1 from __future__ import absolute_import
2 from __future__ import division
3 from future import standard_library
4 standard_library.install_aliases()
5 from builtins import str
6 from builtins import range
25 import arvados.commands.put as arv_put
26 from . import arvados_testutil as tutil
28 from .arvados_testutil import ArvadosBaseTestCase, fake_httplib2_response
29 from . import run_test_server
31 class ArvadosPutResumeCacheTest(ArvadosBaseTestCase):
35 ['/dev/null', '--filename', 'empty'],
40 super(ArvadosPutResumeCacheTest, self).tearDown()
42 self.last_cache.destroy()
43 except AttributeError:
46 def cache_path_from_arglist(self, arglist):
47 return arv_put.ResumeCache.make_path(arv_put.parse_arguments(arglist))
49 def test_cache_names_stable(self):
50 for argset in self.CACHE_ARGSET:
51 self.assertEqual(self.cache_path_from_arglist(argset),
52 self.cache_path_from_arglist(argset),
53 "cache name changed for {}".format(argset))
55 def test_cache_names_unique(self):
57 for argset in self.CACHE_ARGSET:
58 path = self.cache_path_from_arglist(argset)
59 self.assertNotIn(path, results)
62 def test_cache_names_simple(self):
63 # The goal here is to make sure the filename doesn't use characters
64 # reserved by the filesystem. Feel free to adjust this regexp as
65 # long as it still does that.
66 bad_chars = re.compile(r'[^-\.\w]')
67 for argset in self.CACHE_ARGSET:
68 path = self.cache_path_from_arglist(argset)
69 self.assertFalse(bad_chars.search(os.path.basename(path)),
70 "path too exotic: {}".format(path))
72 def test_cache_names_ignore_argument_order(self):
74 self.cache_path_from_arglist(['a', 'b', 'c']),
75 self.cache_path_from_arglist(['c', 'a', 'b']))
77 self.cache_path_from_arglist(['-', '--filename', 'stdin']),
78 self.cache_path_from_arglist(['--filename', 'stdin', '-']))
80 def test_cache_names_differ_for_similar_paths(self):
81 # This test needs names at / that don't exist on the real filesystem.
83 self.cache_path_from_arglist(['/_arvputtest1', '/_arvputtest2']),
84 self.cache_path_from_arglist(['/_arvputtest1/_arvputtest2']))
86 def test_cache_names_ignore_irrelevant_arguments(self):
87 # Workaround: parse_arguments bails on --filename with a directory.
88 path1 = self.cache_path_from_arglist(['/tmp'])
89 args = arv_put.parse_arguments(['/tmp'])
91 path2 = arv_put.ResumeCache.make_path(args)
92 self.assertEqual(path1, path2,
93 "cache path considered --filename for directory")
95 self.cache_path_from_arglist(['-']),
96 self.cache_path_from_arglist(['-', '--max-manifest-depth', '1']),
97 "cache path considered --max-manifest-depth for file")
99 def test_cache_names_treat_negative_manifest_depths_identically(self):
100 base_args = ['/tmp', '--max-manifest-depth']
102 self.cache_path_from_arglist(base_args + ['-1']),
103 self.cache_path_from_arglist(base_args + ['-2']))
105 def test_cache_names_treat_stdin_consistently(self):
107 self.cache_path_from_arglist(['-', '--filename', 'test']),
108 self.cache_path_from_arglist(['/dev/stdin', '--filename', 'test']))
110 def test_cache_names_identical_for_synonymous_names(self):
112 self.cache_path_from_arglist(['.']),
113 self.cache_path_from_arglist([os.path.realpath('.')]))
114 testdir = self.make_tmpdir()
115 looplink = os.path.join(testdir, 'loop')
116 os.symlink(testdir, looplink)
118 self.cache_path_from_arglist([testdir]),
119 self.cache_path_from_arglist([looplink]))
121 def test_cache_names_different_by_api_host(self):
122 config = arvados.config.settings()
123 orig_host = config.get('ARVADOS_API_HOST')
125 name1 = self.cache_path_from_arglist(['.'])
126 config['ARVADOS_API_HOST'] = 'x' + (orig_host or 'localhost')
127 self.assertNotEqual(name1, self.cache_path_from_arglist(['.']))
