Contrail Command - kratna888/Contrail_Command GitHub Wiki

  1. Welcome to the Contrail Command install wiki!
Resource : A VM or Physical Server with 8 vCPU, 64G RAM, 300G disk. Out of this 300G, 256G should be allocated to /root directory. Internet access is needed from this server ( Let's call this server Contrail-Command)

Pre-Requisites

We recommend installing Contrail Command in an x86 server with Centos 7.5 as the baseOS.

Deployment Steps :

Step 1: On your Contrail-Command Server install the below mentioned packages - epel, python-pip, git, ansible. python-pip MUST be installed after epel-release is installed when using CentOS7. These packages are needed to automate the deployment of Contrail-Command software. The 2.4.2.0 Ansible version is mandatory for this installation. 

yum -y install epel-release 
yum -y install python-pip
yum -y install git 
yum -y install ansible-2.4.2.0

Step 2: The Contrail-Command server comprised two docker containers : Contrail-Command and sql containers. These containers images are hosted on hub.juniper.net. To download these containers , customer must request for username/password from appropriate Juniper channel. ( AI - Vineet Gupta /Richard Hendricks )

The containers will be posted on hub.juniper.net. External customers need to get the username /password from appropriate channels

If it's internal - then do a git clone of the below mentioned link 

git clone https://github.com/Juniper/contrail-command-deployer.git

Please Note: The container images needed for contrail-command-deployer will be published in hub.juniper.net with the General Availability version. The path to this image needs to be provided in file config/command_server.yml

Step 3: Then Goto Contrail-Command-Deployer directory ; Update the username /password in file contrail-command-deployer/config/command_server.yml;

default login: admin password: contrail123 

cd Contrail-Command-Deployer

Step 4: Run the Playbook deploy.yml to deploy the Contrail-Command server 

ansible-playbook playbooks/deploy.yml

step 5: Once the playbook completes, log into Contrail-Command using 

https://< Contrail-Command-Server-IP-Address > : 9091

Use the same username password that was provided in Step3.

default login: admin password: contrail123

The UI for Contrail-Command will be shows as below:

![](https://github.com/kratna888/OVS_Contrail/blob/master/Contrail-Command.png)

Troubleshooting Notes

Issue 0: If you cannot log into the new UI - Contrail Command - https://:9091 then,

iptables -F

Compute Nodes

The compute nodes needs to have kernel version of = kernel-3.10.0-862.3.2.el7.x86_64

wget http://mirror.centos.org/centos/7/updates/x86_64/Packages/kernel-3.10.0-862.3.2.el7.x86_64.rpm

yum install kernel-3.10.0-862.3.2.el7.x86_64.rpm

Step b: Ping Test. From each nodes, ping every remote nodes via mgmt interface as well as data interface.

Issue 0+ : If you cannot launch the new UI after the cluster deployment is complete,

Check the NTP sync.

Issue 1: If you encounter - UNREACHABLE issues when you are provisioning the cluster

Step 1: Log into the Contrail Command Server

ssh  [email protected]
[email protected]'s password:

Step 2: List the dockers

docker ps

Step 3: Log into the docker - contrail-command

docker exec -it contrail-command bash

Step 4: goto directory /var/tmp/contrail_cluster

[root@localhost contrail_cluster]# ls -lrt
total 0
drwxr-xr-x 3 root root 50 Jun  7 05:41 config
drwxr-xr-x 4 root root 90 Jun  7 05:41 edadd8ac-c7d9-4f6c-9a04-564477a97227
[root@localhost contrail_cluster]#

Step 5: you will find a cluster directory - edadd8ac-c7d9-4f6c-9a04-564477a97227

cd edadd8ac-c7d9-4f6c-9a04-564477a97227

Step 6: find the instances.yml file

[root@localhost edadd8ac-c7d9-4f6c-9a04-564477a97227]# ls
contrail-ansible-deployer  contrail-kolla-ansible  instances.yml
[root@localhost edadd8ac-c7d9-4f6c-9a04-564477a97227]#

Step 7: vi instances.yml file

provider_config:
  bms:
    ssh_user: root
    ssh_pwd: c0ntrail123

Step 8 : Verify if the password you had provided here in this section - provider_config is correct for your target nodes.

Issue 3: If you loose connectivity to compute node or If the Ansible playbook fails due to network restart is failing ( service network restart )

Issue 4: if the cluster deployment completes however endpoints are not created , please follow the workaround.

It failed since the keystone token had expired since the cluster deployment took longer than expected.

Step 1: Set the provisioning_state = CREATED - send a curl request to set the provisioning_state to CREATED.

Sample curl request below. Please update cluster resource UUID and x-auth-token to yours

curl --insecure -X PUT
https://:9091/contrail-cluster/43f4388a-4d86-49fe-af0c-59e3c1bca451
-H 'Content-Type: application/json'
-H 'x-auth-token: 3191f984-9738-49d8-ab78-3dc89c4b8209'
-d '{ "contrail-cluster": { "provisioning_state": "CREATED", "uuid": "43f4388a-4d86-49fe-af0c-59e3c1bca451" } }' To get the auth-token from the local keystone, use the following sample curl request.

curl -i -k
-H "Content-Type: application/json"
-d ' { "auth": { "identity": { "methods": ["password"], "password": { "user": { "name": "admin", "domain": { "id": "default" }, "password": "contrail123" } } }, "scope": { "project": { "name": "admin", "domain": { "id": "default" } } } } }'
"https://localhost:9091/keystone/v3/auth/tokens" ; echo

Create the following endpoints:

add following endpoint from infrastructure/cluster Advanced Options -> Endpoints

name: keystone Public URL: http://:5000

name: compute Public URL: http://:8774

name: baremetal Public URL: http://:6385

name: glance Public URL: http://:9292

name: swift Public URL: http://:8080

name: config Public URL: http://:8082

name: telemetry Public URL: http://:8081

name: nodejs Public URL: https://:8143

Logging into sql

Step 1 : log into the contrail_mysql

Step 2:

root@demo-server-11:/# mysql -uroot -pcontrail123

delete from endpoint where name='keystone';

==============================================

BMS usecase

when the ens1f0 gets an IP, the default route is changes to default via 192.168.10.1 dev ens1f0

hence lost external connectivity -

====================================

⚠️ **GitHub.com Fallback** ⚠️