It is possible for events from the nb/sb to fire before the opposite
db connection is made. These events can call back into driver code
which tries to access the other db before it is connected.
Closes-Bug: #1929197
Closes-Bug: #1928794
Closes-Bug: #1929633
Change-Id: If947581b90ced42981c4611c32de8f428a052c69
This patch switches the code over to the payload style of callbacks [1]
for TRUNK and SUBPORTS events. As needed existing callbacks are shimmed
to support both payload and kwarg style callbacks. These shims will be
removed once all callbacks are switched over to payloads.
Also the neutron.services.trunk.callback module is removed as consumers
will no longer need the TrunkPayload therein.
NeutronLibImpact
[1]
https://docs.openstack.org/neutron-lib/latest/contributor/callbacks.html
Change-Id: Ie302b48b283f8780072b5c9e2bc8787d87c11794
During the live migration trunk subports where updated only based
on the "host_id" field. But when rpc message to update subports
is send trunk's port is still bound to the old host and has
"migrating_to" field in binding:profile set to the new host.
Because of that binding:host_id for the subport's port wasn't updated
proberly and port was set to DOWN on the new host.
That could even cause connectivity break if L2population is used in the
cloud.
This patch fixes that by updating subport's binding:host_id field based
on the migrating_to field if that is available and not empty.
Closes-Bug: #1914747
Change-Id: I98e55242d381ada642ca0729e9aefdea7628c945
assertItemsEqual was removed from Python's unittest.TestCase in
Python 3.3 [1][2]. We have been able to use them since then, because
testtools required unittest2, which still included it. With testtools
removing Python 2.7 support [3][4], we will lose support for
assertItemsEqual, so we should switch to use assertCountEqual.
NOTE(dmllr): added hacking check
[1] - https://bugs.python.org/issue17866
[2] - https://hg.python.org/cpython/rev/d9921cb6e3cd
[3] - testing-cabal/testtools#286
[4] - testing-cabal/testtools#277
Change-Id: I7c20fec08e5dc9f67b34100c925ea6724bbd25f0
Test test_handle_trunk_remove_trunk_manager_failure is trying
to connect to tcp:127.0.0.1:6640. This behaviour can be visible
by netcat while test is running. As this is unit test, it should
use mock, not real ovsdb. This patch is fixing this.
$ nc -l -p 6640
{"id":0,"method":"get_schema","params":["Open_vSwitch"]}
Change-Id: I25161913d701e15d127c70d71ac025e4573e81f1
Now that we are python3 only, we should move to using the built
in version of mock that supports all of our testing needs and
remove the dependency on the "mock" package.
This patch moves all references to "import mock" to
"from unittest import mock". It also cleans up some new line
inconsistency.
Fixed an inconsistency in the OVSBridge.deferred() definition
as it needs to also have an *args argument.
Fixed an issue where an l3-agent test was mocking
functools.partial, causing a python3.8 failure.
Unit tests only, removing from tests/base.py affects
functional tests which need additional work.
Change-Id: I40e8a8410840c3774c72ae1a8054574445d66ece
This reverts commit 8ebc635a18.
The reverted commit does mass update on all subports of a trunk.
This is not in line with the original design since it causes huge
api-side performance effects.
I think that's the reason why we started seeing gate failures of
rally_openstack.task.scenarios.neutron.trunk.CreateAndListTrunks
in neutron-rally-task.
Change-Id: I6f0fd91c62985207af8dbf29aae463b2b478d5d2
Closes-Bug: #1870110
Related-Bug: #1700428
Introduce a callback function in trunk plugin to receive after_create
and after_delete events for the trunk and subport resources This
function will set the subport device_id to trunk_id to prevent nova and
neutron from reusing it.
Closes-Bug: #1700428
Co-Authored-By: Nate Johnston <nate.johnston@redhat.com>
Change-Id: I97fe59752f9ef2cbe0376cd6ee0739660f5937ad
This is a tweak to changes done to fix bug 1834637. Specifically,
this change addresses scaling. The previous gerrit change had
modifications to all OVN sub-ports performed as a single
transaction. That did not account for race-condition on neutron
DB queries, which leads to timeouts under heavy loads.