129 if orig_host is None:
130 del config['ARVADOS_API_HOST']
132 config['ARVADOS_API_HOST'] = orig_host
134 @mock.patch('arvados.keep.KeepClient.head')
135 def test_resume_cache_with_current_stream_locators(self, keep_client_head):
136 keep_client_head.side_effect = [True]
138 thing['_current_stream_locators'] = ['098f6bcd4621d373cade4e832627b4f6+4', '1f253c60a2306e0ee12fb6ce0c587904+6']
139 with tempfile.NamedTemporaryFile() as cachefile:
140 self.last_cache = arv_put.ResumeCache(cachefile.name)
141 self.last_cache.save(thing)
142 self.last_cache.close()
143 resume_cache = arv_put.ResumeCache(self.last_cache.filename)
144 self.assertNotEqual(None, resume_cache)
146 @mock.patch('arvados.keep.KeepClient.head')
147 def test_resume_cache_with_finished_streams(self, keep_client_head):
148 keep_client_head.side_effect = [True]
150 thing['_finished_streams'] = [['.', ['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_error_on_head(self, keep_client_head):
160 keep_client_head.side_effect = Exception('Locator not found')
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)
169 self.assertRaises(None, resume_cache.check_cache())
171 def test_basic_cache_storage(self):
172 thing = ['test', 'list']
173 with tempfile.NamedTemporaryFile() as cachefile:
174 self.last_cache = arv_put.ResumeCache(cachefile.name)
175 self.last_cache.save(thing)
176 self.assertEqual(thing, self.last_cache.load())
178 def test_empty_cache(self):
179 with tempfile.NamedTemporaryFile() as cachefile:
180 cache = arv_put.ResumeCache(cachefile.name)
181 self.assertRaises(ValueError, cache.load)
183 def test_cache_persistent(self):
184 thing = ['test', 'list']
185 path = os.path.join(self.make_tmpdir(), 'cache')
186 cache = arv_put.ResumeCache(path)
189 self.last_cache = arv_put.ResumeCache(path)
190 self.assertEqual(thing, self.last_cache.load())
192 def test_multiple_cache_writes(self):
193 thing = ['short', 'list']
194 with tempfile.NamedTemporaryFile() as cachefile:
195 self.last_cache = arv_put.ResumeCache(cachefile.name)
196 # Start writing an object longer than the one we test, to make
197 # sure the cache file gets truncated.
198 self.last_cache.save(['long', 'long', 'list'])
199 self.last_cache.save(thing)
200 self.assertEqual(thing, self.last_cache.load())
202 def test_cache_is_locked(self):
203 with tempfile.NamedTemporaryFile() as cachefile:
204 cache = arv_put.ResumeCache(cachefile.name)
205 self.assertRaises(arv_put.ResumeCacheConflict,
206 arv_put.ResumeCache, cachefile.name)
208 def test_cache_stays_locked(self):
209 with tempfile.NamedTemporaryFile() as cachefile:
210 self.last_cache = arv_put.ResumeCache(cachefile.name)
211 path = cachefile.name
212 self.last_cache.save('test')
213 self.assertRaises(arv_put.ResumeCacheConflict,
214 arv_put.ResumeCache, path)
216 def test_destroy_cache(self):
217 cachefile = tempfile.NamedTemporaryFile(delete=False)
219 cache = arv_put.ResumeCache(cachefile.name)
223 arv_put.ResumeCache(cachefile.name)
224 except arv_put.ResumeCacheConflict:
225 self.fail("could not load cache after destroying it")
226 self.assertRaises(ValueError, cache.load)
228 if os.path.exists(cachefile.name):
229 os.unlink(cachefile.name)
231 def test_restart_cache(self):
232 path = os.path.join(self.make_tmpdir(), 'cache')
233 cache = arv_put.ResumeCache(path)
236 self.assertRaises(ValueError, cache.load)
237 self.assertRaises(arv_put.ResumeCacheConflict,
238 arv_put.ResumeCache, path)
241 class ArvPutUploadJobTest(run_test_server.TestCaseWithServers,
242 ArvadosBaseTestCase):
245 super(ArvPutUploadJobTest, self).setUp()
246 run_test_server.authorize_with('active')
247 # Temp files creation
248 self.tempdir = tempfile.mkdtemp()
249 subdir = os.path.join(self.tempdir, 'subdir')
251 data = "x" * 1024 # 1 KB
252 for i in range(1, 5):
253 with open(os.path.join(self.tempdir, str(i)), 'w') as f:
255 with open(os.path.join(subdir, 'otherfile'), 'w') as f:
257 # Large temp file for resume test
258 _, self.large_file_name = tempfile.mkstemp()
259 fileobj = open(self.large_file_name, 'w')
260 # Make sure to write just a little more than one block
261 for _ in range((arvados.config.KEEP_BLOCK_SIZE>>20)+1):
262 data = random.choice(['x', 'y', 'z']) * 1024 * 1024 # 1 MiB
265 # Temp dir containing small files to be repacked
266 self.small_files_dir = tempfile.mkdtemp()
267 data = 'y' * 1024 * 1024 # 1 MB
268 for i in range(1, 70):
269 with open(os.path.join(self.small_files_dir, str(i)), 'w') as f:
270 f.write(data + str(i))
271 self.arvfile_write = getattr(arvados.arvfile.ArvadosFileWriter, 'write')
272 # Temp dir to hold a symlink to other temp dir
273 self.tempdir_with_symlink = tempfile.mkdtemp()
274 os.symlink(self.tempdir, os.path.join(self.tempdir_with_symlink, 'linkeddir'))
275 os.symlink(os.path.join(self.tempdir, '1'),
276 os.path.join(self.tempdir_with_symlink, 'linkedfile'))
279 super(ArvPutUploadJobTest, self).tearDown()
280 shutil.rmtree(self.tempdir)
281 os.unlink(self.large_file_name)
282 shutil.rmtree(self.small_files_dir)
283 shutil.rmtree(self.tempdir_with_symlink)
285 def test_symlinks_are_followed_by_default(self):
286 cwriter = arv_put.ArvPutUploadJob([self.tempdir_with_symlink])
287 cwriter.start(save_collection=False)
288 self.assertIn('linkeddir', cwriter.manifest_text())
289 self.assertIn('linkedfile', cwriter.manifest_text())
290 cwriter.destroy_cache()
292 def test_symlinks_are_not_followed_when_requested(self):
293 cwriter = arv_put.ArvPutUploadJob([self.tempdir_with_symlink],
295 cwriter.start(save_collection=False)
296 self.assertNotIn('linkeddir', cwriter.manifest_text())
297 self.assertNotIn('linkedfile', cwriter.manifest_text())
298 cwriter.destroy_cache()
300 def test_passing_nonexistant_path_raise_exception(self):
301 uuid_str = str(uuid.uuid4())
302 cwriter = arv_put.ArvPutUploadJob(["/this/path/does/not/exist/{}".format(uuid_str)])
303 with self.assertRaises(arv_put.PathDoesNotExistError):
304 cwriter.start(save_collection=False)
306 def test_writer_works_without_cache(self):
307 cwriter = arv_put.ArvPutUploadJob(['/dev/null'], resume=False)
308 cwriter.start(save_collection=False)
309 self.assertEqual(". d41d8cd98f00b204e9800998ecf8427e+0 0:0:null\n", cwriter.manifest_text())
311 def test_writer_works_with_cache(self):
312 with tempfile.NamedTemporaryFile() as f:
315 cwriter = arv_put.ArvPutUploadJob([f.name])
316 cwriter.start(save_collection=False)
317 self.assertEqual(0, cwriter.bytes_skipped)
318 self.assertEqual(3, cwriter.bytes_written)
319 # Don't destroy the cache, and start another upload
320 cwriter_new = arv_put.ArvPutUploadJob([f.name])
321 cwriter_new.start(save_collection=False)
322 cwriter_new.destroy_cache()
323 self.assertEqual(3, cwriter_new.bytes_skipped)
324 self.assertEqual(3, cwriter_new.bytes_written)
326 def make_progress_tester(self):
328 def record_func(written, expected):
329 progression.append((written, expected))
330 return progression, record_func
332 def test_progress_reporting(self):
333 with tempfile.NamedTemporaryFile() as f:
336 for expect_count in (None, 8):
337 progression, reporter = self.make_progress_tester()
338 cwriter = arv_put.ArvPutUploadJob([f.name],
339 reporter=reporter, bytes_expected=expect_count)
340 cwriter.start(save_collection=False)
341 cwriter.destroy_cache()
342 self.assertIn((3, expect_count), progression)
344 def test_writer_upload_directory(self):
345 cwriter = arv_put.ArvPutUploadJob([self.tempdir])
346 cwriter.start(save_collection=False)
347 cwriter.destroy_cache()
348 self.assertEqual(1024*(1+2+3+4+5), cwriter.bytes_written)
350 def test_resume_large_file_upload(self):
351 def wrapped_write(*args, **kwargs):
353 # Exit only on last block
354 if len(data) < arvados.config.KEEP_BLOCK_SIZE:
355 # Simulate a checkpoint before quitting. Ensure block commit.
