When we deploy with SSL we do so via haproxy. So just as we do for
the undercloud services IP, we need to add that one to the no_proxy
list.
Change-Id: If158033663878e621c8683c7754b87c227550c75
Puppet OpenStack CI and RDO CI doesn't need EPEL, so this patch removes
EPEL dependency in TripleO CI.
Change-Id: I7adebb9ff6b544289bb138b5aa8736f0ded0da49
Depends-On: I91b070324eb39af351c2304e6786c395509d08dc
Depends-On: Iceff0d5bedd9816adfd2990970e7c216b67b6bd0
Being able to set a different $TRIPLEO_ROOT other than /opt/stack/new
makes the toci scripts more useful when running locally. There are also
lots of places where /opt/stack/new was used directly, when we
could/should be using $TRIPLEO_ROOT.
Change-Id: Iac212866fc343f3f3f930fad57b3a52236513d34
Currently we always build from source, but the latest master is now
available via delorean-current since o-p-m was broken into per-module
packages.
This should reduce the time taken to install everything from source,
and also provide a better path wnen upgrading as a simple
openstack undercloud install or yum update will update to the latest
puppet modules (currently failures happen if you forget to run via
tripleo.sh because the puppet modules aren't upgraded but things
that depend on the puppet modules are).
Also disable usage of the overcloud image cache as I have no idea what
will happen if we mix an image with puppet modules deployed from source
with ci scripts trying to install packaged modules.
Change-Id: Ie4cae6b884b2e539545371b9f716903b8fe5cf1f
Closes-Bug: #1597438
Temporarily remove epel-release, and don't enable epel via tripleo.sh.
Also remove nodejs and nodejs-devel. This is a quick fix for:
https://bugs.launchpad.net/tripleo/+bug/1606685
We have a CI dependency on EPEL too with moreutils, so that needs
to be installed before we remove epel. Since everything is running
on the Jenkins slave that should be okay for now.
The proper fix will be to remove epel more completely going forward:
https://review.openstack.org/347499https://review.openstack.org/347513
Co-Authored-By: Ben Nemec <bnemec@redhat.com>
Change-Id: I91eff35586547dd278e8a32668f06efb840fc8c6
Partial-Bug: #1606685
This broke the nonha job, which runs SSL on the overcloud and
relies on having a predictable IP for the public VIP. Since
that jobs doesn't do net-iso, it uses the provisioning address
as the VIP.
We also can't just sed the template because the certificate
has to match, so let's just revert this for now and figure out
a safe way to do it once CI is unblocked.
This reverts commit 2ffce235be.
Change-Id: I9bf6e832c4d914ac5d7dcef1070e1ca894b2e2b5
At the moment I'm proposing this based on a hunch, I think
nodes may be failing to deploy because of ARP traffic leaking
between provisioning networks. I'd like to see if this improves
the current pass rate of Jobs.
Related-Bug: #1599858
Change-Id: I6befe6886fdf6cd9e1c1f82b6fe15f64f3377e5c
This commit refactors out some reusable pieces of tripleo-ci into
common_functions.sh and common_vars.bash. This is in preparation for
adding ovb and multinode support.
Change-Id: I2154e07d0433ffbd5d91884c9b1c00adf3ac534d
JOB_NAME parameter is not defined and it's required in periodic
jobs, extracting it from $WORKSPACE
Closes-Bug: #1596425
Change-Id: I191fea6a8825fe862c5122b0d8362e0230f7403c
Running this at the end of each CI job will ensure
disk space used by the CI tests is released. Also creating a new
qcow image for overcloud nodes for each test run should ensure that
this disks internal sparseness is restored.
Change-Id: I0bef4f0a6fa334f894d69616829196fe8fc4d768
This is need once we switch to centos-7 jenkins slaves, since the
default kernel in centos-7 does not support nbd (to mount qcow2
images).
Change-Id: I53df7a535d04b67c3ff149b3e300b344ea773bdc
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
Doing this on stable jobs causes packages to be pulled in from
the delorean deps repository as it contains openstack packages and
yum-pluging-priorities hasn't been installed yet.