Another cleanup done by this change is to fold an additional
update on neutron db into ovn trunk driver. That saves
update from doing another database transaction, thus making
it faster. The no longer needed function in mech_driver was
called _update_subport_host_if_needed
By breaking the iteration into multiple transactions, the
change in time is marginal:
Service-level agreement
NeutronTrunks :: neutron.create_trunk
from 34.2 sec to 35.6 for 50%ile
from 35.6 sec to 36.1 for 95%ile
This patch is a cherry pick from networking-ovn stable/train [1], that
is now part of networking-ovn migration to neutron repo.
[1]: https://review.opendev.org/#/c/698863/
Change-Id: I118f93cdd34a1d25327a0dc61367720ac6559001
Closes-Bug: #1834637
Co-authored-by: Maciej Józefczyk <mjozefcz@redhat.com>
(cherry picked from commit c418dd720b9be9047d779f32407c474e90cb0002)
In patch [1] as partial fix for bug 1828375 retries mechanism
was proposed.
We noticed that sometimes in have loaded environments 3 retries
defined in [1] can be not enough.
So this patch switches to use neutron_lib.db.api.MAX_RETRIES constant
as number of retries when processing trunk subport bindings.
This MAX_RETRIES constant is set to 20 and in our cases it "fixed"
problem.
[1] https://review.opendev.org/#/c/662236/
Change-Id: I016ef3d7ccbb89b68d4a3d509162b3046a9c2f98
Related-Bug: #1828375
Removed E125 (continuation line does not distinguish itself
from next logical line) from the ignore list and fixed all
the indentation issues. Didn't think it was going to be
close to 100 files when I started.
Change-Id: I0a6f5efec4b7d8d3632dd9dbb43e0ab58af9dff3
This is an approximate partial fix to #1828375.
update_trunk_status and update_subport_bindings rpc messages are
processed concurrently and possibly out of order on the server side.
Therefore they may race with each other.
The status update race combined with
1) the versioning feature of sqlalchemy used in the standardattributes
table and
2) the less than serializable isolation level of some DB backends (like
MySQL InnoDB)
does raise StaleDataErrors and by that leaves some trunk subports in
DOWN status.
This change retries the trunk status update (to BUILD) blindly when
StaleDataError was caught. In my local testbed this practically
fixes #1828375.
However theoretically the retry may cover up other real errors (when the
cause of the StaleDataError was a different status not just a different
revision count).
To the best of my understanding a proper fix would entail guaranteeing
the in order processing of the above rpc messages - which likely won't
ever happen.
I'm not sure at all if this change is worth merging - let me know what
you think.
Change-Id: Ie581809f24f9547b55a87423dac7db933862d66a
Partial-Bug: #1828375
The trunk constants now live in neutron-lib. This patch consumes them
by removing neutron.services.trunk.constants and using them from
neutron-lib instead.
Depends-On: https://review.opendev.org/#/c/650372/
NeutronLibImpact
Change-Id: I4445c44c7e321d0fc35976d4d855c148bb9a3b18
The trunk callback resource names now live in neutron-lib [1].
This patch switches over to use those constants from lib and they
will be removed from neutron in a later patch once consumers switch
over to lib as well.
[1] https://review.openstack.org/#/c/635209/
Change-Id: I79dfe76c1b698fe634393e469157014f1914cc8e
The neutron.common.rpc module has been in neutron-lib for awhile now and
neutron is shimmed to use neutron-lib already.
This patch removes neutron.common.rpc and switches the code over to use
neutron-lib's implementation where needed.
NeutronLibImpact
Change-Id: I733f07a8c4a2af071b3467bd710290eee11a4f4c
IPWrapper.get_devices_info(), implemented using pyroute2, retrieves the
device information including the VLAN tag and the parent name and index.
This patch replaces Plumber._get_vlan_children() shell 'ip' commands in
favor of this method.
Change-Id: Ib5cad35d5261ab9391f82a22440338d852894a1d
Closes-Bug: #1804274
When a deployment has instance ports that are neutron trunk ports with
DPDK vhu in vhostuserclient mode, when the instance reboots nova will
delete the ovs port and then recreate when the host comes back from
reboot. This quick transition change can trigger a race condition that
causes the tbr trunk bridge to be deleted after the port has been
recreated. See the bug for more details.
This change mitigates the race condition by adding a check for active
service ports within the trunk port deletion function.
