Age | Commit message (Collapse) | Author | Files | Lines |
|
Galera installation in CentOS fails because the mysqld.sock file is in a
different location, despite what the my.cnf configuration tells. This
has been worked around in upstream OSA-galera_server:
https://git.openstack.org/cgit/openstack/openstack-ansible-galera_server/
commit/?id=f2bfbd38513ac8d61ba4e02a4d5ef6cbbca259cc
Change-Id: Ied11571189eff8cbab0f1990a7599b51cad9ea2b
Signed-off-by: Tapio Tallgren <tapio.tallgren@nokia.com>
|
|
Using 'installer' to describe the tool that will deploy the foundations
of a particular XCI scenario is more appropriate than NFVI which
normally describes both the physical and virtual resources needed by
an NFV deployment.
Change-Id: Ib8b1aac58673bf705ce2ff053574fd10cb390d71
Signed-off-by: Markos Chandras <mchandras@suse.de>
|
|
The upstream patchset has been merged, so drop the SUSE
specific workaround when obtaining the status of the galera
cluster.
Link: https://review.openstack.org/#/c/508154/
Change-Id: Ic990e6939fb794a532caf8af2759969198fa5d3a
Signed-off-by: Markos Chandras <mchandras@suse.de>
|
|
A scenario may only support certain NFVIs or distributions so we
need a mapping with all these things so we know what can be tested
and where.
Change-Id: Ibf1b640b762085f58627e05e1d2ca13edfc4d716
Signed-off-by: Markos Chandras <mchandras@suse.de>
|
|
Introduce a new XCI_DISTRO variable to select the distribution to deploy
on the VMs in order to make deployments more flexible and decouple the
VM OS selection from the host one. The default value for this new
variable is to match the host OS but users can always set it to one
of the supported distributions. We can now simply execute the
install-ansible.sh script instead of sourcing it in order to keep
the environment as clean as possible.
Change-Id: Ia74eb0422f983848cde0fb7b220ea1035dfa78bc
Signed-off-by: Markos Chandras <mchandras@suse.de>
|
|
It was committed by accident in c74cfd33fc3c ("xci: OSA: Move all the
OSA specific playbooks to the NFVI directory") so drop it.
Change-Id: I5de91f0aad750a9cb723a46baa63e75beb3180de
Signed-off-by: Markos Chandras <mchandras@suse.de>
|
|
|
|
we maybe introduce other NFVI in the future in XCI.
it is necessary to put the nfvi files to corresponding directory xci/nfvi/$NFVI/files,
otherwise the files directory will be confused.
Change-Id: Iea98167ff0bc8d338a94fe1c064ac0ab396c53d3
Signed-off-by: wutianwei <wutianwei1@huawei.com>
|
|
The location where we were putting the scenarios was wrong as it was
placing them in xci/playbooks/roles/$scenario/$scenario so in the end
the Ansible couldn't find the role we were looking for. As a result,
the role was never executed. Moreover, we can use the synchronize
module instead of the shell which should be both faster and safer.
Change-Id: I3d996652bb6a62d23e259da5674d94596dad4d8d
Signed-off-by: Markos Chandras <mchandras@suse.de>
|
|
|
|
|
|
|
|
Make sure that any local changes to scenarios are being propagated to
the checkouts. This should only happen when we are testing the 'master'
branch of the scenario since it may be desirable to test a specific SHA
in which case we shouldn't modify the checkout code.
Change-Id: I4197f09a3e4d5a54c86905bc556b06e08948fbc2
Signed-off-by: Markos Chandras <mchandras@suse.de>
|
|
|
|
The commit 6b529de0315fe6cd12f6e78c00a5f2f2d3a01e28 in
openstack-ansible-lxc_hosts reverts an earlier change that introduced
systemd-networkd to lxc containers that run OpenStack services. This
prevents the containers from being stopped (ie, lxc-stop fails until it
times out), since systemd-networkd restarts. lxc-stop sends the SIGRT+3
message to the container which the container interprets as "halt".
systemd-networkd conflicts with shutdown, so when the system tries to
enter halt, systemd-networkd gets started again. As a result, the
container does not shut down cleanly.
