1 # -*- coding: utf-8 -*-
3 # Copyright (C) The Arvados Authors. All rights reserved.
5 # SPDX-License-Identifier: Apache-2.0
7 from __future__ import absolute_import
8 from __future__ import division
9 from future import standard_library
10 standard_library.install_aliases()
11 from builtins import str
12 from builtins import range
13 from functools import partial
37 import arvados.commands.put as arv_put
38 from . import arvados_testutil as tutil
40 from .arvados_testutil import ArvadosBaseTestCase, fake_httplib2_response
41 from . import run_test_server
43 class ArvadosPutResumeCacheTest(ArvadosBaseTestCase):
47 ['/dev/null', '--filename', 'empty'],
52 super(ArvadosPutResumeCacheTest, self).tearDown()
54 self.last_cache.destroy()
55 except AttributeError:
58 def cache_path_from_arglist(self, arglist):
59 return arv_put.ResumeCache.make_path(arv_put.parse_arguments(arglist))
61 def test_cache_names_stable(self):
62 for argset in self.CACHE_ARGSET:
63 self.assertEqual(self.cache_path_from_arglist(argset),
64 self.cache_path_from_arglist(argset),
65 "cache name changed for {}".format(argset))
67 def test_cache_names_unique(self):
69 for argset in self.CACHE_ARGSET:
70 path = self.cache_path_from_arglist(argset)
71 self.assertNotIn(path, results)
74 def test_cache_names_simple(self):
75 # The goal here is to make sure the filename doesn't use characters
76 # reserved by the filesystem. Feel free to adjust this regexp as
77 # long as it still does that.
78 bad_chars = re.compile(r'[^-\.\w]')
79 for argset in self.CACHE_ARGSET:
80 path = self.cache_path_from_arglist(argset)
81 self.assertFalse(bad_chars.search(os.path.basename(path)),
82 "path too exotic: {}".format(path))
84 def test_cache_names_ignore_argument_order(self):
86 self.cache_path_from_arglist(['a', 'b', 'c']),
87 self.cache_path_from_arglist(['c', 'a', 'b']))
89 self.cache_path_from_arglist(['-', '--filename', 'stdin']),
90 self.cache_path_from_arglist(['--filename', 'stdin', '-']))
92 def test_cache_names_differ_for_similar_paths(self):
93 # This test needs names at / that don't exist on the real filesystem.
95 self.cache_path_from_arglist(['/_arvputtest1', '/_arvputtest2']),
96 self.cache_path_from_arglist(['/_arvputtest1/_arvputtest2']))
98 def test_cache_names_ignore_irrelevant_arguments(self):
99 # Workaround: parse_arguments bails on --filename with a directory.
100 path1 = self.cache_path_from_arglist(['/tmp'])
101 args = arv_put.parse_arguments(['/tmp'])
102 args.filename = 'tmp'
103 path2 = arv_put.ResumeCache.make_path(args)
104 self.assertEqual(path1, path2,
105 "cache path considered --filename for directory")
107 self.cache_path_from_arglist(['-']),
108 self.cache_path_from_arglist(['-', '--max-manifest-depth', '1']),
109 "cache path considered --max-manifest-depth for file")
111 def test_cache_names_treat_negative_manifest_depths_identically(self):
112 base_args = ['/tmp', '--max-manifest-depth']
114 self.cache_path_from_arglist(base_args + ['-1']),
115 self.cache_path_from_arglist(base_args + ['-2']))
117 def test_cache_names_treat_stdin_consistently(self):
119 self.cache_path_from_arglist(['-', '--filename', 'test']),
120 self.cache_path_from_arglist(['/dev/stdin', '--filename', 'test']))
122 def test_cache_names_identical_for_synonymous_names(self):
124 self.cache_path_from_arglist(['.']),
125 self.cache_path_from_arglist([os.path.realpath('.')]))
126 testdir = self.make_tmpdir()
127 looplink = os.path.join(testdir, 'loop')
128 os.symlink(testdir, looplink)
130 self.cache_path_from_arglist([testdir]),
131 self.cache_path_from_arglist([looplink]))
133 def test_cache_names_different_by_api_host(self):
134 config = arvados.config.settings()
135 orig_host = config.get('ARVADOS_API_HOST')
137 name1 = self.cache_path_from_arglist(['.'])
138 config['ARVADOS_API_HOST'] = 'x' + (orig_host or 'localhost')
139 self.assertNotEqual(name1, self.cache_path_from_arglist(['.']))
141 if orig_host is None:
142 del config['ARVADOS_API_HOST']
144 config['ARVADOS_API_HOST'] = orig_host
146 @mock.patch('arvados.keep.KeepClient.head')
147 def test_resume_cache_with_current_stream_locators(self, keep_client_head):
148 keep_client_head.side_effect = [True]
150 thing['_current_stream_locators'] = ['098f6bcd4621d373cade4e832627b4f6+4', '1f253c60a2306e0ee12fb6ce0c587904+6']
151 with tempfile.NamedTemporaryFile() as cachefile:
152 self.last_cache = arv_put.ResumeCache(cachefile.name)
153 self.last_cache.save(thing)
154 self.last_cache.close()
155 resume_cache = arv_put.ResumeCache(self.last_cache.filename)
156 self.assertNotEqual(None, resume_cache)
158 @mock.patch('arvados.keep.KeepClient.head')
159 def test_resume_cache_with_finished_streams(self, keep_client_head):
160 keep_client_head.side_effect = [True]
162 thing['_finished_streams'] = [['.', ['098f6bcd4621d373cade4e832627b4f6+4', '1f253c60a2306e0ee12fb6ce0c587904+6']]]
163 with tempfile.NamedTemporaryFile() as cachefile:
164 self.last_cache = arv_put.ResumeCache(cachefile.name)
165 self.last_cache.save(thing)
166 self.last_cache.close()
167 resume_cache = arv_put.ResumeCache(self.last_cache.filename)
168 self.assertNotEqual(None, resume_cache)
170 @mock.patch('arvados.keep.KeepClient.head')
171 def test_resume_cache_with_finished_streams_error_on_head(self, keep_client_head):
172 keep_client_head.side_effect = Exception('Locator not found')
174 thing['_finished_streams'] = [['.', ['098f6bcd4621d373cade4e832627b4f6+4', '1f253c60a2306e0ee12fb6ce0c587904+6']]]
175 with tempfile.NamedTemporaryFile() as cachefile:
176 self.last_cache = arv_put.ResumeCache(cachefile.name)
177 self.last_cache.save(thing)
178 self.last_cache.close()
179 resume_cache = arv_put.ResumeCache(self.last_cache.filename)
180 self.assertNotEqual(None, resume_cache)
181 resume_cache.check_cache()
183 def test_basic_cache_storage(self):
184 thing = ['test', 'list']
185 with tempfile.NamedTemporaryFile() as cachefile:
186 self.last_cache = arv_put.ResumeCache(cachefile.name)
187 self.last_cache.save(thing)
188 self.assertEqual(thing, self.last_cache.load())
190 def test_empty_cache(self):
191 with tempfile.NamedTemporaryFile() as cachefile:
192 cache = arv_put.ResumeCache(cachefile.name)
193 self.assertRaises(ValueError, cache.load)
195 def test_cache_persistent(self):
196 thing = ['test', 'list']
197 path = os.path.join(self.make_tmpdir(), 'cache')
198 cache = arv_put.ResumeCache(path)
201 self.last_cache = arv_put.ResumeCache(path)
202 self.assertEqual(thing, self.last_cache.load())
204 def test_multiple_cache_writes(self):
205 thing = ['short', 'list']
206 with tempfile.NamedTemporaryFile() as cachefile:
207 self.last_cache = arv_put.ResumeCache(cachefile.name)
208 # Start writing an object longer than the one we test, to make
209 # sure the cache file gets truncated.
