As part of the split of functional and unit tests we need to isolate
the unit tests into a separate directory for having multiple test
targets in a sane way.
Part of bp:functional-tests-for-nova
Change-Id: Id42ba373c1bda6a312b673ab2b489ca56da8c628
build_sphinx has a lot of spurious tracebacks, We don't
really need this check for document generation
Change-Id: Id9442c5902918592c4c242cc66975062cf9f2461
We don't need to have the vi modelines in each source file,
it can be set in a user's vimrc if required.
Also a check is added to hacking to detect if they are re-added.
Change-Id: I347307a5145b2760c69085b6ca850d6a9137ffc6
Closes-Bug: #1229324
Previous _ was monkey patched into builtins whenever
certain modules were imported. This removes that and
simply imports it when it is needed.
Change-Id: I0af2c6d8a230e94440d655d13cab9107ac20d13c
E125 continuation line does not distinguish itself from next logical
line.
E125 makes code more readable, and fixing all the existing issues allows
us to gate on this, making this one less thing reviewers have to check
for.
Fix was made using https://github.com/hhatto/autopep8
Change-Id: Ie569fd74fb84d220ceb87e37b1b1f39143bba80f
All of this is because if dnspython is present in your environment then
eventlet monkeypatches socket.getaddrinfo() with an implementation
which doesn't work for IPv6. To turn this off you need to set an
environment variable before you do the import.
Resolves bug 1164822.
Change-Id: I2382db7b3ac4a2d940e0c45dfe3d48f499de2f16
There are global fixtures that get set up for testing in
nova/tests/__init__.py, even though there is a base test class. As we move
towards parallel testing, we want to convert all of these into proper fixtures.
As a step towards that, move them into nova/test.py so that it's clear that
they are part of the global test setup. From there we'll move them into
proper fixture classes.
Part of blueprint grizzly-testtools
Change-Id: Ia0b0ae380069a212bc20e009d1088c2bad7d9177
With a few minor exceptions, the sql config options are used solely
within the nova.db.sqlalchemy.session module so it makes sense to move
their declaration into that module.
Change-Id: Iea9c2bb000cd713b01750ab3e796132ebeaa4ca8
The only reason for importing nova.config now is where one of the
options defined in that file is needed. Rather than importing
nova.config using an import statement, use CONF.import_opt() so
that it is clear which option we actually require.
In future, we will move many options out of nova.config so many
of these import_opt() calls will either go away or cause a module
other than nova.config to be imported.
Change-Id: I0646efddecdf2530903afd50c1f4364cb1d5dce1
Modules import nova.config for two reasons right now - firstly, to
reference nova.config.CONF and, secondly, if they use one of the
options defined in nova.config.
Often modules import nova.openstack.common.cfg and nova.config
which is a bit pointless since they could just use cfg.CONF if
they just want to nova.config in order to reference CONF.
Let's just use cfg.CONF everywhere and we can explicitly state
where we actually require options defined in nova.config.
Change-Id: Ie4184a74e3e78c99658becb18dce1c2087e450bb
Now that options have all moved from nova.flags to nova.config, we can
safely remove the nova.flags imports and replace them with nova.config
imports.
Change-Id: Ic077a72dd6419bbf1e1babe71acfa43c4e8b55c8
Use the global CONF variable instead of FLAGS. This is purely a cleanup
since FLAGS is already just another reference to CONF.
We leave the nova.flags imports until a later cleanup commit since
removing them may cause unpredictable problems due to config options not
being registered.
Change-Id: Ic0168188dfe214fc81af04c8a9644d2d9f31600d
The moduelauthor tags are used by the docs to see who wrote what.
But many of them are outdated and wrong.
fixes bug 1044578
Change-Id: I9cc27430c906d1418faf28e1b53f6306c62e8f98
This resolves issues when running individual tests that require
a database. With this fix I can once again run individual test
modules that make use of the database. For example:
./run_tests.sh test_libvirt
Previously this would fail with database errors.
Fixes LP Bug #1032738.
Change-Id: Icce7ac9414f0e19eece44819a217634947de7f73
Implements blueprint multi-process-api-service
This is based on Huang Zhiteng's patch.