356 self.writer._update(final=True)
357 raise SystemExit("Simulated error")
358 return self.arvfile_write(*args, **kwargs)
360 with mock.patch('arvados.arvfile.ArvadosFileWriter.write',
361 autospec=True) as mocked_write:
362 mocked_write.side_effect = wrapped_write
363 writer = arv_put.ArvPutUploadJob([self.large_file_name],
364 replication_desired=1)
365 # We'll be accessing from inside the wrapper
367 with self.assertRaises(SystemExit):
368 writer.start(save_collection=False)
369 # Confirm that the file was partially uploaded
370 self.assertGreater(writer.bytes_written, 0)
371 self.assertLess(writer.bytes_written,
372 os.path.getsize(self.large_file_name))
374 writer2 = arv_put.ArvPutUploadJob([self.large_file_name],
375 replication_desired=1)
376 writer2.start(save_collection=False)
377 self.assertEqual(writer.bytes_written + writer2.bytes_written - writer2.bytes_skipped,
378 os.path.getsize(self.large_file_name))
379 writer2.destroy_cache()
382 # Test for bug #11002
383 def test_graceful_exit_while_repacking_small_blocks(self):
384 def wrapped_commit(*args, **kwargs):
385 raise SystemExit("Simulated error")
387 with mock.patch('arvados.arvfile._BlockManager.commit_bufferblock',
388 autospec=True) as mocked_commit:
389 mocked_commit.side_effect = wrapped_commit
390 # Upload a little more than 1 block, wrapped_commit will make the first block
392 # arv-put should not exit with an exception by trying to commit the collection
393 # as it's in an inconsistent state.
394 writer = arv_put.ArvPutUploadJob([self.small_files_dir],
395 replication_desired=1)
397 with self.assertRaises(SystemExit):
398 writer.start(save_collection=False)
399 except arvados.arvfile.UnownedBlockError:
400 self.fail("arv-put command is trying to use a corrupted BlockManager. See https://dev.arvados.org/issues/11002")
401 writer.destroy_cache()
403 def test_no_resume_when_asked(self):
404 def wrapped_write(*args, **kwargs):
406 # Exit only on last block
407 if len(data) < arvados.config.KEEP_BLOCK_SIZE:
408 # Simulate a checkpoint before quitting.
409 self.writer._update()
410 raise SystemExit("Simulated error")
411 return self.arvfile_write(*args, **kwargs)
413 with mock.patch('arvados.arvfile.ArvadosFileWriter.write',
414 autospec=True) as mocked_write:
415 mocked_write.side_effect = wrapped_write
416 writer = arv_put.ArvPutUploadJob([self.large_file_name],
417 replication_desired=1)
418 # We'll be accessing from inside the wrapper
420 with self.assertRaises(SystemExit):
421 writer.start(save_collection=False)
422 # Confirm that the file was partially uploaded
423 self.assertGreater(writer.bytes_written, 0)
424 self.assertLess(writer.bytes_written,
425 os.path.getsize(self.large_file_name))
426 # Retry the upload, this time without resume
427 writer2 = arv_put.ArvPutUploadJob([self.large_file_name],
428 replication_desired=1,
430 writer2.start(save_collection=False)
431 self.assertEqual(writer2.bytes_skipped, 0)
432 self.assertEqual(writer2.bytes_written,
433 os.path.getsize(self.large_file_name))
434 writer2.destroy_cache()
437 def test_no_resume_when_no_cache(self):
438 def wrapped_write(*args, **kwargs):
440 # Exit only on last block
441 if len(data) < arvados.config.KEEP_BLOCK_SIZE:
442 # Simulate a checkpoint before quitting.
443 self.writer._update()
444 raise SystemExit("Simulated error")
445 return self.arvfile_write(*args, **kwargs)
447 with mock.patch('arvados.arvfile.ArvadosFileWriter.write',
448 autospec=True) as mocked_write:
449 mocked_write.side_effect = wrapped_write
450 writer = arv_put.ArvPutUploadJob([self.large_file_name],
451 replication_desired=1)
452 # We'll be accessing from inside the wrapper
454 with self.assertRaises(SystemExit):
455 writer.start(save_collection=False)
456 # Confirm that the file was partially uploaded
457 self.assertGreater(writer.bytes_written, 0)
458 self.assertLess(writer.bytes_written,
459 os.path.getsize(self.large_file_name))
460 # Retry the upload, this time without cache usage
461 writer2 = arv_put.ArvPutUploadJob([self.large_file_name],
462 replication_desired=1,
465 writer2.start(save_collection=False)
466 self.assertEqual(writer2.bytes_skipped, 0)
467 self.assertEqual(writer2.bytes_written,
468 os.path.getsize(self.large_file_name))
469 writer2.destroy_cache()
472 def test_dry_run_feature(self):
473 def wrapped_write(*args, **kwargs):
475 # Exit only on last block
476 if len(data) < arvados.config.KEEP_BLOCK_SIZE:
477 # Simulate a checkpoint before quitting.
