In Puppet 3 there was a need to use absolute
names to prevent issues. Since Puppet 4 everything
is absolute by default which makes this not needed.
We need to change this everywhere so that we can
revert the pin in [1] that now prevents us from
using the latest version of the puppet-lint check
puppet-lint-absolute_classname_check that properly
checks that we dont use absolute names.
[1] https://review.opendev.org/#/c/697742/
Change-Id: I78b74fbeb08be7234189e4d266412fb7fb7a73c0
nova conductor and scheduler need to discover placement service when it
starts so let's make sure endopints are discoverable before starting the nova
conductor and scheduler services.
Change-Id: Iaf70c646f29a326d39471e562c61d553121fa3ff
This parameter is deprecated and will be removed in [1].
Removing this will make nova use the default behavior of
looking up the endpoint for talking to neutron. Deployments
that want to override that behavior can set the
neutron_endpoint_override parameter or the use the
neutron_valid_interfaces parameter to let nova know what
type of endpoint type to use.
[1] https://review.opendev.org/#/c/658400/
Change-Id: Id892bf2a2998cb11a5337f36ff55a05dc1f34601
Following on from I5260b4d3478008a76fbe53fdbb716004d11b22df in Stein and
ahead of the removal of the Placement deployment logic from puppet-nova
we can now remove nova_placement from POI. We can also switch the one
remaining job using this class to the puppet-placement based placement
class.
Change-Id: Ia56e403b0f7171ee3583e768404d5f65acd61576
This change extracts the existing logic to deploy the Nova Placement
service into a separate class. This should allow POI to continue to test
that puppet-nova is able to deploy the service after puppet-placement is
introduced.
Change-Id: I8e54d851bc433527d010b763e4aad54172a17676
Double the timeout from 300 to 600 and add --debug
to see if we can catch what is causing it to timeout.
Change-Id: I7d6557eb8175a56d5a31e9b017af9fd192779862
* Earlier the sleep was 10 sec, till that time nova-compute does
not startup and register itself in these 10 seconds which leads
to race condition while tempest tests. Increasing sleep will fix
the issue.
Change-Id: I9ad252eee056762fbaacb68f673d07a24329947f
The keymgr_api_class parameter is deprecated, use keymgr_backend instead.
https://review.openstack.org/#/c/515058/
Change-Id: I44da6a0c277a2d4400708270883a5f44348af143
Support for nova metadata api via wsgi was added in dependency.
This changes so that we test with running via wsgi.
With nova metadata api running via wsgi it is ssl terminated, also
neutron metadata agent does not support an ipv6 address for the
metadata_host, so we need to configure the neutron metadata agent
to connect via https to the nova metadata api.
Related-Bug: 1781405
Depends-On: Ic65736cb0e95c400a728cd699ecf06c6aecff832
Depends-On: I3d572dda2e90b7e24f2f8912d704462b9332d807
Change-Id: I03f4dcb7f40055d802b114ee305323c95975e4df
The nova::logging parameter has been deprecated in puppet-nova repo[1].
[1]https://review.openstack.org/#/c/580213/
Change-Id: I656a630570ea0c22dbc8dfc67956324a3b40175b
puppet-nova switched to www_authenticate_uri in
[1], update puppet-openstack-integration for the same.
[1] https://review.openstack.org/#/c/558451/
Change-Id: I79af66fdbbf2f81a3da4900053d0f2e92b0fcaac
Do not attempt to configure apache for nova-placement-api in the
case of Debian, as everything is handled by the package that uses
uwsgi instead of apache.
Needed-By: I081542bfca7f5d4f3c0d8493c0b7c90d782a3e1f
Change-Id: I2808e73576fc90c6c9d5c0c1c020998a521924f2
This patch refactors all creation of RabbitMQ resources such
as users and ACLs to a define called openstack_integration::mq_user.
This will make sure RabbitMQ is automatically installed as well,
as there are times where dependencies to RabbitMQ are added to things
like Keystone which break Beaker tests in other repos (but we don't
know they're broken until the change merges).
In addition, instead of setting up dependencies on the tagged services,
we use the anchors provided which should hopefully test against them
as well.
Change-Id: Ib122ddd105529de5e12389cc9db2e4e09ec4ad54
Introduce configuration parameters to specify the oslo.messaging
rpc or notification backend as one of rabbit, amqp. The default
config is to use rabbit broker for both rpc and notification. Selection
of amqp for rpc will use brokerless qpid-dispatch router.
This patch:
* Adds notification_transport_url parameter across services where needed
* Adds rpc and notification parameters to config
* Adds qdr for rpc amqp1 configuration
* Modifies scenario001 for dual messaging backends
* supports ssl config
Depends-On: Id6ebc4ce8b0ffdb0be92a758dbf89c84c3274725
Change-Id: Ia2a79a2e1482f6f72426bc81c8e6d2a04cb211e3
nova::api::default_floating_pool has been deprecated, Please use
nova::network::neutron::default_floating_pool instead.
