This reverts commit bf25cc9e7b.
The pin of pip/setuptools to such a low version is now causing the
gate jobs to fail to install packages, since it doesn't understand new
environment markers we use. This probably opens Pandora's box of
issues that we never quite fixed?
Change-Id: Ie3f967d27b03a9594d54e8dd413f61daa65a73f3
We are running into errors when attempting to uninstal hiera on trusty
jobs because the hiera package does not exist. We no longer install
puppet as part of our image build process so we shouldn't need to
uninstall puppet, hiera, and facter at all. Simply drop the the removals
to address this as there is nothing to remove any longer.
Change-Id: I1ce79f3f7dd164e0635a4ae45210194f50ab8ad9
The v3 migration script can handle these just fine - but if we go ahead
and remove them the output can be predominantely shell rather than
script.
Change-Id: I440851a0149e1cc7f8c5c6e8a4e3e0b94088ee82
This reverts commit 1f57c9c327.
The logs.openstack.org site is not configured correctly to serve
over https. I think it's actually serving from docs-draft.
We should fix that, but in the interim, switch back to http.
Change-Id: I170b24f01bdd79e4728d509d2a0f654a479b2674
Add a sample python 3.6 job running on a Fedora 26 node.
Create the job macros and definitions for a Python 3.6 job running on
Fedora 26. Add to the low-turnover bashate project so we can run some
tests on a real project.
At this point, I have not templated it out futher. We could add it to
templates like python-jobs and create groups in zuul/layout.yaml to
make it eaiser if required.
Change-Id: I04fab592032621325f8638680fc5c842f316684a
At the moment, all YUM failures are masked because the shell script
is not setup to exit on failures. Once the failed stage is passed,
the script will always return true to handle requests removal.
This patch forces the script to exit if any commands return invalid
exit commands.
Change-Id: Ia9f4e5f8cdb9fa0735d609dd075cdc7d73591acf
In 0b00bd6abd we added back our npmrc
files pointing to our reverse proxy cache. Manual testing confirmed
it to be working. Lets start having our jobs do the same.
This reverts commit 26edd71461.
Change-Id: Id308190ff0e938a4b010455275344aff3af3b2a6
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
When adding a repository in Ubuntu, one must run apt-get update prior
to trying to install a package from the new repository. Additionally
when this new package installs binaries in an alternative path, you
need to ensure the PATH is properly updated for it to work.
Change-Id: I285baa486221e496baa8e205bd7a92c33d07dd2f
The wget parameter to download the file was wrong. It should be -O not
-o. This is currently blocking puppet module releases.
Change-Id: Ief269df165a8a46d794f21a6e7dcb6423556ebc8
openSUSE installs traceroute in /usr/sbin - but it can run as normal
user. Add /usr/sbin to PATH so that the traceroute and traceroute6 calls
work for net-info.
Change-Id: I49874ca395d2fae173c6054aa9e1d90863460ecb
The puppet-module-build job requires puppet be available. Rather than
include it in bindep.txt which would conflict with the other jobs, this
change creates a builder to pull in the latest version of puppet from
puppetlabs.
Change-Id: I2c56d16ee3e0be6982e51e4457c9beb25f175c3e
Since we removed puppet from the nodepool nodes, the puppet parser
validate command in the puppet-syntax job fails since puppet is not
installed. Rather than modifying that job to install puppet, we can
re-use the existing 'rake syntax' job that the puppet-openstack group
has defined for their modules, which installs puppet via gems and is
therefore simpler. We can also now remove the plain puppet-syntax job
since nothing is using it.
Change-Id: I082ea72743634b8ce0106786ab4ec35add874c8d
I0c73034a85db5664e979b9b71cc340860265bf34 switched from using the
rdo-release to using the centos-release-openstack-* packages. Since our
jobs manage the repos, we need to clean out these rpms as part of our
node preparation. Also removing ceph repos, we already manage the repos
in our manifests.