210 self.last_cache.save(['long', 'long', 'list'])
211 self.last_cache.save(thing)
212 self.assertEqual(thing, self.last_cache.load())
214 def test_cache_is_locked(self):
215 with tempfile.NamedTemporaryFile() as cachefile:
216 _ = arv_put.ResumeCache(cachefile.name)
217 self.assertRaises(arv_put.ResumeCacheConflict,
218 arv_put.ResumeCache, cachefile.name)
220 def test_cache_stays_locked(self):
221 with tempfile.NamedTemporaryFile() as cachefile:
222 self.last_cache = arv_put.ResumeCache(cachefile.name)
223 path = cachefile.name
224 self.last_cache.save('test')
225 self.assertRaises(arv_put.ResumeCacheConflict,
226 arv_put.ResumeCache, path)
228 def test_destroy_cache(self):
229 cachefile = tempfile.NamedTemporaryFile(delete=False)
231 cache = arv_put.ResumeCache(cachefile.name)
235 arv_put.ResumeCache(cachefile.name)
236 except arv_put.ResumeCacheConflict:
237 self.fail("could not load cache after destroying it")
238 self.assertRaises(ValueError, cache.load)
240 if os.path.exists(cachefile.name):
241 os.unlink(cachefile.name)
243 def test_restart_cache(self):
244 path = os.path.join(self.make_tmpdir(), 'cache')
245 cache = arv_put.ResumeCache(path)
248 self.assertRaises(ValueError, cache.load)
249 self.assertRaises(arv_put.ResumeCacheConflict,
250 arv_put.ResumeCache, path)
253 class ArvPutUploadJobTest(run_test_server.TestCaseWithServers,
254 ArvadosBaseTestCase):
257 super(ArvPutUploadJobTest, self).setUp()
258 run_test_server.authorize_with('active')
259 # Temp files creation
260 self.tempdir = tempfile.mkdtemp()
261 subdir = os.path.join(self.tempdir, 'subdir')
263 data = "x" * 1024 # 1 KB
264 for i in range(1, 5):
265 with open(os.path.join(self.tempdir, str(i)), 'w') as f:
267 with open(os.path.join(subdir, 'otherfile'), 'w') as f:
269 # Large temp file for resume test
270 _, self.large_file_name = tempfile.mkstemp()
271 fileobj = open(self.large_file_name, 'w')
272 # Make sure to write just a little more than one block
273 for _ in range((arvados.config.KEEP_BLOCK_SIZE>>20)+1):
274 data = random.choice(['x', 'y', 'z']) * 1024 * 1024 # 1 MiB
277 # Temp dir containing small files to be repacked
278 self.small_files_dir = tempfile.mkdtemp()
279 data = 'y' * 1024 * 1024 # 1 MB
280 for i in range(1, 70):
281 with open(os.path.join(self.small_files_dir, str(i)), 'w') as f:
282 f.write(data + str(i))
283 self.arvfile_write = getattr(arvados.arvfile.ArvadosFileWriter, 'write')
284 # Temp dir to hold a symlink to other temp dir
285 self.tempdir_with_symlink = tempfile.mkdtemp()
286 os.symlink(self.tempdir, os.path.join(self.tempdir_with_symlink, 'linkeddir'))
287 os.symlink(os.path.join(self.tempdir, '1'),
288 os.path.join(self.tempdir_with_symlink, 'linkedfile'))
291 super(ArvPutUploadJobTest, self).tearDown()
292 shutil.rmtree(self.tempdir)
293 os.unlink(self.large_file_name)
294 shutil.rmtree(self.small_files_dir)
295 shutil.rmtree(self.tempdir_with_symlink)
297 def test_symlinks_are_followed_by_default(self):
298 self.assertTrue(os.path.islink(os.path.join(self.tempdir_with_symlink, 'linkeddir')))
299 self.assertTrue(os.path.islink(os.path.join(self.tempdir_with_symlink, 'linkedfile')))
300 cwriter = arv_put.ArvPutUploadJob([self.tempdir_with_symlink])
301 cwriter.start(save_collection=False)
302 self.assertIn('linkeddir', cwriter.manifest_text())
303 self.assertIn('linkedfile', cwriter.manifest_text())
304 cwriter.destroy_cache()
306 def test_symlinks_are_not_followed_when_requested(self):
307 self.assertTrue(os.path.islink(os.path.join(self.tempdir_with_symlink, 'linkeddir')))
308 self.assertTrue(os.path.islink(os.path.join(self.tempdir_with_symlink, 'linkedfile')))
309 cwriter = arv_put.ArvPutUploadJob([self.tempdir_with_symlink],
311 cwriter.start(save_collection=False)
312 self.assertNotIn('linkeddir', cwriter.manifest_text())
313 self.assertNotIn('linkedfile', cwriter.manifest_text())
314 cwriter.destroy_cache()
315 # Check for bug #17800: passed symlinks should also be ignored.
316 linked_dir = os.path.join(self.tempdir_with_symlink, 'linkeddir')
317 cwriter = arv_put.ArvPutUploadJob([linked_dir], follow_links=False)
318 cwriter.start(save_collection=False)
319 self.assertNotIn('linkeddir', cwriter.manifest_text())
320 cwriter.destroy_cache()
322 def test_no_empty_collection_saved(self):
323 self.assertTrue(os.path.islink(os.path.join(self.tempdir_with_symlink, 'linkeddir')))
324 linked_dir = os.path.join(self.tempdir_with_symlink, 'linkeddir')
325 cwriter = arv_put.ArvPutUploadJob([linked_dir], follow_links=False)
326 cwriter.start(save_collection=True)
327 self.assertIsNone(cwriter.manifest_locator())
328 self.assertEqual('', cwriter.manifest_text())
329 cwriter.destroy_cache()
331 def test_passing_nonexistant_path_raise_exception(self):
332 uuid_str = str(uuid.uuid4())
333 with self.assertRaises(arv_put.PathDoesNotExistError):
334 arv_put.ArvPutUploadJob(["/this/path/does/not/exist/{}".format(uuid_str)])
336 def test_writer_works_without_cache(self):
337 cwriter = arv_put.ArvPutUploadJob(['/dev/null'], resume=False)
338 cwriter.start(save_collection=False)
339 self.assertEqual(". d41d8cd98f00b204e9800998ecf8427e+0 0:0:null\n", cwriter.manifest_text())
341 def test_writer_works_with_cache(self):
342 with tempfile.NamedTemporaryFile() as f:
345 cwriter = arv_put.ArvPutUploadJob([f.name])
346 cwriter.start(save_collection=False)
347 self.assertEqual(0, cwriter.bytes_skipped)
348 self.assertEqual(3, cwriter.bytes_written)
349 # Don't destroy the cache, and start another upload
350 cwriter_new = arv_put.ArvPutUploadJob([f.name])
351 cwriter_new.start(save_collection=False)
352 cwriter_new.destroy_cache()
353 self.assertEqual(3, cwriter_new.bytes_skipped)
354 self.assertEqual(3, cwriter_new.bytes_written)
356 def make_progress_tester(self):
358 def record_func(written, expected):
359 progression.append((written, expected))
360 return progression, record_func
362 def test_progress_reporting(self):
363 with tempfile.NamedTemporaryFile() as f:
366 for expect_count in (None, 8):
367 progression, reporter = self.make_progress_tester()
368 cwriter = arv_put.ArvPutUploadJob([f.name],
370 cwriter.bytes_expected = expect_count
371 cwriter.start(save_collection=False)
372 cwriter.destroy_cache()
373 self.assertIn((3, expect_count), progression)
375 def test_writer_upload_directory(self):
376 cwriter = arv_put.ArvPutUploadJob([self.tempdir])
377 cwriter.start(save_collection=False)
378 cwriter.destroy_cache()
379 self.assertEqual(1024*(1+2+3+4+5), cwriter.bytes_written)
381 def test_resume_large_file_upload(self):
382 def wrapped_write(*args, **kwargs):
384 # Exit only on last block
385 if len(data) < arvados.config.KEEP_BLOCK_SIZE:
386 # Simulate a checkpoint before quitting. Ensure block commit.
