propose-puppet-openstack-rdo-promote was useful to submit changes every
night to promote repositories in Puppet OpenStack CI when deploying with
RDO.
Now, we have the process automated in RDO Infra, and we can use a static
URL that is auto-promoted if RDO CI jobs pass with new packaging.
In other words:
- Before:
Puppet OpenStack was testing OpenStack for a repo older than 24h and
manually promoted repositories every day.
- Now:
Puppet OpenStack will use the same repository that is automatically
updated in RDO every 3 hours if jobs are passing.
We're improving the CI/CD toolchain.
It depends on this change in puppet-openstack-integration:
https://review.openstack.org/382940
Change-Id: I52b6098163bd95aae5d8bc69bf12499fa0ddacc9
Depends-On: Ia1ac4b55f8542d94c68dc761a9af04c447d50e3e
Now that we have support for GPG keys, add that ability to
deb-openstack repository.
Change-Id: I1db01e497f0a97f7e0435a99d89dfd04f8375153
Depends-On: Idc64ac00b41d9a4009036890c8194346a1881350
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
Add an initial job to exercise the signing node, which can retrieve
tarballs and wheels, sign them and upload those signatures. Only use
it for the git-review project to start, while we iron out any bugs.
Change-Id: Ic8b69f5ee2c5f203bb443263bca10218e8d20be7
Do not offline the signing node. Right now we only run tag-releases on
that node.
Add also a note to our REVIEWING list about these long-lived nodes.
Change-Id: I7ddea40069add66dc07ee754092c8c236c62b53c
We don't want to offline our nodes when we run reprepro, since these
are on static workers.
Change-Id: I5ecf83dfea0329e79d59ccdc334c7a64463252c7
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
As discussed on the mailing list [1], merging tags between branches
confuses the history as reno sees it and results in invalid release
notes builds.
[1] http://lists.openstack.org/pipermail/openstack-dev/2016-June/096895.html
Change-Id: Ie09158ca0d522905b229251564ff5141e59e2b6c
Signed-off-by: Doug Hellmann <doug@doughellmann.com>
Background: during a cycle, we deploy OpenStack on centos7 from trunk
using RDO repositories, automatically built at each commit in OpenStack.
Currently, Puppet OpenStack CI manually promotes
puppet-openstack-integration every week, to latest URL. We would like to
automate this process with a daily periodic job so every day we would
have a promotion and we could detect which commit in OpenStack would
break our CI, and we could work on it quicker than usual, because
feedback would be daily.
This patch adds a periodic job, called
"propose-puppet-openstack-rdo-promote" that will execute
jenkins/scripts/propose_rdo_promote.sh script. The script has been
tested and already used manually.
Change-Id: I9b19a440325f609746a197676d3799aa0f60ce66
Do not offline the proposal slave after
propose-puppet-openstack-constraints is run: Enhance proposal regex for
this.
Change-Id: I1b1408fe8a5de46e9df2611d11158562d4723713
The mitaka translatation proposal jobs were turning off
the slave due to 'update-mitaka' not matching the regex.
Change-Id: Ic578a49f0c48289e61d97121e0e3cbf306f0977c
The new propose-devstack-plugin-list job runs on the proposal slave but
doesn't match the regex for reusable nodes. Update the regex so that
this job does not offline the slave.
Change-Id: Ib86875492adb959456d7db1b6866f968365f3df4
TOX_TESTENV_PASSENV is a variable that lists env variables that
tox should allow through to test environments. Make sure that
the ZUUL_ variables are always on the list so that zuul-cloner
may always be used inside of a tox environment.
Change-Id: Iacba1611e5c50fbccd6ae700b3456028b142fc5c
With no longer setting a node default, Zuul won't pass it as a
parameter to the set_node_options() parameter function. Drop it from
the required arguments for this function.
Change-Id: I2aa8f7179ecb0bdf6bd8785653f1726321bbcb7f
Now that there are no jobs declaring both .*-precise and .*-trusty
nodes at the same time, it is safe to remove the associated mapping
in Zuul and let it trust that any worker registering the job name
is capable of running it sanely.