This patch uses a version of openstack-ansible-lxc_hosts that does not
use systemd-networkd in the containers.
Change-Id: Ie85d2117ce2e6e8afa87810d690eab78e3b3be3d
Signed-off-by: Tapio Tallgren <tapio.tallgren@nokia.com>
|
|
We need to set the 'pipefail' option so if an SSH command fails we stop
the deployment instead of silently ignoring it just because the right
leg of the pipe succeeded.
Change-Id: Iebc720fbd3e6e96fc7135a6cf15e3c0941974d2c
Signed-off-by: Markos Chandras <mchandras@suse.de>
|
|
1. the "+" and "'" will be part of the string in the path parameter of ansible.
So the directories we create are not right. we need remove these char.
2. it miss the "{{" in ansible variable
Change-Id: I30c9804450588c63f5f18c63c1d90cd869fad90e
Signed-off-by: wutianwei <wutianwei1@huawei.com>
|
|
|
|
Move all the playbooks that only make sense for OpenStack-Ansible
deployments to the NFVI/OSA directory where they belong. This further
disassociates XCI from OSA.
Change-Id: Iab8b6dc81d9025a1d85608a98fb1eee0f1c6a69f
Signed-off-by: Markos Chandras <mchandras@suse.de>
Signed-off-by: Fatih Degirmenci <fdegir@gmail.com>
|
|
We need to use proper jinja2 code to set the fact properly.
Change-Id: I903233c0e059edf2af973ea9c08b8bd80795e1a9
Signed-off-by: Markos Chandras <mchandras@suse.de>
|
|
|
|
Bump the bifrost SHA so we can get rid of the testvm.xml.j2 template
which has now made it upstream.
Change-Id: Ia9cdd999a5702f37125d38e0ba68bb31920d2117
Signed-off-by: Markos Chandras <mchandras@suse.de>
|
|
|
|
Upstream change modifying policy.json results in failures with
several heat test case failures with functest healthcheck.
https://review.openstack.org/#/c/526306/
Using the policy.json from the earlier commit seems to solve
the failures which is this
https://git.openstack.org/cgit/openstack/openstack-ansible-os_heat/commit/?id=0b3eb9348d55d6b1cf077a2c45b297f9a1be730d
Change-Id: Iaac0911e73743d1b628cc8d4abe460a0a2a66a84
Signed-off-by: Fatih Degirmenci <fdegir@gmail.com>
|
|
Rest of the OPNFV projects use the variable DEPLOY_SCENARIO so
XCI should be aligned with them as well even though OPNFV_SCENARIO
fits better than DEPLOY_SCENARIO.
Change-Id: Id48c41fa8a1fa9493cfc7a4906f64b6d8ed27d64
Signed-off-by: Fatih Degirmenci <fdegir@gmail.com>
|
|
This is needed in order to update the files before we attempt to
run osa-periodic job.
Change-Id: Ied5a53c294fd753039a6f79fa0ebce24b515cdba
Signed-off-by: Fatih Degirmenci <fdegir@gmail.com>
|
|
|
|
|
|
|
|
|
|
The default CPU model of QEMU GenuineIntel/QEMU Virtual CPU is not the
best option for vCPU in terms of performance. As such, lets use the
real host CPU. This has also been submitted upstream and the template
should be removed once the upstream patch has been merged and the SHA
for bifrost is bumped.
Change-Id: I23d86933604c290d5650b73f420972d0a2b23dc8
Link: https://review.openstack.org/#/c/528677/
Signed-off-by: Markos Chandras <mchandras@suse.de>
|
|
These options seem to help with the functest execution
within XCI VM.
Change-Id: I3b9b6b02fc571feb42543025f8ab179c6db18e0a
Signed-off-by: Fatih Degirmenci <fdegir@gmail.com>
|
|
Tacker is already part of the upstream setup-openstack.yml but as releng-xci
overwrites it, tacker does not get deployed. Adding it to the
setup-openstack.yml of releng-xci does not disturb the other scenarios as it
will only be deployed when the tacker container is defined which only happens
when deploying odl-sfc
Change-Id: Ifd76fc49369c114c55e966eb20658d82c99e14cb
Signed-off-by: Manuel Buil <mbuil@suse.com>
|
|
The host-passthrough mode will allow the VM to use all the host cpu
features so it should provide the best possible CPU performance.
Change-Id: I6ecd0363d046aaefa08f9e97144aa9ca9f1afae5
Signed-off-by: Markos Chandras <mchandras@suse.de>
|
|
In preparation for supporting more NFVIs in XCI, we need to make the
main script NFVI agnostic. Right now, our NFVI is being deployed using
bifrost+OSA so we keep bifrost in the main script since the hardware
provisioning is NFVI agnostic (at least for now) but we move OSA to
its own directory. This should allow more NFVI to plug into XCI by
simply providing a nfvi-deploy.sh script. The user can select the
preferred NFVI using the XCI_NFVI variable.
Change-Id: Ieeacd3b22a64d363feff5d43ff23a80c39038837
Signed-off-by: Markos Chandras <mchandras@suse.de>
|
|
|
|
|
|
Scenarios may not support all XCI flavors so record that information
per scenario basis. This will resolve failures when we try to deploy
a scenario that doesn't support the selected flavor.
JIRA: RELENG-338
Change-Id: I5f73b139ee6c1831896aa32147c4dc1368673d92
Signed-off-by: Markos Chandras <mchandras@suse.de>
|
|
This ensures that the virtio paravirtualization driver is used for both
disk and network I/O which should provide the best performance since we
no longer need to emulate a real bus or driver.
Change-Id: Ic46a38f51b06cb4b848d87d890d23fec022b2f15
Signed-off-by: Markos Chandras <mchandras@suse.de>
|
|
The configure-opnfvhost playbook which is used for all flavors except
AIO can also be used for AIO if it's adapted to not configure networking
on the host and also do not fail if there aren't any OSA variables files
available so lets do all that in order to avoid duplicating code across
flavors.
Change-Id: I58379d8b52094294b9349494753ffcdd44982013
Signed-off-by: Markos Chandras <mchandras@suse.de>
|
|
|
|
The OPENSTACK_OSA_PATH only makes sense on localhost. As such, when we
use it on playbooks that operate on remote hosts, the result is not
predictable. However, we rsync the entire releng-xci repository to the
opfnv host so we can make everything predictable by simply clone
everything in advance in the .cache directory. That directory is then
rsync'd to the opnfv host. As such, we can repurpose the
OPENSTACK_OSA_PATH to point to the path into the OPNFV host. Moreover,
all external repositories are being cloned to .cache/repos so we can
eliminate some variables in order to simplify the code. Finally, we
bring back the ability to use an external OSA repository for
development purposes.
Change-Id: Ieef3e22ae2085f6735185634d555cfc0d4b69b39
Signed-off-by: Markos Chandras <mchandras@suse.de>
|
|
There is too much noise when XCI is cloning repositories, installing
packages etc so lets make the console output somewhat more readable.
Change-Id: I15667e0349f943ad5488daf0e3cea9336c9166d1
Signed-off-by: Markos Chandras <mchandras@suse.de>
|
|
The variable was accidentally removed in ab3c9ad0f9a4 so bring it back.
Change-Id: I48e4b8dd68969c8e5336e5e571a73f037fb65988
Signed-off-by: Markos Chandras <mchandras@suse.de>
|
|
|
|
|
|
This change fixes the scenario os-nosdn-nofeature by
moving directories/files from os-nosdn-ovs to os-nosdn-nofeature.
The contents of the files and variables are also either adjusted
or removed in order to prevent impacts on CI.
Change-Id: Icfde27f413c8f93d097f9f262c8cb1230b7fe59d
Signed-off-by: Fatih Degirmenci <fdegir@gmail.com>
|
|
Change-Id: I5a3f533f61b6572d8070fe5f0a306ef9d418ced8
Signed-off-by: Manuel Buil <mbuil@suse.com>
|
|
|
|
|