Public bug reported:
Hi,
>From neutron-dhcp-agent.log, I can see it's trying to access:
/var/lib/neutron/external/pids/*.pid.haproxy
It used to be that these files where having the unix rights (at least in
Debian 11, aka Bullseye):
-rw-r--r--
However, in Debian 12 (aka Bookworm), for a
Public bug reported:
Running unit tests of Horizon in Debian Unstable leads to many unit test
failures like this one below.
My instinct tells me that this is Python 3.12 related, but I'm not sure.
Here's a pip freeze output with the installed build-dependency of the
package:
alabaster==0.7.12
-provider
Please fix the unit tests, and let me know where to find the patch (zigo
on IRC).
Cheers,
Thomas Goirand (zigo)
** Affects: neutron
Importance: Undecided
Status: New
--
You received this bug notification because you are a member of Yahoo!
Engineering Team, which
Public bug reported:
Hi,
In Debian, we're trying to switch to 3.11 before the Bookworm freeze in
January.
Rebuilding Neutron under Python 3.11 fails:
FAIL:
neutron.tests.unit.plugins.ml2.drivers.ovn.mech_driver.test_mech_driver.TestOVNMechanismDriverPortsV2.test__port_provisioned_no_binding
Public bug reported:
Looks like a bunch of OVN unit tests are highly depending on the test
order.
When rebuilding the Debian Zed package of Neutron under Debian Unstable,
I get 200+ unit test failures like what's below. Using tox, if running:
tox -e py3 --
Public bug reported:
Without the latest os-traits, we get unit test failures like below.
==
FAIL:
nova.tests.unit.compute.test_pci_placement_translator.TestTranslator.test_trait_normalization_09
a
maintainers: for the Debian package, I'll just skip these 2 unit tests.
Cheers,
Thomas Goirand (zigo)
** Affects: nova
Importance: Undecided
Status: New
--
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute
Public bug reported:
Looks like nova/db/api is missing an __init__.py, which breaks *at
least* my Debian packaging.
** Affects: nova
Importance: Undecided
Status: In Progress
--
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is
Public bug reported:
Hi,
As per this Debian bug entry:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=976954
a number of unit tests are failing under ppc64el arch. Please fix these
or exclude the tests on this arch.
** Affects: nova
Importance: Undecided
Status: New
--
You
Public bug reported:
Hi,
Running test_get_disk_mapping_stable_rescue_ide_cdrom on arm64 or
ppcel64 results in failure:
> ==
> FAIL:
>
ugh, you can also try directly in Debian Unstable
(that's the same packages which I upload there, and maintain as
backports for Debian Stable).
Cheers,
Thomas Goirand
** Affects: glance
Importance: Undecided
Status: New
** Summary changed:
- openstack image-download / glance im
Public bug reported:
When I add a speaker to a dragent, the table
bgp_speaker_dragent_bindings gets 2 entries instead of one. One is the
one I am expecting, and the 2nd one is another dragent which has nothing
to do with the one I'm trying to configure.
When there's only a single agent
Public bug reported:
When doing something like this:
su nova -s /bin/sh -c "nova-manage cell_v2 discover_hosts"
I see lots of deprecation warnings. There should be a way to disable the
warnings, or having them off by default. Discussion should be open on
how to fix this.
** Affects: nova
Public bug reported:
Looks like the generated policy.json is broken by default and can't be
used by operators as-is, as it doesn't include the deprecated options
which are unfortunately needed for it to work.
With the default policy.json as generated by the nova namespace, the
admin user can't
Public bug reported:
It looks like Glance is broken regarding using a custom CA certificate
for the Swift API. Indeed, even if I set a CA file in
swift_store_cacert, Glance wouldn't upload an image to Swift, unless I
set swift_store_auth_insecure to True.
** Affects: glance
Importance:
Public bug reported:
Rebuilding Glance 19.0.0 in Debian Sid gives the below result:
FAIL:
glance.tests.unit.v2.test_tasks_resource.TestTasksController.test_create_with_live_time
glance.tests.unit.v2.test_tasks_resource.TestTasksController.test_create_with_live_time
on
compute-1 has no effect, which means that there's a problem there.
Please help us fixing the bug.
