| Age | Commit message (Collapse) | Author | Files | Lines |
|
|
|
Now we use flunetd version is v12.0
we use fluent-plugin-grok-parser plugin shoud < 1.0.0.
https://github.com/fluent/fluent-plugin-grok-parser
Change-Id: I013e1e6d1070c4ba050be5dd75be7be4df652e4e
Closes-Bug: #1669405
|
|
|
|
neutron-metadata-agent image expected the /var/lib/neutron/kolla
directory to be present but it was not created as part of the image
building. As a result, the extend_start.sh script fails with:
stat: cannot stat '/var/lib/neutron/kolla': No such file or directory
This commit defers setting the right ownership for files and directories
to the kolla config API via a config.json file.
Kolla-ansible already mounts the directory as the
neutron_metadata_socket named volume and uses kolla API to set the
ownership and permissions [1].
[1] https://github.com/openstack/kolla-ansible/blob/master/ansible/roles/neutron/templates/neutron-metadata-agent.json.j2
Closes-Bug: #1669487
Change-Id: Ib521b0fc69ab6a72295a2fde3309a001bc3063c2
|
|
Change-Id: I2b2ef84589ca6cbb0d27934000f9640a8eb202a3
Closes-Bug: #1660092
|
|
|
|
|
|
|
|
|
|
|
|
Shade has new functionality we need for sanity checks. This also
requires a newer os-client-config.
Change-Id: If6d10f2ce90c7fc72353566554352b0310942f30
Partially-Implements: blueprint sanity-check-container
|
|
|
|
Heka is removed and replaced by fluentd.
Change-Id: I4ed0f61b02bea5dcb02b61a9eb99e3ff3f460682
|
|
|
|
/etc/openstack-dashboard/local_settings.py is deleted in the followup
command, the sed is never used.
Change-Id: Ife02321a4077544ae13428968c4bfe27eb7a7a48
|
|
Change-Id: Idb5a40f080937c74008caf2650ba82d82ab520c4
|
|
The location of the binaries in these images can be overridden by
setting ENV in the header blocks, but if an operator wants to use a
different install method all together (e.g. rpm), these blocks will
allow that.
Also import the macros in kube-base to allow their use in overrides.
Change-Id: I3e52e315f0103d8a1c95d4489fa3841900fdff75
|
|
When built with the binary install type (at least on CentOS), Ironic
inspector fails to start and the container remains in a restarting
state. The log file shows that it is failing to execute iptables, and
analysis found that this was due to an incorrect rootwrap
configuration. The RDO ironic inspector RPM expects Ironic inspector
to be run as the ironic-inspector user, however Kolla uses the ironic
user. This means that neither of the packaged ironic nor
ironic-inspector sudoers configuration files works for us.
Kolla currently installs a sudoers file pointing to the rootwrap
script in the virtualenv of the source install, but of course this
only makes sense for source installs, and should not be installed for
binary installs.
This change adds a second sudoers file that will work for the binary
install type, and installs the correct sudoers file for the install
type.
Change-Id: I8ecd0b658b8df8f38ddf717fa9443d4dc2896984
Closes-Bug: #1624457
|
|
|
|
|
|
Allow operators to customise these using the usual Kolla mechanisms.
Change-Id: I59d5eae43fb12f8301b26f5b693c104fca635762
|
|
|
|
|
|
Allow overriding of kubernetes-helm to facilitate an alternate location.
Change-Id: I630d1e1368ddbbbc795048a73ecc182952da751d
|
|
Allow operators to customise the location k8s binaries are pulled from
in kubernetes-entrypoint & kubetoolbox
Change-Id: I8bab3fe7dfb9ac23dcfbfd2c709cfddaaab40534
|
|
These packages are needed by ironic console.
Change-Id: Ieff5142935fb55cc9860d6cd4f195da095a01449
Closes-Bug: #1666131
|
|
Installs `uwsgi-plugin-python` and dependencies in `barbican-api` image.