Closes-Bug: #1585937
Change-Id: I27451be61a3bc6235a6fa5e021e88d21e8842b72
get_host_info sometimes freezes for hours and causes job to fail
by timeout. Added killing it after 5 or with -9 after 10 minutes
Closes-Bug: #1583676
Change-Id: I7f414751709e6dd965f1218c5b6c1e571d94e54b
Our existing tars are not available for logstash to index, so
let's extract the logs out into the workspace log directory.
Change-Id: I76656bb0b362f37938bd22d1377bbf5dc95dcd86
- Reinstall glibc to fix locale bug
- Only add fedora mirror if we are running fedora
- Make sure python-gear is installed
- Fix DLRN dependency issues
- Use raw image since CentOS doesn't support nbd.
Change-Id: I9da52eff1976678691e19f8ab4bdb26e819c71d0
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
Make use of the automatic certificate generation in the dependent
patch to test that installing the undercloud with ssl endpoints
works.
Change-Id: I0ad97df8a05646ac48be937cdc5b97ae77c7aa71
Depends-On: I7616b82913d16f2ba837298309d836c4811c2a44
We need this in order to upload it to the mirror server after periodic
jobs, we only removed it to save disk space, so find space elsewhere.
Change-Id: Ib8a3daa0873562a7a13c68f7cfe3592a7c6b3840
Since adding the use of the API cached image, a lot of jobs have
started failing because the jenkins slave is now running out of
disk space. Delete some artifacts after we're done with them to
avoid the increase in disk usage.
Change-Id: Ia1d9c9b41156ba5b508c86ac8a258a2003e4b7e1
Without this it looks like postci just isn't doing anything, unless
you already know that the output has been moved to a log file.
Change-Id: I3da396fdd8c264527c6fe3b8ef6b3164bfa87d29
To allow the toci scripts to run twice on the same machine we were
removing all packages installed from a delorean repository. This
no longer works as delorean-deps now contains python-six and removing
it fails because yum depends on it. Instead only remove from the
delorean repositories that contain openstack packages.
Change-Id: I4c80fa7a91768747dcb5d1d8f7020f01ec78f626
Only request a test environment from geard if we don't
have one defined. This will help allow the toci script
to be run without the need for a geard managing the
environments. Also if the testenv wasn't requested from
geard leave it running to allow debugging.
Change-Id: Id2b03456be86e7f2900388f40668d79861858bf6
Save the overcloud images to the cache following a successful
periodic job. They can then be used in subsequent ci runs.
Change-Id: Ib8d451500f7cfa0f7798fd65a0f90b9b8c6316f0
For projects that can't use the cached qcow2 image, this output
spams the logs with almost entirely uninteresting output. Let's
send it to a log file like we do for other image builds.
Change-Id: I249a94ea851995a784f37209ee369b49a2d2f30f
As we we make more use of the cache the logic to decide
if we can use the cache will be better placed in a single
function where is can be easier understood.
Change-Id: Ic1149a61c46f6d1d4bb9649da4710543962cd285
Track wall time metrics for:
-tripleo.undercloud.install.seconds
-tripleo.overcloud.images.seconds
-tripleo.register.nodes.seconds
-tripleo.introspect.seconds
-tripleo.overcloud.{job_type}.deploy.seconds
-tripleo.overcloud.{job_type}.settle.seconds
-tripleo.overcloud.{job_type}.ping_test.seconds
-tripleo.testenv.wait.seconds
-tripleo.ci.total.seconds
Additionally we track the image size in MB for:
-tripleo.overcloud.image.size_mb
Currently sending these to Dan's graphite server.
Change-Id: I9307ca50bf5fb8398fd22ccbe45e79524b8837c8
Adding repositories to mirrored.list makes CI try and use them, but
they wont exist until the patch has merged and everything is synced.
Allow for git repositories that are in the list but not actually on
the mirror server yet.
Change-Id: I8e9e0b90660d5791377b8bcc112b1ae068718ead
We're seeing lots of failures with DNS requests in CI jobs,
it looks like the problem is happening outside of our overcloud
deployment. This is particularly bad for DNS requests presumably
because its UDP. So switch our primary DNS server to point at named
on the bastion server, while we try and find the root cause.
Change-Id: Ic59e338e4d3abee15c6d6614dd377ca81479850b