The var overcloud_image defaults to overcloud-full, and will until the
switchover to the whole-disk overcloud.
Change-Id: I64c0d9f97b34e9145be70af7c489bb5b706baf8a
Blueprint: whole-disk-default
https://review.opendev.org/c/openstack/tripleo-quickstart/+/790701
removes tripleo-operator-ansible from TQE requires and will be get
installed from rpm as a part for undercloud deps.
In current setup, dlrn build the rpm packages just after undercloud
rpm packages installation but tripleo-operator-ansible is used before
that so that dlrn needs to build the changes early before undercloud
deps installation so that we can consume the changes in the deployment
via updated rpm installed from gating repo.
In order to do that, it moves build-test-packages roles task after repo
setup to install latest packages.
It also removes tripleo-operator-ansible from cloning on nodepool.
Testing different scenarios:
1. https://review.opendev.org/c/openstack/tripleo-heat-templates/+/801296
2. https://review.rdoproject.org/r/c/testproject/+/34676
Patch for moving remove_tags to skip_tag_list:
Depends-On: https://review.opendev.org/c/openstack/tripleo-ci/+/803919
Signed-off-by: Chandan Kumar (raukadah) <chkumar@redhat.com>
Change-Id: I4f9f3f13346761d3dec35844234cb6ae03ad4107
Makes those files conformant with current linting rules and avoids
linting errors when we need to toch them again.
Previous doing "pre-commit run -a" uncovered these errors, now is no
longer reporting any other errors.
Change-Id: Ie4cf229c8f11c2b55b323eac23c89483b26d3781
OVB failures due to
- ansible_user variable was used instead of ansible_user_id, the first
one being unreliable as is may only be user defined (or undefined).
- lack of become on tasks that needed it
Fixes problem identified by molecule testing
- lack of idempotence on few tasks
- failure when ansible_user variale was not defined
Change-Id: I11adec7104a59bfa29b00c702eb9b7eb6a5c6766
Depends-On: Iade08ee92d7f48e07a625678d2a1c6ac1cd1982b
Partial-Bug: https://tree.taiga.io/project/tripleo-ci-board/task/358
Closes-Bug: #1802825
Use atop[1] tool to monitor the whole job process.
Atop generates binary output that could be downloaded
and then investigated locally.
Using atop -r /path/to/atop.bin you can read the file
and by pressing "t" to move 10 seconds futher or by
pressing "b" to jump to specific time in job and to see
what happened on host in this time. It allows to track
all resources in specific time.
It allows also tracking of containers separately.
For more info you can visit the site[1]
If atop installation fails it shouldn't fail the job,
so ignore_errors is added.
Currently it's for undercloud in OVB and all nodes in
multinode.
[1] https://www.atoptool.nl/
Change-Id: I7e17db3e376218f620a18db7ea7ca82d7578f618
Depends-On: Ibcdcfb4d8c5c94e1a06c7e635b0b6778ad318094
Currently we have a few possible places for building package in
the job. It requires logic change in order to differentiate
between injection cases.
build-test-packages role can run and not to build anything bc of
various reasons, in this case we need to avoid running
install-built-repo role, that's why we set compressed_gating_repo
to empty string and check in every case if it's empty or not.
If compressed_gating_repo is emptry - we didn't built *last* time
anything, so nothing should run.
Change-Id: Ib19addc66678804abb86f36006fb35f3f1f1d471
Don't update IPA image, use the provided from promotion.
Introspection fails when we unpack and pack again IPA images,
let's use promoted one in jobs until bug 1770972 if fixed.
Related-Bug: 1770972
Change-Id: Iab5565d1743d7a6d9aa9fdb5473b3b3e5fea7c62
Set undercloud inventory before it's used in next playbook.
Fix images update issues and set centos images to newest.
Set 'libguestfs_mode: false' to avoid running virt-customize
on initramfs.
Closes-bug: #1730997
Change-Id: I46dce77166ecb828d4186b3f2ffa56cf9aa9d31e
This continues the work to merge the OVB playbook in tripleo-ci with
the OVB playbook here. This should again be a no-op change, with the
only new behavior turned on by setting use_testenv_broker=true.
This change completes the differences between the two playbooks before
the undercloud is installed. There are also slight differences in what
we do after the undercloud is installed in the
baremetal-quickstart-extras playbook. Those will be addressed in the
final patch of this series.
Change-Id: Id24b8c4866f46dcb37a98b682f54101a341f9adb
In order to combine the tripleo-ci OVB playbook with what we have
here, we need to implement the logic around when to build images[1].
We default to fetching pre-built images. However, in the case we are testing
a change to a project that can affect the image build or the case we are
running a periodic job, we want to build the images.
[1] https://github.com/openstack-infra/tripleo-ci/blob/master/scripts/to_build
Change-Id: I72169d4c6ae891397916f831e56eb883d047c89b
This is the first step in combining the upstream playbook used
for OVB jobs with this one. This should be a no-op change in terms
of what gets run, however the part that is split out is common to
the upstream jobs which create the stack via the testenv broker.
We can then iterate on adding in the testenv broker steps (guarded
by some option) in follow-up patches.
Change-Id: I7cad64e83ed0fbf791b26bdf0e381599d177de7b
Ghanshyam Mann