Commit Graph

9694 Commits

Author SHA1 Message Date
binhong.hua
3d3f5f1613 add ulimit support for kolla_docker
By default, docker containers inherit ulimit from limits of docker
deamon. On CentOS 7, docker daemon default NOFILE is 1048576.
It can found in /usr/lib/systemd/system/docker.service.
The big limit will cause many problem. we should control it in
production environment.

Change-Id: Iab962446a94ef092977728259d9818b86cfa7f68
2019-01-24 10:54:14 +08:00
Zuul
cd0d8cca9b Merge "Ensure keystone endpoint is updatable" 2019-01-17 09:42:22 +00:00
YI-JIE,SYU
d39ba689ac Update tacker CLI to openstack CLI in cleanup-tacker
Change to openstack CLI

Change-Id: I75db9548b94800b6a451a658931d11a333a8e054
2019-01-16 21:12:48 +08:00
Zuul
20c58a9c53 Merge "Run Zun job on kuryr's changes" 2019-01-15 11:06:16 +00:00
Zuul
0abf5c8e43 Merge "kuryr: use internal neutron endpoint" 2019-01-14 14:45:39 +00:00
Zuul
d6c395ce98 Merge "Fix command connect to replica set primary" 2019-01-14 09:28:42 +00:00
Hongbin Lu
54761726c8 Run Zun job on kuryr's changes
Zun depends on Kuryr so let's trigger the Zun job on kuryr's change.

Change-Id: If95a6bfb5f305c2f69808a781ff9d8513bf2006c
2019-01-13 22:43:47 +00:00
Hongbin Lu
a98c06f1fd kuryr: use internal neutron endpoint
Kuryr-neutron communication should use internal endpoint because
public endpoint might be protected by SSL which fails the API call.

Change-Id: I3f98f14d4f481c17a4368cc5c2a184323c548193
Closes-Bug: #1811262
2019-01-13 22:39:31 +00:00
zhangshj
199133a9cc Correct two wrong words
Change-Id: I89d9f7a90d72e6c5df7dde078ed7407eedd6f92f
2019-01-13 10:27:26 +00:00
Zuul
aab2dd4669 Merge "Use kolla-ansible version to tag images" 2019-01-11 16:12:23 +00:00
Michal Nasiadka
03d27cb8d9 Add ceph docs note about rgw keystone integration.
Change-Id: If87cdb132875c4e2c5dff2cc427e00f7a95927de
2019-01-10 14:59:55 +01:00
Zuul
4aa89776ba Merge "Doc: Update Kolla Ceph document for mixed Ceph OSD deployment" 2019-01-10 09:11:46 +00:00
Mark Goddard
65aa706a44 Use kolla-ansible version to tag images
If building images, kolla currently uses its own version for their tag
by default.  This might not match the version of kolla-ansible, which is
what is used by default in kolla-ansible. This leads to the image not
being found.

This change uses kolla-ansible's version to tag locally built images.

Change-Id: I49c7afb9167d1e3f8a4470b6ebbb6ad6798d4996
2019-01-09 18:13:27 +00:00
ZijianGuo
30f8511cd5 Fix command connect to replica set primary
Without this option would cause non master error.

Change-Id: I5da653227fb1b2af7a36f39789265584ac0146da
Signed-off-by: ZijianGuo <guozijn@gmail.com>
2019-01-08 20:12:26 +08:00
Zuul
beee4f7c96 Merge "Make Rabbitmq config files overridable" 2019-01-08 10:34:54 +00:00
tone.zhang
4ff993c45e Doc: Update Kolla Ceph document for mixed Ceph OSD deployment
Since Rocky release Kolla Ceph supports both bluestore and
filestore Ceph OSD. In multinode mode, users can overwrite
“ceph_osd_store_type” in the inventory file and deploy the mixed
Ceph OSD.

Change-Id: I4582861c81fdea5fb67114cf3ac1e48f7812f0d7
Signed-off-by: tone.zhang <tone.zhang@arm.com>
2019-01-08 16:28:07 +08:00
Zuul
e9fda8664e Merge "Add Kolla Ceph filestore CI jobs" 2019-01-07 11:35:54 +00:00
jacky06
d55129d126 Update the remain hacking version for lower-constraints
It missing from https://review.openstack.org/#/c/627149/, this ps
to add it.

