Container sync no longer sending swift_bytes value

Container sync had a bug where it'd send out the trailing
"; swift_bytes=xxx" part of the content-type header. That trailing part
is just for internal cluster usage by SLO. Since that needed to be
stripped in two places now, I separated it out to a function that both
spots call.

Change-Id: Ibd6035d7a6b78205344bcc9d98bc1b7a9d463427

swift/common/utils.py

@@ -2658,6 +2658,14 @@ def override_bytes_from_content_type(listing_dict, logger=None):
     listing_dict['content_type'] = content_type
 
 
+def clean_content_type(value):
+    if ';' in value:
+        left, right = value.rsplit(';', 1)
+        if right.lstrip().startswith('swift_bytes='):
+            return left
+    return value
+
+
 def quote(value, safe='/'):
     """
     Patched version of urllib.quote that encodes utf-8 strings before quoting

swift/container/sync.py

@@ -29,9 +29,10 @@ from swift.common.direct_client import direct_get_object
 from swift.common.internal_client import delete_object, put_object
 from swift.common.exceptions import ClientException
 from swift.common.ring import Ring
-from swift.common.utils import audit_location_generator, get_logger, \
-    hash_path, config_true_value, validate_sync_to, whataremyips, \
-    FileLikeIter, urlparse, quote
+from swift.common.utils import (
+    audit_location_generator, clean_content_type, config_true_value,
+    FileLikeIter, get_logger, hash_path, quote, urlparse, validate_sync_to,
+    whataremyips)
 from swift.common.daemon import Daemon
 from swift.common.http import HTTP_UNAUTHORIZED, HTTP_NOT_FOUND
 
@@ -403,6 +404,9 @@ class ContainerSync(Daemon):
                         del headers[key]
                 if 'etag' in headers:
                     headers['etag'] = headers['etag'].strip('"')
+                if 'content-type' in headers:
+                    headers['content-type'] = clean_content_type(
+                        headers['content-type'])
                 headers['x-timestamp'] = row['created_at']
                 if realm and realm_key:
                     nonce = uuid.uuid4().hex

swift/proxy/controllers/obj.py

@@ -35,9 +35,10 @@ from eventlet import GreenPile
 from eventlet.queue import Queue
 from eventlet.timeout import Timeout
 
-from swift.common.utils import ContextPool, normalize_timestamp, \
-    config_true_value, public, json, csv_append, GreenthreadSafeIterator, \
-    quorum_size, GreenAsyncPile, normalize_delete_at_timestamp
+from swift.common.utils import (
+    clean_content_type, config_true_value, ContextPool, csv_append,
+    GreenAsyncPile, GreenthreadSafeIterator, json,
+    normalize_delete_at_timestamp, normalize_timestamp, public, quorum_size)
 from swift.common.bufferedhttp import http_connect
 from swift.common.constraints import check_metadata, check_object_creation, \
     check_copy_from_header
@@ -206,11 +207,8 @@ class ObjectController(Controller):
             req.swift_entity_path)
 
         if ';' in resp.headers.get('content-type', ''):
-            # strip off swift_bytes from content-type
-            content_type, check_extra_meta = \
-                resp.headers['content-type'].rsplit(';', 1)
-            if check_extra_meta.lstrip().startswith('swift_bytes='):
-                resp.content_type = content_type
+            resp.content_type = clean_content_type(
+                resp.headers['content-type'])
         return resp
 
     @public

test/unit/common/test_utils.py

@@ -1879,6 +1879,22 @@ cluster_dfw1 = http://dfw1.host/v1/
         self.assertEquals(listing_dict['content_type'],
                           'text/plain;hello="world"')
 
+    def test_clean_content_type(self):
+        subtests = {
+            '': '', 'text/plain': 'text/plain',
+            'text/plain; someother=thing': 'text/plain; someother=thing',
+            'text/plain; swift_bytes=123': 'text/plain',
+            'text/plain; someother=thing; swift_bytes=123':
+                'text/plain; someother=thing',
+            # Since Swift always tacks on the swift_bytes, clean_content_type()
+            # only strips swift_bytes if it's last. The next item simply shows
+            # that if for some other odd reason it's not last,
+            # clean_content_type() will not remove it from the header.
+            'text/plain; swift_bytes=123; someother=thing':
+                'text/plain; swift_bytes=123; someother=thing'}
+        for before, after in subtests.items():
+            self.assertEqual(utils.clean_content_type(before), after)
+
     def test_quote(self):
         res = utils.quote('/v1/a/c3/subdirx/')
         assert res == '/v1/a/c3/subdirx/'

test/unit/container/test_sync.py

@@ -751,13 +751,15 @@ class TestContainerSync(unittest.TestCase):
                         'US abcdef ef62c64bb88a33fa00722daa23d5d43253164962',
                         'x-timestamp': '1.2',
                         'etag': 'etagvalue',
-                        'other-header': 'other header value'})
+                        'other-header': 'other header value',
+                        'content-type': 'text/plain'})
                 else:
                     self.assertEquals(headers, {
                         'x-container-sync-key': 'key',
                         'x-timestamp': '1.2',
                         'other-header': 'other header value',
-                        'etag': 'etagvalue'})
+                        'etag': 'etagvalue',
+                        'content-type': 'text/plain'})
                 self.assertEquals(contents.read(), 'contents')
                 self.assertEquals(proxy, 'http://proxy')
 
@@ -770,7 +772,8 @@ class TestContainerSync(unittest.TestCase):
             def fake_direct_get_object(node, part, account, container, obj,
                                        resp_chunk_size=1):
                 return ({'other-header': 'other header value',
-                         'etag': '"etagvalue"', 'x-timestamp': '1.2'},
+                         'etag': '"etagvalue"', 'x-timestamp': '1.2',
+                         'content-type': 'text/plain; swift_bytes=123'},
                         iter('contents'))
 
             sync.direct_get_object = fake_direct_get_object
@@ -790,7 +793,8 @@ class TestContainerSync(unittest.TestCase):
                          'last-modified': 'last modified value',
                          'x-timestamp': '1.2',
                          'other-header': 'other header value',
-                         'etag': '"etagvalue"'},
+                         'etag': '"etagvalue"',
+                         'content-type': 'text/plain; swift_bytes=123'},
                         iter('contents'))
 
             sync.direct_get_object = fake_direct_get_object