387 self.writer._update(final=True)
388 raise SystemExit("Simulated error")
389 return self.arvfile_write(*args, **kwargs)
391 with mock.patch('arvados.arvfile.ArvadosFileWriter.write',
392 autospec=True) as mocked_write:
393 mocked_write.side_effect = wrapped_write
394 writer = arv_put.ArvPutUploadJob([self.large_file_name],
395 replication_desired=1)
396 # We'll be accessing from inside the wrapper
398 with self.assertRaises(SystemExit):
399 writer.start(save_collection=False)
400 # Confirm that the file was partially uploaded
401 self.assertGreater(writer.bytes_written, 0)
402 self.assertLess(writer.bytes_written,
403 os.path.getsize(self.large_file_name))
405 writer2 = arv_put.ArvPutUploadJob([self.large_file_name],
406 replication_desired=1)
407 writer2.start(save_collection=False)
408 self.assertEqual(writer.bytes_written + writer2.bytes_written - writer2.bytes_skipped,
409 os.path.getsize(self.large_file_name))
410 writer2.destroy_cache()
413 # Test for bug #11002
414 def test_graceful_exit_while_repacking_small_blocks(self):
415 def wrapped_commit(*args, **kwargs):
416 raise SystemExit("Simulated error")
418 with mock.patch('arvados.arvfile._BlockManager.commit_bufferblock',
419 autospec=True) as mocked_commit:
420 mocked_commit.side_effect = wrapped_commit
421 # Upload a little more than 1 block, wrapped_commit will make the first block
423 # arv-put should not exit with an exception by trying to commit the collection
424 # as it's in an inconsistent state.
425 writer = arv_put.ArvPutUploadJob([self.small_files_dir],
426 replication_desired=1)
428 with self.assertRaises(SystemExit):
429 writer.start(save_collection=False)
430 except arvados.arvfile.UnownedBlockError:
431 self.fail("arv-put command is trying to use a corrupted BlockManager. See https://dev.arvados.org/issues/11002")
432 writer.destroy_cache()
434 def test_no_resume_when_asked(self):
435 def wrapped_write(*args, **kwargs):
437 # Exit only on last block
438 if len(data) < arvados.config.KEEP_BLOCK_SIZE:
439 # Simulate a checkpoint before quitting.
440 self.writer._update()
441 raise SystemExit("Simulated error")
442 return self.arvfile_write(*args, **kwargs)
444 with mock.patch('arvados.arvfile.ArvadosFileWriter.write',
445 autospec=True) as mocked_write:
446 mocked_write.side_effect = wrapped_write
447 writer = arv_put.ArvPutUploadJob([self.large_file_name],
448 replication_desired=1)
449 # We'll be accessing from inside the wrapper
451 with self.assertRaises(SystemExit):
452 writer.start(save_collection=False)
453 # Confirm that the file was partially uploaded
454 self.assertGreater(writer.bytes_written, 0)
455 self.assertLess(writer.bytes_written,
456 os.path.getsize(self.large_file_name))
457 # Retry the upload, this time without resume
458 writer2 = arv_put.ArvPutUploadJob([self.large_file_name],
459 replication_desired=1,
461 writer2.start(save_collection=False)
462 self.assertEqual(writer2.bytes_skipped, 0)
463 self.assertEqual(writer2.bytes_written,
464 os.path.getsize(self.large_file_name))
465 writer2.destroy_cache()
468 def test_no_resume_when_no_cache(self):
469 def wrapped_write(*args, **kwargs):
471 # Exit only on last block
472 if len(data) < arvados.config.KEEP_BLOCK_SIZE:
473 # Simulate a checkpoint before quitting.
474 self.writer._update()
475 raise SystemExit("Simulated error")
476 return self.arvfile_write(*args, **kwargs)
478 with mock.patch('arvados.arvfile.ArvadosFileWriter.write',
479 autospec=True) as mocked_write:
480 mocked_write.side_effect = wrapped_write
481 writer = arv_put.ArvPutUploadJob([self.large_file_name],
482 replication_desired=1)
483 # We'll be accessing from inside the wrapper
485 with self.assertRaises(SystemExit):
486 writer.start(save_collection=False)
487 # Confirm that the file was partially uploaded
488 self.assertGreater(writer.bytes_written, 0)
489 self.assertLess(writer.bytes_written,
490 os.path.getsize(self.large_file_name))
491 # Retry the upload, this time without cache usage
492 writer2 = arv_put.ArvPutUploadJob([self.large_file_name],
493 replication_desired=1,
496 writer2.start(save_collection=False)
497 self.assertEqual(writer2.bytes_skipped, 0)
498 self.assertEqual(writer2.bytes_written,
499 os.path.getsize(self.large_file_name))
500 writer2.destroy_cache()
503 def test_dry_run_feature(self):
504 def wrapped_write(*args, **kwargs):
506 # Exit only on last block
507 if len(data) < arvados.config.KEEP_BLOCK_SIZE:
508 # Simulate a checkpoint before quitting.
509 self.writer._update()
510 raise SystemExit("Simulated error")
511 return self.arvfile_write(*args, **kwargs)
513 with mock.patch('arvados.arvfile.ArvadosFileWriter.write',
514 autospec=True) as mocked_write:
515 mocked_write.side_effect = wrapped_write
516 writer = arv_put.ArvPutUploadJob([self.large_file_name],
517 replication_desired=1)
518 # We'll be accessing from inside the wrapper
520 with self.assertRaises(SystemExit):
521 writer.start(save_collection=False)
522 # Confirm that the file was partially uploaded
523 self.assertGreater(writer.bytes_written, 0)
524 self.assertLess(writer.bytes_written,
525 os.path.getsize(self.large_file_name))
526 with self.assertRaises(arv_put.ArvPutUploadIsPending):
527 # Retry the upload using dry_run to check if there is a pending upload
528 writer2 = arv_put.ArvPutUploadJob([self.large_file_name],
529 replication_desired=1,
531 # Complete the pending upload
532 writer3 = arv_put.ArvPutUploadJob([self.large_file_name],
533 replication_desired=1)
534 writer3.start(save_collection=False)
535 with self.assertRaises(arv_put.ArvPutUploadNotPending):
536 # Confirm there's no pending upload with dry_run=True
537 writer4 = arv_put.ArvPutUploadJob([self.large_file_name],
538 replication_desired=1,
541 with self.assertRaises(arv_put.ArvPutUploadIsPending):
542 arv_put.ArvPutUploadJob([self.large_file_name],
543 replication_desired=1,
547 with self.assertRaises(arv_put.ArvPutUploadIsPending):
548 arv_put.ArvPutUploadJob([self.large_file_name],
549 replication_desired=1,
554 class CachedManifestValidationTest(ArvadosBaseTestCase):
555 class MockedPut(arv_put.ArvPutUploadJob):
556 def __init__(self, cached_manifest=None):
557 self._state = arv_put.ArvPutUploadJob.EMPTY_STATE
558 self._state['manifest'] = cached_manifest
559 self._api_client = mock.MagicMock()
560 self.logger = mock.MagicMock()
563 def datetime_to_hex(self, dt):
564 return hex(int(time.mktime(dt.timetuple())))[2:]
567 super(CachedManifestValidationTest, self).setUp()
568 self.block1 = "fdba98970961edb29f88241b9d99d890" # foo
569 self.block2 = "37b51d194a7513e45b56f6524f2d51f2" # bar
570 self.template = ". "+self.block1+"+3+Asignature@%s "+self.block2+"+3+Anothersignature@%s 0:3:foofile.txt 3:6:barfile.txt\n"
572 def test_empty_cached_manifest_is_valid(self):
573 put_mock = self.MockedPut()
574 self.assertEqual(None, put_mock._state.get('manifest'))
575 self.assertTrue(put_mock._cached_manifest_valid())
576 put_mock._state['manifest'] = ''
577 self.assertTrue(put_mock._cached_manifest_valid())
579 def test_signature_cases(self):
580 now = datetime.datetime.utcnow()
581 yesterday = now - datetime.timedelta(days=1)
582 lastweek = now - datetime.timedelta(days=7)
583 tomorrow = now + datetime.timedelta(days=1)
584 nextweek = now + datetime.timedelta(days=7)
586 def mocked_head(blocks={}, loc=None):
587 blk = loc.split('+', 1)[0]
590 raise arvados.errors.KeepRequestError("mocked error - block invalid")
592 # Block1_expiration, Block2_expiration, Block1_HEAD, Block2_HEAD, Expectation
594 # All expired, reset cache - OK
595 (yesterday, lastweek, False, False, True),