Cheers,
Thomas Goirand (zigo)
** Affects: neutron
Importance: Undecided
Status: New
--
You received this bug notification because you are a member of Yahoo!
Engineering Team, which
Public bug reported:
Hi,
Previously on our deployment, we had DHCP agents on the compute nodes.
We decided to add network nodes, and later, move the DHCP agents from
the compute nodes to the network nodes.
When removing the network from the DHCP agent, it looks like the L3
rules were also
, it probably looks more like a bug in oslo.db or in SQLAlchemy to
me, with tables not being created before the unit tests start or
something.
Any clue what's going on here?
Cheers,
Thomas Goirand (zigo)
** Affects: neutron
Importance: Undecided
Status: New
--
You received this bug
*** This bug is a duplicate of bug 1838666 ***
https://bugs.launchpad.net/bugs/1838666
Public bug reported:
When rebuilding Nova from Stein in Debian Sid, I get 3 unit test errors,
probably due to a more recent libvirt (ie: 5.6.0). See for example, on
this first one:
we get bus= and dev=
Public bug reported:
Hi,
I'm trying to build Stein Neutron from Buster to Sid.
There's the following 4 unit tests failing when building the debian
package in Sid, all located in neutron.tests.unit.extensions:
Public bug reported:
Running unit tests of neutron-dynamic-routing with SQLAlchemy 1.3.1
leads to these failures. Please fix this before the Buster release, as
SQLAlchemy 1.3.x will be uploaded to Sid instead of Experimental, and
anyway, we can't upload the Stein release to Debian Experimental
Public bug reported:
We've started an instance last Wednesday, and the compute where it ran
failed (maybe hardware issue?). Since the networking looked wrong (ie:
missing network interfaces), I tried to migrate the instance.
According to Matt, it looked like the request_spec entry for the
t;set --disable" process.
Cheers,
Thomas Goirand (zigo)
** Affects: neutron
Importance: Undecided
Status: New
--
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to neutron.
https://bugs.launchpad.net/bugs/1825345
Title:
a
Public bug reported:
Hi,
Trying to evaluate if we can upgrade Buster to SQLAlchemy 1.3.0, doing
this in my PoC:
openstack server add floating ip demo-server 192.168.105.101
leads to this stack dump below. Obviously, there's something wrong that
needs fixing. Best would be before Stein is out.
Public bug reported:
In our setup, we run with DVR, and really a lot of VMs in the same
tenant/project (we have currently between 1500 and 2000 VMs). In such
setup, the internal function _set_subnet_arp_info of
neutron/agent/l3/dvr_local_router.py is taking a way too long. Indeed,
what it does
Public bug reported:
When browsing to Admin -> System -> System information, as an admin
user, if there's no Neutron installed, then Horizon just crashes:
[ ... snip ... ]
File "/usr/lib/python3/dist-packages/horizon/utils/memoized.py" in wrapped
176. args.insert(request_index,
Public bug reported:
Building the Nova 18.0.0 package in Sid (ie: Python 3.7), I get the
below unit test failure with hacking tests:
==
FAIL: nova.tests.unit.test_hacking.HackingTestCase.test_check_doubled_words
Public bug reported:
Building Nova 18.0.0 in Debian Sid (ie: Python 3.7), I get the below
unit test failures.
==
FAIL:
nova.tests.unit.test_api_validation.PatternPropertiesTestCase.test_validate_patternProperties_fails
Public bug reported:
Building Nova 18.0.0 in Debian Sid makes the below unit tests fail. Note
that it is probably related to OpenSSL 1.1.1 (though I didn't
investigate further and disabled running test tests).