Change-Id: If852e608429df25edf4dd4676d35d3957cdeee3e
Closes-Bug: #1810530
2019-01-04 10:06:28 -05:00
Nick Jones
683f65aaf3
Fix Telegraf configuration template
Remove the erroneous curly braces that prevent the Telegraf
configuration template from being rendered.

Change-Id: Ibb8738b258245b4e43d5d877b156ed296fc66e4e
Closes-Bug: 1810505
2019-01-04 12:11:39 +00:00
Zuul
2ed4649ebd Merge "Use <project>_install_type instead of kolla_install_type" 2019-01-04 10:00:18 +00:00
Zuul
1ffed6f2d8 Merge "Add documentation for bootstrap-servers command" 2019-01-04 08:34:52 +00:00
Kien Nguyen
043943117d Use <project>_install_type instead of kolla_install_type
Use <project>_install_type instead of kolla_install_type
to set python_path. For example, general kolla_install_type
is 'binary', but user wants to deploy Horizon from 'source'.
Horizon templates still use python_path=/usr/share/openstack-dashboard,
it is wrong.

Change-Id: Ide6a24e17b1f8ab6506aa5e53f70693706830418
2019-01-04 14:33:46 +07:00
Zuul
335c27c726 Merge "Add customizable gnocchi volume" 2019-01-03 12:06:13 +00:00
Mark Goddard
91e17b23d4 Add documentation for bootstrap-servers command
Change-Id: I3bd50cb1e7db85acbf8ab20fc7bdd945b5232aaa
2019-01-03 10:29:05 +00:00
Zuul
f939b710c6 Merge "Allow separate listen port in haproxy-config templates" 2019-01-03 09:52:46 +00:00
ZijianGuo
6d876934ea Add customizable gnocchi volume
Change-Id: I8d21546847a2f5fa418e07cfe79016a349af4783
Signed-off-by: ZijianGuo <guozijn@gmail.com>
2019-01-02 11:49:41 +08:00
tone.zhang
6786bc5b77 Add Kolla Ceph filestore CI jobs
Since Rocky bluestore is the default store type used by Kolla Ceph.
Filestore is still used by the customers. The patch allows deploy
both Ceph bluestore OSD and Ceph filestore OSD in Kolla-ansible CI
jobs.

Change-Id: Iaa7c1b2ff35eb649c3a6d9f7266463d8961cdc9e
Signed-off-by: tone.zhang <tone.zhang@arm.com>
2019-01-01 13:32:16 +08:00
Zuul
69666b6665 Merge "Support Docker CE in bootstrap-servers" 2018-12-28 12:21:26 +00:00
Jim Rollenhagen
3ea51cefe9 Allow separate listen port in haproxy-config templates
This allows a service to pass in a different listening port than haproxy
listens on for the service. This allows haproxy to live on the same
network interface as the services, and will be used by services in
subsequent changes.

Change-Id: Id1d2429a4c800628cb43e0fb7d2b69131e71ee1c
Implements: blueprint service-hostnames
2018-12-27 15:45:50 -05:00
Dai, Dang Van
8d5355dbc1 Fix bootstrap prometheus container location
This change to fix the case that
I won't use prometheus-mysqld-exporter

Change-Id: I1936bbae0172f4e65605d71066dced837bc30f7a
2018-12-27 12:46:22 +00:00
Zuul
4ded787e79 Merge "Allow custom horizon policy files for disabled openstack projects." 2018-12-27 10:19:58 +00:00
Zuul
4cad49284b Merge "Update hacking version" 2018-12-27 09:46:23 +00:00
confi-surya
b41309ef16 Update hacking version
Use latest release 1.1.0 and compatible changes w.r.t pep8

Change-Id: Ie5e5417a1d8481bb59d516a320bd7cc1dd89591e
2018-12-27 04:44:49 +00:00
wu.chunyang
4939fc28d2 remove cinder v1 endpoint
now,cinder does not support v1 endpoint,so no need to create it in
keystone

Change-Id: I4b3bbf24c19a996e73f4b360beeb20a110907a6d
2018-12-24 14:39:43 +00:00
Zuul
54da9fa92c Merge "Add become for Octavia configuration tasks" 2018-12-24 08:36:02 +00:00
Zuul
04fb9981e4 Merge "Fix Vitrage wrong configuration" 2018-12-24 07:36:17 +00:00
Kien Nguyen
90fd7aa03d Fix Vitrage wrong configuration
* In service_credentials section, project_name should be
admin and auth_url should use keystone public port (5000)
according [1].
* User vitrage should have role admin in project admin [2].

