repozo: integrate a quick option for the incremental recover

The default behavior becomes to check the integrality of the existing
recovered Data.fs before incrementally restore on it. The quick option
allows to verify only the latest chunck previously recovered on the
Data.fs before restoring incrementally. This saves many reads.

Author: Sebatyne

src/ZODB/scripts/repozo.py

@@ -79,6 +79,13 @@
         ZODB file. A full recover will always be done if a pack has occured
         since the last incremental backup.
 
+    -Q / --quick
+        Verify via md5 checksum only the last incremental recovered of the
+        output file.  This reduces the disk i/o at the (theoretical) cost of
+        inconsistency.  This is a probabilistic way of determining whether a
+        full recover is necessary.  This argument is ignored when -F / --full
+        is used.
+
     -w
     --with-verify
         Verify on the fly the backup files on recovering. This option runs
@@ -267,6 +274,9 @@ class Options:
         if options.killold:
             log('--kill-old-on-full option is ignored in recover mode')
             options.killold = False
+        if options.full and options.quick:
+            log('--quick option is ignored if --full option is used')
+            options.quick = None
     else:
         assert options.mode == VERIFY
         if options.date is not None:
@@ -752,42 +762,49 @@ def do_incremental_recover(options, repofiles):
     with open(options.output, 'r+b') as outfp:
         outfp.seek(0, 2)
         initial_length = outfp.tell()
-    with open(datfile) as fp:
-        previous_chunk = None
+
+    error = ''
+    previous_chunk = None
+    with open(datfile) as fp, open(options.output, 'r+b') as outfp:
         for line in fp:
-            fn, startpos, endpos, _ = chunk = line.split()
+            fn, startpos, endpos, check_sum = chunk = line.split()
             startpos = int(startpos)
             endpos = int(endpos)
             if endpos > initial_length:
                 break
+            if not options.quick:
+                if check_sum != checksum(outfp, endpos - startpos):
+                    error = ('Target file is not consistent with backup %s, '
+                             'falling back to a full recover.') % fn
+                    break
             previous_chunk = chunk
-
-    if previous_chunk is None:
+    if error:
+        log(error)
+        return do_full_recover(options, repofiles)
+    elif previous_chunk is None:
         log('Target file smaller than full backup, '
             'falling back to a full recover.')
         return do_full_recover(options, repofiles)
-    if endpos < initial_length:
+    elif endpos < initial_length:
         log('Target file is larger than latest backup, '
             'falling back to a full recover.')
         return do_full_recover(options, repofiles)
-    check_startpos = int(previous_chunk[1])
-    check_endpos = int(previous_chunk[2])
-    with open(options.output, 'r+b') as outfp:
-        outfp.seek(check_startpos)
-        check_sum = checksum(outfp, check_endpos - check_startpos)
-    if endpos == initial_length and chunk[3] == check_sum:
+    if options.quick:
+        check_startpos = int(previous_chunk[1])
+        check_endpos = int(previous_chunk[2])
+        with open(options.output, 'r+b') as outfp:
+            outfp.seek(check_startpos)
+            if previous_chunk[3] != checksum(
+                    outfp, check_endpos - check_startpos):
+                error = ('Target file is not consistent with backup %s, '
+                         'falling back to a full recover.' % previous_chunk[0])
+        if error:
+            log(error)
+            return do_full_recover(options, repofiles)
+    if endpos == initial_length:
         log('Target file is same size as latest backup, '
             'doing nothing.')
         return
-    elif previous_chunk[3] != check_sum:
-        if endpos == initial_length:
-            log('Target file is not consistent with latest backup, '
-                'falling back to a full recover.')
-            return do_full_recover(options, repofiles)
-        else:
-            log('Last whole common chunk checksum did not match with backup, '
-                'falling back to a full recover.')
-            return do_full_recover(options, repofiles)
 
     filename = os.path.join(options.repository,
                             os.path.basename(fn))

