openstack
/
glance
Code Issues Proposed changes

Handling scrubber's exit in non-daemon mode. 

Currently, Scrubber exits with a status code of zero
upon job fetching errors both in daemon and non daemon
mode. It is more appropriate to have scrubber exit with
status code 1 if it is not in daemon mode.

This patch causes scrubber to exit with status code 1
(RuntimeError)if the scrubber is not running in daemon
mode, facilitating accurate exits.

Co-Authored-By: Steve Lewis <steve.lewis@rackspace.com>
Co-Authored-By: Dharini Chandrasekar <dharini.chandrasekar@intel.com>

Closes-Bug: #1548289

Change-Id: Ib75676dc90349f008395fb118978f4e37fa876ea
This commit is contained in:
Dharini Chandrasekar 2016-10-06 03:50:42 +00:00
parent e842cffc12
commit 91434be160
8 changed files with 81 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
glance/common/exception.py
View File

@ -292,6 +292,11 @@ class ImageSizeLimitExceeded(GlanceException):
message = _("The provided image is too large.")
class FailedToGetScrubberJobs(GlanceException):
message = _("Scrubber encountered an error while trying to fetch "
"scrub jobs.")
class ImageMemberLimitExceeded(LimitExceeded):
message = _("The limit has been exceeded on the number of allowed image "
"members for this image. Attempted: %(attempted)s, "

11
glance/scrubber.py
View File

@ -27,7 +27,7 @@ from glance.common import crypt
from glance.common import exception
from glance import context
import glance.db as db_api
from glance.i18n import _, _LE, _LI, _LW
from glance.i18n import _, _LC, _LE, _LI, _LW
import glance.registry.client.v1.api as registry
LOG = logging.getLogger(__name__)
@ -409,9 +409,12 @@ class Scrubber(object):
try:
records = self.db_queue.get_all_locations()
except Exception as err:
LOG.error(_LE("Can not get scrub jobs from queue: %s") %
encodeutils.exception_to_unicode(err))
return {}
# Note(dharinic): spawn_n, in Daemon mode will log the
# exception raised. Otherwise, exit 1 will occur.
msg = (_LC("Can not get scrub jobs from queue: %s") %
encodeutils.exception_to_unicode(err))
LOG.critical(msg)
raise exception.FailedToGetScrubberJobs()
delete_jobs = {}
for image_id, loc_id, loc_uri in records:

17
glance/tests/functional/__init__.py
View File

@ -603,6 +603,10 @@ class FunctionalTest(test_utils.BaseTestCase):
self.api_protocol = 'http'
self.api_port, api_sock = test_utils.get_unused_port_and_socket()
self.registry_port, reg_sock = test_utils.get_unused_port_and_socket()
# NOTE: Scrubber is enabled by default for the functional tests.
# Please disbale it by explicitly setting 'self.include_scrubber' to
# False in the test SetUps that do not require Scrubber to run.
self.include_scrubber = True
self.tracecmd = tracecmd_osmap.get(platform.system())
@ -799,11 +803,11 @@ class FunctionalTest(test_utils.BaseTestCase):
self.start_with_retry(self.api_server, 'api_port', 3, **kwargs)
exitcode, out, err = self.scrubber_daemon.start(**kwargs)
self.assertEqual(0, exitcode,
"Failed to spin up the Scrubber daemon. "
"Got: %s" % err)
if self.include_scrubber:
exitcode, out, err = self.scrubber_daemon.start(**kwargs)
self.assertEqual(0, exitcode,
"Failed to spin up the Scrubber daemon. "
"Got: %s" % err)
def ping_server(self, port):
"""
@ -919,7 +923,8 @@ class FunctionalTest(test_utils.BaseTestCase):
# Spin down the API and default registry server
self.stop_server(self.api_server, 'API server')
self.stop_server(self.registry_server, 'Registry server')
self.stop_server(self.scrubber_daemon, 'Scrubber daemon')
if self.include_scrubber:
self.stop_server(self.scrubber_daemon, 'Scrubber daemon')
self._reset_database(self.registry_server.sql_connection)