This patch adds support for running services as multiple processes. This
is primarily intended to be used with the API service as a way to provide
more concurrency than eventlet can sometimes provide.
A SIGTERM or SIGINT signal will cause the parent process to gracefully
terminate the child processes, allowing them to finish processing the
requests currently being processed. The parent will wait for the
children to finish before exiting.
Change-Id: Ie6d6802626eb42d5e64c4167be363fbf6cea2a1b
Fixes bug #1013770
1. Reorder imports by full module path
2. Insert two blank lines after the last import
Change-Id: I294ac3ab528f17a72811392d1732158e9487f3bf
Shrink tox.ini to the new short version.
Fix the test cases to be able to be run in nosetets plus the
openstack.nose_plugin, which finally removes the need for
nova/testing/runner.py
Also, now we'll just output directly to stdout, which will
make nose collect the trace logging directly and either output
it at the end of the run, or inject it into the xunit output
appropriately.
Change-Id: I1456e18a11a840145492038108bdfe812c8230d1
Part of blueprint common-rpc.
This patch makes the rpc code use the global config object from
openstack-common. Based on some recent discussions on the mailing list,
this may not be the final way configuration handling is done here, but
it is certainly better than the register_opts() hack that is removed by
this patch.
Change-Id: Id128126e0bc064a2a1c710c6bd32fb3d137dc7f6
Compacts the pre-Folsom database migrations into
a single migration (082_essex.py).
Pre-Folsom users will need to upgrade to Essex before
running any Folsom migrations.
Implements blueprint db-migration-cleanup.
Testing notes available at:
http://wiki.openstack.org/database_migration_testing
Change-Id: I64c06a3adcf83d6d723c4c11001544ba97668413
This patch does a couple of things:
1) Remove the dependency of nova.rpc on nova.flags. This is a step
toward decoupling nova.rpc from the rest of nova so that it can be moved
to openstack-common.
2) Refactor nova.rpc so that a configuration object is passed around as
needed instead of depending on nova.flags.FLAGS.
This was done by avoiding changing the nova.rpc API as much as possible
so that existing usage of nova.rpc would not have to be touched. So,
instead, a config object gets registered, cached, and then passed into
the rpc implementations as needed. Getting rid of this global config
reference in nova.rpc will require changing the public API and I wanted
to avoid doing that until there was a better reason than this.
Change-Id: I9a7fa67bd12ced877c83e48e31f5ef7263be6815
Fail early if you can't import mox, rather than return
a cryptic and undecipherable error. Fixes bug 810424.
Change-Id: Ie56d3e7447bbba841fdc882cb1f8f70e234812ef
part 1 of blueprint separate-nova-metadata
* adds api/metadata/ and moves code from ec2
* moves metadata into separate binary
* changes metadata forward to use metadata host and port
* moves the metadata accept rule to the metadata api
* adds nova-api-* to setup.py
Change-Id: I7f5d8e6cafc55b5c383cd88991f29c6059fb8d82
This introduces a new flag "quantum_use_dhcp=<boolean>" which indicates whether
or not to enable dhcp for all of the networks. If it is set then we start
dnsmasq (and provide it with the IP/MACs from Melange) similar to how this was
done in linux_net before.
Prior to this if you enabled dhcp then you would also get a gateway device..
some people may not want that so we now require that you specify the gateway
when creating the network in order to end up with a device that will act as a
gateway. If you're using Melange IPAM and you don't specify the gateway you
still end up with one because it doesn't allow you to not have one. This lays
the groundwork for the option of not having one in the future, at least :)
Also, fix quantum/melange ipam interaction
We now query for the subnets by net_id/vif_id instead of searching through all
the blocks to find the right one. Both of the allocate and deallocate for
instance calls are now using the vif_id -> network_id mapping instead of
searching the quantum networks. get_port_by_attachment was also changed to
take a net_id so that we don't have to search through all of the quantum
networks to find the corresponding port.
Change-Id: I6a84da35237b6c5f5cdee91ada92642103439a97
Sean Dague