596 (lastweek, yesterday, False, False, True),
597 # All non-expired valid blocks - OK
598 (tomorrow, nextweek, True, True, True),
599 (nextweek, tomorrow, True, True, True),
600 # All non-expired invalid blocks - Not OK
601 (tomorrow, nextweek, False, False, False),
602 (nextweek, tomorrow, False, False, False),
603 # One non-expired valid block - OK
604 (tomorrow, yesterday, True, False, True),
605 (yesterday, tomorrow, False, True, True),
606 # One non-expired invalid block - Not OK
607 (tomorrow, yesterday, False, False, False),
608 (yesterday, tomorrow, False, False, False),
611 b1_expiration, b2_expiration, b1_valid, b2_valid, outcome = case
613 self.block1: b1_valid,
614 self.block2: b2_valid,
616 cached_manifest = self.template % (
617 self.datetime_to_hex(b1_expiration),
618 self.datetime_to_hex(b2_expiration),
620 arvput = self.MockedPut(cached_manifest)
621 with mock.patch('arvados.collection.KeepClient.head') as head_mock:
622 head_mock.side_effect = partial(mocked_head, head_responses)
623 self.assertEqual(outcome, arvput._cached_manifest_valid(),
624 "Case '%s' should have produced outcome '%s'" % (case, outcome)
626 if b1_expiration > now or b2_expiration > now:
627 # A HEAD request should have been done
628 head_mock.assert_called_once()
630 head_mock.assert_not_called()
633 class ArvadosExpectedBytesTest(ArvadosBaseTestCase):
634 TEST_SIZE = os.path.getsize(__file__)
636 def test_expected_bytes_for_file(self):
637 writer = arv_put.ArvPutUploadJob([__file__])
638 self.assertEqual(self.TEST_SIZE,
639 writer.bytes_expected)
641 def test_expected_bytes_for_tree(self):
642 tree = self.make_tmpdir()
643 shutil.copyfile(__file__, os.path.join(tree, 'one'))
644 shutil.copyfile(__file__, os.path.join(tree, 'two'))
646 writer = arv_put.ArvPutUploadJob([tree])
647 self.assertEqual(self.TEST_SIZE * 2,
648 writer.bytes_expected)
649 writer = arv_put.ArvPutUploadJob([tree, __file__])
650 self.assertEqual(self.TEST_SIZE * 3,
651 writer.bytes_expected)
653 def test_expected_bytes_for_device(self):
654 writer = arv_put.ArvPutUploadJob(['/dev/null'], use_cache=False, resume=False)
655 self.assertIsNone(writer.bytes_expected)
656 writer = arv_put.ArvPutUploadJob([__file__, '/dev/null'])
657 self.assertIsNone(writer.bytes_expected)
660 class ArvadosPutReportTest(ArvadosBaseTestCase):
661 def test_machine_progress(self):
662 for count, total in [(0, 1), (0, None), (1, None), (235, 9283)]:
663 expect = ": {} written {} total\n".format(
664 count, -1 if (total is None) else total)
666 arv_put.machine_progress(count, total).endswith(expect))
668 def test_known_human_progress(self):
669 for count, total in [(0, 1), (2, 4), (45, 60)]:
670 expect = '{:.1%}'.format(1.0*count/total)
671 actual = arv_put.human_progress(count, total)
672 self.assertTrue(actual.startswith('\r'))
673 self.assertIn(expect, actual)
675 def test_unknown_human_progress(self):
676 for count in [1, 20, 300, 4000, 50000]:
677 self.assertTrue(re.search(r'\b{}\b'.format(count),
678 arv_put.human_progress(count, None)))
681 class ArvPutLogFormatterTest(ArvadosBaseTestCase):
682 matcher = r'\(X-Request-Id: req-[a-z0-9]{20}\)'
685 super(ArvPutLogFormatterTest, self).setUp()
686 self.stderr = tutil.StringIO()
687 self.loggingHandler = logging.StreamHandler(self.stderr)
688 self.loggingHandler.setFormatter(
689 arv_put.ArvPutLogFormatter(arvados.util.new_request_id()))
690 self.logger = logging.getLogger()
691 self.logger.addHandler(self.loggingHandler)
692 self.logger.setLevel(logging.DEBUG)
695 self.logger.removeHandler(self.loggingHandler)
698 super(ArvPutLogFormatterTest, self).tearDown()
700 def test_request_id_logged_only_once_on_error(self):
701 self.logger.error('Ooops, something bad happened.')
702 self.logger.error('Another bad thing just happened.')
703 log_lines = self.stderr.getvalue().split('\n')[:-1]
704 self.assertEqual(2, len(log_lines))
705 self.assertRegex(log_lines[0], self.matcher)
706 self.assertNotRegex(log_lines[1], self.matcher)
708 def test_request_id_logged_only_once_on_debug(self):
709 self.logger.debug('This is just a debug message.')
710 self.logger.debug('Another message, move along.')
711 log_lines = self.stderr.getvalue().split('\n')[:-1]
712 self.assertEqual(2, len(log_lines))
713 self.assertRegex(log_lines[0], self.matcher)
714 self.assertNotRegex(log_lines[1], self.matcher)
716 def test_request_id_not_logged_on_info(self):
717 self.logger.info('This should be a useful message')
718 log_lines = self.stderr.getvalue().split('\n')[:-1]
719 self.assertEqual(1, len(log_lines))
720 self.assertNotRegex(log_lines[0], self.matcher)
722 class ArvadosPutTest(run_test_server.TestCaseWithServers,
724 tutil.VersionChecker):
726 Z_UUID = 'zzzzz-zzzzz-zzzzzzzzzzzzzzz'
728 def call_main_with_args(self, args):
729 self.main_stdout.seek(0, 0)
730 self.main_stdout.truncate(0)
731 self.main_stderr.seek(0, 0)
732 self.main_stderr.truncate(0)
733 return arv_put.main(args, self.main_stdout, self.main_stderr)
735 def call_main_on_test_file(self, args=[]):
736 with self.make_test_file() as testfile:
738 self.call_main_with_args(['--stream', '--no-progress'] + args + [path])
740 os.path.exists(os.path.join(os.environ['KEEP_LOCAL_STORE'],
741 '098f6bcd4621d373cade4e832627b4f6')),
742 "did not find file stream in Keep store")
745 super(ArvadosPutTest, self).setUp()
746 run_test_server.authorize_with('active')
747 arv_put.api_client = None
748 self.main_stdout = tutil.StringIO()
749 self.main_stderr = tutil.StringIO()
750 self.loggingHandler = logging.StreamHandler(self.main_stderr)
751 self.loggingHandler.setFormatter(
752 arv_put.ArvPutLogFormatter(arvados.util.new_request_id()))
753 logging.getLogger().addHandler(self.loggingHandler)
756 logging.getLogger().removeHandler(self.loggingHandler)
757 for outbuf in ['main_stdout', 'main_stderr']:
758 if hasattr(self, outbuf):
759 getattr(self, outbuf).close()
760 delattr(self, outbuf)
761 super(ArvadosPutTest, self).tearDown()
763 def test_version_argument(self):
764 with tutil.redirected_streams(
765 stdout=tutil.StringIO, stderr=tutil.StringIO) as (out, err):
766 with self.assertRaises(SystemExit):
767 self.call_main_with_args(['--version'])
768 self.assertVersionOutput(out, err)
770 def test_simple_file_put(self):
771 self.call_main_on_test_file()
773 def test_put_with_unwriteable_cache_dir(self):
774 orig_cachedir = arv_put.ResumeCache.CACHE_DIR
775 cachedir = self.make_tmpdir()
776 os.chmod(cachedir, 0o0)
777 arv_put.ResumeCache.CACHE_DIR = cachedir
779 self.call_main_on_test_file()
781 arv_put.ResumeCache.CACHE_DIR = orig_cachedir
782 os.chmod(cachedir, 0o700)
784 def test_put_with_unwritable_cache_subdir(self):
785 orig_cachedir = arv_put.ResumeCache.CACHE_DIR
786 cachedir = self.make_tmpdir()
787 os.chmod(cachedir, 0o0)
788 arv_put.ResumeCache.CACHE_DIR = os.path.join(cachedir, 'cachedir')
790 self.call_main_on_test_file()
792 arv_put.ResumeCache.CACHE_DIR = orig_cachedir
793 os.chmod(cachedir, 0o700)
795 def test_put_block_replication(self):
796 self.call_main_on_test_file()
797 with mock.patch('arvados.collection.KeepClient.local_store_put') as put_mock:
798 put_mock.return_value = 'acbd18db4cc2f85cedef654fccc4a4d8+3'
799 self.call_main_on_test_file(['--replication', '1'])
800 self.call_main_on_test_file(['--replication', '4'])
801 self.call_main_on_test_file(['--replication', '5'])
803 [x[-1].get('copies') for x in put_mock.call_args_list],
806 def test_normalize(self):
807 testfile1 = self.make_test_file()
808 testfile2 = self.make_test_file()
809 test_paths = [testfile1.name, testfile2.name]
810 # Reverse-sort the paths, so normalization must change their order.