27
glance/tests/functional/test_scrubber.py
View File

@ -274,6 +274,33 @@ class TestScrubber(functional.FunctionalTest):
self.stop_servers()
def test_scrubber_app_queue_errors_not_daemon(self):
"""
test that the glance-scrubber exits with an exit code > 0 when it
fails to lookup images, indicating a configuration error when not
in daemon mode.
Related-Bug: #1548289
"""
# Don't start the registry server to cause intended failure
# Don't start the api server to save time
exitcode, out, err = self.scrubber_daemon.start(
delayed_delete=True, daemon=False, registry_port=28890)
self.assertEqual(0, exitcode,
"Failed to spin up the Scrubber daemon. "
"Got: %s" % err)
# Run the Scrubber
exe_cmd = "%s -m glance.cmd.scrubber" % sys.executable
cmd = ("%s --config-file %s" %
(exe_cmd, self.scrubber_daemon.conf_file_name))
exitcode, out, err = execute(cmd, raise_error=False)
self.assertEqual(1, exitcode)
self.assertIn('Can not get scrub jobs from queue', err)
self.stop_server(self.scrubber_daemon, 'Scrubber daemon')
def wait_for_scrub(self, path, headers=None):
"""
NOTE(jkoelker) The build servers sometimes take longer than 15 seconds

6
glance/tests/functional/v2/test_images.py
View File

@ -40,6 +40,7 @@ class TestImages(functional.FunctionalTest):
def setUp(self):
super(TestImages, self).setUp()
self.cleanup()
self.include_scrubber = False
self.api_server.deployment_flavor = 'noauth'
self.api_server.data_api = 'glance.db.sqlalchemy.api'
for i in range(3):
@ -3028,6 +3029,7 @@ class TestImagesIPv6(functional.FunctionalTest):
# Setting up monkey patch (2), after object is ready...
self.ping_server_ipv4 = self.ping_server
self.ping_server = self.ping_server_ipv6
self.include_scrubber = False
def tearDown(self):
# Cleaning up monkey patch (2).
@ -3083,6 +3085,7 @@ class TestImageDirectURLVisibility(functional.FunctionalTest):
def setUp(self):
super(TestImageDirectURLVisibility, self).setUp()
self.cleanup()
self.include_scrubber = False
self.api_server.deployment_flavor = 'noauth'
def _url(self, path):
@ -3292,6 +3295,7 @@ class TestImageLocationSelectionStrategy(functional.FunctionalTest):
def setUp(self):
super(TestImageLocationSelectionStrategy, self).setUp()
self.cleanup()
self.include_scrubber = False
self.api_server.deployment_flavor = 'noauth'
for i in range(3):
ret = test_utils.start_http_server("foo_image_id%d" % i,
@ -3395,6 +3399,7 @@ class TestImageMembers(functional.FunctionalTest):
def setUp(self):
super(TestImageMembers, self).setUp()
self.cleanup()
self.include_scrubber = False
self.api_server.deployment_flavor = 'fakeauth'
self.registry_server.deployment_flavor = 'fakeauth'
self.start_servers(**self.__dict__.copy())
@ -3761,6 +3766,7 @@ class TestQuotas(functional.FunctionalTest):
def setUp(self):
super(TestQuotas, self).setUp()
self.cleanup()
self.include_scrubber = False
self.api_server.deployment_flavor = 'noauth'
self.registry_server.deployment_flavor = 'trusted-auth'
self.user_storage_quota = 100

1
glance/tests/functional/v2/test_tasks.py
View File

@ -143,3 +143,4 @@ class TestTasksWithRegistry(TestTasks):
self.api_server.data_api = (
'glance.tests.functional.v2.registry_data_api')
self.registry_server.deployment_flavor = 'trusted-auth'
self.include_scrubber = False

12
glance/tests/unit/test_scrubber.py
View File

@ -16,12 +16,14 @@
import uuid
import glance_store
import mock
from mock import patch
from mox3 import mox
from oslo_config import cfg
# NOTE(jokke): simplified transition to py3, behaves like py2 xrange
from six.moves import range
from glance.common import exception
from glance import scrubber
from glance.tests import utils as test_utils
@ -110,6 +112,16 @@ class TestScrubber(test_utils.BaseTestCase):
scrub._scrub_image(id, [(id, '-', uri)])
self.mox.VerifyAll()
def test_scrubber_exits(self):
# Checks for Scrubber exits when it is not able to fetch jobs from
# the queue
scrub_jobs = scrubber.ScrubDBQueue.get_all_locations
scrub_jobs = mock.MagicMock()
scrub_jobs.side_effect = exception.NotFound
scrub = scrubber.Scrubber(glance_store)
self.assertRaises(exception.FailedToGetScrubberJobs,
scrub._get_delete_jobs)
class TestScrubDBQueue(test_utils.BaseTestCase):

12
releasenotes/notes/scrubber-exit-e5d77f6f1a38ffb7.yaml Normal file
View File

@ -0,0 +1,12 @@
---
fixes:
- |
Please note a change in the Scrubber's behavior in case
of job fetching errors:
* If configured to work in daemon mode, the Scrubber
will log an error message at level critical, but
will not exit the process.
* If configured to work in non-daemon mode, the Scrubber
will log an error message at level critical and exit
with status one.