This should not be approved until manual jenkins-jobs updates have
been completed and confirmed successful on all Jenkins masters.
Change-Id: I3b40021ffbf89842ed3a53412dfffa119f311797
There are no jobs called icehouse-dibtest anymore, remove special
handling from openstack_functions.
Change-Id: Id8bf7929af54b3a122766aaf6b391766a9d983d7
The prior change (Ifa6367488165df6cad18725a9191b426a41c08bf) has
disabled F22 nodes.
This adds F23 nodes (just one for inital testing) and modifies the
fedora-22 based jobs to use the new nodes.
(note the depends-on, required for building f23 fedora-minimal, is
available in dib 1.4.0, so this is ready to go)
Depends-On: I1a6e45d04755515286b3d49f8280c16b527e2f48
Change-Id: I3de17c2d670d8cb0f28ffb5e1541751bc7f4d73e
Due to the length of the job name and the tox target we run into an
uncommon limitation; the virtualenv that tox launches is nested in a
path that is too long. This leads to an error on our longest named
job which prevents tox from running properly.
This limitation is the limit for the line length of the first line
in a shell script. In this case <virtualenv>/bin/pip has a path that
exceeds that limit by a few characters. This results in a very
confusing "OSError: [Errno 2] No such file or directory" error.
See `man execve` for more info. A quote from that manpage:
'A maximum line length of 127 characters is allowed for the first line
in a #! executable shell script.'
To reproduce this locally you can just create a very long path and
then run tox in it. Example: http://paste.openstack.org/show/478559/
Change-Id: I51f8f67d9ab139f1d539fce05ce3cc4766aeedff
Depends-On: I43fba2a5ff1890d699045496c9eaee5e849f3e75
clarkb suggested dsvm be in the job names for Kolla to follow the
job naming convention. In the case of the experimental job we have
recently merged, that it as a requirement for the name of the job.
To keep consistency within the project and within the OpenStack
community we are renaming all of the gates that use the dsvm's to the
convention that all other jobnames follow.
Since the tox paramaters do not change in the jobs/kolla.yaml, this
requires no change on the Kolla side.
Change-Id: I9a22434d99b4deb7936ecaf7b714d0dc41666367
A recent change [1] added new gates which arent doing their job
properly due to not matching the regex which this patch fixes
[1] I2c94eac72c61385fe6472a95ea3c9cafa33bc22d
Change-Id: I3459d7f36cd2c31810f2df9293b599012ececdab
Sadly our fedora names are not constent yet :(
Change-Id: I9fa840975b2f3136e9d52dae352f33e692296694
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
The tripleo-ci jobs were using a regex that just looked for
*-tripleo* in the job name, which also matches tripleo-common and
python-tripleoclient. This causes the py34 jobs for those projects
to sometimes get scheduled to precise nodes that don't have a py34
interpreter installed.
This changes the tripleo-ci jobs to be named tripleo-ci-* and
updates the regex appropriately so it will no longer match non-toci
jobs.
Change-Id: Ie4d803574221b936b249673436db36fd75940587
We need to be explicit that we want the npm upload job to run on the
release slave otherwise gearman will attempt to run it on the default
slave label which will fail because that job won't be registered.
To do this rename the job so that it matches existing convention which
makes the regex easier to read/write. Then update the reuseable slave
regex to include the job which will force it to run on release slave
without offlining the slave.
Change-Id: I26355680f2d8fc3fa5f514370d1586deb701c21c
Add fedora 22 node matching to openstack_functions.py. Although we
could make the re even less specific, I think it's best to just match
on what we have.
This is my own silly fault for not noticing this in
I2be12566985ac992a2da0c8174b4c9064347d91f
Change-Id: I98eecf21b8cee5114ffccef4d23384a4c4cc6774
Enhance regex to handle the translation liberty jobs like
nova-propose-translation-update-liberty.