811 test_paths.sort(reverse=True)
812 self.call_main_with_args(['--stream', '--no-progress', '--normalize'] +
814 manifest = self.main_stdout.getvalue()
815 # Assert the second file we specified appears first in the manifest.
816 file_indices = [manifest.find(':' + os.path.basename(path))
817 for path in test_paths]
818 self.assertGreater(*file_indices)
820 def test_error_name_without_collection(self):
821 self.assertRaises(SystemExit, self.call_main_with_args,
822 ['--name', 'test without Collection',
823 '--stream', '/dev/null'])
825 def test_error_when_project_not_found(self):
826 self.assertRaises(SystemExit,
827 self.call_main_with_args,
828 ['--project-uuid', self.Z_UUID])
830 def test_error_bad_project_uuid(self):
831 self.assertRaises(SystemExit,
832 self.call_main_with_args,
833 ['--project-uuid', self.Z_UUID, '--stream'])
835 def test_error_when_excluding_absolute_path(self):
836 tmpdir = self.make_tmpdir()
837 self.assertRaises(SystemExit,
838 self.call_main_with_args,
839 ['--exclude', '/some/absolute/path/*',
842 def test_api_error_handling(self):
843 coll_save_mock = mock.Mock(name='arv.collection.Collection().save_new()')
844 coll_save_mock.side_effect = arvados.errors.ApiError(
845 fake_httplib2_response(403), b'{}')
846 with mock.patch('arvados.collection.Collection.save_new',
848 with self.assertRaises(SystemExit) as exc_test:
849 self.call_main_with_args(['/dev/null'])
850 self.assertLess(0, exc_test.exception.args[0])
851 self.assertLess(0, coll_save_mock.call_count)
852 self.assertEqual("", self.main_stdout.getvalue())
854 def test_request_id_logging_on_error(self):
855 matcher = r'\(X-Request-Id: req-[a-z0-9]{20}\)\n'
856 coll_save_mock = mock.Mock(name='arv.collection.Collection().save_new()')
857 coll_save_mock.side_effect = arvados.errors.ApiError(
858 fake_httplib2_response(403), b'{}')
859 with mock.patch('arvados.collection.Collection.save_new',
861 with self.assertRaises(SystemExit):
862 self.call_main_with_args(['/dev/null'])
864 self.main_stderr.getvalue(), matcher)
867 class ArvPutIntegrationTest(run_test_server.TestCaseWithServers,
868 ArvadosBaseTestCase):
870 KEEP_SERVER = {'blob_signing': True}
871 PROJECT_UUID = run_test_server.fixture('groups')['aproject']['uuid']
875 super(ArvPutIntegrationTest, cls).setUpClass()
876 cls.ENVIRON = os.environ.copy()
877 cls.ENVIRON['PYTHONPATH'] = ':'.join(sys.path)
879 def datetime_to_hex(self, dt):
880 return hex(int(time.mktime(dt.timetuple())))[2:]
883 super(ArvPutIntegrationTest, self).setUp()
884 arv_put.api_client = None
886 def authorize_with(self, token_name):
887 run_test_server.authorize_with(token_name)
888 for v in ["ARVADOS_API_HOST",
889 "ARVADOS_API_HOST_INSECURE",
890 "ARVADOS_API_TOKEN"]:
891 self.ENVIRON[v] = arvados.config.settings()[v]
892 arv_put.api_client = arvados.api('v1')
894 def current_user(self):
895 return arv_put.api_client.users().current().execute()
897 def test_check_real_project_found(self):
898 self.authorize_with('active')
899 self.assertTrue(arv_put.desired_project_uuid(arv_put.api_client, self.PROJECT_UUID, 0),
900 "did not correctly find test fixture project")
902 def test_check_error_finding_nonexistent_uuid(self):
903 BAD_UUID = 'zzzzz-zzzzz-zzzzzzzzzzzzzzz'
904 self.authorize_with('active')
906 result = arv_put.desired_project_uuid(arv_put.api_client, BAD_UUID,
908 except ValueError as error:
909 self.assertIn(BAD_UUID, str(error))
911 self.assertFalse(result, "incorrectly found nonexistent project")
913 def test_check_error_finding_nonexistent_project(self):
914 BAD_UUID = 'zzzzz-tpzed-zzzzzzzzzzzzzzz'
915 self.authorize_with('active')
916 with self.assertRaises(apiclient.errors.HttpError):
917 arv_put.desired_project_uuid(arv_put.api_client, BAD_UUID,
920 def test_short_put_from_stdin(self):
921 # Have to run this as an integration test since arv-put can't
922 # read from the tests' stdin.
923 # arv-put usually can't stat(os.path.realpath('/dev/stdin')) in this
924 # case, because the /proc entry is already gone by the time it tries.
925 pipe = subprocess.Popen(
926 [sys.executable, arv_put.__file__, '--stream'],
927 stdin=subprocess.PIPE, stdout=subprocess.PIPE,
928 stderr=subprocess.STDOUT, env=self.ENVIRON)
929 pipe.stdin.write(b'stdin test\xa6\n')
931 deadline = time.time() + 5
932 while (pipe.poll() is None) and (time.time() < deadline):
934 returncode = pipe.poll()
935 if returncode is None:
937 self.fail("arv-put did not PUT from stdin within 5 seconds")
938 elif returncode != 0:
939 sys.stdout.write(pipe.stdout.read())
940 self.fail("arv-put returned exit code {}".format(returncode))
941 self.assertIn('1cb671b355a0c23d5d1c61d59cdb1b2b+12',
942 pipe.stdout.read().decode())
944 def test_sigint_logs_request_id(self):
945 # Start arv-put, give it a chance to start up, send SIGINT,
946 # and check that its output includes the X-Request-Id.
947 input_stream = subprocess.Popen(
949 stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
950 pipe = subprocess.Popen(
951 [sys.executable, arv_put.__file__, '--stream'],
952 stdin=input_stream.stdout, stdout=subprocess.PIPE,
953 stderr=subprocess.STDOUT, env=self.ENVIRON)
954 # Wait for arv-put child process to print something (i.e., a
955 # log message) so we know its signal handler is installed.
956 select.select([pipe.stdout], [], [], 10)
957 pipe.send_signal(signal.SIGINT)
958 deadline = time.time() + 5
959 while (pipe.poll() is None) and (time.time() < deadline):
961 returncode = pipe.poll()
962 input_stream.terminate()
963 if returncode is None:
965 self.fail("arv-put did not exit within 5 seconds")
966 self.assertRegex(pipe.stdout.read().decode(), r'\(X-Request-Id: req-[a-z0-9]{20}\)')
968 def test_ArvPutSignedManifest(self):
969 # ArvPutSignedManifest runs "arv-put foo" and then attempts to get
970 # the newly created manifest from the API server, testing to confirm
971 # that the block locators in the returned manifest are signed.
