14012: Rephrases error message.
[arvados.git] / sdk / python / tests / test_arv_put.py
index e11c30959a05cb2f473eb93239d6b8a6d0b8ca43..642f64ec5a54fd23d77ea37eb8901d056509c7d1 100644 (file)
@@ -18,11 +18,12 @@ import os
 import pwd
 import random
 import re
+import select
 import shutil
+import signal
 import subprocess
 import sys
 import tempfile
-import threading
 import time
 import unittest
 import uuid
@@ -548,7 +549,7 @@ class ArvadosExpectedBytesTest(ArvadosBaseTestCase):
                          writer.bytes_expected)
 
     def test_expected_bytes_for_device(self):
-        writer = arv_put.ArvPutUploadJob(['/dev/null'])
+        writer = arv_put.ArvPutUploadJob(['/dev/null'], use_cache=False, resume=False)
         self.assertIsNone(writer.bytes_expected)
         writer = arv_put.ArvPutUploadJob([__file__, '/dev/null'])
         self.assertIsNone(writer.bytes_expected)
@@ -729,6 +730,11 @@ class ArvadosPutTest(run_test_server.TestCaseWithServers,
                           self.call_main_with_args,
                           ['--project-uuid', self.Z_UUID, '--stream'])
 
+    def test_error_when_multiple_storage_classes_specified(self):
+        self.assertRaises(SystemExit,
+                          self.call_main_with_args,
+                          ['--storage-classes', 'hot,cold'])
+
     def test_error_when_excluding_absolute_path(self):
         tmpdir = self.make_tmpdir()
         self.assertRaises(SystemExit,
@@ -857,6 +863,30 @@ class ArvPutIntegrationTest(run_test_server.TestCaseWithServers,
         self.assertIn('4a9c8b735dce4b5fa3acf221a0b13628+11',
                       pipe.stdout.read().decode())
 
+    def test_sigint_logs_request_id(self):
+        # Start arv-put, give it a chance to start up, send SIGINT,
+        # and check that its output includes the X-Request-Id.
+        input_stream = subprocess.Popen(
+            ['sleep', '10'],
+            stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
+        pipe = subprocess.Popen(
+            [sys.executable, arv_put.__file__, '--stream'],
+            stdin=input_stream.stdout, stdout=subprocess.PIPE,
+            stderr=subprocess.STDOUT, env=self.ENVIRON)
+        # Wait for arv-put child process to print something (i.e., a
+        # log message) so we know its signal handler is installed.
+        select.select([pipe.stdout], [], [], 10)
+        pipe.send_signal(signal.SIGINT)
+        deadline = time.time() + 5
+        while (pipe.poll() is None) and (time.time() < deadline):
+            time.sleep(.1)
+        returncode = pipe.poll()
+        input_stream.terminate()
+        if returncode is None:
+            pipe.terminate()
+            self.fail("arv-put did not exit within 5 seconds")
+        self.assertRegex(pipe.stdout.read().decode(), r'\(X-Request-Id: req-[a-z0-9]{20}\)')
+
     def test_ArvPutSignedManifest(self):
         # ArvPutSignedManifest runs "arv-put foo" and then attempts to get
         # the newly created manifest from the API server, testing to confirm
@@ -908,7 +938,7 @@ class ArvPutIntegrationTest(run_test_server.TestCaseWithServers,
         self.assertEqual(1, len(collection_list))
         return collection_list[0]
 
-    def test_expired_token_invalidates_cache(self):
+    def test_all_expired_signatures_invalidates_cache(self):
         self.authorize_with('active')
         tmpdir = self.make_tmpdir()
         with open(os.path.join(tmpdir, 'somefile.txt'), 'w') as f:
@@ -944,7 +974,89 @@ class ArvPutIntegrationTest(run_test_server.TestCaseWithServers,
         (out, err) = p.communicate()
         self.assertRegex(
             err.decode(),
-            r'WARNING: Uploaded file .* access token expired, will re-upload it from scratch')
+            r'INFO: Cache expired, starting from scratch.*')
+        self.assertEqual(p.returncode, 0)
+
+    def test_invalid_signature_invalidates_cache(self):
+        self.authorize_with('active')
+        tmpdir = self.make_tmpdir()
+        with open(os.path.join(tmpdir, 'somefile.txt'), 'w') as f:
+            f.write('foo')
+        # Upload a directory and get the cache file name
+        p = subprocess.Popen([sys.executable, arv_put.__file__, tmpdir],
+                             stdout=subprocess.PIPE,
+                             stderr=subprocess.PIPE,
+                             env=self.ENVIRON)
+        (out, err) = p.communicate()
+        self.assertRegex(err.decode(), r'INFO: Creating new cache file at ')
+        self.assertEqual(p.returncode, 0)
+        cache_filepath = re.search(r'INFO: Creating new cache file at (.*)',
+                                   err.decode()).groups()[0]
+        self.assertTrue(os.path.isfile(cache_filepath))
+        # Load the cache file contents and modify the manifest to simulate
+        # an invalid access token
+        with open(cache_filepath, 'r') as c:
+            cache = json.load(c)
+        self.assertRegex(cache['manifest'], r'\+A\S+\@')
+        cache['manifest'] = re.sub(
+            r'\+A.*\@',
+            "+Aabcdef0123456789abcdef0123456789abcdef01@",
+            cache['manifest'])
+        with open(cache_filepath, 'w') as c:
+            c.write(json.dumps(cache))
+        # Re-run the upload and expect to get an invalid cache message
+        p = subprocess.Popen([sys.executable, arv_put.__file__, tmpdir],
+                             stdout=subprocess.PIPE,
+                             stderr=subprocess.PIPE,
+                             env=self.ENVIRON)
+        (out, err) = p.communicate()
+        self.assertRegex(
+            err.decode(),
+            r'ERROR: arv-put: Cache seems to contain invalid data.*')
+        self.assertEqual(p.returncode, 1)
+
+    def test_single_expired_signature_reuploads_file(self):
+        self.authorize_with('active')
+        tmpdir = self.make_tmpdir()
+        with open(os.path.join(tmpdir, 'foofile.txt'), 'w') as f:
+            f.write('foo')
+        # Write a second file on its own subdir to force a new stream
+        os.mkdir(os.path.join(tmpdir, 'bar'))
+        with open(os.path.join(tmpdir, 'bar', 'barfile.txt'), 'w') as f:
+            f.write('bar')
+        # Upload a directory and get the cache file name
+        p = subprocess.Popen([sys.executable, arv_put.__file__, tmpdir],
+                             stdout=subprocess.PIPE,
+                             stderr=subprocess.PIPE,
+                             env=self.ENVIRON)
+        (out, err) = p.communicate()
+        self.assertRegex(err.decode(), r'INFO: Creating new cache file at ')
+        self.assertEqual(p.returncode, 0)
+        cache_filepath = re.search(r'INFO: Creating new cache file at (.*)',
+                                   err.decode()).groups()[0]
+        self.assertTrue(os.path.isfile(cache_filepath))
+        # Load the cache file contents and modify the manifest to simulate
+        # an expired access token
+        with open(cache_filepath, 'r') as c:
+            cache = json.load(c)
+        self.assertRegex(cache['manifest'], r'\+A\S+\@')
+        a_month_ago = datetime.datetime.now() - datetime.timedelta(days=30)
+        # Make one of the signatures appear to have expired
+        cache['manifest'] = re.sub(
+            r'\@.*? 3:3:barfile.txt',
+            "@{} 3:3:barfile.txt".format(self.datetime_to_hex(a_month_ago)),
+            cache['manifest'])
+        with open(cache_filepath, 'w') as c:
+            c.write(json.dumps(cache))
+        # Re-run the upload and expect to get an invalid cache message
+        p = subprocess.Popen([sys.executable, arv_put.__file__, tmpdir],
+                             stdout=subprocess.PIPE,
+                             stderr=subprocess.PIPE,
+                             env=self.ENVIRON)
+        (out, err) = p.communicate()
+        self.assertRegex(
+            err.decode(),
+            r'WARNING: Uploaded file \'.*barfile.txt\' access token expired, will re-upload it from scratch')
         self.assertEqual(p.returncode, 0)
         # Confirm that the resulting cache is different from the last run.
         with open(cache_filepath, 'r') as c2:
@@ -1036,6 +1148,18 @@ class ArvPutIntegrationTest(run_test_server.TestCaseWithServers,
                                        '--project-uuid', self.PROJECT_UUID])
         self.assertEqual(link_name, collection['name'])
 
+    def test_put_collection_with_storage_classes_specified(self):
+        collection = self.run_and_find_collection("", ['--storage-classes', 'hot'])
+
+        self.assertEqual(len(collection['storage_classes_desired']), 1)
+        self.assertEqual(collection['storage_classes_desired'][0], 'hot')
+
+    def test_put_collection_without_storage_classes_specified(self):
+        collection = self.run_and_find_collection("")
+
+        self.assertEqual(len(collection['storage_classes_desired']), 1)
+        self.assertEqual(collection['storage_classes_desired'][0], 'default')
+
     def test_exclude_filename_pattern(self):
         tmpdir = self.make_tmpdir()
         tmpsubdir = os.path.join(tmpdir, 'subdir')