Age | Commit message (Collapse) | Author | Files | Lines |
|
Change-Id: Ifcf1d4bd48b64bc9175c3b0d594f897f4f3f5e7d
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
It also fixes few docstrings
Change-Id: I10bae7b7fab40edd931dd55ab8859946e3b38a71
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
It conforms with the flavor extra specs model [1].
[1] https://gerrit.opnfv.org/gerrit/#/c/60713/
Change-Id: I1555a313aaf56ae47f9ad2a4529eb8b94e636c00
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
It takes all new testcases into account.
It allows running vnfs vs dpdk scenarios too.
Change-Id: I789f6bb1c3df891b1a94f2a674c15085f81f2f23
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
XCI will be able to deploy this feature
on top of pharos lab with GPU so we have test it
This test can be improved in the future with OpenCL tests etc
| DEPLOY_SCENARIO | os-nosdn-vgpu-ha |
+-------------------+-------------------+------------------+----------------+
| TEST CASE | PROJECT | DURATION | RESULT |
+-------------------+-------------------+------------------+----------------+
| vgpu | functest | 01:00 | PASS |
+-------------------+-------------------+------------------+----------------+
Change-Id: I50618f47097a0257b1834b3966054d20458715df
Signed-off-by: Valentin Boucher <valentin.boucher@kontron.com>
|
|
It makes difficult to find the issues (floating ip, boot, ssh
connection, etc...) for a small benefit regarding the current
testcase implementations.
It increases the default floating ip timeout because it has been
reached when running Functest multiple times in parallel. It may
be the root cause when running singlevm vs XCI (in that case,
neutron api and agents are hosted in a nested vm).
Change-Id: Ia97e04902321644c207c59987283754318aee5e5
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
|
|
It now runs vtmp via a new admin member. Else it fails due to cidr
already used.
It only creates a router and both tenant networks are still created
by VMTP.
Change-Id: I05ed41a3b7a01ee97c720176c404929bf8ea97a0
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
It completes "Switch create_server timeout to default value" [1]
which forbids waiting more that 180s (create_server was asynchronous).
[1] https://gerrit.opnfv.org/gerrit/#/c/60423/2
Change-Id: Icbf9ee48024fe9e10571c2bb7ca895d5828fb561
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
Change-Id: Ibe131e0cf02493f0f3a82c1e179848605bd0a292
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
The timeout has been reached when running multiple vnfs in parallel.
Only few additional seconds seemed required here.
Change-Id: I0d6e38e827c581edb617bb0cc13249645468a2aa
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
Change-Id: I152ac898661a0e99a6232df86689c70c468ceaa5
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
60s was successfully reached when running all containers (3
containers * 3 branches) in parallel.
It also removes the wait attribute (the current impl will meet
a possible next asynchronous design) which is useless here.
For the time being, only cloudify needs an extra timeout.
Change-Id: Ic7741b4c1a3223be076fb69ef70a3cb6c05456c2
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
All the vnfs are now inherating from scenarios.
Change-Id: I63509102067676ca0676b773db542086caa02d84
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
40 Gb is enough
Change-Id: Iae6a72fcf91fda67bbd036f478e344478d48264f
Signed-off-by: Valentin Boucher <valentin.boucher@kontron.com>
|
|
It tries to prevent race conditions when running cloudify_ims.
Traceback (most recent call last):
File "/usr/lib/python2.7/site-packages/functest/core/singlevm.py", line 411, in run
if not self.execute():
File "/usr/lib/python2.7/site-packages/functest/opnfv_tests/vnf/ims/cloudify_ims.py", line 120, in execute
self.cfy_client.secrets.create(k, val)
File "/usr/lib/python2.7/site-packages/cloudify_rest_client/secrets.py", line 86, in create
response = self.api.put('/secrets/{0}'.format(key), data=data)
File "/usr/lib/python2.7/site-packages/cloudify_rest_client/client.py", line 263, in put
timeout=timeout)
File "/usr/lib/python2.7/site-packages/cloudify_rest_client/client.py", line 233, in do_request
'SSL, but the client does'.format(e)
ConnectionError: ('Connection aborted.', BadStatusLine("''",))
Change-Id: Id2603bbda462b50d3140b6f45894c8810d542e4c
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
It allows checking when cloud-init is finished (see heat_ims).
