import json import logging from data import model from data.database import ImageStorage from app import app, storage as store from data.database import db from gzip import GzipFile from StringIO import StringIO logger = logging.getLogger(__name__) def backfill_sizes_from_json(): query = (ImageStorage .select() .where(ImageStorage.uncompressed_size == None, ImageStorage.uploading == False) .limit(100)) total = 0 missing = 0 batch_processed = 1 while batch_processed > 0: batch_processed = 0 with app.config['DB_TRANSACTION_FACTORY'](db): for image_storage in query.clone(): total += 1 batch_processed += 1 if (total - 1) % 100 == 0: logger.debug('Storing entry: %s', total) # Lookup the JSON for the image. uuid = image_storage.uuid with_locations = model.get_storage_by_uuid(uuid) json_string = store.get_content(with_locations.locations, store.image_json_path(uuid)) json_data = json.loads(json_string) size = json_data.get('Size', json_data.get('size', -1)) if size == -1: missing += 1 logger.debug('Missing entry %s (%s/%s)', uuid, missing, total) image_storage.uncompressed_size = size image_storage.save() def backfill_sizes_from_data(): storage_ids = list(ImageStorage .select(ImageStorage.uuid) .where(ImageStorage.uncompressed_size == -1, ImageStorage.uploading == False)) counter = 0 for uuid in [s.uuid for s in storage_ids]: counter += 1 # Load the storage with locations. logger.debug('Loading entry: %s (%s/%s)', uuid, counter, len(storage_ids)) with_locations = model.get_storage_by_uuid(uuid) layer_size = -2 # Read the layer from backing storage. logger.debug('Reading entry: %s (%s bytes)', uuid, with_locations.image_size) try: layer_data = store.get_content(with_locations.locations, store.image_layer_path(uuid)) except Exception as ex: logger.debug('Could not read entry: %s. Reason: %s', uuid, ex) continue # Calculate the uncompressed size. try: layer_stream = StringIO(layer_data) with GzipFile(fileobj=layer_stream, mode='rb') as gzip_file: layer_size = len(gzip_file.read()) logger.debug('%s' % gzip_file.size) layer_stream.close() except Exception as ex: logger.debug('Could not gunzip entry: %s. Reason: %s', uuid, ex) continue # Write the size to the image storage. We do so under a transaction AFTER checking to # make sure the image storage still exists and has not changed. logger.debug('Writing entry: %s. Size: %s', uuid, layer_size) with app.config['DB_TRANSACTION_FACTORY'](db): try: current_record = model.get_storage_by_uuid(uuid) except: # Record no longer exists. continue if not current_record.uploading and current_record.uncompressed_size == -1: current_record.uncompressed_size = layer_size current_record.save() if __name__ == "__main__": logging.basicConfig(level=logging.DEBUG) logging.getLogger('boto').setLevel(logging.CRITICAL) backfill_sizes_from_data()