478 self.writer._update()
479 raise SystemExit("Simulated error")
480 return self.arvfile_write(*args, **kwargs)
482 with mock.patch('arvados.arvfile.ArvadosFileWriter.write',
483 autospec=True) as mocked_write:
484 mocked_write.side_effect = wrapped_write
485 writer = arv_put.ArvPutUploadJob([self.large_file_name],
486 replication_desired=1)
487 # We'll be accessing from inside the wrapper
489 with self.assertRaises(SystemExit):
490 writer.start(save_collection=False)
491 # Confirm that the file was partially uploaded
492 self.assertGreater(writer.bytes_written, 0)
493 self.assertLess(writer.bytes_written,
494 os.path.getsize(self.large_file_name))
495 # Retry the upload using dry_run to check if there is a pending upload
496 writer2 = arv_put.ArvPutUploadJob([self.large_file_name],
497 replication_desired=1,
499 with self.assertRaises(arv_put.ArvPutUploadIsPending):
500 writer2.start(save_collection=False)
501 # Complete the pending upload
502 writer3 = arv_put.ArvPutUploadJob([self.large_file_name],
503 replication_desired=1)
504 writer3.start(save_collection=False)
505 # Confirm there's no pending upload with dry_run=True
506 writer4 = arv_put.ArvPutUploadJob([self.large_file_name],
507 replication_desired=1,
509 with self.assertRaises(arv_put.ArvPutUploadNotPending):
510 writer4.start(save_collection=False)
511 writer4.destroy_cache()
513 with self.assertRaises(arv_put.ArvPutUploadIsPending):
514 arv_put.ArvPutUploadJob([self.large_file_name],
515 replication_desired=1,
519 with self.assertRaises(arv_put.ArvPutUploadIsPending):
520 arv_put.ArvPutUploadJob([self.large_file_name],
521 replication_desired=1,
526 class ArvadosExpectedBytesTest(ArvadosBaseTestCase):
527 TEST_SIZE = os.path.getsize(__file__)
529 def test_expected_bytes_for_file(self):
530 self.assertEqual(self.TEST_SIZE,
531 arv_put.expected_bytes_for([__file__]))
533 def test_expected_bytes_for_tree(self):
534 tree = self.make_tmpdir()
535 shutil.copyfile(__file__, os.path.join(tree, 'one'))
536 shutil.copyfile(__file__, os.path.join(tree, 'two'))
537 self.assertEqual(self.TEST_SIZE * 2,
538 arv_put.expected_bytes_for([tree]))
539 self.assertEqual(self.TEST_SIZE * 3,
540 arv_put.expected_bytes_for([tree, __file__]))
542 def test_expected_bytes_for_device(self):
543 self.assertIsNone(arv_put.expected_bytes_for(['/dev/null']))
544 self.assertIsNone(arv_put.expected_bytes_for([__file__, '/dev/null']))
547 class ArvadosPutReportTest(ArvadosBaseTestCase):
548 def test_machine_progress(self):
549 for count, total in [(0, 1), (0, None), (1, None), (235, 9283)]:
550 expect = ": {} written {} total\n".format(
551 count, -1 if (total is None) else total)
553 arv_put.machine_progress(count, total).endswith(expect))
555 def test_known_human_progress(self):
556 for count, total in [(0, 1), (2, 4), (45, 60)]:
557 expect = '{:.1%}'.format(1.0*count/total)
558 actual = arv_put.human_progress(count, total)
559 self.assertTrue(actual.startswith('\r'))
560 self.assertIn(expect, actual)
562 def test_unknown_human_progress(self):
563 for count in [1, 20, 300, 4000, 50000]:
564 self.assertTrue(re.search(r'\b{}\b'.format(count),
565 arv_put.human_progress(count, None)))
568 class ArvadosPutTest(run_test_server.TestCaseWithServers,
570 tutil.VersionChecker):
572 Z_UUID = 'zzzzz-zzzzz-zzzzzzzzzzzzzzz'
574 def call_main_with_args(self, args):
575 self.main_stdout = tutil.StringIO()
576 self.main_stderr = tutil.StringIO()
577 return arv_put.main(args, self.main_stdout, self.main_stderr)
579 def call_main_on_test_file(self, args=[]):
580 with self.make_test_file() as testfile:
582 self.call_main_with_args(['--stream', '--no-progress'] + args + [path])
584 os.path.exists(os.path.join(os.environ['KEEP_LOCAL_STORE'],
585 '098f6bcd4621d373cade4e832627b4f6')),
586 "did not find file stream in Keep store")
589 super(ArvadosPutTest, self).setUp()
590 run_test_server.authorize_with('active')
591 arv_put.api_client = None
594 for outbuf in ['main_stdout', 'main_stderr']:
595 if hasattr(self, outbuf):
596 getattr(self, outbuf).close()
597 delattr(self, outbuf)
598 super(ArvadosPutTest, self).tearDown()
600 def test_version_argument(self):
601 with tutil.redirected_streams(
602 stdout=tutil.StringIO, stderr=tutil.StringIO) as (out, err):
603 with self.assertRaises(SystemExit):
604 self.call_main_with_args(['--version'])
605 self.assertVersionOutput(out, err)
607 def test_simple_file_put(self):
608 self.call_main_on_test_file()
610 def test_put_with_unwriteable_cache_dir(self):
611 orig_cachedir = arv_put.ResumeCache.CACHE_DIR
612 cachedir = self.make_tmpdir()
613 os.chmod(cachedir, 0o0)
614 arv_put.ResumeCache.CACHE_DIR = cachedir
616 self.call_main_on_test_file()
618 arv_put.ResumeCache.CACHE_DIR = orig_cachedir
619 os.chmod(cachedir, 0o700)
621 def test_put_with_unwritable_cache_subdir(self):
622 orig_cachedir = arv_put.ResumeCache.CACHE_DIR
623 cachedir = self.make_tmpdir()
624 os.chmod(cachedir, 0o0)
625 arv_put.ResumeCache.CACHE_DIR = os.path.join(cachedir, 'cachedir')
627 self.call_main_on_test_file()
629 arv_put.ResumeCache.CACHE_DIR = orig_cachedir
630 os.chmod(cachedir, 0o700)
632 def test_put_block_replication(self):
633 self.call_main_on_test_file()
634 with mock.patch('arvados.collection.KeepClient.local_store_put') as put_mock:
635 put_mock.return_value = 'acbd18db4cc2f85cedef654fccc4a4d8+3'
636 self.call_main_on_test_file(['--replication', '1'])
637 self.call_main_on_test_file(['--replication', '4'])
638 self.call_main_on_test_file(['--replication', '5'])
640 [x[-1].get('copies') for x in put_mock.call_args_list],
643 def test_normalize(self):
644 testfile1 = self.make_test_file()
645 testfile2 = self.make_test_file()
646 test_paths = [testfile1.name, testfile2.name]
647 # Reverse-sort the paths, so normalization must change their order.
648 test_paths.sort(reverse=True)
649 self.call_main_with_args(['--stream', '--no-progress', '--normalize'] +
651 manifest = self.main_stdout.getvalue()
652 # Assert the second file we specified appears first in the manifest.
653 file_indices = [manifest.find(':' + os.path.basename(path))
654 for path in test_paths]
655 self.assertGreater(*file_indices)
657 def test_error_name_without_collection(self):
658 self.assertRaises(SystemExit, self.call_main_with_args,
659 ['--name', 'test without Collection',
660 '--stream', '/dev/null'])
662 def test_error_when_project_not_found(self):
663 self.assertRaises(SystemExit,
664 self.call_main_with_args,
665 ['--project-uuid', self.Z_UUID])
667 def test_error_bad_project_uuid(self):
668 self.assertRaises(SystemExit,
669 self.call_main_with_args,
670 ['--project-uuid', self.Z_UUID, '--stream'])
672 def test_api_error_handling(self):
673 coll_save_mock = mock.Mock(name='arv.collection.Collection().save_new()')
674 coll_save_mock.side_effect = arvados.errors.ApiError(
675 fake_httplib2_response(403), b'{}')
676 with mock.patch('arvados.collection.Collection.save_new',
678 with self.assertRaises(SystemExit) as exc_test:
679 self.call_main_with_args(['/dev/null'])
680 self.assertLess(0, exc_test.exception.args[0])
681 self.assertLess(0, coll_save_mock.call_count)
682 self.assertEqual("", self.main_stdout.getvalue())
685 class ArvPutIntegrationTest(run_test_server.TestCaseWithServers,
686 ArvadosBaseTestCase):
687 def _getKeepServerConfig():
688 for config_file, mandatory in [
689 ['application.yml', False], ['application.default.yml', True]]:
690 path = os.path.join(run_test_server.SERVICES_SRC_DIR,
691 "api", "config", config_file)
692 if not mandatory and not os.path.exists(path):
694 with open(path) as f:
695 rails_config = yaml.load(f.read())
696 for config_section in ['test', 'common']:
698 key = rails_config[config_section]["blob_signing_key"]
699 except (KeyError, TypeError):
702 return {'blob_signing_key': key,
703 'enforce_permissions': True}
704 return {'blog_signing_key': None, 'enforce_permissions': False}
707 KEEP_SERVER = _getKeepServerConfig()
708 PROJECT_UUID = run_test_server.fixture('groups')['aproject']['uuid']
712 super(ArvPutIntegrationTest, cls).setUpClass()
713 cls.ENVIRON = os.environ.copy()
714 cls.ENVIRON['PYTHONPATH'] = ':'.join(sys.path)
717 super(ArvPutIntegrationTest, self).setUp()
718 arv_put.api_client = None
720 def authorize_with(self, token_name):
721 run_test_server.authorize_with(token_name)
722 for v in ["ARVADOS_API_HOST",
723 "ARVADOS_API_HOST_INSECURE",
724 "ARVADOS_API_TOKEN"]:
725 self.ENVIRON[v] = arvados.config.settings()[v]
726 arv_put.api_client = arvados.api('v1')
728 def current_user(self):
729 return arv_put.api_client.users().current().execute()
731 def test_check_real_project_found(self):
732 self.authorize_with('active')
733 self.assertTrue(arv_put.desired_project_uuid(arv_put.api_client, self.PROJECT_UUID, 0),
734 "did not correctly find test fixture project")
736 def test_check_error_finding_nonexistent_uuid(self):
737 BAD_UUID = 'zzzzz-zzzzz-zzzzzzzzzzzzzzz'
738 self.authorize_with('active')
740 result = arv_put.desired_project_uuid(arv_put.api_client, BAD_UUID,
742 except ValueError as error:
743 self.assertIn(BAD_UUID, str(error))
745 self.assertFalse(result, "incorrectly found nonexistent project")
747 def test_check_error_finding_nonexistent_project(self):
748 BAD_UUID = 'zzzzz-tpzed-zzzzzzzzzzzzzzz'
749 self.authorize_with('active')
750 with self.assertRaises(apiclient.errors.HttpError):
751 result = arv_put.desired_project_uuid(arv_put.api_client, BAD_UUID,
754 def test_short_put_from_stdin(self):
755 # Have to run this as an integration test since arv-put can't
756 # read from the tests' stdin.
757 # arv-put usually can't stat(os.path.realpath('/dev/stdin')) in this
758 # case, because the /proc entry is already gone by the time it tries.
759 pipe = subprocess.Popen(
760 [sys.executable, arv_put.__file__, '--stream'],
761 stdin=subprocess.PIPE, stdout=subprocess.PIPE,
762 stderr=subprocess.STDOUT, env=self.ENVIRON)
763 pipe.stdin.write(b'stdin test\n')
765 deadline = time.time() + 5
766 while (pipe.poll() is None) and (time.time() < deadline):
768 returncode = pipe.poll()
769 if returncode is None:
771 self.fail("arv-put did not PUT from stdin within 5 seconds")
772 elif returncode != 0:
773 sys.stdout.write(pipe.stdout.read())
774 self.fail("arv-put returned exit code {}".format(returncode))
775 self.assertIn('4a9c8b735dce4b5fa3acf221a0b13628+11',
776 pipe.stdout.read().decode())
778 def test_ArvPutSignedManifest(self):
779 # ArvPutSignedManifest runs "arv-put foo" and then attempts to get
780 # the newly created manifest from the API server, testing to confirm
781 # that the block locators in the returned manifest are signed.
782 self.authorize_with('active')
784 # Before doing anything, demonstrate that the collection
785 # we're about to create is not present in our test fixture.
786 manifest_uuid = "00b4e9f40ac4dd432ef89749f1c01e74+47"
787 with self.assertRaises(apiclient.errors.HttpError):
788 notfound = arv_put.api_client.collections().get(
789 uuid=manifest_uuid).execute()
791 datadir = self.make_tmpdir()
792 with open(os.path.join(datadir, "foo"), "w") as f:
793 f.write("The quick brown fox jumped over the lazy dog")
794 p = subprocess.Popen([sys.executable, arv_put.__file__,
795 os.path.join(datadir, 'foo')],
796 stdout=subprocess.PIPE,
797 stderr=subprocess.PIPE,
799 (out, err) = p.communicate()
800 self.assertRegex(err.decode(), r'INFO: Collection saved as ')
801 self.assertEqual(p.returncode, 0)
803 # The manifest text stored in the API server under the same
804 # manifest UUID must use signed locators.
