Age | Commit message (Collapse) | Author | Files | Lines |
|
switch dpdk interface driver and bring
up corresponding ovs bridges.
Change-Id: I07bd9b332978482507711cfbd8ad29802a8b1a8c
Signed-off-by: Harry Huang <huangxiangyu5@huawei.com>
(cherry picked from commit 849d51833bdb16e8f04370270cd3e082ad9db928)
|
|
JIRA: -
For offline deployment, some packages' versions need to be
upgraded and add the version limitation of tacker and tacker
client.
Change-Id: I8b55fe270998c716df8791f1909ea4f4ab654a6c
Signed-off-by: Yifei Xue <xueyifei@huawei.com>
(cherry picked from commit 2b48cb8ce734459ccfe4c106d0c398d7e55c972f)
|
|
JIRA: -
Add tacker support in compass euphrates. Upsteam OSA currently doesn't
support tacker in Ocata. But OPNFV needs to use tacker in SFC scenario.
So we refactor os_tacker of Pike to satisfy OpenStack Ocata.
Change-Id: I40bb90949d8cd3b9db166527b0d85ff42ad5cdb9
Signed-off-by: Yifei Xue <xueyifei@huawei.com>
(cherry picked from commit d0886a8dcf47ef0d27a0b7960dd152596f0a3a88)
|
|
|
|
When storage nodes reboot or shutdown,
the partitions of loop device will lose.
We add the command partprobe -s {{loopdevice}} to rc.local,
it will reload the partitions, when boot up the storage nodes.
Change-Id: I31dfca953aa254fa516421a494318b01cd39675c
Signed-off-by: wutianwei <wutianwei1@huawei.com>
|
|
When not define ceph role in DHA, there are somes issue on config-osa.
This patch add the judgement to check the parameter define or not.
Change-Id: If221fd3d02c87fc60f545b44f94043f38e34d283
Signed-off-by: wutianwei <wutianwei1@huawei.com>
(cherry picked from commit 660b8bfbb0546322f35f8c7f54d62e1bf5d5dd95)
|
|
OSA don't maintain the ceilometer component.
We change the OSA code to make ceilometer work.
1. create ceilometer database
2. change the ceilometer policy
3. change the ceilometer config
4. change the OS_AUTH_TYPE in openrc
Change-Id: I2d8630a0c50f5c6bd1e86171cae1fa8f9934334b
Signed-off-by: wutianwei <wutianwei1@huawei.com>
(cherry picked from commit f55635085312ab2c367d1ae37d0979874b2a7d82)
|
|
JIRA: -
In the Euphrates release, compass has supported kubernates
deployment now. However, there is no k8s related description
in docs of compass4nfv. So we add k8s related description in
this patch including the k8s architeture, k8s networking in
Compass, k8s virtual deployment and k8s bare metel deployment.
Change-Id: I3bea3ae06b1c73076ba44d10caf89602928b7021
Signed-off-by: Yifei Xue <xueyifei@huawei.com>
(cherry picked from commit 4f90858ffbe3e566f30f8bf58309d28989212d0e)
|
|
JIRA: -
Currently Compass support 3+2 highly available Kubernetes cluster.
This yaml file define 1+1 Kubernetes cluster for k8-nosdn-nofeature-noha scenario
Change-Id: Iaeb4ef854a88657813fab9d255120ac93854dc6f
Signed-off-by: hu xinhui <xinhui_hu@foxmail.com>
(cherry picked from commit 2e3e4dcfd50ad29f80f675fcefff3c3a579338e5)
|
|
the 0.21.0rc1 version of pandas will block the gnocchi install,
so we use 0.20.3. we replace the pandas package in repo
Change-Id: Ic8d9fa3a691b64a0783034ffbf0d958427e30d16
Signed-off-by: wutianwei <wutianwei1@huawei.com>
(cherry picked from commit 05aafa14c76537222fac2f93fc6ab7bec741cff5)
|
|
JIRA: -
There is some hardcode for kargo inventory.This patch is to generate dynamic inventory for kargo
Change-Id: If4a7c5f973e520e4ca24c3c9e0bfa97725a5a44c
Signed-off-by: hu xinhui <xinhui_hu@foxmail.com>
(cherry picked from commit 179cd6db9f1bd9be1606187cf74e128bede7a8e6)
|
|
1. set the log level debug to true
2. add oslo_middleware enable_proxy_headers_parsing: True
in cinder.conf
Change-Id: I555345ab2e1ea2ff0b56f0b9279ed7141bde2282
Signed-off-by: wutianwei <wutianwei1@huawei.com>
(cherry picked from commit b3aec6773447e8ab0b532723b08fd76ebddf9d8c)
|
|
JIRA: -
Currently Compass uses the gateway of install network as the only
DNS server of host1 to 5. Sometimes it doesn't work in some
environments. This patch we add a variable named USER_NAMESERVER
which can be exported before deployment. It is used to store a
user defined DNS nameserver.
