Age | Commit message (Collapse) | Author | Files | Lines |
|
While at it, tweak regex for verify job to match 'recheck' right
before EOF.
Change-Id: I3e8e1eeaf2ca947425fa2443b29e150c808eea07
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
- disable functest smoke on arm-virtual* PODs, since we don't perform
a full deployment on those PODs, only a basic integrity check;
Change-Id: I0bcd6da32118ea256652dc74977ef0ee7f2525ab
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Duplicate fuel-verify jobs based on cluster type (virtual, baremetal)
and allow (re)checking changes using a specific scenario:
- ALL changes will trigger a virtual POD CI cycle (build, deploy,
smoke) on submit, rebase etc., just like they used to;
- change comments containing a substring formatted like:
* `recheck: os-odl-nofeature-ha`
* `reverify: os-odl-nofeature-noha`
* `recheck`
will trigger only virtual or baremetal deployments of the required
scenario (default os-nosdn-nofeature-noha) on:
* virtual PODs if scenario is NOHA;
* baremetal PODs if scenario is HA;
- limit arm64 verify job running on arm-virtual2 to the installation
of infrastructure VMs and skip cluster setup to prevent POD time
starvation;
Limitations:
- HA scenarios can only be verified on manual request via a change
comment (e.g. 'recheck: os-odl-ovs-ha');
JIRA: FUEL-390
Change-Id: I6368cfd045180d6a7202b8d817b17a4b123ba2af
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
- cleanup deploy tooling for old-style artifacts (ISO-based);
- remove danube stream, related jobs and hacks required for them;
Change-Id: Ib2486173df1ec83f2dbc08a5e9faf950bffb4b06
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
We don't currently support ONOS and KVM-based scenarios.
Change-Id: I9869498c1f726293d571562291e850a817bf3e18
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Change-Id: I64aae5719f1a0a667e56722bab9e5012a813571f
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
|
|
This reverts commit b4a65c5e6777a53be5ea817c967caeee51285a7f.
Chain of DOCKER-ISOLATION already has required rule.
Change-Id: Ic7d3e93a2814a42764b4822ef96e0841f68b774d
Signed-off-by: Michael Polenchuk <mpolenchuk@mirantis.com>
|
|
JIRA: ARMBAND-401
Change-Id: Icd32874ecb38c3f7469e71bd461794b65f50494c
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Docker build automatically appends 'amd64-<TAG>' or 'arm64-<TAG>',
which are normally combined using manifest tool. However, for verify
we don't push a manifest, so use the explicit arch-specific tag.
Change-Id: Idc036216919164e71d3da5032a79f7f47d115710
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Change-Id: I6b641c87e2672a9e879865baebfa3a58129823f7
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
|
|
Rework Fuel verify jobs:
- split 'fuel-verify' into per-arch jobs (amd64 and arm64):
* for now, arm64 jobs are non-voting;
- add 'docker-build' phase job:
* the docker build job will run on the same virtual POD jump host
as the deploy, to bypass the need of pushing to a (public)
Docker hub registry - this might be refactored later to resemble
the old Fuel ISO build workflow where the artifacts built during
verify are the ones actually promoted - with the current method
the final build & tag (ran on OPNCI CI build servers) might yield
different results due to different docker build cache states;
* the locally built container image is tagged with the 'verify'
designator and not pushed to a public registry, only consumed
during verify's job 'deploy-virtual' phase;
JIRA: ARMBAND-401
JIRA: FUEL-383
Change-Id: I9b0d8079e759cdc895260219a8a4989411c995ff
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Change-Id: I2d0a71fc40a4a6cd1f5b1e63ca429c5b83b7ad00
Signed-off-by: Cristina Pauna <cristina.pauna@enea.com>
|
|
Keep running Fraser CI scenarios so Dovetail gets to run against its
intended baseline.
To keep POD usage down, schedule only baremetal HA scenarios,
3 Fraser scenarios per week.
JIRA: FUEL-381
Change-Id: I0cedf4ed38a146b4e6ad0e10bd3030ab0c1e833f
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
|
|
JIRA: FUEL-383
Change-Id: I80049726c256d2891ae1cfbacbc54781d3d81c70
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Added the triggers for both ha and noha versions of the scenario
Change-Id: Ibba782d82712b69b755592a878fc69bf6696cba7
Signed-off-by: Cristina Pauna <cristina.pauna@enea.com>
|
|
Change-Id: I24cc328cdefcd7ac5f34292f0cf1635f797f8f00
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Now healthcheck includes vping_ssh, so remove it as a dedicated test
for verifying patches on gerrit to avoid double checking.
Change-Id: I4daec677e7d1fdc65997f7a9ab580d5f93ece8a8
Signed-off-by: Michael Polenchuk <mpolenchuk@mirantis.com>
|
|
Change-Id: I6a2b571ca2696f9466a4588b7e4deef13b947eae
Signed-off-by: xudan <xudan16@huawei.com>
|
|
JIRA: FUEL-375
Change-Id: I01f17a28fdb1bd692d1997b0a32be66de8856a8b
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Functest Docker has issues connecting to Openstack clusters on
Ericsson virtual PODs due to iptables rules interfering.
Apply old iptables workaround from functest, at least until we
rootcause and fix the issue on those virtual PODs.
JIRA: FUNCTEST-985
Change-Id: I01b6788e5dd7a9fd64e80e3e78d7cd8828570cb1
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
|
|
Change-Id: I2946684b8e5941201865c123793b7f5800b5421b
Signed-off-by: Michael Polenchuk <mpolenchuk@mirantis.com>
|
|
We will try to reuse the job template
for deploying from armband project.