==
FAIL:
Public bug reported:
After I just finished packaging Rocky, I wanted to test it with puppet-
openstack. Then I couldn't boot VMs after the puppet run, because
allocation_ration in placement is set to 0.0 by default:
# openstack resource provider list
Public bug reported:
Hi,
Building Horizon in Debian Stretch is no problem, and all unit test
pass. Unfortunately, under Python 3.7 in Debian Sid, it doesn't even
start to run unit tests:
+ http_proxy=127.0.0.1:9 https_proxy=127.0.0.9:9 HTTP_PROXY=127.0.0.1:9
HTTPS_PROXY=127.0.0.1:9
Public bug reported:
While testing puppet-openstack and Neutron with Debian packages (ie:
running with Stretch and Queens), I had to use neutron-api using uwsgi,
as /usr/bin/neutron-server would not work with Eventlet + Python 3
(which is famously broken). Therefore, I did a setup with uwsgi and
Public bug reported:
Hi,
Building the Nova Queens package with OpenSSL 1.1.1 leads to unit test
problems. This was reported to Debian at: https://bugs.debian.org/898807
The new openssl 1.1.1 is currently in experimental [0]. This package
failed to build against this new package [1] while it
Public bug reported:
Because of an error in the packaging of Glance in Debian, I had:
# openstack image create --container-format bare --disk-format qcow2 --file
debian-testing-openstack-amd64.qcow2 debian-buster-amd64
410 Gone: Error in store configuration. Adding images to store is disabled.
Public bug reported:
While building Keystone in Debian Sid + Python 3.6, I get the below
stack dump. Obviously, this is a broken test, not a broken code. Notice
the final . (ie: dot) after "status 1" that is the cause of the test
failure.
Everything else seems to pass in my environment.
Public bug reported:
FAIL:
neutron_lbaas.tests.unit.drivers.haproxy.test_jinja_cfg.TestHaproxyCfg.test_render_template_https
neutron_lbaas.tests.unit.drivers.haproxy.test_jinja_cfg.TestHaproxyCfg.test_render_template_https
--
Public bug reported:
While building the neutron-lbaas package in Debian Sid with Python 3.6,
I get the below failure.
FAIL:
neutron_lbaas.tests.unit.drivers.radware.test_v2_plugin_driver.TestLBaaSDriverRestClient.test_recover_was_called
Public bug reported:
While building the neutron-lbaas package in Debian Sid with Python 3.6,
I get the below failure. As this looks like a broken test rather than a
software bug, I've disabled running the unit test at package build time,
however, it'd be nice to have it fixed.
FAIL:
Public bug reported:
While building the neutron-lbaas package in Debian Sid with Python 3.6,
I get the below failure. As this looks like a broken test rather than a
software bug, I've disabled running the unit test at package build time,
however, it'd be nice to have it fixed.
My wet finger
Public bug reported:
When generating the sphinx doc in Debian Sid with Python 3.6 for the
Queens RC1 release of Nova, I get the below stack dump, though it passes
under Python 2.7. A fix would be more than welcome, cause I'm removing
all traces of Python 2.7, including sphinx stuff and all
Public bug reported:
Hi,
When generating metadata_agent.ini, the metadata_workers directive
default value is filled with the number of CPUs used when building the
file. This makes the whole Neutron package not reproducible.
The config code is like this (from
Public bug reported:
I'm trying to (tempest) validate OpenStack Pike for Debian. So I'm
running Pike (nova 16.0.3) on Debian Sid. My environment for running
tempest is a Debian Live system which I just boot once, install all of
OpenStack on, and run tempest.
As a consequence, my filesystem is a
Public bug reported:
In this commit:
https://github.com/openstack/nova/commit/4647f418afb9ced223c089f9d49cd686eccae9e2
nova starts using the modified_eui64() function which isn't available in
netaddr 0.7.12. It is available in version 0.7.18, which is what the
upper-constraints.txt has. I
Public bug reported:
In distributions, it's not possible to ship twice a daemon in /usr/bin
with 2 different packages. This means that /usr/bin/neutron-l3-agent in
neutron-fwaas is clashing with /usr/bin/neutron-l3-agent from python-
neutron. In Debian, we will rename /usr/bin/neutron-l3-agent as
Public bug reported:
In nova/conf/paths.py, I can read:
path_opts = [
cfg.StrOpt('pybasedir',
default=os.path.abspath(os.path.join(os.path.dirname(__file__),
'../../')),
help='Directory where the nova python
Public bug reported:
Keystone Mitaka b1 fails to build because of a missing testresources in
test-requirements.txt.