Change-Id: I7a82f9804e2fe52bb1da8bcfa682c0685771af48
There is ruby-devel package installed here for CentOS but there is no
ruby-dev for Ubuntu. That commit fix that problem.
CentOS change-id: Ib504f3fea4713f14bf3dc2e9659fad0e999f8146
Change-Id: I29e50121f7e5fa48adb47aeec82c68e83282ded5
Closes-Bug: #1681851
Remove the unused builders and publishers for zuul-swift, remove unused
script to upload, and remove unused swift-logs nodepool element.
Change-Id: Idfa83d435a951dcfe6f6545a2515aec281ed43cc
Remove branch-git-prep, we can use zuul-cloner everywhere.
Rework zuul-git-branch-prep to follow the logic of
zuul-git-branch-prep-upper-constraints. Note that zuul-git-branch-prep was not
used prior to this patch.
Change occurences of branch-git-prep to zuul-git-branch-prep.
Change-Id: Id938a858a0ff1967b97293405ef41b9b281f407f
We do not need gerrit-git-prep anymore, zuul-git-prep works on all our
slaves and has the advantage of using zuul-cloner so that the local
cache is used and Depends-On works.
Remove gerrit-git-prep and use zuul-git-prep instead everywhere.
Change-Id: I2f106af58f78fceaf3c497fe43e6f5a002695645
Nova has subproject called Placement API. It will be separated
openstack project in the future but until some moment it's in Nova.
Placement API has its own api-ref run by `tox -e placement-api-ref`.
This change adds new jenkins job gate-placement-api-ref-nv run for Nova
project only. It uploads built api-ref to separate folder on docs-draft.
When Placement API has separate code base this changes should be reverted.
Change-Id: I891a2512efe067f3f45a24cc4ff19460a568046e
- renames chef-test-jobs to chef-rake-jobs for consistency
- adds multinode integration jobs
- changes gerrit-git-prep to zuul-git-prep as zuul-git-prep is preferred
- removes aio periodic gates
Change-Id: I0c316804a2eea2574681778678a8599203007c0d
There has been a few scenarios where eth1 doesn't get an IP
address which results in subnodes being unaccessible and
builds failing.
This patch prints out the information consumed by Glean in
order to help troubleshoot the issue and figure out the
root cause of the issue if it occurs again.
Change-Id: Ifd50e83fe3d73064b9141ed62380ec57277241d7
We are seeing some issues restarting SSHD for centos. Because we are
doing config file changes, and SSHD is running, we can use the reload
command.
Reload is a little different then restart, specifically reload does a
SIGUP on the process. In some local testing, this seems to be more
reliable to restart, which causes some bind address errors.
Change-Id: Ic5e00e1e1eed1de87c430180a6f00d30410a95b4
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
bifrost sets up the mysql database itself, we can remove the setup from
project-config now.
Note that change I8485751eda745841c1436ecd766c9bb95446f407 has been
merged to all branches excpet mitaka since that branch is broken.
Proposal is to move forward without change in mitaka - and add it there
if critical changes need to go in.
Change-Id: I339d2a301a76ad34ea4bcbfb26c13a21ecd86b6f
The trove repo now uses tools/test-setup.sh we do not need the db jobs
anymore, change to the standard python unit jobs.
Also use standard pylint job-template which does the same as the tox
template.
Remove all templates and job-groups for the db-jobs since it's last user
has been removed.
Only leave mysql setup jobs for bifrost in.
Change-Id: I92e6e6502c2c516babf2bf66f3514875f77c460e
Depends-On: I3463e75057d0d4544f6a0212da888759ab5e171b
While trying to add centos-7 gate job to swift tests
this script failed calling mkfs.xfs. Updating to use the full path
so that it can be used both on centos and ubuntu
Change-Id: I18adb1c1daa11d4f84518e449662ffe2b4e9dbf7
gate-openstack-chef-repo job was removed in:
https://review.openstack.org/#/c/276106
With the removal of this job, the following builders are not
needed anymore:
ruby-191-prep
chef-bundler-prep
chef-berkshelf-prep
Change-Id: I8ea78c43b91977a38d35512e66b5bf4ccc5ed455
old chef-jobs have been removed in:
https://review.openstack.org/#/c/264130
The following builders are no longer needed:
chef-cookbook-chefspec
chef-cookbook-foodcritic
chef-cookbook-rubocop
Change-Id: Ie8cda6d175fe5307537e1187393bcf2bcd440736
swagger job, is not used anymore. it was removed in:
https://review.openstack.org/#/c/364076
The publisher is not used any more as well.