Change-Id: I70b9c26990e6902f8888449bfd7483c25e5bff46
Closes-Bug: #1807239
For agent such as linux bridge which allows creating trunk on
bound port, this patch provides a fix to allow deleting trunk
without unbound port first. And it will help to keep the port
(trunk's parent port) working while deleting the trunk.
Co-Authored-By: Allain Legacy <Allain.legacy@windriver.com>
Closes-Bug: #1794424
Story: 2003889
Change-Id: Iae2ae535bf3ba1548136bf3fe4306a42bad4e635
The common rpc and exceptions were rehomed into
neutron-lib with [1]. This patch shims those rehomed
modules in neutron to switch over to neutron-lib's
versions under the covers.
To do so:
- The rpc and common exceptions are changed to
reference their counterpart in neutron-lib effectively
swapping the impl over to neutron-lib.
- The fake_notifier is removed from neutron and lib's
version is used instead.
- The rpc tests are removed; they live in lib now.
- A few unit test related changes are required
including changing mock.patch to mock.patch.object,
changing the mock checks for a few UTs as they don't
quite work the same with the shim in place.
- Using the RPC fixture from neutron-lib rather than
that setup in neutron's base test class.
With this shim in place, consumers are effectively using
neutron-lib's RPC plumbing and thus we can move consumers
over to neutron-lib's version at will. Once all
consumers are moved over we can come back and remove
the RPC logic from neutron and follow-up with a consumption
patch.
NeutronLibImpact
[1] https://review.openstack.org/#/c/319328/
Change-Id: I87685be8764a152ac24366f13e190de9d4f6f8d8
This incorporates flake8 2.6.x and pycodestyle will be used
instead of older pep8. This ensures future python3 compatibility
and a bit better code styling.
Change-Id: Ia7c7c5a44727f615a151e1e68dd94c7ed42f974f
A bulk of the public APIs that are part of neutron.plugins.common.utils
were rehomed into neutron-lib with [1] and the remaining with [2].
This patch consumes [1] by:
- Removing the rehomed code from neutron.
- Removing the UTs that are no longer applicable.
- Leaving the functions in [2][3] in neutron.plugins.common.utils until
we release [2][3] and can consume it at which point we should be able to
remove the utils module.
NeutronLibImpact
[1] Iabb155b5d2d0ec6104ebee5dd42cf292bdf3ec61
[2] I2c0e4ef03425ba0bb2651ae3e68d6c8cde7b8f90
[3] I73f5e8ad7a1a83392094db846d18964d811b8bb2
Change-Id: I1d63cbea463e92e1d2e053f8e1a564ed52cb84f8
patch set of [1] added callback priority so that a new argument, priority,
was added to subscribe method. Thus its signature was changed.
It caused to break some unit tests.[2] Those test cases check arguments passed
to subscribe method.
The fix to those test cases needs to be compatible with older version of
neutron-lib.
So add a helper method, get_subscribe_args, to produce argument to subscribe
by checking if neutron-lib supports callback priority or not. and apply it
to broken test cases.
[1] https://review.openstack.org/#/c/541766/
[2] http://logs.openstack.org/periodic/git.openstack.org/openstack/neutron/master/openstack-tox-py35-with-neutron-lib-master/66d328a/testr_results.html.gz
test_capabilities.CapabilitiesTest.test_register
test_ovs_capabilities.CapabilitiesTest.test_register
test_backend.ServerSideRpcBackendTest.test___init__
Change-Id: I94066ff8a8dae5c5637478aa52883b3b451f2857
This patch switches callbacks over to the payload object style events
[1] for PRECOMMIT_UPDATE based notifications. To do so a DBEventPayload
object is used with the publish() method to pass along the related data.
In addition a few UTs are updated to work with the changes. Finally
a few shims are put into place to allow PRECOMMIT_UPDATE based events to
use payloads while still supporting the existing kwarg style events.
NeutronLibImpact
[1] https://docs.openstack.org/neutron-lib/latest/contributor/callbacks.html#event-payloads
Change-Id: Ie6d27df01cd7b87894efc80946d41eb1ebe25bef
While things may have become generally slower for reasons still
to be determined at the time this patch was developed, we noticed
that enforcing MTU rules between subports and trunk parent was
done by performing one DB lookup per subport. This patch optmizes
the performance by piggybacking on the logic that is used to
retrieve the subport underlying network's segmentation type when
the segmentation type must be inherited.