972 self.authorize_with('active')
974 # Before doing anything, demonstrate that the collection
975 # we're about to create is not present in our test fixture.
976 manifest_uuid = "00b4e9f40ac4dd432ef89749f1c01e74+47"
977 with self.assertRaises(apiclient.errors.HttpError):
978 arv_put.api_client.collections().get(
979 uuid=manifest_uuid).execute()
981 datadir = self.make_tmpdir()
982 with open(os.path.join(datadir, "foo"), "w") as f:
983 f.write("The quick brown fox jumped over the lazy dog")
984 p = subprocess.Popen([sys.executable, arv_put.__file__,
985 os.path.join(datadir, 'foo')],
986 stdout=subprocess.PIPE,
987 stderr=subprocess.PIPE,
989 (_, err) = p.communicate()
990 self.assertRegex(err.decode(), r'INFO: Collection saved as ')
991 self.assertEqual(p.returncode, 0)
993 # The manifest text stored in the API server under the same
994 # manifest UUID must use signed locators.
995 c = arv_put.api_client.collections().get(uuid=manifest_uuid).execute()
998 r'^\. 08a008a01d498c404b0c30852b39d3b8\+44\+A[0-9a-f]+@[0-9a-f]+ 0:44:foo\n')
1000 os.remove(os.path.join(datadir, "foo"))
1003 def run_and_find_collection(self, text, extra_args=[]):
1004 self.authorize_with('active')
1005 pipe = subprocess.Popen(
1006 [sys.executable, arv_put.__file__] + extra_args,
1007 stdin=subprocess.PIPE, stdout=subprocess.PIPE,
1008 stderr=subprocess.PIPE, env=self.ENVIRON)
1009 stdout, stderr = pipe.communicate(text.encode())
1010 self.assertRegex(stderr.decode(), r'INFO: Collection (updated:|saved as)')
1011 search_key = ('portable_data_hash'
1012 if '--portable-data-hash' in extra_args else 'uuid')
1013 collection_list = arvados.api('v1').collections().list(
1014 filters=[[search_key, '=', stdout.decode().strip()]]
1015 ).execute().get('items', [])
1016 self.assertEqual(1, len(collection_list))
1017 return collection_list[0]
1019 def test_all_expired_signatures_invalidates_cache(self):
1020 self.authorize_with('active')
1021 tmpdir = self.make_tmpdir()
1022 with open(os.path.join(tmpdir, 'somefile.txt'), 'w') as f:
1024 # Upload a directory and get the cache file name
1025 p = subprocess.Popen([sys.executable, arv_put.__file__, tmpdir],
1026 stdout=subprocess.PIPE,
1027 stderr=subprocess.PIPE,
1029 (_, err) = p.communicate()
1030 self.assertRegex(err.decode(), r'INFO: Creating new cache file at ')
1031 self.assertEqual(p.returncode, 0)
1032 cache_filepath = re.search(r'INFO: Creating new cache file at (.*)',
1033 err.decode()).groups()[0]
1034 self.assertTrue(os.path.isfile(cache_filepath))
1035 # Load the cache file contents and modify the manifest to simulate
1036 # an expired access token
1037 with open(cache_filepath, 'r') as c:
1038 cache = json.load(c)
1039 self.assertRegex(cache['manifest'], r'\+A\S+\@')
1040 a_month_ago = datetime.datetime.now() - datetime.timedelta(days=30)
1041 cache['manifest'] = re.sub(
1043 "@{} ".format(self.datetime_to_hex(a_month_ago)),
1045 with open(cache_filepath, 'w') as c:
1046 c.write(json.dumps(cache))
1047 # Re-run the upload and expect to get an invalid cache message
1048 p = subprocess.Popen([sys.executable, arv_put.__file__, tmpdir],
1049 stdout=subprocess.PIPE,
1050 stderr=subprocess.PIPE,
1052 (_, err) = p.communicate()
1055 r'INFO: Cache expired, starting from scratch.*')
1056 self.assertEqual(p.returncode, 0)
1058 def test_invalid_signature_in_cache(self):
1059 for batch_mode in [False, True]:
1060 self.authorize_with('active')
1061 tmpdir = self.make_tmpdir()
1062 with open(os.path.join(tmpdir, 'somefile.txt'), 'w') as f:
1064 # Upload a directory and get the cache file name
1065 arv_put_args = [tmpdir]
1067 arv_put_args = ['--batch'] + arv_put_args
1068 p = subprocess.Popen([sys.executable, arv_put.__file__] + arv_put_args,
1069 stdout=subprocess.PIPE,
1070 stderr=subprocess.PIPE,
1072 (_, err) = p.communicate()
1073 self.assertRegex(err.decode(), r'INFO: Creating new cache file at ')
1074 self.assertEqual(p.returncode, 0)
1075 cache_filepath = re.search(r'INFO: Creating new cache file at (.*)',
1076 err.decode()).groups()[0]
1077 self.assertTrue(os.path.isfile(cache_filepath))
1078 # Load the cache file contents and modify the manifest to simulate
1079 # an invalid access token
1080 with open(cache_filepath, 'r') as c:
1081 cache = json.load(c)
1082 self.assertRegex(cache['manifest'], r'\+A\S+\@')
1083 cache['manifest'] = re.sub(
1085 "+Aabcdef0123456789abcdef0123456789abcdef01@",
1087 with open(cache_filepath, 'w') as c:
1088 c.write(json.dumps(cache))
1089 # Re-run the upload and expect to get an invalid cache message
1090 p = subprocess.Popen([sys.executable, arv_put.__file__] + arv_put_args,
1091 stdout=subprocess.PIPE,
1092 stderr=subprocess.PIPE,
1094 (_, err) = p.communicate()
1098 r'ERROR: arv-put: Resume cache contains invalid signature.*')
1099 self.assertEqual(p.returncode, 1)
1103 r'Invalid signatures on cache file \'.*\' while being run in \'batch mode\' -- continuing anyways.*')
1104 self.assertEqual(p.returncode, 0)
1106 def test_single_expired_signature_reuploads_file(self):
1107 self.authorize_with('active')
1108 tmpdir = self.make_tmpdir()
1109 with open(os.path.join(tmpdir, 'foofile.txt'), 'w') as f:
1111 # Write a second file on its own subdir to force a new stream
1112 os.mkdir(os.path.join(tmpdir, 'bar'))
1113 with open(os.path.join(tmpdir, 'bar', 'barfile.txt'), 'w') as f:
1115 # Upload a directory and get the cache file name
1116 p = subprocess.Popen([sys.executable, arv_put.__file__, tmpdir],
1117 stdout=subprocess.PIPE,
1118 stderr=subprocess.PIPE,
1120 (_, err) = p.communicate()
1121 self.assertRegex(err.decode(), r'INFO: Creating new cache file at ')
1122 self.assertEqual(p.returncode, 0)
1123 cache_filepath = re.search(r'INFO: Creating new cache file at (.*)',
1124 err.decode()).groups()[0]
1125 self.assertTrue(os.path.isfile(cache_filepath))
1126 # Load the cache file contents and modify the manifest to simulate
1127 # an expired access token
1128 with open(cache_filepath, 'r') as c:
1129 cache = json.load(c)
1130 self.assertRegex(cache['manifest'], r'\+A\S+\@')
1131 a_month_ago = datetime.datetime.now() - datetime.timedelta(days=30)
1132 # Make one of the signatures appear to have expired
1133 cache['manifest'] = re.sub(
1134 r'\@.*? 3:3:barfile.txt',
1135 "@{} 3:3:barfile.txt".format(self.datetime_to_hex(a_month_ago)),
1137 with open(cache_filepath, 'w') as c:
1138 c.write(json.dumps(cache))
1139 # Re-run the upload and expect to get an invalid cache message
1140 p = subprocess.Popen([sys.executable, arv_put.__file__, tmpdir],
1141 stdout=subprocess.PIPE,
1142 stderr=subprocess.PIPE,
1144 (_, err) = p.communicate()
1147 r'WARNING: Uploaded file \'.*barfile.txt\' access token expired, will re-upload it from scratch')
1148 self.assertEqual(p.returncode, 0)
1149 # Confirm that the resulting cache is different from the last run.
