Age | Commit message (Collapse) | Author | Files | Lines |
|
JIRA: RELENG-254
Change-Id: Ibfda27a637ee448862ae2df5884b88093210ed4d
Signed-off-by: Trevor Bramwell <tbramwell@linuxfoundation.org>
|
|
|
|
The following alpine containers will be launched one after the others:
- healthcheck
- smoke
- features
- vnf
Old container is kept for suite (which could be refactored later)
Jenkins exit condition (return_value) and push_test_logs to be checkeded.
Regarding the return condition, I did not reapply the mechanism that returns false
when one of the test suite is FAIL. This was misleading for lots of people as
for them, red means functest is FAIL.The argument was, if functest jenkins status
is blue, people believe that tests are PASS and do not consider the results, which
may be FAIL, which may be also misleading...
Due to the refactoring I think we can be more Functest Centric and consider
that if we succeeded in running the test, our CI status shall be PASS (same
behavior than other test projects..HA test FAIL in yardstick do not prevent blue
status).
Scenario owner must have a look at the reportign page to get more details.
Functest PASS means that tests have been succesfully launched, so they know that
there are available results
To summarize
previously Functest OK but 1 or more tests FAIL => red status in Jenkins
currently Functest OK but 1 or more tests FAIL => blue status in Jenkins
Regarding the push to artifacts, it should be transparent (volume) but the last
log links listed on huawei-virtual5 are broken [1]
[1]: https://build.opnfv.org/ci/view/functest/job/functest-alpine-compass-huawei-virtual5-daily-master/27/consoleFull
Change-Id: I4ed036dda3daae3d46726e5946f9d04897bab9bb
Signed-off-by: Morgan Richomme <morgan.richomme@orange.com>
|
|
|
|
Change-Id: I63e361eebe1345a34086ecbe99aba4504cb0f5bc
Signed-off-by: xudan <xudan16@huawei.com>
|
|
This change fixes the following issues:
1) 'nounset' will cause the script to fail when activating a virtualenv
since the activation of a virtualenv unsets variables
2) There are not artifacts available from releng-merge-jjb and copying
them would save minimal time over re-running jjb
Change-Id: Ifc4297947c8fe235987a507ead932841480165a7
Signed-off-by: Trevor Bramwell <tbramwell@linuxfoundation.org>
|
|
Change-Id: I34a555697d963efc38064a7b53f25191d6bf1382
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
|
|
|
|
Previous commit extended the SUDO variable to include '-E', but
failed to properly quote the full `sudo -E` string, resulting in
"command not found" errors.
Change-Id: I266aa5259f09c3605e1c5a7ad165397ce83be0bb
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
|
|
Due to bug with forbidden file paths conflicting with file paths this
explicitly defines all the file paths in apex. This isn't the greatest
solution in the case where someone adds a new folder, but chances are
they will also modify one of the other files as well. Note it also
fixes a bug where unit tests woudl run twice because it would be
triggered as part of apex-verify-master and apex-unit-tests-master.
Change-Id: I6b10f8eb70c2cd7c8dd5c01363e06b7458c00127
Signed-off-by: Tim Rozet <trozet@redhat.com>
|
|
|
|
|
|
|
|
Functest daily is based on the functest alpine dockers:
- healthcheck
- smoke
- features
- vnf
Change-Id: I922f3a5946e8bc06229fb03a959abd2bfead132b
Signed-off-by: Morgan Richomme <morgan.richomme@orange.com>
|
|
|
|
|
|
1. Jobs on Compass often timeout
2. Apex jobs on huawei-pod4 will trigger 2 jobs at the same time. They will be
conflict.
JIRA: DOVETAIL-491
Change-Id: I5297bfb01c45c15cce7ab1438d8eed3dd9c7ff07
Signed-off-by: xudan <xudan16@huawei.com>
|
|
AArch64 build runs inside a VM, which tends to be slow.
Until we move all AArch64 builders to baremetal, let's increase
the timeout from 30 min to 60.
Change-Id: I0e16d32ac1cbe8f13b24e3cadb87715f12360d7a
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Change-Id: I19fff409b5e8335ff44f0d9ba8167d3e882a9ad5
Signed-off-by: Tim Rozet <trozet@redhat.com>
|
|
Each time a patchset is created that has an effect on Jenkins jobs, a
non-voting job will trigger that comments to Gerrit which jobs are
added, modified, or deleted.
This should help developers judge the overall impact an individual
patchset may have on Jenkins since JJB macros can be expanded to
any number of jobs.