Closes-bug: #1741954
Change-Id: Ib18427ef7a3509088eaf8fa0cfed3fb659b6baea
This patch switches callbacks over to the payload object style events
[1] for BEFORE_READ based notifications. To do so an EventPayload object
is used with the publish() method to pass along the API related data.
In addition a few UTs are updated to work with the changes.
NeutronLibImpact
[1] https://docs.openstack.org/neutron-lib/latest/contributor/callbacks.html#event-payloads
Change-Id: Iff3e96c56867b4bf4272fed676f39cd6796d756c
In reviews we usually check import grouping but it is boring.
By using flake8-import-order plugin, we can avoid this.
It enforces loose checking so it sounds good to use it.
This flake8 plugin is already used in tempest.
Note that flake8-import-order version is pinned to avoid unexpected
breakage of pep8 job.
Setup for unit tests of hacking rules is tweaked to disable
flake8-import-order checks. This extension assumes an actual file exists
and causes hacking rule unit tests.
Change-Id: Ib51bd97dc4394ef2b46d4dbb7fb36a9aa9f8fe3d
There were a few places left in the code (with TODOs) that were still
mocking out the callback manager. This patch switches them over to
the neutron-lib callback fixture.
Change-Id: I9e710db6a5103436b0f098e8f73625e3941df492
The ml2 MechanismDriver is now in neutron-lib along with its associated
constants. This patch switches over to the lib versions of those, but
leaves a shim of the MechanismDriver that just ref's the driver from
lib. This shim allows our broad consumer base of the driver to switch
over at their leisure.
NeutronLibImpact
Change-Id: I99e3de6d933a1bb341394f85415fb07306a82a01
The callback modules have been available in neutron-lib since commit [1]
and are ready for consumption.
As the callback registry is implemented with a singleton manager
instance, sync complications can arise ensuring all consumers switch to
lib's implementation at the same time. Therefore this consumption has
been broken down:
1) Shim neutron's callbacks using lib's callback system and remove
existing neutron internals related to callbacks (devref, UTs, etc.).
2) Switch all neutron's callback imports over to neutron-lib's.
3) Have all sub-projects using callbacks move their imports over to use
neutron-lib's callbacks implementation.
4) Remove the callback shims in neutron-lib once sub-projects are moved
over to lib's callbacks.
5) Follow-on patches moving our existing uses of callbacks to the new
event payload model provided by neutron-lib.callback.events
This patch implements #2 from above, moving all neutron's callback
imports to use neutron-lib's callbacks.
There are also a few places in the UT code that still patch callbacks,
we can address those in step #4 which may need [2].
NeutronLibImpact
[1] fea8bb64ba7ff52632c2bd3e3298eaedf623ee4f
[2] I9966c90e3f90552b41ed84a68b19f3e540426432
Change-Id: I8dae56f0f5c009bdf3e8ebfa1b360756216ab886
This patch introduces support for requests where the user does
not know the segmentation details of a subport and by specifying
segmentation_type=inherit will let the trunk plugin infer these
details from the network to which the subport is connected to, thus
ignoring the segmentation_id in case it were to be specified.
This type of request is currently expected to have correct results
when the network segmentation type is 'vlan', and the network has
only one segment (provider-net extension use case).
DocImpact: Extend trunk documentation to include Ironic use case.
Closes-bug: #1648129
Depends-on: Ib510aade1716e6ca92940b85245eda7d0c84a070
Change-Id: I3be2638fddf3a9723dd852a3f9ea9f64eb1d0dd6
Neutron-lib 1.1.0 is now out and contains the portbindings
API definition (as per commit [1]). This patch moves neutron
references over to the neutron-lib version.
NeutronLibImpact
- Consumers using the public constants within neutron's
portbindings API extension must now use the values
from neutron-lib.
[1] 87e42f993c07ae320159d5123662ee9f3bd4d903
Change-Id: I669af9b4c712877772d91a03857ab108714001d4
Matching the MAC of the tap interface on the hypervisor
to the MAC that the VM will actually use to send traffic
through the tap causes the following error messages:
brq9e974900-cd: received packet on tapa8a6ce4a-e8 with
own address as source address
This is triggering a warning because there are now two
forwarding entries on the bridge, a static one from the
tap device, and a learned one from the packets received.[1]
Due to the presence of a static MAC entry, this will break
things like allowed_address_pair use cases that allow a MAC
to be used by multiple ports.