Change-Id: I951aabff29457d7b7544a809f9fce1f83706035e
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
It reverts the subnet pool to 192.168.120.0/24.
It avoids conflicts with default CPE private subnets.
It also increases cloudify spawning timeout.
Change-Id: I70183cede0690c2cbc74f9c30a38892d829db2c9
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
It allows testing controllers which implements other RBAC policies.
Defining NEW_USER_ROLE=admin will easily raise side effects for
orchestrators such as Juju and Cloudify.
Change-Id: Ia78a25381aeb8aa81fd1872661629e1c08432b8f
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
It takes times to prepare disk on compute nodes (spawning).
The default value is enough for all cirros VMs but it has to be
increased for all cloudify VMs (4.1G).
Change-Id: I3f1fc0142a73c586e9219a2718bfc467cc48e275
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
Cloudify is considered as running by the client but sometimes it
failed when listing secrets right after. It adds that check in
cloudify.py.
Traceback (most recent call last):
File "/usr/lib/python2.7/site-packages/functest/core/singlevm.py", line 391, in run
if not self.execute():
File "/usr/lib/python2.7/site-packages/functest/opnfv_tests/vnf/ims/cloudify_ims.py", line 119, in execute
secrets_list = self.cfy_client.secrets.list()
File "/usr/lib/python2.7/site-packages/cloudify_rest_client/secrets.py", line 113, in list
response = self.api.get('/secrets', params=params)
File "/usr/lib/python2.7/site-packages/cloudify_rest_client/client.py", line 252, in get
timeout=timeout)
File "/usr/lib/python2.7/site-packages/cloudify_rest_client/client.py", line 233, in do_request
'SSL, but the client does'.format(e)
ConnectionError: ('Connection aborted.', BadStatusLine("''",))
Change-Id: Ie1b25ff529cfbf9f7476821499da369a46126e81
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
It also defines 2 new methods to avoid duplicating code:
- TenantNetwork1.get_public_auth_url()
- VmReady1.publish_image_alt()
Shaker and Tempest are updated as well.
Change-Id: Ie9095c57c580b95f067d5a8275ad71818b4df335
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
Apex creates member as default role which conflicts with the default
swift_operator_role [1]. It detects the default role in lowercase as
well and write it in rally.conf if required.
[1] https://build.opnfv.org/ci/view/functest/job/functest-apex-virtual-suite-master/116/console
Change-Id: Ic4dc98c793207fc32ff35852472742c92db6699c
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
Else it raises side effects when both project.id and project.name are
defined (see Apex [1])
[1] https://build.opnfv.org/ci/view/functest/job/functest-apex-baremetal-daily-master/148/console
Change-Id: I98665ed55a55fdd02e25d8ad17b78d84548b0ec8
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
It completes "Try to get Member or member role" [1][2].
[1] https://gerrit.opnfv.org/gerrit/#/c/59661/
[2] https://build.opnfv.org/ci/view/functest/job/functest-apex-baremetal-daily-master/144/console
Change-Id: Ib38b513906b08d6f85834e03f5415e83b66b733b
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
It checks all openstack resources before cleaning.
It allows removing images even if the test fail quickly.
Change-Id: I1763b8a70fa7ffeb03c435e8e775c7afd1a7adac
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
Change-Id: If63c228f2844194279b560515f4ab24a4ee88be0
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
Keystone roles are case insensitive and Apex declares member which
couldn't be detected. Then conflict occured when creating the
mandatory role [1].
It now searches the default role in lower case too.
[1] https://build.opnfv.org/ci/view/functest/job/functest-apex-baremetal-daily-master/141/console
Change-Id: I687ecfe768ab4717fdb42a135c4bf9b76733ffd9
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
It eases deploying any vnf via Cloudify.
It also stops duplicating code between vims and vrouter.
Co-Authored-By: Valentin Boucher <valentin.boucher@kontron.com>
Change-Id: I5fe06d91804f4781f6b1301da12862bac793ffbb
Signed-off-by: Cédric Ollivier <ollivier.cedric@gmail.com>
|
|
It uses keypair name instead of id as expected by Shade.
It also increases timers as it boots an ubuntu VM.
