| Age | Commit message (Collapse) | Author | Files | Lines |
|
This reverts commit 931d401887252d8830ebfb4d0b5d352f77094e9a.
Reason for revert: monasca-grafana is no more
Change-Id: I483db6d6eddd84b376251b38509e74d4a5c2a10b
Depends-on: https://review.opendev.org/c/openstack/kolla/+/784902
|
|
|
|
Infra mirrors get their indices rebuilt to avoid broken
indices (due to partial update).
Unfortunately, this wipes out the cryptographic signatures.
Our approach so far was disabling apt security features globally.
However, this is not a valid choice for external repos.
It hid an issue we introduced with new RabbitMQ repos missing
proper keys installed in the image.
This caused permanent failures outside of our CI.
Our process should be as close as possible to users' experience.
This patch makes CI trust only the mirrors that have their indices
rebuilt (so infra mirrors).
Change-Id: Ic5abc4b87fd76f87aba383abf43e95ba70629fcb
|
|
[[ is bash only
In many distributions /bin/sh == bash but not in Debian and derived
where it is dash (very simple POSIX shell).
Change-Id: I2b084ea78b236623b174473f411bae04e624a3e9
|
|
Bullseye is now in hard freeze cycle and goes for release.
https://release.debian.org/bullseye/freeze_policy.html
Co-Authored-By: Michal Nasiadka <mnasiadka@gmail.com>
Change-Id: I543965a2741cebfa759576a4c75669a7bacd4208
|
|
|
|
Based on the discussions from the past 3 Kolla meetings [1][2][3],
we established that, to avoid the DockerHub pull limit issues
in our CI, we will try using quay.io.
We still publish weekly to the DockerHub.
[1] http://eavesdrop.openstack.org/meetings/kolla/2021/kolla.2021-03-17-15.00.log.html#l-102
[2] http://eavesdrop.openstack.org/meetings/kolla/2021/kolla.2021-03-24-15.00.log.html#l-137
[3] http://eavesdrop.openstack.org/meetings/kolla/2021/kolla.2021-03-31-15.01.log.html#l-50
Change-Id: I7c1f0c5fc643678b29f963b6769d8f900a03b726
Depends-On: Id113a53c083d105ac9200474ff20faacef6decaf
|
|
The monasca-grafana image is currently failing to build.
INFO:kolla.common.utils.monasca-grafana:go install -v ./pkg/cmd/grafana-server
INFO:kolla.common.utils.monasca-grafana:[91mgo: cannot find main module, but found vendor/vendor.json in /grafana/src/github.com/grafana/grafana
INFO:kolla.common.utils.monasca-grafana: to create a module there, run:
INFO:kolla.common.utils.monasca-grafana: go mod init
Let's unblock the gate and diagnose asynchronously.
Change-Id: Id113a53c083d105ac9200474ff20faacef6decaf
|
|
This fixes comparisons when files are not Unicode-encoded.
A relevant unit test is included.
It can be used as a base for other _cmp_file method unit tests
if the need arises.
Change-Id: Ic638516eb92d24ad247a7866fd1b5e2ac0400388
Closes-Bug: #1913952
|
|
|
|
Change-Id: I799bf8edfef6ff68382d651dca6de96fa80ed733
|
|
Change-Id: I0c4d5e1ea16c6677813a334b38b3a22ad811531d
|
|
When building images in CI, we use local package mirrors to reduce
external network traffic. These mirrors are not necessarily reliable nor
are they guaranteed to be accessible outside of the CI environment
(although many are).
While it is an antipattern to install packages in containers at runtime,
the Bifrost container does just that. If the mirrors configured in the
image are unavailable, Bifrost deployment can fail.
This change fixes the issue by reverting to public package mirrors as a
final step of the image build process.
Closes-Bug: #1902101
Change-Id: I2592a736206eaec811290e9fbdbf2540c0518ffe
|
|
Login to DockerHub was failing because of trailing newlines in
the encrypted credentials.
This patch makes the playbook trim them before usage.
Change-Id: I411f91af927ffa4d04c8021ddb5c3d584d90bdfd
|
|
After a recent change [1] publish jobs are failing on missing docker python
SDK.
[1]: https://review.opendev.org/c/openstack/kolla/+/763146
Change-Id: I63662dd903b914fe06165b45e0a0bb5ca74b21b9
|
|
|
|
Currently we use a for loop in shell that performs push of images
to Docker Hub - that doesn't really seem to fail properly on push
errors (e.g. related to new Docker Hub limits).