This patch removes all of the logic to set the MAC
address on the TAP device because it didn't serve any
other purpose than easy to correlate output from
ip link show commands and neutron port MAC addresses.
TAP devices will now just get whatever MAC address the kernel
selects for them.
1. https://lists.linuxfoundation.org/pipermail/bridge/2004-January/003740.html
Closes-Bug: #1668209
Change-Id: I0ff46f9550a79f486063a8e2810ed3b1140a4769
The patch introduces new abstract method to the API abstract class. The
method is supposed to return a new Transaction object. Each
API object is capable of store one nested transaction which is returned
by context manager in case some transaction already exists.
As there are no projects in OpenStack that use inheritance directly from
API abstract class, it's safe to make this new create_transaction() abstract method.
Only projects that currenlty use ovsdb API are networking-ovn,
dragonflow and networking-l2gw. OVN and Dragonflow use only IDL
implementation and L2GW copies the code of API abstract class.
Closes-bug 1653517
Change-Id: I55dd417cae7ebbe0668ba5606949ce4ab045d251
Objects must use project_id and not tenant_id. The object framework
ensures that tenant_id is added as an extra field for backward
compatibility.
This patch reverts the workaround implemented in change
I4ec9340094bc51cd8aa6e5112bf8114aa26c2982 and implements a proper fix
by explicitly updating the objects.
Co-Authored-By: Artur Korzeniewski <artur.korzeniewski@intel.com>
Co-Authored-By: Darek Smigiel <smigiel.dariusz@gmail.com>
Closes-Bug: #1630748
Change-Id: Iab90bcab41655b2e210aea0e7581eb00b94ce5e5
By default, wait_until_true uses default exception from eventlet which
is eventlet.TimeoutError. This class is not subclass of Exception but
BaseException. In case wait_until_true times out in any test, the whole
test executor worker is stopped leaving scheduled tests not executed.
This patch replaces eventlet.TimeoutError with new WaitTimeout
exception, that inherits from Exception and thus won't break execution
of other test cases in case it's raised.
Related-Bug: 1625221
Change-Id: I44c0c22f427f61d84963e6e59393b90fbaa8f058
If a failure occurs while unwiring stale ports, we currently
ignore the outcome of the operation, whereas we should at least
warn the user that something is not quite as it is supposed
to be.
This patch makes sure that trunk status is accurately reflected
after a sequence of unwire+wire operations for a given trunk.
Closes-bug: #1649887
Change-Id: I3b6ed57e00c0146babe23ea6ed0ca14e83020d26
Maintaining the context is important for keeping the request ID
and subsequently operator/developer sanity while debugging.
The resource_type is also helpful to have since a function could be
subscribed for multiple resources.
This maintains and deprecates the existing 'subscribe' method for
backwards compatibility with callbacks that don't support receiving
the context and resource type. A new 'register' method is added
for callbacks to use that are compatible with receiving the context.
Change-Id: I06c8302951c99039b532acd9f2a68d5b989fdab5
Neutron Manager is loaded at the very startup of the neutron
server process and with it plugins are loaded and stored for
lookup purposes as their references are widely used across the
entire neutron codebase.
Rather than holding these references directly in NeutronManager
this patch refactors the code so that these references are held
by a plugin directory.
This allows subprojects and other parts of the Neutron codebase
to use the directory in lieu of the manager. The result is a
leaner, cleaner, and more decoupled code.
Usage pattern [1,2] can be translated to [3,4] respectively.
[1] manager.NeutronManager.get_service_plugins()[FOO]
[2] manager.NeutronManager.get_plugin()
[3] directory.get_plugin(FOO)
[4] directory.get_plugin()
The more entangled part is in the neutron unit tests, where the
use of the manager can be simplified as mocking is typically
replaced by a call to the directory add_plugin() method. This is
safe as each test case gets its own copy of the plugin directory.
That said, unit tests that look more like API tests and that rely on
the entire plugin machinery, need some tweaking to avoid stumbling
into plugin loading failures.
Due to the massive use of the manager, deprecation warnings are
considered impractical as they cause logs to bloat out of proportion.
Follow-up patches that show how to adopt the directory in neutron
subprojects are tagged with topic:plugin-directory.
NeutronLibImpact
Partially-implements: blueprint neutron-lib
Change-Id: I7331e914234c5f0b7abe836604fdd7e4067551cf
Terry Wilson