Change-Id: I8cd131e7c7a94c8f6b0fcc54a4d75ef8af31b807
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
|
|
It seems there is no common role configuration between OPNFV
installers. Then it creates/removes the expected role if required.
Change-Id: I38a92e2522a401f46a59813b9e9443a11d6c7060
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
It also fixes clean().
Change-Id: I42d1305d5cdb5a717cfa7fce26798ac7102310f7
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
Role name _member_ is no longer available
The test returnes OpenstackCloudException that
role _member_ is not found
Change-Id: I37c8dfe6c4da04009ccabbc49e12c2123d92832b
Signed-off-by: Delia Popescu <delia.popescu@enea.com>
|
|
https://build.opnfv.org/ci/view/functest/job/functest-apex-baremetal-daily-master/127/console
Change-Id: I233cef05560eba5f238b0eea3e2c56cb3c882931
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
It now leverages on the new user created and gets the public endpoint.
Change-Id: Iead106e614a73708cb13e33b49a2f438766a3056
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
Change-Id: I71f79b9ac8314ac979d4b1009f662baa7c93a4fd
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
Change-Id: I070e1404a4b0dba4733f38e2de32e1fa77254ddc
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
https://build.opnfv.org/ci/view/functest/job/functest-compass-baremetal-daily-master/172/console
Change-Id: I0fd3bb833d306626c0eb6393fff3d0fca0579100
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
os_client_config creates a singleton which forbids the previous
processing. It also cleans remaining resources and allows
publishing public images (needed by Tempest)
Change-Id: I3a0ad8664b263d3ee8599efe2973465006d5750b
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
It will simplify code in juju_epc, tempest, etc.
Change-Id: I195ae1d7d75f1c2f2eba76ae3c5307153e9b9759
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
They ease creating all ressources to execute command in a new VM
reachable via ssh. This scenarios should be reused by other
more advanced ones (shaker, vping, cinder, etc.).
Change-Id: If23866827c64a7afbdd8ff596dbc33710ee4ee20
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
They ease creating all tenant network ressources.
These scenarios should be reused by other more complex ones.
Change-Id: I4ce1872065bde92408d34f2ca8823ce98275d8d5
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
It also print snaps_creds to ease debugging [1].
[1] https://build.opnfv.org/ci/view/functest/job/functest-joid-baremetal-daily-master/22/console
Change-Id: I5b0d0dcffe3e45d8fa7d1d7d8ce19172fc42aaa1
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
It removes all the files which have moved to Xtesting.
Vnf inheritances and env management will be improved in other
changes.
It keeps the same tree thanks to a symlink to allow publishing
artifacts.
Change-Id: I551bbd3f344cdab0158a50b7b09e541576695631
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
|
|
For the moment, VNF works well only if
OS_PROJECT_DOMAIN_NAME=Default
OS_USER_DOMAIN_NAME=Default
It's now possible to use domain != Default
JIRA: FUNCTEST-935
Change-Id: I779a93538662ff32906d91d7cf3b2c5c3e872765
Signed-off-by: Valentin Boucher <valentin.boucher@kontron.com>
|
|
env_file can't be defined in a config file simply because it's an
entry point (Jenkins jobs, end users)
Change-Id: Ie23c0ef90efc839d60f0f3a9754c58746b1f3a00
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
It defines env.get() as an unique way to get Functest env vars or
their default values. It can be considered as a wrapper above os.environ.
It enforces backward compatibility via CONST which mustn't be used
for that purpose. It should be noted that it also stops using CONST
for getting OpenStack env vars.
Change-Id: I333dc1afbc0123166a7eaff8b551370098efa341
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
All vnf testcases have to add suffixes in OpenStack resources names
simply because the password is now generated before any run [1].
Snaps mainly creates the resource only if unfound in all tenants [2].
Then one remaining resource could easily raise permissions denied.
It avoid conflicts between vrouter and cloudify_ims which are using the
same names.
[1] https://gerrit.opnfv.org/gerrit/#/c/50747/
[2] https://jira.opnfv.org/browse/SNAPS-257
Depends-On: If56cc44fe35ca06bc9033af620ddd3c0586877fe
Change-Id: I1800886e575189b64fdc580a5ee6b59c81fcde5c
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|