def test_lo_chunk_reuse(self):
"""Test that chunks that are reused are only downloaded once"""
target = casync.parse_caibx(self.manifest_fn)
sources = [('target', casync.FileChunkReader(self.target_lo),
casync.build_chunk_dict(target))]
sources += [('remote', casync.RemoteChunkReader(self.store_fn),
casync.build_chunk_dict(target))]
stats = casync.extract(target, sources, self.target_lo)
with open(self.target_lo, 'rb') as f:
self.assertEqual(f.read(len(self.contents)), self.contents)
self.assertLess(stats['remote'], len(self.contents))
def extract_casync_image(target_slot_number: int, partition: dict, cloudlog):
path = get_partition_path(target_slot_number, partition)
seed_path = path[:-1] + ('b' if path[-1] == 'a' else 'a')
target = casync.parse_caibx(partition['casync_caibx'])
sources: List[Tuple[str, casync.ChunkReader, casync.ChunkDict]] = []
# First source is the current partition.
try:
raw_hash = get_raw_hash(seed_path, partition['size'])
caibx_url = f"{CAIBX_URL}{partition['name']}-{raw_hash}.caibx"
try:
cloudlog.info(f"casync fetching {caibx_url}")
sources += [
('seed', casync.FileChunkReader(seed_path),
casync.build_chunk_dict(casync.parse_caibx(caibx_url)))
]
except requests.RequestException:
cloudlog.error(f"casync failed to load {caibx_url}")
except Exception:
cloudlog.exception("casync failed to hash seed partition")
# Second source is the target partition, this allows for resuming
sources += [('target', casync.FileChunkReader(path),
casync.build_chunk_dict(target))]
# Finally we add the remote source to download any missing chunks
sources += [('remote', casync.RemoteChunkReader(partition['casync_store']),
casync.build_chunk_dict(target))]
last_p = 0
def progress(cur):
nonlocal last_p
p = int(cur / partition['size'] * 100)
if p != last_p:
last_p = p
print(f"Installing {partition['name']}: {p}", flush=True)
stats = casync.extract(target, sources, path, progress)
cloudlog.error(f'casync done {json.dumps(stats)}')
os.sync()
if not verify_partition(
target_slot_number, partition, force_full_check=True):
raise Exception(f"Raw hash mismatch '{partition['hash_raw'].lower()}'")
def test_already_done(self):
"""Test that an already flashed target doesn't download any chunks"""
target = casync.parse_caibx(self.manifest_fn)
with open(self.target_fn, 'wb') as f:
f.write(self.contents)
sources = [('target', casync.FileChunkReader(self.target_fn),
casync.build_chunk_dict(target))]
sources += [('remote', casync.RemoteChunkReader(self.store_fn),
casync.build_chunk_dict(target))]
stats = casync.extract(target, sources, self.target_fn)
with open(self.target_fn, 'rb') as f:
self.assertEqual(f.read(), self.contents)
self.assertEqual(stats['target'], len(self.contents))
def test_seed(self):
target = casync.parse_caibx(self.manifest_fn)
# Populate seed with half of the target contents
with open(self.seed_fn, 'wb') as seed_f:
seed_f.write(self.contents[:len(self.contents) // 2])
sources = [('seed', casync.FileChunkReader(self.seed_fn),
casync.build_chunk_dict(target))]
sources += [('remote', casync.RemoteChunkReader(self.store_fn),
casync.build_chunk_dict(target))]
stats = casync.extract(target, sources, self.target_fn)
with open(self.target_fn, 'rb') as target_f:
self.assertEqual(target_f.read(), self.contents)
self.assertGreater(stats['seed'], 0)
self.assertLess(stats['remote'], len(self.contents))
def test_simple_extract(self):
target = casync.parse_caibx(self.manifest_fn)
sources = [('remote', casync.RemoteChunkReader(self.store_fn),
casync.build_chunk_dict(target))]
stats = casync.extract(target, sources, self.target_fn)
with open(self.target_fn, 'rb') as target_f:
self.assertEqual(target_f.read(), self.contents)
self.assertEqual(stats['remote'], len(self.contents))
if __name__ == "__main__":
parser = argparse.ArgumentParser(
description='Compute overlap between two casync manifests')
parser.add_argument('frm')
parser.add_argument('to')
args = parser.parse_args()
frm = casync.parse_caibx(args.frm)
to = casync.parse_caibx(args.to)
remote_url = args.to.replace('.caibx', '')
most_common = collections.Counter(t.sha for t in to).most_common(1)[0][0]
frm_dict = casync.build_chunk_dict(frm)
# Get content-length for each chunk
with multiprocessing.Pool() as pool:
szs = list(tqdm(pool.imap(get_chunk_download_size, to), total=len(to)))
chunk_sizes = {t.sha: sz for (t, sz) in zip(to, szs)}
sources: Dict[str, List[int]] = {
'seed': [],
'remote_uncompressed': [],
'remote_compressed': [],
}
for chunk in to:
# Assume most common chunk is the zero chunk
if chunk.sha == most_common:
