Before switching the default machine type of instances to q35 in
tripleo-heat-templates via I9f60a73577ae7cd712e2a8285abc0c0788906112 we
first need to ensure that Cirros is updated to 0.5.2 that includes the
ahci module [1][2].
As set out in I12e0bdb3699e5343592ab834468ba6b2fcdcaaf4 for devstack and
Iad1adbc23b31dd54a96299e7a8a4b622c15eed8d for Nova this is required to
allow Cirros based instances to see the now SATA based config drive
cdroms attached to them during a Tempest run.
[1] https://github.com/cirros-dev/cirros/releases/tag/0.5.2
[2] https://github.com/cirros-dev/cirros/pull/65
Change-Id: Ia3f839a3c5e4e4b59898c11561fe7ef7126bba5f
The current cirros image ssh setup fails due to lack of disk space
causing config-drive setup to fail forcing fallback to metadata
server. It is currently seen in OSA and devstack gates.
It might affect tripleo also.
Change-Id: I84a335921f3dd55c8e319c6bed260aa41201cdbb
Related-Bug: #1808010
Upgrades yamllint to latest version and adots use of its strict
checking.
Fix all known problems reported by yamllint so we don't have to do
that while touching these files.
Change-Id: I4bdc520d9e2aff086c4b463718bc1e053261a4f5
Story: https://tree.taiga.io/project/tripleo-ci-board/task/381
"echo >... <<EOF" idiom doesn't work as echo doesn't take its input from stdin.
This was just creating a file with an empty line. Using cat fixes it.
Change-Id: I09f01ea854df38eea7bb6e99f43e4a792e458abe
It was already using the openstack CLI unconditionally in one place, so
let's use it everywhere. This also fixes the documentation link.
Change-Id: Ie510b10429e5d734830d2815212a3ae565914667
The v3 file is going to be removed as it is no longer needed.
Related-Bug: #1733640
Needed-By: I3612c58c356f8955bd44655cf5aacc20be532cbc
Change-Id: I5e779f48a02c5389cebf84e9d3899250d63cace8
After Zuulv3 migration the path of cirros image changed on disk.
This patch fixes it in order to use cirros on image and not to
download it from net every job.
Change-Id: I36870180222d4ab13a34db3e978e9108644491e6
This fixes errors during the pingtest, as the v2 API is not available on
Queens overcloud.
Change-Id: I600ee25a9a29192b6b1755ec7032ed52bd9ae394
Closes-bug: #1727698
Currently the wait_for call on the validate stack does not included a
--fail-match argument. This means that if the validate heat stack goes
to CREATE_FAILED, wait_for will continue polling until --walltime.
Related-Bug: #1709327
Change-Id: I473df0dc267134d472d69ab43459ffbbd872ffa0
If a task is skipped, the associated variable to be registered is in any
case defined. Here, when validate_template is not "tenantvm_floatingip"
the task is skipped but template_check is registered anyway, as
"skipped", and it does not contain any "stat" attribute. So the next
conditionals fail.
Adding the original conditional to ensure we're checking for
template_check.stat.exists when we're sure that template_check.stat
exists too
Change-Id: I62f8c1556c94c132215b760d9597488659762e6f
Infra provides us ready cirros images on disk, so we can add option
to use them.
Don't remove cirros image after test.
Bump version of cirros up.
Change-Id: Iaceb8142cdffddc2dad2999d5766ea8c68839f79
When the ping test runs (in the undercloud) to ping to the
floating ip of the VM, ping count of 1 is not sufficient
and it could fail for some SDN controllers. In the case of
OVN (wherein it has native L3 support and unlike neutron
reference implementation it doesn't create router namespaces),
when the VM sends the icmp (ping) response, the OVN router
pipeline generates an ARP packet to learn the mac address
of the next hop (if it hasn't already learnt yet, which is always
the case for the fist packet) causing the first ping reply packet
to be lost. Also there is no harm in increasing the count to 3.
Change-Id: I4879f2f3cecced8b96e4ee230e4de7b210a8f7f8
We need test in subunit format first of all for tracking periodic
jobs in openstack-health, which identifies only jobs with subunit.
It could be usable for other statistics systems, which can parse
subunit format.
Change-Id: I28507aac37998cfa2e48dd1697822288f60e71fe
Add option to validate-simple role to skip pingtest deletion:
both for debugging purposes or saving time in CI.
