There is an ansible_managed variable that can be referenced in templates
to add a comment about a file being managed by ansible. Use it.
Change-Id: I60c27cf0eb35dcd4805c3517b3b7e0c7d9a0b3ea
There is no need to have this disabled (seems to be by default). We
already have ensured our SSH hostkeys have been setup properly.
Change-Id: I5f0ffea13b4adb053c10d56728bcf02c13255513
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
This will be used to replace our current confgure_mirror.sh script
that we today as a ready-script in nodepool.
This only adds support for ubuntu today. As we move forward and
configuration openstack-infra from JJB to ansible, we'll need to add
more distros.
Attach the role to the unittests base job for now. It should ultimately
be attached to the base job, but adding it to the unittests job lets us
test it more easily.
Co-Authored-By: Monty Taylor <mordred@inaugust.com>
Change-Id: I9bfa28c87390c09bb2c4cd0de6ce4c7890f8d81a
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
Start gating on documentation changes. Eventually add support for
publishing them.
Change-Id: I9e75fee12e7d74db51661b6a1071a510788cde74
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
Put the tasks into a role so that we can set zuul_work_dir in
defaults.yaml so that this also works if we're setting a directory in a
variable.
Change-Id: I638f7682d086781178f2dc7517eb0fb09a5f2acd
Putting it all directly into logs is a bit noisy.
Depends-On: I9f03ea408085e3da3ddeb231264e45b1c1fb6896
Change-Id: I82289ec0b66502995278eb897257f345b6c44878
If neither traceroute works, this is a host with broken networking and
doing anything else is a waste of time.
Change-Id: I042357c2d55596296ce2d570aa54433195a31b82
Copy in the openstack info, generalize it, and get it to stop
spamming stdout. We'll use a python module to do this so that we can
control what's going on.
Change-Id: Iec3bcec75f28451284c3c7d94593bfca3830bc8b
There's still a chunk of shell, because it's honestly better than doing
the equiv with package manager modules would be.
Also, make zuul_work_dir a parameter we pass in from the playbook. The
bindep role itself is otherwise totally non-zuul specific.
Change-Id: I95e78dd25a1a54b3ce5f2ddb18228a183d3c06ad
OpenStack nodes pre-install bindep, but other nodes may not. Check to
see if bindep is on the node. If so, use it. If not, make a temp dir
that we remove at the end of the role and install bindep into a virtualenv
in that tempdir.
Change-Id: I3f34c178254add2143dc7f1b9758844480d7d914
We'd like to be able to make some jobs that run tox on other projects
for ease of testing. This should allow us to do that.
Change-Id: I774a76eaaea16d8ea218b439a099b5c9335c2fc5
The original idea was that we'd have a "python27" job. But it turns out
that "python27" isn't actually as universal as you might think.
"tox-py27" on the other hand is a thing that can and should always work
for people who want to run tox with the py27 env.
Port in the jobs we had in openstack-zuul-jobs.
Also, copy in the text of the old shell scripts so that we can iterate on
them and make them better piecemeal.
These are named zuul-tox not tox because we already have tox jobs
defined elsewhere. We'll need to do the naming dance while we work on
them.
Change-Id: I813f3f2ae138c07918556bc81655518023527131
Before we revoke sudo, we need to run bindep in the project. Add
the run-bindep role.
Copy the content of the jenkins script for now. This way we can iterate
on the content in smaller chunks.
Change-Id: Ib893b06d05039af078e6eea22e882f6f1efae226
Also add a .gitignore and establish a documentation pattern.
Change-Id: Ib447ac13d03de4ab61171d50b19117ff654ad329
Co-Authored-By: Paul Belanger <pabelanger@redhat.com>