1150 with open(cache_filepath, 'r') as c2:
1151 new_cache = json.load(c2)
1152 self.assertNotEqual(cache['manifest'], new_cache['manifest'])
1154 def test_put_collection_with_later_update(self):
1155 tmpdir = self.make_tmpdir()
1156 with open(os.path.join(tmpdir, 'file1'), 'w') as f:
1157 f.write('Relaxing in basins at the end of inlets terminates the endless tests from the box')
1158 col = self.run_and_find_collection("", ['--no-progress', tmpdir])
1159 self.assertNotEqual(None, col['uuid'])
1160 # Add a new file to the directory
1161 with open(os.path.join(tmpdir, 'file2'), 'w') as f:
1162 f.write('The quick brown fox jumped over the lazy dog')
1163 updated_col = self.run_and_find_collection("", ['--no-progress', '--update-collection', col['uuid'], tmpdir])
1164 self.assertEqual(col['uuid'], updated_col['uuid'])
1165 # Get the manifest and check that the new file is being included
1166 c = arv_put.api_client.collections().get(uuid=updated_col['uuid']).execute()
1167 self.assertRegex(c['manifest_text'], r'^\..* .*:44:file2\n')
1169 def test_put_collection_with_utc_expiring_datetime(self):
1170 tmpdir = self.make_tmpdir()
1171 trash_at = (datetime.datetime.utcnow() + datetime.timedelta(days=90)).strftime('%Y%m%dT%H%MZ')
1172 with open(os.path.join(tmpdir, 'file1'), 'w') as f:
1173 f.write('Relaxing in basins at the end of inlets terminates the endless tests from the box')
1174 col = self.run_and_find_collection(
1176 ['--no-progress', '--trash-at', trash_at, tmpdir])
1177 self.assertNotEqual(None, col['uuid'])
1178 c = arv_put.api_client.collections().get(uuid=col['uuid']).execute()
1179 self.assertEqual(ciso8601.parse_datetime(trash_at),
1180 ciso8601.parse_datetime(c['trash_at']))
1182 def test_put_collection_with_timezone_aware_expiring_datetime(self):
1183 tmpdir = self.make_tmpdir()
1184 trash_at = (datetime.datetime.utcnow() + datetime.timedelta(days=90)).strftime('%Y%m%dT%H%M-0300')
1185 with open(os.path.join(tmpdir, 'file1'), 'w') as f:
1186 f.write('Relaxing in basins at the end of inlets terminates the endless tests from the box')
1187 col = self.run_and_find_collection(
1189 ['--no-progress', '--trash-at', trash_at, tmpdir])
1190 self.assertNotEqual(None, col['uuid'])
1191 c = arv_put.api_client.collections().get(uuid=col['uuid']).execute()
1193 ciso8601.parse_datetime(trash_at).replace(tzinfo=None) + datetime.timedelta(hours=3),
1194 ciso8601.parse_datetime(c['trash_at']).replace(tzinfo=None))
1196 def test_put_collection_with_timezone_naive_expiring_datetime(self):
1197 tmpdir = self.make_tmpdir()
1198 trash_at = (datetime.datetime.utcnow() + datetime.timedelta(days=90)).strftime('%Y%m%dT%H%M')
1199 with open(os.path.join(tmpdir, 'file1'), 'w') as f:
1200 f.write('Relaxing in basins at the end of inlets terminates the endless tests from the box')
1201 col = self.run_and_find_collection(
1203 ['--no-progress', '--trash-at', trash_at, tmpdir])
1204 self.assertNotEqual(None, col['uuid'])
1205 c = arv_put.api_client.collections().get(uuid=col['uuid']).execute()
1207 offset = datetime.timedelta(seconds=time.altzone)
1209 offset = datetime.timedelta(seconds=time.timezone)
1211 ciso8601.parse_datetime(trash_at) + offset,
1212 ciso8601.parse_datetime(c['trash_at']).replace(tzinfo=None))
1214 def test_put_collection_with_expiring_date_only(self):
1215 tmpdir = self.make_tmpdir()
1216 trash_at = '2140-01-01'
1217 end_of_day = datetime.timedelta(hours=23, minutes=59, seconds=59)
1218 with open(os.path.join(tmpdir, 'file1'), 'w') as f:
1219 f.write('Relaxing in basins at the end of inlets terminates the endless tests from the box')
1220 col = self.run_and_find_collection(
1222 ['--no-progress', '--trash-at', trash_at, tmpdir])
1223 self.assertNotEqual(None, col['uuid'])
1224 c = arv_put.api_client.collections().get(uuid=col['uuid']).execute()
1226 offset = datetime.timedelta(seconds=time.altzone)
1228 offset = datetime.timedelta(seconds=time.timezone)
1230 ciso8601.parse_datetime(trash_at) + end_of_day + offset,
1231 ciso8601.parse_datetime(c['trash_at']).replace(tzinfo=None))
1233 def test_put_collection_with_invalid_absolute_expiring_datetimes(self):
1234 cases = ['2100', '210010','2100-10', '2100-Oct']
1235 tmpdir = self.make_tmpdir()
1236 with open(os.path.join(tmpdir, 'file1'), 'w') as f:
1237 f.write('Relaxing in basins at the end of inlets terminates the endless tests from the box')
1238 for test_datetime in cases:
1239 with self.assertRaises(AssertionError):
1240 self.run_and_find_collection(
1242 ['--no-progress', '--trash-at', test_datetime, tmpdir])
1244 def test_put_collection_with_relative_expiring_datetime(self):
1246 dt_before = datetime.datetime.utcnow() + datetime.timedelta(days=expire_after)
1247 tmpdir = self.make_tmpdir()
1248 with open(os.path.join(tmpdir, 'file1'), 'w') as f:
1249 f.write('Relaxing in basins at the end of inlets terminates the endless tests from the box')
1250 col = self.run_and_find_collection(
1252 ['--no-progress', '--trash-after', str(expire_after), tmpdir])
1253 self.assertNotEqual(None, col['uuid'])
1254 dt_after = datetime.datetime.utcnow() + datetime.timedelta(days=expire_after)
1255 c = arv_put.api_client.collections().get(uuid=col['uuid']).execute()
1256 trash_at = ciso8601.parse_datetime(c['trash_at']).replace(tzinfo=None)
1257 self.assertTrue(dt_before < trash_at)
1258 self.assertTrue(dt_after > trash_at)
1260 def test_put_collection_with_invalid_relative_expiring_datetime(self):
1261 expire_after = 0 # Must be >= 1
1262 tmpdir = self.make_tmpdir()
1263 with open(os.path.join(tmpdir, 'file1'), 'w') as f:
1264 f.write('Relaxing in basins at the end of inlets terminates the endless tests from the box')
1265 with self.assertRaises(AssertionError):
1266 self.run_and_find_collection(
1268 ['--no-progress', '--trash-after', str(expire_after), tmpdir])
1270 def test_upload_directory_reference_without_trailing_slash(self):
1271 tmpdir1 = self.make_tmpdir()
1272 tmpdir2 = self.make_tmpdir()
1273 with open(os.path.join(tmpdir1, 'foo'), 'w') as f:
1274 f.write('This is foo')
1275 with open(os.path.join(tmpdir2, 'bar'), 'w') as f:
1276 f.write('This is not foo')
1277 # Upload one directory and one file
1278 col = self.run_and_find_collection("", ['--no-progress',
1280 os.path.join(tmpdir2, 'bar')])
1281 self.assertNotEqual(None, col['uuid'])
1282 c = arv_put.api_client.collections().get(uuid=col['uuid']).execute()
1283 # Check that 'foo' was written inside a subcollection
1284 # OTOH, 'bar' should have been directly uploaded on the root collection
1285 self.assertRegex(c['manifest_text'], r'^\. .*:15:bar\n\./.+ .*:11:foo\n')
1287 def test_upload_directory_reference_with_trailing_slash(self):
1288 tmpdir1 = self.make_tmpdir()
1289 tmpdir2 = self.make_tmpdir()
1290 with open(os.path.join(tmpdir1, 'foo'), 'w') as f:
1291 f.write('This is foo')
1292 with open(os.path.join(tmpdir2, 'bar'), 'w') as f:
1293 f.write('This is not foo')
1294 # Upload one directory (with trailing slash) and one file
1295 col = self.run_and_find_collection("", ['--no-progress',
1297 os.path.join(tmpdir2, 'bar')])
1298 self.assertNotEqual(None, col['uuid'])
1299 c = arv_put.api_client.collections().get(uuid=col['uuid']).execute()
1300 # Check that 'foo' and 'bar' were written at the same level
1301 self.assertRegex(c['manifest_text'], r'^\. .*:15:bar .*:11:foo\n')
1303 def test_put_collection_with_high_redundancy(self):
1304 # Write empty data: we're not testing CollectionWriter, just
1305 # making sure collections.create tells the API server what our
1306 # desired replication level is.