Change-Id: Id74889c20cf09ec7e25d882b56ab6331ee64cdcf
Signed-off-by: Yifei Xue <xueyifei@huawei.com>
(cherry picked from commit 7b110b32e0688896967703dc2bfe7192d469164f)
|
|
JIRA: COMPASS-563
The switch controls offline deploy missing
default value which cause passing a empty
value into ansible and lead to condition
check error
Change-Id: I0af4df0556dcbe30dacc743456728e8bdee698e8
Signed-off-by: Harry Huang <huangxiangyu5@huawei.com>
(cherry picked from commit 7e70daca6c7ac54e03fc72f6e27940c778ed5521)
|
|
JIRA: COMPASS-562
Because of the inventory file modification, the sfc switch cannot be
enabled. Add a sperate sfc role to fix this bug.
Change-Id: If583d10ebbe78d0abe9078add24beb5f506dbda8
Signed-off-by: Yifei Xue <xueyifei@huawei.com>
(cherry picked from commit 1c0f6a907f57e75529dcef959763220c10e7e1f0)
|
|
Due to pod2 virtual3 and virtual4 run danube before,
we need to change the network and DHA to run euphrates
Change-Id: I720138cb960cc6f3a57bf5f777c0debb80f64eba
Signed-off-by: wutianwei <wutianwei1@huawei.com>
(cherry picked from commit f70d3aa8ffe7878e86b8f12c91c99018924a74b3)
|
|
|
|
Add k8s DHA file k8s-nosdn-nofeature-ha.yml
when run k8s-nosdn-nofeature-ha scenario, export two variables
COMPASS_OS_VERSION=centos7
KUBERNETES_VERSION="v1.7.3"
Change-Id: I9012598e6d46b20f325e93aba92f03da444d97a2
Signed-off-by: wutianwei <wutianwei1@huawei.com>
|
|
Add network configuration files and scenario files
to fix running error in Jenkins
https://build.opnfv.org/ci/job/compass-deploy-baremetal-daily-master/2888/console
Change-Id: Ie0e930e143d520c553eac41e56cf4b2504374e30
Signed-off-by: chigang <chigang@huawei.com>
|
|
|
|
JIRA: COMPASS-550
add OVS-DPDK plugin, there are two roles in this plugin.
1. ins_dpdk is for dpdk installation and hugepages setting.
2. ins_ovs is for openvswitch installation.
add dpdk scenario networking configuration example called "network_cfg_dpdk.yaml"
Change-Id: Ifd8c1aadc218753f99bc26bb530e7cf9962ad8e3
Signed-off-by: chigang <chigang@huawei.com>
|
|
JIRA: -
Currently the latest tag is used in building tarball.
As OPNFV Euphrates will be released soon, we change the
tags of some docker images to euphrates.
Change-Id: Ic7a746f3a3338cf7f5cacd2f932816b3776e7de6
Signed-off-by: Yifei Xue <xueyifei@huawei.com>
|
|
JIRA: -
Add a new k8s scenario for compass
Change-Id: Ic5f58a6152315333684e4f2752aaa0d5d870d9ee
Signed-off-by: hu xinhui <xinhui_hu@foxmail.com>
|
|
Change-Id: I2d1cbc6948d2cea1ccb852ed23033d5ddc6393e4
Signed-off-by: wutianwei <wutianwei1@huawei.com>
|
|
JIRA: COMPASS-559
1. support multiple mapping in provider_net_mappings
of network.yml and create corresponding ovs bridges
2. support seperate config in sys_intf_mappings
of network.yml
3. remove linux bridges in compute nodes for ovs
port binding convenience
4. support openstack vlan tenant network
5. modify odl and odl_sfc roles to use configurable
provider mapping
6. remove some hard coding
Change-Id: Ib57484ce60d029f89c647fd5baf2c7af37c85d0b
Signed-off-by: Harry Huang <huangxiangyu5@huawei.com>
|
|
|
|
JIRA: -
Add a repo docker which includes ubuntu PPA and python packages
used for offline deployment.
Change-Id: Iad9b0df7f33f5a7ab974c5deffb408fc93b3f3ad
Signed-off-by: Yifei Xue <xueyifei@huawei.com>
|
|
Netmask of install network is parsed from
netwok.yml now. A var named INSTALL_NETWORK
will be generated in script/deploy_config.sh
and exported. But var of install mask in
deploy/network.sh is called INSTALL_MASK.
This mismatch cause empty netmask of install
network.
Change-Id: I2a468844a74a9c38481615d6353149fa74ccbcd9
Signed-off-by: Harry Huang <huangxiangyu5@huawei.com>
|
|
|
|
Docker install only exist in off-line deploy
flow. Add it into normal deploy flow.