This will resolve problems bootstrapping barbican-api with binary Centos
images which otherwise fail to start due to:
```
Running command: 'uwsgi --master --emperor /etc/barbican/vassals --logto /var/log/kolla/barbican/barbican-api.log'
/usr/local/bin/kolla_start: line 24: exec: uwsgi: not found
```
Tested with:
```
docker run --rm -ti --entrypoint=/bin/bash
kolla/centos-binary-barbican-api:3.0.3
$ uwsgi
*** Starting uWSGI 2.0.14 (64bit) on [Sat Feb 18 09:32:12 2017] ***
compiled with version: 4.8.5 20150623 (Red Hat 4.8.5-11) on 13 January 2017 23:06:53
os: Linux-4.9.4-moby #1 SMP Wed Jan 18 17:04:43 UTC 2017
nodename: b9063757e276
machine: x86_64
clock source: unix
pcre jit disabled
...
```
Closes-Bug: 1634010
Signed-off-by: Magnus Lööf <magnus.loof@basalt.se>
Change-Id: I77613bbeb19d89e6a25e52f24f3275d309ceb032
|
|
helm-repository Dockerfile"
|
|
|
|
Dockerfile
Allow overriding the pip installation so that it can
be easily customised if needed.
Change-Id: Ib60e2e26841d011aef7fdf1fab936bcdc79eb336
|
|
|
|
1. Enable customization of pip packages in source
branch of most images
2. All pip packages install uniformly through
install-pip macro, user can easily customize his
own pip command (For example using a mirror)
Co-Authored-By: Mauricio Lima <mauriciolimab@gmail.com>
Change-Id: If09582039f690fa4136e8f33200d5da15e092da7
|
|
|
|
Bifrost now targets the play in the install.yml playbook at the
target Ansible group instead of localhost.
This change uses the target inventory file to pick up this group
and avoid a 'noop' playbook.
Change-Id: I1b351ccfea5684aeb1394e5a12ee848eb62447fe
Closes-Bug: #1665413
|
|
|
|
|
|
TripleO makes use of a heat-all container for undercloud
installation. This container includes the API, Engine, and
a new heat-all launcher (included in the monolith package).
Implements: blueprint heat-all
Change-Id: If1cc3e959b63f775e15e6eeef4b54981fb3793e9
|
|
grafana package is signed now.
Change-Id: I5e0cfe138198b37328c7a7f56413f48999118cfa
Closes-Bug: #1664272
|
|
mariadb.log mode is 660, which can not be read by fluentd. Add td-agent
user into mysql group to allow fluent to read the log.
Closes-Bug: #1664867
Change-Id: If624002909d5fcd0bb82a5bc9937945518ec84b9
|
|
|
|
|
|
There are two issue in kolla gate now. Since source building is voting,
we have to squash these two fixes into one
Disable gpgcheck in grafana repo
Seems there is something wrong in grafana repo gpg key. Disable gpgcheck
until it is fixed
Partial-Bug: #1664272
Add VENV variable in bifrost base Dockerfile
* ANSIBLE_FROM_PYPI is removed in bifrost.
* Only when VENV exist, ansible is installed in global rather then root
user
Change-Id: Ib96433e6d2a03941cace8ed15dafbb314c4551eb
|
|
Updating kubernetes entrypoint image to version 0.1.1 which includes
important for kolla-kubernetes fix and updated kubectl client.
Closes-Bug: #1664210
Change-Id: I0cf6fd2754c146c98aea1f44d24f5efccd00a310
|
|
This change fixes one warning:
INFO:kolla.image.build.base:W: GPG error: http://download.ceph.com jessie InRelease: The following signatures couldn't be verified because the public key is not available: NO_PUBKEY E84AC2C0460F3994
Change-Id: I2a1109e787de9a342a7c359c90619cb277c7c2df
Signed-off-by: Marcin Juszkiewicz <marcin.juszkiewicz@linaro.org>
|
|
|
|
|
|
|
|
The deployment of a service requires the creation of a new service
endpoint into Keystone. To create a new service endpoint the
kolla_keystone_service tests if the service is already registered and
updates the URL if so, or create the new endpoint in other cases. In the
actual implementation, the register test only looks at the service id
and interface which makes the test behaving wrong in multi-regions
context. In multi-regions, we deploy the same service many times, and
thus, actual implementation considers that this is the same service and
updates its endpoint rather than creating a new one. This fix adds a
condition to the test that verifies the region to distinguish two
identical services deployed into two different regions.
Change-Id: Ie7ca8800fe3c5a9585e4b0bbbc38ac628ec753bc
|
|
|
|
Change-Id: I4c22803f2d7fbea88ad2d2d4648cf498905d4013
|