1307 collection = self.run_and_find_collection("", ['--replication', '4'])
1308 self.assertEqual(4, collection['replication_desired'])
1310 def test_put_collection_with_default_redundancy(self):
1311 collection = self.run_and_find_collection("")
1312 self.assertEqual(None, collection['replication_desired'])
1314 def test_put_collection_with_unnamed_project_link(self):
1315 link = self.run_and_find_collection(
1316 "Test unnamed collection",
1317 ['--portable-data-hash', '--project-uuid', self.PROJECT_UUID])
1318 username = pwd.getpwuid(os.getuid()).pw_name
1321 r'^Saved at .* by {}@'.format(re.escape(username)))
1323 def test_put_collection_with_name_and_no_project(self):
1324 link_name = 'Test Collection Link in home project'
1325 collection = self.run_and_find_collection(
1326 "Test named collection in home project",
1327 ['--portable-data-hash', '--name', link_name])
1328 self.assertEqual(link_name, collection['name'])
1329 my_user_uuid = self.current_user()['uuid']
1330 self.assertEqual(my_user_uuid, collection['owner_uuid'])
1332 def test_put_collection_with_named_project_link(self):
1333 link_name = 'Test auto Collection Link'
1334 collection = self.run_and_find_collection("Test named collection",
1335 ['--portable-data-hash',
1336 '--name', link_name,
1337 '--project-uuid', self.PROJECT_UUID])
1338 self.assertEqual(link_name, collection['name'])
1340 def test_put_collection_with_storage_classes_specified(self):
1341 collection = self.run_and_find_collection("", ['--storage-classes', 'hot'])
1342 self.assertEqual(len(collection['storage_classes_desired']), 1)
1343 self.assertEqual(collection['storage_classes_desired'][0], 'hot')
1345 def test_put_collection_with_multiple_storage_classes_specified(self):
1346 collection = self.run_and_find_collection("", ['--storage-classes', ' foo, bar ,baz'])
1347 self.assertEqual(len(collection['storage_classes_desired']), 3)
1348 self.assertEqual(collection['storage_classes_desired'], ['foo', 'bar', 'baz'])
1350 def test_put_collection_without_storage_classes_specified(self):
1351 collection = self.run_and_find_collection("")
1352 self.assertEqual(len(collection['storage_classes_desired']), 1)
1353 self.assertEqual(collection['storage_classes_desired'][0], 'default')
1355 def test_exclude_filename_pattern(self):
1356 tmpdir = self.make_tmpdir()
1357 tmpsubdir = os.path.join(tmpdir, 'subdir')
1359 for fname in ['file1', 'file2', 'file3']:
1360 with open(os.path.join(tmpdir, "%s.txt" % fname), 'w') as f:
1361 f.write("This is %s" % fname)
1362 with open(os.path.join(tmpsubdir, "%s.txt" % fname), 'w') as f:
1363 f.write("This is %s" % fname)
1364 col = self.run_and_find_collection("", ['--no-progress',
1365 '--exclude', '*2.txt',
1366 '--exclude', 'file3.*',
1368 self.assertNotEqual(None, col['uuid'])
1369 c = arv_put.api_client.collections().get(uuid=col['uuid']).execute()
1370 # None of the file2.txt & file3.txt should have been uploaded
1371 self.assertRegex(c['manifest_text'], r'^.*:file1.txt')
1372 self.assertNotRegex(c['manifest_text'], r'^.*:file2.txt')
1373 self.assertNotRegex(c['manifest_text'], r'^.*:file3.txt')
1375 def test_exclude_filepath_pattern(self):
1376 tmpdir = self.make_tmpdir()
1377 tmpsubdir = os.path.join(tmpdir, 'subdir')
1379 for fname in ['file1', 'file2', 'file3']:
1380 with open(os.path.join(tmpdir, "%s.txt" % fname), 'w') as f:
1381 f.write("This is %s" % fname)
1382 with open(os.path.join(tmpsubdir, "%s.txt" % fname), 'w') as f:
1383 f.write("This is %s" % fname)
1384 col = self.run_and_find_collection("", ['--no-progress',
1385 '--exclude', 'subdir/*2.txt',
1386 '--exclude', './file1.*',
1388 self.assertNotEqual(None, col['uuid'])
1389 c = arv_put.api_client.collections().get(uuid=col['uuid']).execute()
1390 # Only tmpdir/file1.txt & tmpdir/subdir/file2.txt should have been excluded
1391 self.assertNotRegex(c['manifest_text'],
1392 r'^\./%s.*:file1.txt' % os.path.basename(tmpdir))
1393 self.assertNotRegex(c['manifest_text'],
1394 r'^\./%s/subdir.*:file2.txt' % os.path.basename(tmpdir))
1395 self.assertRegex(c['manifest_text'],
1396 r'^\./%s.*:file2.txt' % os.path.basename(tmpdir))
1397 self.assertRegex(c['manifest_text'], r'^.*:file3.txt')
1399 def test_unicode_on_filename(self):
1400 tmpdir = self.make_tmpdir()
1401 fname = u"iā¤arvados.txt"
1402 with open(os.path.join(tmpdir, fname), 'w') as f:
1403 f.write("This is a unicode named file")
1404 col = self.run_and_find_collection("", ['--no-progress', tmpdir])
1405 self.assertNotEqual(None, col['uuid'])
1406 c = arv_put.api_client.collections().get(uuid=col['uuid']).execute()
1407 self.assertTrue(fname in c['manifest_text'], u"{} does not include {}".format(c['manifest_text'], fname))
1409 def test_silent_mode_no_errors(self):
1410 self.authorize_with('active')
1411 tmpdir = self.make_tmpdir()
1412 with open(os.path.join(tmpdir, 'test.txt'), 'w') as f:
1413 f.write('hello world')
1414 pipe = subprocess.Popen(
1415 [sys.executable, arv_put.__file__] + ['--silent', tmpdir],
1416 stdin=subprocess.PIPE, stdout=subprocess.PIPE,
1417 stderr=subprocess.PIPE, env=self.ENVIRON)
1418 stdout, stderr = pipe.communicate()
1419 # No console output should occur on normal operations
1420 self.assertNotRegex(stderr.decode(), r'.+')
1421 self.assertNotRegex(stdout.decode(), r'.+')
1423 def test_silent_mode_does_not_avoid_error_messages(self):
1424 self.authorize_with('active')
1425 pipe = subprocess.Popen(
1426 [sys.executable, arv_put.__file__] + ['--silent',
1427 '/path/not/existant'],
1428 stdin=subprocess.PIPE, stdout=subprocess.PIPE,
1429 stderr=subprocess.PIPE, env=self.ENVIRON)
1430 stdout, stderr = pipe.communicate()
1431 # Error message should be displayed when errors happen
1432 self.assertRegex(stderr.decode(), r'.*ERROR:.*')
1433 self.assertNotRegex(stdout.decode(), r'.+')
1436 if __name__ == '__main__':