src/ZODB/scripts/tests/test_repozo.py

@@ -220,13 +220,16 @@ def test_recover_ignored_args(self):
         from ZODB.scripts import repozo
         options = repozo.parseargs(['-R', '-r', '/tmp/nosuchdir', '-v',
                                     '-f', '/tmp/ignored.fs',
-                                    '-k'])
+                                    '-k', '--full', '--quick'])
         self.assertEqual(options.file, None)
         self.assertIn('--file option is ignored in recover mode',
                       sys.stderr.getvalue())
         self.assertEqual(options.killold, False)
         self.assertIn('--kill-old-on-full option is ignored in recover mode',
                       sys.stderr.getvalue())
+        self.assertEqual(options.quick, None)
+        self.assertIn('--quick option is ignored if --full option is used',
+                      sys.stderr.getvalue())
 
     def test_verify_ignored_args(self):
         from ZODB.scripts import repozo
@@ -1040,6 +1043,7 @@ def tearDown(self):
     def _makeOptions(self, **kw):
         options = super()._makeOptions(**kw)
         options.full = False
+        options.quick = kw.get('quick', False)
         return options
 
     def _createRecoveredDataFS(self, output, options):
@@ -1092,6 +1096,33 @@ def test_w_incr_recover_from_incr_backup(self):
         self.assertEqual(_read_file(output), b'AAABBBCCCDDD')
         self.assertFalse(os.path.exists(output + '.part'))
 
+    def test_w_quick_incr_recover_from_incr_backup(self):
+        import tempfile
+        dd = self._data_directory = tempfile.mkdtemp(prefix='zodb-test-')
+        output = os.path.join(dd, 'Data.fs')
+        options = self._makeOptions(date='2010-05-15-13-30-57',
+                                    output=output,
+                                    quick=True,
+                                    withverify=False)
+        self._createRecoveredDataFS(output, options)
+        # Create 2 more .deltafs, to prove the code knows where to pick up
+        self._makeFile(6, 7, 8, '.deltafs', 'CCC')
+        self._makeFile(8, 9, 10, '.deltafs', 'DDD')
+        self._makeFile(
+            2, 3, 4, '.dat',
+            '/backup/2010-05-14-02-03-04.fs 0 3 e1faffb3e614e6c2fba74296962386b7\n'  # noqa: E501 line too long
+            '/backup/2010-05-14-04-05-06.deltafs 3 6 2bb225f0ba9a58930757a868ed57d9a3\n'  # noqa: E501 line too long
+            '/backup/2010-05-14-06-07-08.deltafs 6 9 defb99e69a9f1f6e06f15006b1f166ae\n'  # noqa: E501 line too long
+            '/backup/2010-05-14-08-09-10.deltafs 9 12 45054f47ac3305a2a33e9bcceadff712\n')  # noqa: E501 line too long
+        os.unlink(
+            os.path.join(self._repository_directory,
+                         '2010-05-14-04-05-06.deltafs'))
+        self._callFUT(options)
+        self.assertNotIn('falling back to a full recover.',
+                         sys.stderr.getvalue())
+        self.assertEqual(_read_file(output), b'AAABBBCCCDDD')
+        self.assertFalse(os.path.exists(output + '.part'))
+
     def test_w_incr_backup_with_verify_sum_inconsistent(self):
         import tempfile
         dd = self._data_directory = tempfile.mkdtemp(prefix='zodb-test-')
@@ -1192,14 +1223,14 @@ def test_w_incr_backup_switch_auto_to_full_recover_if_chunk_is_wrong(self):  # n
         self._makeFile(6, 7, 8, '.deltafs', 'CCC')
         self._makeFile(
             2, 3, 4, '.dat',
-            '/backup/2010-05-14-02-03-04.fs 0 3 e1faffb3e614e6c2fba74296962386b7\n'  # noqa: E501 line too long
-            '/backup/2010-05-14-04-05-06.deltafs 3 6 2bb225f0ba9a58930757a868ed57d9a4\n'  # noqa: E501 line too long
+            '/backup/2010-05-14-02-03-04.fs 0 3 e1faffb3e614e6c2fba74296962386b8\n'  # noqa: E501 line too long
+            '/backup/2010-05-14-04-05-06.deltafs 3 6 2bb225f0ba9a58930757a868ed57d9a3\n'  # noqa: E501 line too long
             '/backup/2010-05-14-06-07-08.deltafs 6 9 defb99e69a9f1f6e06f15006b1f166ae\n')  # noqa: E501 line too long
         self._callFUT(options)
         self.assertEqual(_read_file(output), b'AAABBBCCC')
         self.assertFalse(os.path.exists(output + '.part'))
         self.assertIn(
-            "Last whole common chunk checksum did not match with backup, falling back to a full recover.",  # noqa: E501 line too long
+            "Target file is not consistent with backup /backup/2010-05-14-02-03-04.fs, falling back to a full recover.",  # noqa: E501 line too long
             sys.stderr.getvalue())
 