** Affects: keystone
Importance: Undecided
Assignee: Thomas Goirand (thomas-goirand)
Status: In Progress
--
You received this bug notification because you
** Changed in: neutron
Status: Expired => Confirmed
--
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to neutron.
https://bugs.launchpad.net/bugs/1566972
Title:
Neutron unit tests are failing with SQLAlchemy 1.0.11, 1.0.12
Public bug reported:
When running the unit tests (when building the Debian package for
Neutron Mitaka RC3), Neutron fails more than 500 unit tests. Upgrading
from SQLAlchemy 1.0.11 to 1.0.12 fixed the issue.
Example of failed run:
** Also affects: nova
Importance: Undecided
Status: New
--
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to neutron.
https://bugs.launchpad.net/bugs/1548433
Title:
neutron returns objects other than oslo_config.cfg.Opt
Public bug reported:
When building the Debian package of Nova for Mitaka b3 (ie: Nova
13.0.0~b3), I get the below unit test failures. Please help me to fix
this. I'm available on IRC if you need more details and a way to
reproduce (but basically, try to build the package in Sid + Experimental
Public bug reported:
When generating glance-api.conf and glance-registry.conf, the workers
directive, even though commented, is populated with a value which
depends on the number of CPU (or core?) of the machine generating the
file. This means that the build of Glance isn't reproducible (ie:
Public bug reported:
Hi,
I have 3 unit test failures when building the Glance Mitaka b2 package,
as per below. Please help me to fix them.
==
FAIL: glance.tests.functional.test_reload.TestReload.test_reload
Public bug reported:
When building the Debian package of Neutron for the Mitaka b2 release, I
get the below unit test failures. All other tests are ok (6451 tests).
Please help me to fix these last 3.
==
FAIL:
Public bug reported:
Horizon uses the TemplateDoesNotExist which was removed from Django 1.9
** Affects: horizon
Importance: Undecided
Assignee: Thomas Goirand (thomas-goirand)
Status: In Progress
--
You received this bug notification because you are a member of Yahoo
Public bug reported:
The class django.forms.util was renamed django.forms.utils (note the
added s) in Django 1.9. Horizon must follow.
** Affects: horizon
Importance: Undecided
Status: New
--
You received this bug notification because you are a member of Yahoo!
Engineering Team,
Public bug reported:
Horizon still uses django.utils.importlib which is removed from Django
1.9. We should use:
from importlib import import_module
instead of:
from django.utils.importlib import import_module
** Affects: horizon
Importance: Undecided
Status: New
--
You
Public bug reported:
We should use logging.NullHandler, as this is removed from Django 1.9
** Affects: horizon
Importance: Undecided
Status: New
--
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Dashboard
I've reopened the issue, as there's no sign that it was fixed.
** Changed in: horizon
Status: Expired => Confirmed
--
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Dashboard (Horizon).
Public bug reported:
As glance is building its config, it needs swiftclient. And it isn't
defined in test-requirements.txt, meaning that some options may be
missing for swift. I've added python-swiftclient in the Build-Depends-
Indep: of the Debian package, though it should IMO also go into the
Public bug reported:
Testing Glance Liberty RC1 over on top of Jessie, I have found out that
if I don't write a value for filesystem_store_datadir in both glance-
api.conf and glance-registry.conf, Glance simply doesn't work, despite
what is written here:
Public bug reported:
Hi,
Trying to see if it is possible to debug Horizon in production, one of
my colleague tried to disable compress. Then the result isn't nice at
all. Setting COMPRESS_ENABLED = False and restarting Apache leads to
every xstatic library being NOT FOUND, and loading of pages
Public bug reported:
The files generated using oslo-config-generator are completely wrong.
For example, it is missing [keystone_authtoken] and many more. This
shows on the example config in git (ie: etc/glance-api.conf in Glance's
git repo).
I believe the generator's config files is missing
Public bug reported:
Hi there!
I perfectly know that Nova is a required component for Horizon, however,
for all pages, if only Keystone is installed (ie: Nova isn't), there's
only a RED error message on the top-right of the web page, saying that
compute isn't installed. However, going to
Public bug reported:
When trying Horizon from Liberty b3, I can login into Horizon, but then
I get the below error all the time in the JS console of Iceweasel:
Error: [$injector:modulerr] Failed to instantiate module horizon.app due to:
[$injector:modulerr] Failed to instantiate module
Public bug reported:
Hi,
Building Keystone in Jessie poses no problem, but it looks like in Sid,
Keystone doesn't like SQLAlchemy 1.0.6. Here's a full build log:
http://sid.gplhost.com/keystone_8.0.0~b1-1_amd64.build
Just in case if that file wasn't available, here's an example crash.