[1] https://github.com/openstack/vitrage/blob/master/devstack/plugin.sh#L164
[2] https://docs.openstack.org/vitrage/latest/install/install-rdo.html#initialize-vitrage

Change-Id: I8bf028c702df5659e4c58df0e17600ab5be14eea
2018-12-23 15:46:23 +00:00
Kien Nguyen
b1d7a8ad22 Add missing transport_url config
In oslo_messaging_notification section - vitrage conf,
kolla-ansible is missing transport_url.

Change-Id: I0100af9771dfdb9cd40ce5effff2a60c7bd43575
2018-12-23 22:32:38 +07:00
dommgifer
5c771df7d1 Add become for Octavia configuration tasks
This is required to support execution as a non-root user.

Change-Id: Idd2deec2918e5eb7743e60112a91cb31edf5aad3
Close-bug: #1809527
2018-12-22 15:04:13 +08:00
Martin Chlumsky
18dd74ac8d Allow custom horizon policy files for disabled openstack projects.
Currently, you cannot deploy custom policy files in horizon for disabled
openstack projects that don't have a horizon plugin.

This patch allows customizing disabled openstack projects policy files
used by horizon.  For services that do not have horizon plugins, it
creates variables of the form `enable_*_horizon_policy_file` and sets
these to the relevant `enable_*` variables by default. This ensures
backwards compatibility.

This patch is useful when you are migrating to kolla-ansible from
another deployment method one openstack project at a time and you have
custom policy files deployed by the old method.

Partially-Implements: blueprint docs-migrating-to-kolla
Closes-Bug: #1809314
Change-Id: Ifc33ef65759fa4831c24d177b272af5b45f80931
2018-12-21 11:39:53 -05:00
Zuul
17f9ff5a10 Merge "Add NFV orchestration services CI job" 2018-12-21 13:15:50 +00:00
Zuul
a3628f63a8 Merge "Add become for Prometheus configuration tasks" 2018-12-21 11:13:27 +00:00
Eduardo Gonzalez
535aba79a9 Add NFV orchestration services CI job
Change-Id: I33484d0961f1cd17ed05dd124eba68d99b4db83f
2018-12-21 11:43:22 +01:00
dommgifer
69823f8692 Add become for Prometheus configuration tasks
This is required to support execution as a non-root user.

Change-Id: I60d224407c2828d6b9f1701f7637385a25fbcced
Closes-Bug: #1809233
2018-12-21 16:59:18 +08:00
confi-surya
e73fe77d7d Use template for lower-constraints
Small cleanups:

* Use openstack-lower-constraints-jobs template, remove individual
  jobs.
* Sort list of templates

Change-Id: I67199fabe6a9f7b1fd38dac77a6157bf4fb465b9
Needed-By: https://review.openstack.org/623229
2018-12-21 07:33:19 +00:00
Zuul
d6ad05ca07 Merge "Support list of controller ip for the amphorae heartbeat receiver" 2018-12-20 17:34:46 +00:00
Zuul
de71cf2f21 Merge "Test zun container creation in gates" 2018-12-20 09:31:49 +00:00
Zuul
acda7ba592 Merge "Configure vitrage notification topic" 2018-12-20 03:26:54 +00:00
Eduardo Gonzalez
362b6ee40e Test zun container creation in gates
Change-Id: If5b4ba975a65e07d2704eb6bdb9d841d6a9c3d42
2018-12-19 19:50:59 +01:00
Duc Nguyen Cong
144d9bbc2d Support list of controller ip for the amphorae heartbeat receiver
In multi controller deployment, kolla will generate
"controller_ip_port_list option" in [health_manager] section with
ONLY IP of that node instead of a list of controller ip.
Therefor, "amphora-agent.conf" file of amphora instance will
contain IP of ONLY ONE controller node.
In case of that node fail, amphora agent won't send heartbeat
message to other health manager node, and the loadbalancer will
go to ERROR state.

Change-Id: I102ed6ba3fff2c12cc6d37f81ad59508eacc859c
Co-Authored-By: Hieu LE <hieulq2@viettel.com.vn>
2018-12-19 16:30:09 +00:00