This change introduces usage of Ansible docker_* modules for doing the
same job and improves logging (so it's easier to find out which image
failed to publish).
Change-Id: I083ce70772edd1abb873ad43f6beb73aae01383f
|
|
Change-Id: I702f8fba37e22a6bb45eae14f1fbb251bcdb4d8a
|
|
Fixes a FIXME.
Change-Id: I2e6e07d6f466a7485ce6481e6e43b034fdc27f41
|
|
Following this change it is possible to add 'Depends-On' lines to a
kolla commit message that point to reviews in other projects. For
example, this could be used to test building and deploying container
images containing a change to the nova project that is currently in
review.
It also allows to introduce jobs testing against Kolla in other
projects.
The kolla-build-config Ansible role can also be used by other projects,
such as kolla-ansible, to generate the necessary configuration.
This takes us closer to the Kolla Builder - Next Generation concept [1].
[1] http://lists.openstack.org/pipermail/openstack-discuss/2020-April/014255.html
Change-Id: Ia2efa6b5d7278b75ab8dec23207f92d6bd7c58f0
Co-Authored-By: Radosław Piliszek <radoslaw.piliszek@gmail.com>
|
|
The imp module is deprecated[1] since version 3.4, use importlib to
instead
1: https://docs.python.org/3/library/imp.html
Change-Id: I79dfd923531dd3e00d23e1682865f56f0cf437d2
|
|
There are several issues with kolla_set_configs --check:
1. We calculate the destination path incorrectly when comparing a file
in a directory, due to passing arguments to os.path.relpath in the
wrong order
2. For directories that have not changed, we also attempt to compare
them as files, which fails when they are open()ed.
3. If the config JSON does not have a config_files key, it fails with a
KeyError.
The first two issues affect the fluentd container, which specifies
directories as the source, without using a glob. The third affects OVN
containers.
This patch fixes these issues.
Closes-Bug: #1890567
Change-Id: I8921befe51da4282121443849177a7ca5ebe8822
|
|
|
|
Since Kolla is run in the wild against PyPI mirrors which do not
include the OpenDev-provided wheels, it might be nice to be able
to test these scenarios in the CI.
These jobs are added to the experimental queue so that they do
not consume resources when not needed.
Any Gerrit user may run them now to verify whether the change
passes them.
Change-Id: I9c33eeaf03be037651f82c831f0e07ab9a27752a
|
|
CentOS wheels do not differentiate between minor versions atm.
Change-Id: I5365c813bab5e3205fd34eb64dc3337f192d67d4
|
|
If they are available then we do not need to build them. Especially
AArch64 jobs will speedup (once wheels are built).
Change-Id: I79af6c37950e156018a9204fbcc7417cd7d41012
|
|
Looks like I missed one situation with "base/centos: take care of
whitespaces" change.
Change-Id: Id7b5e8ddf23f9ab13806fb02042e76183bcddf05
|
|
Let's get rid of this warning:
Change-Id: I121e2ecb30ec6e8d1b6d88d9b921894438c3e094
INFO:kolla.common.utils.base:ESC[91m[WARNING]: Empty continuation line found in:
INFO:kolla.common.utils.base: RUN dnf -y install centos-release-opstools epel-release dnf-plugins-core && dnf config-manager --enable PowerTools && dnf config-manager --disable centos-ceph-nautilus && dnf config-manager --disable centos-nfs-ganesha28 && dnf config-manager --disable influxdb && rpm --import /etc/pki/rpm-gpg/RPM-GPG-KEY-CentOS-SIG-OpsTools && dnf -y distro-sync --security --sec-severity=Important --sec-severity=Critical
INFO:kolla.common.utils.base:[WARNING]: Empty continuation lines will become errors in a future release.
INFO:kolla.common.utils.base:ESC[0m
INFO:kolla.common.utils.base:Step 1/39 : FROM centos:8
|
|
Failed to find required executable virtualenv in paths:
/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games
Change-Id: Ie6094e1c5de5c89e0180c42edf5ef94c1c9964b7
|
|
|
|
With the move to RHEL/CentOS 8 we no longer have Python 2 in our images
so there is no need for checking which Python version (2.x or 3.x) is
used inside of containers.