There's
Hi Doug. I think you missunderstood. suds-jurko isn't more actively
maintained than suds, the upstream maintainer is *ALSO* dead, so we need
a better replacement. pysimplesoap is one of the possible alternatives.
** Changed in: nova
Status: Invalid = Confirmed
--
You received this bug
Public bug reported:
Suds is a library which isn't maintained upstream. Please switch to
something else. As you may have seen, someone filed a bug against the
nova package in Debian because of this:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=788081
As mentioned on the bug report: Please
Public bug reported:
Releasing Kilo in Debian, I found out that Keystone just broke with
pysaml2 2.0.0, and in fact needs 2.4.0. The unit tests just passed, but
with pysaml2 2.0.0 Keystone just crashes with a stack dump.
Out of this, 2 remarks:
- requirements.txt is wrong and should ask for
Public bug reported:
There's no python-ordereddict package anymore in Debian, as this is
normally included in Python 2.7. I have therefore patched
requirements.txt to remove ordereddict. However, even after this, I get
some bad unit test errors about it. This must be fixed upstream, because
build log may be found on my jenkins at:
https://kilo-jessie.pkgs.mirantis.com/job/neutron/
Cheers,
Thomas Goirand (zigo)
FAIL:
neutron.tests.unit.ml2.test_ext_portsecurity.PSExtDriverTestCase.test_create_network_with_portsecurity_mac
neutron.tests.unit.ml2
a Sbuild chroot, and the
full build log may be found on my jenkins at:
https://kilo-jessie.pkgs.mirantis.com/job/neutron/
Cheers,
Thomas Goirand (zigo)
FAIL:
neutron.tests.unit.ml2.test_ext_portsecurity.PSExtDriverTestCase.test_create_network_with_portsecurity_mac
neutron.tests.unit.ml2
Public bug reported:
Hi,
I get the bellow issues when building Keystone in Debian (Jessie chroot
using sbuild). Help from the keystone team would be appreciate to
resolve this. Cheers! :)
==
FAIL:
Public bug reported:
neutron-lbaas is trying to read neutron.conf from /usr/lib/python2.7
/dist-packages/etc/neutron.conf. Of course, in the context of building a
Debian package, this fails miserably.
Also, I'm really not fan of neutron-lbaas and neutron being so
interdependent. The way it is
Public bug reported:
Depending on what's in the flavor directive of /etc/glance/glance-
registry.conf, glance-registry searches for a different section of
/etc/glance/glance-registry-paste.ini. For example, if I put flavor =
keystone+caching, then glance-registry searches for a section [pipeline
Public bug reported:
When building Horizon Juno RC1, I have the below Python stack dump. If
you need more details on how to reproduce this in Debian Sid, please let
me know, and I'll explain, though the package is currently in Debian
Experimental, so just doing dpkg-buildpackage using it should
: Undecided
Assignee: Thomas Goirand (thomas-goirand)
Status: In Progress
--
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Dashboard (Horizon).
https://bugs.launchpad.net/bugs/1374919
Title
Public bug reported:
This bug is the 2nd instance of
https://bugs.launchpad.net/keystone/+bug/948495, 2 years later.
Its looking like keystone is fetching python-keystoneclient from git, in
order to run its unit tests. This is not possible when running during
the build of the Debian package.
Public bug reported:
As per the subject, horizon/workflows/base.py contains a definition of
add_error(). Unfortunately, this now a function name used by Django 1.7.
This conflicts with it, and leads to unit test errors when running with
Django 1.7 installed
** Affects: horizon
Importance:
Public bug reported:
==
FAIL:
keystone.tests.test_exception.ExceptionTestCase.test_invalid_unicode_string
--
Traceback (most recent call last):
Public bug reported:
Keystone fails its unit tests when running with SQLAlchemy 0.9.3, as per
the log below. It is important for Debian that Havana Keystone continues
to work in Sid with SQLA 0.9.
==
ERROR:
81 matches
Mail list logo