Follows the same configuration that was used on
tripleo-quickstart-extras and documented use on tripleo-docs.
Change-Id: Iba8a2db92137f9f6ad28f498627eb1b87039d99f
Story: https://tree.taiga.io/project/tripleo-ci-board/task/381
Some of the CI jobs use very long <hostname>'s,
resulting in the FQDN to be more than 63 characters
long which is the max for linux, see GETHOSTNAME(2).
Change the default to use: <hostname>.ooo.test instead.
Change-Id: Idfbd604e35f85307c666ab940adfe9ea94b29004
Related-Bug: #1803703
Add missing bits to support stable/queens branch in TripleO;
Some adjustments might be done later but this is a first iteration.
Change-Id: I4a369389fbecf6890a75ffaaf695f391cd6b7e2d
Closes-Bug: #1750311
Don't construct CENTOS_MIRROR env var because we have it in the
file mirror_info.sh which is provided from nodepool.
Remove EPEL_MIRROR because it's not used anywhere.
Change-Id: I40fe8f6b2be1864b27bbe369e7da988fa7f1a0bb
Update the code to test upgrades between stable/ocata and stable/pike,
and stable/pike to master (queens).
Change-Id: I61afed141d611b4b6b5ad40b3d48e9ea2346d1e2
This used to rely in the undercloud being clever and assuming the domain
from the undercloud. However, this caused issues for some folks. So
instead we do it explicitly.
Change-Id: I3c1a36fa70e46721bbe21826254171cb0166c87c
Several love fixes needed here:
- We forgot to override STABLE_RELEASE so
the upgrade was failing because it was
deploying repos from Pike instead of Ocata.
- Allow only to check that Ironic API is running
on Newton or Ocata.
In master we can check that ipmi driver is
available.
- On old relases like Newton or Ocata we will
use the old heat calls on get host info.
Change-Id: I558e86f5cec139cef165b0908bb3861434c7c8bc
Depends-On: Ie79dfff667437305ff4109da5a9c42ae358b1a3e
In https://review.openstack.org/#/c/464667 I assumed we wanted to
keep the stable scenario on upgrade, to avoid adding services during
the upgrade process.
It seems we're expecting to run the pingtest for master (or whatever
post-upgrade branch) so were actually depending on the services getting
added during the upgrade.
Change-Id: I83a421a23333a01d56781fb329116ae0036ae48a
Some changes in t-h-t recently mean we've got to support multinode
scenario environments that reference j2 rendered files, but this
only works when the environment is contained within the --templates
tree. So we copy the environment under test to the "new" tree
during an upgrade test, and ensure all paths reference the new
location.
Change-Id: I0288c2b470606d990281f6c37071b632a2d0d423
Closes-Bug: #1687908
This deploys an extra domain whose LDAP server is the FreeIPA node.
Co-Authored-By: Rob Crittenden <rcritten@redhat.com>
Change-Id: I9d64bda89c5551335ea85c3170d1e74c9b7fed6c
In the upgrades job, we should be using the path to the stable release
templates that we are upgrading from for all the environments.
Otherwise, we're mixing environments from master (or the release you're
upgrading *to*) with templates that you're upgrading *from*.
Change-Id: I3e7d21850c4cc5c42fa309e846701ac7e59c6e02
Partial-Bug: #1677186
This removes the explicit setting of the domain and the ipa server in
the undercloud hieradata overrides. These values should be determined
from DNS.
Change-Id: Ie10f9e9d2ffe878ca13aa68ccd2ea66688e86c4b
update-from-deployed-server-$UPGRADE_RELEASE.yaml only exists for newton
and is not needed for ocata. So it's only added in that case.
Change-Id: I9057d47eea15c8ba92ca34717b6b5965d4425ab1
The initial deploy uses UPGRADE_RELEASE branch, which is the previous
release, not the branch under test. This means, because Depends-On
and ZUUL_REFS aren't branch-aware, that we can only test packages for
the current branch, and we have to disable delorean-ci to avoid pulling
in e.g master puppet-tripleo on a newton deployment due to
Depends-On.
Change-Id: I8222da355fa7a7134b695a45d815caa640b8029b
the domain I set initially was causing issues with some operations when
used outside of CI, this uses a more appropriate one.