     def test_w_incr_backup_switch_auto_to_full_recover_after_pack(self):
@@ -1226,9 +1257,55 @@ def test_w_incr_backup_switch_auto_to_full_recover_after_pack(self):
         self.assertEqual(_read_file(output), b'CCDD')
         self.assertFalse(os.path.exists(output + '.part'))
         self.assertIn(
-            'Target file is larger than latest backup, falling back to a full recover.',  # noqa: E501 line too long
+            "Target file is not consistent with backup /backup/2010-05-14-06-07-08.fs, falling back to a full recover.",  # noqa: E501 line too long
             sys.stderr.getvalue())
 
+    def test_w_quick_incr_backup_switch_auto_to_full_recover_if_last_chunk_is_wrong(self):  # noqa: E501 line too long
+        import tempfile
+        dd = self._data_directory = tempfile.mkdtemp(prefix='zodb-test-')
+        output = os.path.join(dd, 'Data.fs')
+        options = self._makeOptions(date='2010-05-15-13-30-57',
+                                    output=output,
+                                    quick=True,
+                                    withverify=False)
+        self._createRecoveredDataFS(output, options)
+        self._makeFile(6, 7, 8, '.deltafs', 'CCC')
+        self._makeFile(
+            2, 3, 4, '.dat',
+            '/backup/2010-05-14-02-03-04.fs 0 3 e1faffb3e614e6c2fba74296962386b7\n'  # noqa: E501 line too long
+            '/backup/2010-05-14-04-05-06.deltafs 3 6 2bb225f0ba9a58930757a868ed57d9a4\n'  # noqa: E501 line too long
+            '/backup/2010-05-14-06-07-08.deltafs 6 9 defb99e69a9f1f6e06f15006b1f166ae\n')  # noqa: E501 line too long
+        self._callFUT(options)
+        self.assertEqual(_read_file(output), b'AAABBBCCC')
+        self.assertFalse(os.path.exists(output + '.part'))
+        self.assertIn(
+            "Target file is not consistent with backup /backup/2010-05-14-04-05-06.deltafs, falling back to a full recover.",  # noqa: E501 line too long
+            sys.stderr.getvalue())
+
+    def test_w_quick_incr_backup_dont_see_old_inconsistencies(self):
+        import tempfile
+        dd = self._data_directory = tempfile.mkdtemp(prefix='zodb-test-')
+        output = os.path.join(dd, 'Data.fs')
+        options = self._makeOptions(date='2010-05-15-13-30-57',
+                                    output=output,
+                                    quick=True,
+                                    withverify=False)
+        self._createRecoveredDataFS(output, options)
+        self._makeFile(6, 7, 8, '.deltafs', 'CCC')
+        self._makeFile(
+            2, 3, 4, '.dat',
+            '/backup/2010-05-14-02-03-04.fs 0 3 e1faffb3e614e6c2fba74296962386b7\n'  # noqa: E501 line too long
+            '/backup/2010-05-14-04-05-06.deltafs 3 6 2bb225f0ba9a58930757a868ed57d9a3\n'  # noqa: E501 line too long
+            '/backup/2010-05-14-06-07-08.deltafs 6 9 defb99e69a9f1f6e06f15006b1f166ae\n')  # noqa: E501 line too long
+        # The ZODB is longer than announced in the .dat file
+        with open(output, 'r+b') as f:
+            f.write(b'ZZZBBBCCC')
+        self._callFUT(options)
+        self.assertEqual(_read_file(output), b'ZZZBBBCCC')
+        self.assertFalse(os.path.exists(output + '.part'))
+        self.assertNotIn(
+            "falling back to a full recover", sys.stderr.getvalue())
+
 
 class Test_do_verify(OptionsTestBase, unittest.TestCase):