Change-Id: I2addf54eb6f74cb8faf799db343d84eaacea9fcb
Signed-off-by: Harry Huang <huangxiangyu5@huawei.com>
|
|
1. add parameter to install keepalived when only one controller
2. change the osa code to start keepalived and haproxy
3. add the noha DHA files
4. fix the setup ovs bug
5. fix the pod1 public vip
Change-Id: I19f4e78b9ca834ae3d454e1c8058728ece055f69
Signed-off-by: wutianwei <wutianwei1@huawei.com>
|
|
|
|
1. ceph.yml define which host install ceph-mon and ceph-osd
2. user_ceph.yml define variables of ceph and cinder backend
3. user_variables_ceph.yml.j2 define variables of ceph for user
Change-Id: Iafea998f4603eb4b5c15a26f2387e6cb04c35d4c
Signed-off-by: wutianwei <wutianwei1@huawei.com>
|
|
JIRA:-
TASK [moon : configure moon.conf for moon slave]
fatal: [host5]: FAILED! => {"failed": true, "msg": "'moon_slave_name' is undefined"}
https://build.opnfv.org/ci/job/compass-deploy-virtual-daily-master/3311/console
Change-Id: Ic6211ef7017e0ed794718f1705ba566af370ea49
Signed-off-by: chigang <chigang@huawei.com>
|
|
1. modify interface file for moon according
to new network structure
2. modify some ips for pods running master
Change-Id: Idfca20c8e101222d27847fcac4f12ba21a1b5a67
Signed-off-by: Harry Huang <huangxiangyu5@huawei.com>
|
|
JIRA:COMPASS-557
1. modify sys_intf_mappings and ip_settings in
network.yml to synchronize with OSA network.
networks compass build will be mgmt, external,
tenant, storage.
2. modify compass_conf to support new netwok.yml
3. fix some hard coding when transfer ip parameters
4. support heterogeneous network
Change-Id: Ib4a8af1f96a40e6456242e3dc2b456e0c8c17c2e
Signed-off-by: Harry Huang <huangxiangyu5@huawei.com>
|
|
|
|
1. functest change the judgement condition, the DOMAIN_NAME need use Default instead of default
2. making the eth0 down and up to change the route in baremetal deploy
Change-Id: I7dacf59cdd0c4db396b2db8b0a5fdd7003829391
Signed-off-by: wutianwei <wutianwei1@huawei.com>
|
|
There is a missing indent in the judgement
code block for OS installation complete which
cause OS installation complete only affected
by the state of host1
Change-Id: I5d470aef93b65ca77d290e596a4fefd62a82517b
Signed-off-by: Harry Huang <huangxiangyu5@huawei.com>
|
|
|
|
Dynamic inventory support for compass4nfv/compass-tasks-osa
is ready, switch docker repo.
Change-Id: I9ad5cd1cc860dd649cf43c6d5f1ea1e7c39e20fc
Signed-off-by: Harry Huang <huangxiangyu5@huawei.com>
|
|
JIRA:-
For KVM scenario, each host should set cpumode pass-through,
otherwise instances will fail to launch.
Change-Id: If58dca1ad590203f6e8de36e44c242252350ad19
Signed-off-by: chigang <chigang@huawei.com>
|
|
The DHA name of SFC scenario in compass4nfv repo is "
os-odl_l3-sfc-ha.yml". However, the scenario name on
CI master is "os-odl-sfc-ha.yml". So CI failed when read
DHA file. THis patch is proposed to fix this problem.
Change-Id: Icb8db6c6c023a794bdd803d0365d719a853dc712
Signed-off-by: Yifei Xue <xueyifei@huawei.com>
|
|
Add SFC support for Ocata and Carbon.
1. Add ovs-nsh package for sfc scenario
2. Add sfc configuration and tasks
3. Add sfc DHA files
Change-Id: Icf8f20220a2c5ac38a5b92d3ca0e24cbb0e89a52
Signed-off-by: Yifei Xue <xueyifei@huawei.com>
|
|
|
|
|
|
when OS_USER_DOMAIN_NAME and OS_PROJECT_DOMAIN_NAME
is Default in openrc, it will cause some issue in
functest.
Change-Id: I39a7a3712ac2311076c45caed70c0c4c2036b561
Signed-off-by: wutianwei <wutianwei1@huawei.com>
|
|
JIRA:-
To deploy moon master, export MOON_CFG env like below compass deploy:
export MOON_CFG="master:flag=Disable,slave:flag=Enable,slave:name=slave1,slave:master_ip=master_ip"
To deploy moon slave, export MOON_CFG env like below compass deploy:
export MOON_CFG="master:flag=Enable,slave:flag=Disable,slave:name=slave1,slave:master_ip=master_ip"
Change-Id: I21c0186a029c508d390c69112b43510d9c5b9c5c
Signed-off-by: chigang <chigang@huawei.com>
|
|
|
|
Add plugin switch in group_vars/all which missed when
upgrade to dynamic inventory
Change-Id: I71927bce552dca89bc6182abc9f721d5222f2be6
Signed-off-by: Harry Huang <huangxiangyu5@huawei.com>
|