Change-Id: I99f9b68bc7933abfa39a941d6e65a73f62e58e56
Signed-off-by: Trevor Bramwell <tbramwell@linuxfoundation.org>
|
|
Change-Id: Ide3c39d508d26be40e73cec13cd00eace984b6cc
Signed-off-by: Tim Rozet <trozet@redhat.com>
|
|
SSH_KEY and other env vars set by Jenkins should be preserved
when using sudo to invoke the Fuel deploy script.
JIRA: FUEL-280
Change-Id: I495589a4a0519b3a5052cd5e5dc3122f9e49c215
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
JIRA: FUEL-280
Change-Id: I95559dcafbd179e5b7305dd0f23a7b3d89056888
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
|
|
|
|
|
|
|
|
this should make forbidden-file-paths work
Change-Id: I18d63bffb92abd916eae54c7aafe9a8e1e98926f
Signed-off-by: agardner <agardner@linuxfoundation.org>
|
|
Change-Id: I360cd5ba894022cd30cf7237e8338ab2e213c940
Signed-off-by: Tim Rozet <trozet@redhat.com>
|
|
JIRA: BOTTLENECK-158
It seems the build_tag is not automatedly passed to docker.
This patch is to try to pass build_tag to Bottlenecks docker.
Change-Id: I783d6f0702040449b0385d11bca709f8cdbdd58a
Signed-off-by: Yang Yu <Gabriel.yuyang@huawei.com>
|
|
Temporary solutions relied on ollivier/functest-* but
now opnfv/functest-* are well produced and can be used as official
reference for CI
Change-Id: I02b5a03929d5b31df0db2a6c4722edf62b920df9
Signed-off-by: Morgan Richomme <morgan.richomme@orange.com>
|
|
|
|
JIRA: FUEL-279
Change-Id: Ie87ecaee17d6bb29776c3056274c59c1264762de
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Using include-raw-escape inside macros adds extra curly braces, which
breaks bash scripts.
JIRA: FUEL-287
Change-Id: If1e5f528e76aff78d2c9b99530b72070b6731f75
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
* changes:
fuel, armband: Move GS_URL to installer params
fuel, armband: Move SSH_KEY outside /tmp
fuel, armband: verify jobs: Enable virtual deploy
|
|
We only care to trigger builds based on changes under xci/, not
including xci/README.rst
Change-Id: I347e0d7ef658c41322b7c3e25740fa4d2f960bdc
Signed-off-by: Trevor Bramwell <tbramwell@linuxfoundation.org>
|
|
|
|
There is no 'README.rst' at the root of the repo, but there are
README.rst under xci, bifrost, etc. Changes to these READMEs shouldn't
trigger builds.
Change-Id: I6a09388cf40b35211d58413b45b5ac8c596a2dc7
Signed-off-by: Trevor Bramwell <tbramwell@linuxfoundation.org>
|
|
Change-Id: I3b62e855be9205326fee1fa8b2b56cb2276c689c
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Change-Id: I2754c7d5502a05ef0fa1dec9dbd14b4c97b70f11
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
|
|
|
|
Change-Id: I0c6e54ce4caf0b6889bac62b25afcee2706ca75f
Signed-off-by: Tim Rozet <trozet@redhat.com>
|
|
Recent rework changed the label for the verify jobs from
"opnfv-build-ubuntu" to just "opnfv-build", trying to make it
more OS-agnostic.
Since verify jobs fail after the rework, try reverting the
label change, hopefully it will fix the failure as well.
Change-Id: I15d139d8dc2d2f940a1af2c1b5fb218d6be19f82
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
If we make changes to apex/tests path it is only unit tests and we do
not want to trigger apex-verify-stream.
Change-Id: I42ef56330883ea6e9643f24567eaee3d5fa36071
Signed-off-by: Tim Rozet <trozet@redhat.com>
|
|
Change-Id: I33f719578baadc70571832c5b7e7f146516c3020
Signed-off-by: Catalina Focsa <catalina.focsa@enea.com>
|
|
Commit 27f5d5c refactored the functest verify jobs, but also
introduced an issue triggering said job by not passing project
and branch parameters to the new trigger macros.
Change-Id: I7a61a1f85df345c090813b5060959fa79d294f9b
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Change-Id: I59523b6233ca48c2eb2d4b4934263f0114d8e7a8
Signed-off-by: Tim Rozet <trozet@redhat.com>
|