Change-Id: I76a783094ce5be1b4aa93585a6b90898d909ca13
Signed-off-by: pava <Paul.Vaduva@enea.com>
|
|
global-jjb only supports the .yaml file ending for jjb jobs. Instead of
waiting for a release we're going to rename the files.
Change-Id: Icf3339eacd2320c583333e02250998cf6b1881f7
Signed-off-by: Trevor Bramwell <tbramwell@linuxfoundation.org>
|
|
|
|
Using these macros for deployment jobs will report the deployment
results of Fuel to the testresults database[1].
[1] http://testresults.opnfv.org/test/#/deployresults
Change-Id: I353631f073578b4a2da0f08ffba4042931e65dca
Signed-off-by: Trevor Bramwell <tbramwell@linuxfoundation.org>
|
|
Enable Functest based gating for jobs:
- fuel-verify-*
- armband-verify-*
This includes Functest suite 'healthcheck' and testcase 'vping_ssh',
similar to Compass4NFV verify jobs.
While at it, remove 'basic' phase from verify jobs, as linting is
handled via separate jobs on Jenkins master and there's no other
reason to keep it around.
JIRA: FUEL-340
Change-Id: I11dce6998d7ac16e68cf2b9d72a1ed6b23ca2483
Signed-off-by: ting wu <ting.wu@enea.com>
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Change-Id: Ib3fba4d6193cfa4dfa422163561cfbc8bcfa087c
Signed-off-by: jenkins-ci <jenkins-opnfv-ci@opnfv.org>
Signed-off-by: Trevor Bramwell <tbramwell@linuxfoundation.org>
|
|
Change-Id: I90f42dc36ed0555788b517572088b5f09c1569fb
Signed-off-by: Fatih Degirmenci <fdegir@gmail.com>
|
|
|
|
Change-Id: I5823f763d2f321a0b2505ca3ecd042e928b5a42c
Signed-off-by: xudan <xudan16@huawei.com>
|
|
While at it, cleanup useless ovn/Euphrates jobs, we didn't have OVN
in E release of Fuel@OPNFV.
JIRA: FUEL-329
Change-Id: I4a9b476c292a05cf0b9cadaa3102ebccf2baa10a
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Fixes: af7df31
Change-Id: I6369ded12a50a2aa7f0268488c5511fba9732a99
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Fixes: af7df31
Change-Id: I13e20591ce82792a12a514ae62e3b3dd4ed50e65
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
1. run test suite default on Fuel on Tuesday for HA scenario
2. run test suite proposed_tests on Fuel on Monday, Wednesday and Friday
for HA scenario
JIRA: DOVETAIL-613
Change-Id: I67ef5d7dddfee01817b933ae3eeccf704924f9b0
Signed-off-by: xudan <xudan16@huawei.com>
|
|
Change-Id: I3fd11c6efb13f46ec74a0237a180497397cdfefe
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Delete Fuel Euphrates CI job on zte-pod1
Change-Id: I93d5ed7b8aa40715fe8e7541d810ab0ee9a89f28
Signed-off-by: zhihui wu <wu.zhihui1@zte.com.cn>
|
|
|
|
Now that vPDF support landed in Pharos and Fuel repos, disable the
'virtual_kvm' generic override, so we can load POD-specific PDF/IDF.
JIRA: FUEL-322
Change-Id: I0964220a25d6fe30afcbeb3ad1f020ae76d38e5a
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Only Danube still uses this param, newer branches rely on IDF.
Going forward, the `-B` deploy arg will be deprecated in Fuel.
Do not yet remove BRIDGE slave params from non-CI PODs, as
other installers (or custom Danube jobs?) might still use them.
Change-Id: I6d259299d1e13e8a5da732aca60b1f56b8472af6
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Instead of cloning securedlab locally and passing it to Fuel deploy
script, rely on Pharos git repo, embedded as a git submodule in Fuel.
Keep old behavior (using securedlab) for stable/{danube,euphrates}.
Depends-on: Ib04e4fb384568a6efd9e78a080857b663521ae88
Change-Id: Ie30879292ae04b641257760de69c900a9eb4ef84
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
|
|
Change-Id: I67574f37a2bb4908b935fdc3e72ecf891ad66def
Signed-off-by: Michael Polenchuk <mpolenchuk@mirantis.com>
|
|
Dovetail only ran against fuel danube release before.
It needs to run against fuel euphrates and master because Dovetail team
is planning the 2nd release which is based on OPNFV Euphrates.
Change-Id: I2c39d53a16ca6dbf8eda31943f09896cf7ee35c6
Signed-off-by: xudan <xudan16@huawei.com>
|
|
- Delete Danube job on zte-pod1
- Modify the triggers on zte-pod1
- Delete master and Euphrates job on zte-pod3
Change-Id: I034dc3b968ee9255aa014c28dc08237fc0803d32
Signed-off-by: zhihui wu <wu.zhihui1@zte.com.cn>
|
|
While at it, limit functest & yardstick to master and
stable/euphrates, trying to work around missing job references
in Danube scenarios on ZTE PODs.
Change-Id: I89de0c3b3c914e1ad735980a995e5cae625238be
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
On rare occasions, Fuel/Armband verify jobs can get into a deadlock
on nodes with exactly 2 executors.
Add a build blocker to prevent such scenarios.
Change-Id: Ia00d88df0dd033264eaeb66e27188c62514b74d8
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|