Change-Id: Ic0bd5c7f7da2152e0f3ce4f698bbf598ff1d2eb2
When pipe is added to command, it should still exit with its
result code, so save it by setting pipeline option before each
command.
Close-Bug: #1676156
Change-Id: Ibbe49b4a15a5b7825447a563fe35af85fd48b3ff
Prepend a timestamp with awk to our deployment and test commands. This
allows html linking from the log servers and also better correlation
with the system logs.
We are not using the "ts" command from the moreutils package to avoid
extra dependencies. The drawback is the lack of sub-second timinig.
It is possible to disable the timestamping by setting the
timestamper_cmd to an empty string.
Also add a README file for the extras-common role.
Closes-Bug: #1673046
Change-Id: I8a2a1e4fde2ba805d3361e2e7a2337e873133ef0
add role var undercloud_network_cidr from undercloud-deploy so it can be
reused by validate-undercloud role for the sanity check script
Removed unnecessary default from other roles.
Change-Id: Id46d7a9e3fdc2028cf345564311c04af323b289f
The main purpose of this commit is to allow custom
tripleo-heat-templates to be used for validation, removing the risk of
getting out of sync with the upsteam templates and letting us validate
the results of the scenario jobs.
In addition to that:
* create a full documentation for the role in the README.md file
* remove and rename role variables to make the naming consistent, remove
variables where overriding them is not possible in the templates
* fix and simplify the validation script, add non-deprecated openstack
client commands where possible
* remove dummy tests and vars files from the role
Change-Id: Iea3c0a3c8f4e10786cb15b60ff28cfb15065165e
wait_for script was removed from incubator repo and added to CI
code in tripleo-ci. We need it for validate pingtest role.
Change-Id: I5aaa0d87fa0229fea8b4e7f32301e72b26f83b09
This will now run bashate on all shell scripts. We could ignore the
following list of bashate errors:
- E006: Line longer than 79 columns (as many scripts use jinja
templating, this is very difficult)
- E040: Syntax error determined using `bash -n` (as many scripts
use jinja templating, this will often fail and the syntax
error will be discovered in execution anyway)
Change-Id: I6f6d454ac54a0ac98d8b8a092fa522cba092a592
Signed-off-by: Gael Chamoulaud <gchamoul@redhat.com>
* silence the console output from curl but keep error output to avoid
garbled download output in the logs
* acquire the floating IP address from the stack output to test the same
way as tripleo-ci
* follow up the heat template differences (key -> key_pair)
Change-Id: I4c2bbbc3d542bf62f375d81f70069dff2490fc36
We need to differentiate local_working_dir from working_dir
as well as decouple the stack user from `ansible_user` var.
Both of these are causing issues as we begin to automate
deployments in more environments.
- Cleanup duplicate variables that are consumed via extras-common
- Note: extras-common depends on the common role in OOOQ
- Cleanup redundant var and superfluous quotes from overcloud-scale
role
- Cleanup redundant comments in <role>/defaults/main.yml
Closes-bug: 1654574
Change-Id: I9c7a3166ed1fc5042c11e420223134ea912b45c5
As more ansible variables are shared or reused across roles it is
important to define these variables in a role that is always
executed. In this case that role is extras-common.
Note: This review is a blocker for https://review.openstack.org/#/c/418998/
Change-Id: I31fd13d7bcb98d73e7f16048c57c027d95faeec5
Currently this will fail if the ansible user is modified, since the
command may be run from a different directory, so the heat template
won't be in $PWD.
Change-Id: I4f44e44f18c11a29b18b271f51fe68e5557353a5
We have built up a lot of cruft over the months within our roles.
This review aims to clean up the unused sections. Now that we have
each role within a single repository, ensuring a clean house
is even more important than when they were separate.
Change-Id: Ibb34b376c516a09ba0f7fdb02580e8803497fc50
The glance cli is having issues with keystone (v2) credentials.
We should be using the openstack cli anyways.
Change-Id: I9a088e8da70871eae90e55ced29ecd361c23c380
Closes-Bug: 1646527
Set validate_image_dir relative to working dir, in case
of different user it shouldn't be defined separately.
Change-Id: Ibfd69cab46b242ff9344da30c0ce78c0af9a7d05