805 c = arv_put.api_client.collections().get(uuid=manifest_uuid).execute()
808 r'^\. 08a008a01d498c404b0c30852b39d3b8\+44\+A[0-9a-f]+@[0-9a-f]+ 0:44:foo\n')
810 os.remove(os.path.join(datadir, "foo"))
813 def run_and_find_collection(self, text, extra_args=[]):
814 self.authorize_with('active')
815 pipe = subprocess.Popen(
816 [sys.executable, arv_put.__file__] + extra_args,
817 stdin=subprocess.PIPE, stdout=subprocess.PIPE,
818 stderr=subprocess.PIPE, env=self.ENVIRON)
819 stdout, stderr = pipe.communicate(text.encode())
820 self.assertRegex(stderr.decode(), r'INFO: Collection (updated:|saved as)')
821 search_key = ('portable_data_hash'
822 if '--portable-data-hash' in extra_args else 'uuid')
823 collection_list = arvados.api('v1').collections().list(
824 filters=[[search_key, '=', stdout.decode().strip()]]
825 ).execute().get('items', [])
826 self.assertEqual(1, len(collection_list))
827 return collection_list[0]
829 def test_put_collection_with_later_update(self):
830 tmpdir = self.make_tmpdir()
831 with open(os.path.join(tmpdir, 'file1'), 'w') as f:
832 f.write('Relaxing in basins at the end of inlets terminates the endless tests from the box')
833 col = self.run_and_find_collection("", ['--no-progress', tmpdir])
834 self.assertNotEqual(None, col['uuid'])
835 # Add a new file to the directory
836 with open(os.path.join(tmpdir, 'file2'), 'w') as f:
837 f.write('The quick brown fox jumped over the lazy dog')
838 updated_col = self.run_and_find_collection("", ['--no-progress', '--update-collection', col['uuid'], tmpdir])
839 self.assertEqual(col['uuid'], updated_col['uuid'])
840 # Get the manifest and check that the new file is being included
841 c = arv_put.api_client.collections().get(uuid=updated_col['uuid']).execute()
842 self.assertRegex(c['manifest_text'], r'^\..* .*:44:file2\n')
844 def test_upload_directory_reference_without_trailing_slash(self):
845 tmpdir1 = self.make_tmpdir()
846 tmpdir2 = self.make_tmpdir()
847 with open(os.path.join(tmpdir1, 'foo'), 'w') as f:
848 f.write('This is foo')
849 with open(os.path.join(tmpdir2, 'bar'), 'w') as f:
850 f.write('This is not foo')
851 # Upload one directory and one file
852 col = self.run_and_find_collection("", ['--no-progress',
854 os.path.join(tmpdir2, 'bar')])
855 self.assertNotEqual(None, col['uuid'])
856 c = arv_put.api_client.collections().get(uuid=col['uuid']).execute()
857 # Check that 'foo' was written inside a subcollection
858 # OTOH, 'bar' should have been directly uploaded on the root collection
859 self.assertRegex(c['manifest_text'], r'^\. .*:15:bar\n\./.+ .*:11:foo\n')
861 def test_upload_directory_reference_with_trailing_slash(self):
862 tmpdir1 = self.make_tmpdir()
863 tmpdir2 = self.make_tmpdir()
864 with open(os.path.join(tmpdir1, 'foo'), 'w') as f:
865 f.write('This is foo')
866 with open(os.path.join(tmpdir2, 'bar'), 'w') as f:
867 f.write('This is not foo')
868 # Upload one directory (with trailing slash) and one file
869 col = self.run_and_find_collection("", ['--no-progress',
871 os.path.join(tmpdir2, 'bar')])
872 self.assertNotEqual(None, col['uuid'])
873 c = arv_put.api_client.collections().get(uuid=col['uuid']).execute()
874 # Check that 'foo' and 'bar' were written at the same level
875 self.assertRegex(c['manifest_text'], r'^\. .*:15:bar .*:11:foo\n')
877 def test_put_collection_with_high_redundancy(self):
878 # Write empty data: we're not testing CollectionWriter, just
879 # making sure collections.create tells the API server what our
880 # desired replication level is.
881 collection = self.run_and_find_collection("", ['--replication', '4'])
882 self.assertEqual(4, collection['replication_desired'])
884 def test_put_collection_with_default_redundancy(self):
885 collection = self.run_and_find_collection("")
886 self.assertEqual(None, collection['replication_desired'])
888 def test_put_collection_with_unnamed_project_link(self):
889 link = self.run_and_find_collection(
890 "Test unnamed collection",
891 ['--portable-data-hash', '--project-uuid', self.PROJECT_UUID])
892 username = pwd.getpwuid(os.getuid()).pw_name
895 r'^Saved at .* by {}@'.format(re.escape(username)))
897 def test_put_collection_with_name_and_no_project(self):
898 link_name = 'Test Collection Link in home project'
899 collection = self.run_and_find_collection(
900 "Test named collection in home project",
901 ['--portable-data-hash', '--name', link_name])
902 self.assertEqual(link_name, collection['name'])
903 my_user_uuid = self.current_user()['uuid']
904 self.assertEqual(my_user_uuid, collection['owner_uuid'])
906 def test_put_collection_with_named_project_link(self):
907 link_name = 'Test auto Collection Link'
908 collection = self.run_and_find_collection("Test named collection",
909 ['--portable-data-hash',
911 '--project-uuid', self.PROJECT_UUID])
912 self.assertEqual(link_name, collection['name'])
915 if __name__ == '__main__':