Change-Id: I354af582c69aac08a6e4666788001beecbb9c47a
Zuul-cloner / zuul-merger have been fixed and can be used again.
This reverts commit 759da2b1cf.
Change-Id: I99bf00b046951ed6ea544572c52138f671e78be0
Depends-On: If22d8cfbe591afe4f5516da1eb3b0be98b2de874
The AFS based publishing and serving via files.openstack.org works fine,
we can stop publishing to docs.openstack.org
Change-Id: I4d6b51c7f8065e5cd21942acc64ae56639bb185b
Across the entirety of our job config corpus, we only ever set the
tarball-site parameter to tarballs.openstack.org. Instead just set
it directly in the publishers where it's needed and reduce
unwarranted confusion for people writing job configs. This also
allows us to get rid of the parameterized sites in the tarball,
wheel and war publishers.
Change-Id: I31109f41177b8c07177d7ed8ef16cbfd135131ef
With the introduction of tools/test-setup.sh, we can move the setup for
searchlight to the repositories and remove the special setup for these
repos and use the normal python jobs instead.
Also, remove gate-searchlight-python27-searchlight-es1, this is broken
since August and PTL agreed that it should be removed.
Note that this is needed to fix liberty branch of searchlight - liberty
needs elasticsearch 1 while newer branches need elasticsearch 2.
Change-Id: Ie038e9c9a38727c12532bded48a76fe2cecb7fd8
Builder added:
- mount-xfs-tmp-space: takes a path, creates a 1G file there,
and loopback-mounts it to a passed in path
- tox-with-tmpdir-passed-in: a copy of the tox job, but with an
extra {new_tmpdir_loc} value passed in and TMPDIR set
projects.yaml changes (for swift):
- replaced 'gate-{name}-tox-{envlist}-{node}' for
'gate-{name}-tox-xfs-tmp-{envlist}-{node}' and placed
'py27' and 'py34' envlist entries under it
scripts updated:
- run-tox.sh: now only sets TMPDIR if it's currently unset
job-template updates:
- create 'gate-{name}-tox-xfs-tmp-{envlist}-{node}' that is a copy
of 'gate-{name}-tox-{envlist}' but calls 'mount-xfs-tmp-space'
before dropping sudo. Then it calls 'tox-with-tmpdir-passed-in'
instead of 'tox'
zuul/layouy.yaml updates:
- added the "with xfs" tox jobs to swift
Change-Id: I3339de0d57726b339dfc6638d225e49d097f7b63
Instead of setting up databases for jobs, allow projects to do this in
tools/test-setup.sh instead - and run this script whenever it exists
in our tox based unit tests.
Note that currently no repo has such a file, change
I9af7c7d09b40c103af09ae8dd0e50afe3762ecba introduces the first one.
Change-Id: Ic6269e0a03c8540569f242ebc78c98a5922c8026
JJB already has a tox module[1]. Defining a macro with the same name is
confusing and may cause problems when attempting to use this macro
elsewhere. Renaming it to something else should clear things up.
This is the warning from JJB:
$ jenkins-jobs test -o out ../project-config/jenkins/jobs
WARNING:jenkins_jobs.registry:You have a macro ('tox') defined for
'builder' component type that is masking an inbuilt definition
[1] http://docs.openstack.org/infra/jenkins-job-builder/builders.html#builders.tox
Change-Id: I6946229f6e94a5f4c10d6a5fc3b7049adeacacb1