Change-Id: Ic8a5296d4d4a754bad28a091493c7627a770e6f1
Depends-On: Ife09171b84e0536b22784875ca4d0d69d16fb7fb
Instead of overriding nova::compute::libvirt::vncserver_listen, just use
the default value, so we can check that nova-compute will bind the
process on 127.0.0.1 and not 0.0.0.0.
It's recommanded for security to use an host IP address, instead of
0.0.0.0.
Change-Id: I724e8485d9fa58c213606ef0b30e2483d8a04802
We've seen on ubuntu that the nova computes are taking a bit longer to
register so we end up running discover_hosts prior to the registration.
This results in a failure when we attempt to use nova. So for p-o-i we
need to add in a sleep to give the computes enough time to register
before the discover_hosts is run. In most down stream consumers of these
modules, this should be handled as part of the overall orchestration
(tripleo/fuel/etc) rather than puppet itself because it's something that
will require cross node orchestration.
Change-Id: I9ad0393a34fc1416a3f4959eb83deb7a9b5aa9af
1. ubuntu: enable placement API in Nova
Until now, we were only deploying Placement API on Ubuntu because
packaging was too old. Though it's becoming a requirement in Ocata RC1
so before our CI breaks when Ubuntu will update their packages, let's
enable it now.
2. ubuntu: enable ::nova::cell_v2::simple_setup
Because we have to.
Change-Id: Id0ef365539a907c131167d745f9ac6eb1cd676fd
Nova team suggest it to avoid situations where nova-compute don't find
the placement service in the service catalog, which could lead to
compute resource not registred.
Change-Id: I14d7079e3830ca4436780ce690dbee5381a5e454
It was suggested by Nova team to not deploying Nova API in WSGI with
Apache in production.
It's causing some issues that we didn't catch until now (see in the bug
report). Until we figure out what was wrong, let's disable it so we can
move forward in the upgrade process.
Note: we also need to fix orchestration in provision to make sure
Keystone is ready before creating nova flavors, and other resources.
Depends-On: I1688eae1369f6da2c7084dc3864d19708d15c78d
Change-Id: I4a0d999d5290785a416bbb11953fee7a5028a00b
Related-Bug: 1661360
nova: do not install virtlockd and virtlog (workaround)
virtlog and virtlockd services are not idempotent with our version of
Puppet when deploying on Ubuntu.
See https://tickets.puppetlabs.com/browse/PUP-6370
Note: we already had this issue 8 months ago: https://review.openstack.org/#/c/324816
But it was fixed after Puppet 4.5. We'll have to investigate.
Change-Id: I2059b83d055d551872f4b63dc952af775f83c020
Disable the deprecation warnings as errors for puppet-syntax
Recently puppet4 started deprecating ruby 2.0 with the following
commit: e9eda7ed56
One way to work-around this (in the absence of a more recent ruby
version) is to not treat this deprecation warnings as fatal when
doing the puppet syntax check
Change-Id: If31ac681c75c6eb361dbb58682233160b31ae113
Related-Bug: #1660943
Co-Authorized-By: Michele Baldessari <michele@acksyn.org>
Cells are being required so we need to setup the default cell.
Change-Id: I119353e5edf343d704ad961b21c5aa466eb7d54d
Depends-On: I591b451197dc3bd0783978f5e3d2b1c830afe54e
It's recommended to have a database dedicated for placement service in
Nova. Let's do it.
Change-Id: Ib113bc341582e1053fa89fb531586a04e5b19df4
Depends-On: I139e1549fb4b33e518892a30bf781856d6cb31bc
Activate a new service in Nova: Placement API
It will be mandatory in the future.
docs.openstack.org/developer/nova/placement.html
Note: testing only on Centos7. Ubuntu package is broken.
Change-Id: I02f5ec6b99f14ec3568934815c45efb83711d706
Cleanup some workarounds for idempotency, the bug should be fixed since
we updated the version of Puppet.
Change-Id: I429b4159dfd26893b5c488e171b26c5bbae53d7d
This change adds in the ability to test the cell_v2 setup for nova.
Change-Id: I08efa36f5f62ae406ab20e8da77c01fbf3d9b11f
Depends-On: Idfc369e9e17f7d5a30ce4ff52beb604dd4a6ac23
Related-Bug: #1649341
1) Pin fast_gettext
The fast_gettext gem 1.2.0 and greater require ruby 2.1 which we don't
have.
2) Pass a generic cpu to qemu instead of trying to pass host-model
We don't need to pass any CPU extensions in the gate, given the
context where we don't use KVM or (proper) nested virtualization.
Truthfully, this also becomes required as there are upcoming changes
in qemu-kvm 2.6.0 which breaks us when trying to pass extensions
while using qemu.