We also no longer have to support yum as a value for
distro_package_manager.
Partially-Implements: blueprint centos-rhel-8
Change-Id: Ie45cf3465fedddbde7856961527421883ba3d5c9
|
|
also simplify Debian job definition
Depends-on: https://review.opendev.org/718413
Depends-on: https://review.opendev.org/718363
Depends-on: https://review.opendev.org/718236
Depends-on: https://review.opendev.org/718238
Change-Id: I2c75dbc2ea063dd3507eaaf93a484c50acae5fb5
|
|
|
|
|
|
|
|
CI is set up using Ansible playbooks now.
This also drops redundant shell scripts.
Co-authored-by: Marcin Juszkiewicz <marcin.juszkiewicz@linaro.org>
Change-Id: I5264746fd04276448f4668c3bb3a218a40e8660b
|
|
We don't need tox nor tell our users to use it when doing kolla builds.
Tox is going away from base infra images (it is already gone from
aarch64 ones).
Depends-on: https://review.opendev.org/713134
Co-authored-by: Marcin Juszkiewicz <marcin.juszkiewicz@linaro.org>
Change-Id: Ib1d97a783951ac42740ebf91bcc6ecaf2bf70853
|
|
|
|
Some CentOS repos (currently one - Ceph NFS Ganesha) use
https://mirror.centos.org instead of http:// and we break
their definition by removing mirrorlist= entry and not inserting baseurl
entry due to a missing match.
Change-Id: I7b29f8c6be80820042023d2b85162d5cc9a65f08
|
|
Now that py2 is gone, oslotest dropped dependency on mock and will
soon affect Ussuri CI [1], let's use unittest.mock built in py3.
This also fixes py38 jobs and proactively prevents py36 and py37
failing due to [1]. This is because we never included mock in
test-requirements (but in lower-constraints where it does not
really belong at all) and instead relied on oslotest to bring
it in.
[1] https://review.opendev.org/716322
Change-Id: Iceedcc7ad0d087414f6f94dfc0235fda4754be63
|
|
The repo is Python 3 now, so update hacking to version 3.0 which
supports Python 3.
Fix problems found by updated hacking version.
Update local hacking checks for new flake8.
Remove hacking and friends from lower-constraints, they are not
needed in installations.
Change-Id: I67e5ca24c550163c218fb06692de0b7c528d8318
|
|
We want to test upgrades on aarch64. For this we need to publish images.
AArch64 images will get "-aarch64" added to tag to not collide with
x86-64 ones.
Change-Id: Ibae2578b4bc18c0ab7fcc1840c39bb86659ab920
|
|
|
|
tox will be removed from the base image. Install it before that happens.
This change is made in a simple way that can be easily backported.
Change-Id: I2a24c87aa44c1e03ba76a2fb0bd71caf54eb4211
|
|
Changes the default base tag and distro for CentOS images from 7 to 8.
Also changes the tag used by the publisher jobs for CentOS 8 images from
'master-centos8' to 'master'.
Change-Id: I5498141ade23219d63f34debc2b8e1330dec2e68
Partially-Implements: blueprint centos-rhel-8
|
|
This avoids issues with IPv6 not enabled in docker.
Change-Id: Iff80392357ef5e2dd3f2e24dbf4df299660805b9
|
|
Reverts temporary.
This reverts commit 063a46a7de4471504535499ca2a972a79e4ce5e5.
Change-Id: Ibf2848db954862f917a23eb0aa1594ccc4e7ef67
|
|
Since virtualenv 20.0 (amongst other changes) six version >1.12.0 is required.
This change adds upgrade of virtualenv and six in pre - to be reverted once
infra CentOS images are sorted out.
Depends-On: https://review.opendev.org/707377
Change-Id: I75249055989309d0afdfc29c894beb6b1fcd1994
|
|
Currently we use CentOS 8 nodepool images for all CI build jobs. This
means we can't use the ansible_distribution_major_version fact to detect
whether we are building CentOS 7 or 8 images. Due to this, all CentOS
images (including CentOS 7) are currently published with a
'master-centos8' tag.
This change fixes the issue by using the base_tag variable which is
defined for CentOS 8 build jobs.
Change-Id: I29b7b8ee742120107a11bd419ac58ab5c09b0b75
Partially-Implements: blueprint centos-rhel-8
|
|
|