib/tapsoob/operation/pull.rb', line 385

def pull_indexes
  log.info "Receiving indexes"

  raw_idxs = Tapsoob::Schema.indexes_individual(database_url)
  idxs     = (raw_idxs && raw_idxs.length >= 2 ? JSON.parse(raw_idxs) : {})

  # Calculate max title width for consistent alignment
  filtered_idxs = apply_table_filter(idxs).select { |table, indexes| indexes.size > 0 }
  Tapsoob::ProgressEvent.indexes_start(filtered_idxs.size)
  max_title_width = filtered_idxs.keys.map { |table| "#{table} indexes".length }.max || 14

  filtered_idxs.each do |table, indexes|
    progress = opts[:progress] ? Tapsoob::Progress::Bar.new("#{table} indexes", indexes.size, STDOUT, max_title_width) : nil
    indexes.each do |idx|
      output = Tapsoob::Utils.export_indexes(dump_path, table, idx)
      puts output if dump_path.nil? && output
      progress.inc(1) if progress
    end
    progress.finish if progress
  end
  Tapsoob::ProgressEvent.indexes_complete(filtered_idxs.size)
end

#pull_partial_data ⇒ Object 



153
154
155
156
157
158
159
160
161
162
163
164
165
 
# File 'lib/tapsoob/operation/pull.rb', line 153

def pull_partial_data
  return if stream_state == {}

  table_name = stream_state[:table_name]
  record_count = tables[table_name.to_s]
  log.info "Resuming #{table_name}, #{format_number(record_count)} records"

  stream = Tapsoob::DataStream::Base.factory(db, stream_state)
  chunksize = stream_state[:chunksize] || default_chunksize
  estimated_chunks = [(record_count.to_f / chunksize).ceil, 1].max
  progress = (opts[:progress] ? Tapsoob::Progress::Bar.new(table_name.to_s, estimated_chunks) : nil)
  pull_data_from_table(stream, progress)
end

#pull_reset_sequences ⇒ Object 



408
409
410
411
412
413
414
415
416
 
# File 'lib/tapsoob/operation/pull.rb', line 408

def pull_reset_sequences
  log.info "Resetting sequences"
  Tapsoob::ProgressEvent.sequences_start

  output = Tapsoob::Utils.schema_bin(:reset_db_sequences, database_url)
  puts output if dump_path.nil? && output

  Tapsoob::ProgressEvent.sequences_complete
end

#pull_schema ⇒ Object 



42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
 
# File 'lib/tapsoob/operation/pull.rb', line 42

def pull_schema
  log.info "Receiving schema"
  Tapsoob::ProgressEvent.schema_start(tables.size)

  progress = opts[:progress] ? Tapsoob::Progress::Bar.new('Schema', tables.size) : nil
  tables.each do |table_name, count|
    # Reuse existing db connection for better performance
    schema_data = Tapsoob::Schema.dump_table(db, table_name, @opts.slice(:indexes, :same_db))
    log.debug "Table: #{table_name}\n#{schema_data}\n"
    output = Tapsoob::Utils.export_schema(dump_path, table_name, schema_data)
    puts output if dump_path.nil? && output
    progress.inc(1) if progress
  end
  progress.finish if progress
  Tapsoob::ProgressEvent.schema_complete(tables.size)

  # Save table order for dependency-aware schema loading during push
  save_table_order(tables.keys) if dump_path
end

#record_count ⇒ Object 



351
352
353
 
# File 'lib/tapsoob/operation/pull.rb', line 351

def record_count
  tables_info.values.inject(:+)
end

#run ⇒ Object 



17
18
19
20
21
22
23
24
25
26
27
28
29
30
 
# File 'lib/tapsoob/operation/pull.rb', line 17

def run
  catch_errors do
    unless resuming?
      initialize_dump_directory if dump_path
      pull_schema if schema?
      pull_indexes if indexes_first? && schema?
    end
    setup_signal_trap
    pull_partial_data if data? && resuming?
    pull_data if data?
    pull_indexes if !indexes_first? && schema?
    pull_reset_sequences
  end
end

#tables ⇒ Object 



342
343
344
345
346
347
348
349
 
# File 'lib/tapsoob/operation/pull.rb', line 342

def tables
  h = {}
  tables_info.each do |table_name, count|
    next if completed_tables.include?(table_name.to_s)
    h[table_name.to_s] = count
  end
  h
end

#tables_info ⇒ Object 



355
356
357
 
# File 'lib/tapsoob/operation/pull.rb', line 355

def tables_info
  opts[:tables_info] ||= fetch_tables_info
end

#to_hash ⇒ Object 



13
14
15
 
# File 'lib/tapsoob/operation/pull.rb', line 13

def to_hash
  super.merge(:remote_tables_info => remote_tables_info)
end