Worth noting that devstack has been setting 'none' [1] as well.
[1]: 1c442eebc8/lib/nova_plugins/hypervisor-libvirt (L42)
Change-Id: I881bd712266ae63b0ac9074ec18bd6a41b946c3a
The rabbit_host, rabbit_hosts, rabbit_userid, rabbit_password,
rabbit_port and rabbit_virtual_host parameters are being deprecated.
This change updates the p-o-i tests to use the default_transport_url for
the rabbit connection instead.
Change-Id: Icb15ad3b5d75a35afcb685f4ae03c5755031fc3d
Depends-On: I43ac3079da9e52179457ca1a34f4faee60d5617b
Depends-On: Ie63457370ae37ef5af5878b1a97b8bafa80a8189
Depends-On: Ia0c5b5072431bbc186c8b96b7f034e68e9248af3
Depends-On: I43500ac7be4296ea534266fa14ca37fc700b0092
Depends-On: If4dfe8b82518fc4e590731a805ee3115b11f4f47
Depends-On: I00d4782693c1f67b82673624d4d36822c9762303
Depends-On: If83c0f0e61a08061334536399a42767a305966b7
nova::memcached_servers is deprecated and we're already using
nova::keystone::authtoken::memcached_servers instead.
Change-Id: I78fd0a1fa421477129d5c9b2670435da65dce9f4
Configure Aodh, Ceilometer, Ironic and Nova authtoken sections with new
method.
Depends-On: I779cb0b76187f742555c5c77f3d6b97d35c6be28
Change-Id: I8f5b7f3d7392bfa4215fccd7defe953857cef3f6
When running puppet-openstack-integration in certain contexts, for
example outside of the gate, we might want to choose a different
hypervisor such as KVM.
Making it a parameter will allow it to be overridden by hiera
easily.
Change-Id: Ie80740b67561ca21e1e81e9e571e66e8d505d940
Ensure RabbitMQ is ready before trying to start Nova services.
it seems like if RabbitMQ is not ready after 16 attempts, some Nova
services just crash.
Change-Id: Iefbb018ccbecd999a124d760b29a36c9e7e8fcb7
Starting from Newton, VMs can be deployed with a metadata server that
runs with SSL enabled.
This patch aims to activate the tests for that, so we'll have a better
coverage of Neutron in SSL & IPv6 environment.
Note: I could not enable SSl for neutron metadata agent, I found this
bug: https://bugs.launchpad.net/neutron/+bug/1514424 and I'm not sure it
actually works.
Change-Id: Ia5b19d22549fc0c891a46cf47a742b35eaec6276
virtlog and virtlockd services are not idempotent when running Puppet 4.5.
See https://tickets.puppetlabs.com/browse/PUP-6370
This is a workaround and will be reverted once Puppet is fixed.
Change-Id: Ib6be58ad527ec875107578fa10c3d17e1d65133d
1) scenario001: disable ceilometer tests on Ubuntu
Because we need this backport in UCA:
https://review.openstack.org/#/c/318503/
And we won't have it before... who knows?
Let's disable the tests for Ceilometer on Ubuntu until UCA provides a
packaging update.
Let's also disable Telemetry at all on Ubuntu, since Ceilometer can't be
tested until UCA does the backport.
2) add RDO Newton dependencies repository, so we have paramiko >= 2.0
and have a more stable gate.
3) Update RDO trunk URL to have a very recent URL.
4) Update nova endpoints to match with what is required upstream (2.1)
5) Un-pin Tempest, so we keep testing master (our desire in our master
gate but we pin it in stable branches).
Change-Id: I10dc1beeca3ec47fd4c3a866aa3e1d3bc9d0aa81
This is a first iteration of testing IPv6 deployment.
This patch will deploy scenario002 on centos7 by using IPv6 network.
To make it work on Ubuntu, it will require more work, that will be done
in future iterations. During my testing, I found some bugs and did not
investigate because I wanted a first scenario working on v6.
Same for scenario001 and scenario003, they'll be tested later.
Change-Id: Ib28b379645b5a3d351438e0408a9de62b9f48302
* Deploy Self-Signed Certificates for both IPv6 & IPv4 deployments.
* Disable IPv6 for RabbitMQ now, for SSL reasons, will be enabled again
later in a next iteration.
* Deploy Ironic API under WSGI instead of eventlet.
* Switch Glance API, Ironic API and Keystone to SSL.
* Configure Tempest with SSL endpoints when needed.
* Reduce the Ironic tests because of [1].
[1] https://bugs.launchpad.net/ironic/+bug/1554237
Note #1: puppet-swift, and puppet-cinder will require some work to support SSL, so it's not
implemented in this patch.
Note #2: we don't enable SSL for Neutron because of
https://bugs.launchpad.net/neutron/+bug/1514424
Change-Id: Ib2b5289b6f5e82f43cf60dee3152b2c2ddd5a014