Change-Id: I592443baa981b44437e09209b933f69a45a85b7c
Now that the Zanata translation jobs have had their -zanata suffix
removed, we no longer need this section of the regex to force the
jobs onto the proposal slave.
This reverts commit e1cdfc4a42450e01d40b2eed5b63ae5e4a0a76ad.
Change-Id: I6be79c60b480f21904312289c5f00f88c3a956d6
The jobs did not run on the proposal node since the regex did not catch
them. Enhance the proposal re.
Change-Id: I12fe718f0fb82d7036a0fe0c7e0233e16ebc8173
In Id3f6175fcdd4a21019eabca56cd337145ff9c2b0 os-ansible-deployment moved
into the openstack namespace and was renamed to
openstack/openstack-ansible. It's specs repository will also be renamed
to openstack/openstack-ansible-specs.
This also switches the publishing of documentation from readthedocs to
docs.openstack.org and includes both doc and spec publishing.
Change-Id: I10588f1118f21793ef9a1a647184cf2e8de07260
This patch adds an ubuntu-trusty node-based non-voting
check to os-ansible-deployment to help determine whether
the differing images between hpcloud and raxcloud are
contributing to the gate check failures seen solely in
hpcloud.
The idea is to use the exact same image across both
clouds to see the results.
Change-Id: I7809e3f4b3012586f4aaa0de9d9bc7b0011e720d
Since we have zuul picking node labels to accomodate for possible branch
specific label needs we need to whitelist the kolla image build jobs as
they always have a single build_host:image_type mapping and branches
don't affect that.
Change-Id: I75e20a53c2091df1571eeff89cb6fd2e927e0a91
Our RTFD hook just does a curl request to upstream. As such, it is
very expensive to nodepool to build a node for the purpose of just
using curl.
Lets move the job to use the proposal node which is long living.
Change-Id: I8f407fdaeb890a4ae6a9b99caefe836b8f2f7819
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
We're really close on the ubuntu-trusty images. Add in an experimental
job so that we can verify before re-switching all the multi-node stuff.
Change-Id: I5311fb8b67abb644435033f2a88ea3dd9453d87f
This reverts commit 2b5b800d4fc48724da66e72e03812222125d3667. These new
images cannot be reliably uploaded to rackspace which makes iterating on
image updates/fixes very difficult. Switch back to the old image (using
new shorter hostnames) to get multinode working again. We can switch
back when we have uploads sorted out.
Change-Id: I46e051a4a566ae400c2fc153ad72596c853af32c
Because we want to move to it anyway, and it lets us fix the hostname
problem with glean.
Depends-On: I67ea62ebf9360f72f121d77acd1cbd35d8db728b
Change-Id: Iabe8b0f0cac9d80180b9542263ade5ffdfe641a2
This job doesn't have "dsvm" in the name, so zuul is calling
default_params_trusty() on it rather than devstack-params(). Correct
the functions to match this expectation.
Change-Id: I25e0e638d3d0dc1be66dfc8e17712ca995aa3a3a
Update the experimental nova Python 2.7 unit test job which runs on
thin "ubuntu-trusty" workers to use bindep for determining its
distro package dependencies, and rename it to something more
appropriate now that it's not actually using devstack-trusty workers
any longer. Also switch it from gerrit-git-prep to the new
zuul-git-prep builder macro, and while at it adjust zuul-git-prep to
have a zuul-branch-git-prep variant rather than requiring a branch
parameter for the more common case.
Change-Id: Ib3f4844c1203f8a3890d9e5d91347d4507b5d08b
This is mainly cleanup to get rid of the old misleading slave label.
The old worker already has the new label applied, to make switching
more seamless.
Change-Id: Iefd6ed3fbb696b2009e8525c4cac1f0e1aca50e3
The new nodes are running in both HP and Rackspace now, so it's time to
make sure that they can run devstack.
Change-Id: I77704f2a3ce79851df677d2c38cd7ad504427a97