Change-Id: I3142b3593d93bcdc1f92d12c1192262b0cc1e557
When upgrade has been completed, run a pingtest to check if we can
actually consume the OpenStack Cloud.
That is a first iteration. The next will be:
1. Run pingtest just before the upgrade
2. Verify that resources survived to the upgrade
3. Run pingtest again when we know resources survived
Change-Id: I83f45efd0b4608ec29489b8f8e717945b22469d4
In order to support upgrade testing from stable to stable +1 releases,
we need to stop hardcoding newton keyword and use $UPGRADE_RELEASE
everywhere we can.
Change-Id: I0257945500400c0ec69a0da99ec0f31178d0f70e
This is now in t-h-t so we should only use that version.
Note this means we use the master version of the env for
defining the service list for the deployed (newton) overcloud.
Change-Id: Ia7cc49daf8e9e26e3dee9db6f1bdb53b543b658e
This deploys novajoin in the fakeha-ca-server job which deploys
TLS-everywhere.
bp novajoin
Depends-On: I4c055e4b4086b02fa706380f01911f499966dfc1
Change-Id: If2d392caa54a0090da5f2654d37742d5aecaa4b0
Discussion on IRC has made clear that for real deployments we can't
avoid the converge step, because we do still need a way to pin/unpin
the RPC (even when using "auto" as the version), and also potentially
some other nova specific steps will be required during the converge.
So align the CI test with how real deployments will be upgraded, even
though this will make the job take a little longer, otherwise the test
isn't representative of the expected upgrade workflow.
Depends-On: If5016b910931364a621b280465420d0bf2617895
Change-Id: Ifcdb048a1081b16a8a33816be98ddbbe07c27856
wait_for is the last script still used by tripleo-ci from
tripleo-incubator. This commit brings the script over into tripleo-ci
(like we did in the past for other scripts) so that we can continue with
removing tripleo-incubator.
Change-Id: Icd62f75df3e27ead780a392b34a9720b127c9b6d
Right now, most metrics are recorded the same regardless of branch,
and some ignore the job type as well. Since many of the metrics
can vary significantly depending on those details (mitaka jobs take
far less time than master, for example), let's separate out the
metrics more.
This change adds a .release.jobtype specifier to most of the
metrics. The notable exception is the testenv wait time because
that is not release-specific. Also, the image metrics do not
actually need to be split based on job type since they don't vary,
but for consistency with the other overcloud metrics I did anyway.
With this change we do not lose the ability to look at broader
aggregate numbers either. We can still combine these newly
separated metrics back together in the Graphite UI.
Change-Id: I1df58f3a0c44708728e6ea1d705a6e8bf49ef330
I'm not sure what the purpose of this sed was, but all it's doing
right now is cutting off the last digit of the image size. So an
image of 1152 MB is being reported as 115.
My best guess is that this was designed to remove a newline from
the output, but in my local testing there is no newline. Maybe
something changed in the output from du/cut.
Change-Id: I225daf1e3abfe627713b1f54fae902236b5c321f
The register-nodes in tripleo.sh is expecting INTROSPECT_NODES to be
set in order to actually introspect the nodes. This sets it, then unsets
after the --register-nodes call so that it doesn't leak anywhere.
Change-Id: I6f39a19f51ad94c9ad541856083991d98ee63a52
Closes-Bug: #1659931
Initially just the predefined ips piece of that functionality
because it is the easiest to implement. Testing the other bits
will come in followup patches.
Change-Id: Ic12b85be917aad1ae8313cb4748b4f11b6b24192
* Prepare undercloud upgrade job to support stable/ocata, though don't
switch yet since it's not released.
* Put some comments in upgrade blocks to mention Newton is hardcoded and
some work needs to be done to be release agnostic.
* Update the distgit branch logic for Mitaka, Newton and Ocata.
Change-Id: Ia2fa8e251439634d162512ccdad6c68529fda03e
We don't have test coverage for these features in the ha and
updates jobs anyway, so there's no point spending time deploying
them in those jobs. Leaving them in the nonha job and the
undercloud job should provide sufficient coverage.
This saves approximately 4.5 minutes per undercloud deploy in my
local environment, which adds up when you're running 750 jobs per
day.
Change-Id: I3eb3324c3291c70544a7b71cee1ff31cec687769
Sagi Shnaidman
yatin