This topic describes how to install Customizer Lite. There are two deployment options,
hereby referred to as Customizer Lite and Customizer Lite HA. Customizer Lite – the most basic
deployment using a single Docker node running a single instance of each required service and
Customizer Lite HA – a more resilient deployment using a pair (or more) of Docker nodes all running
instances of each required service with requests load balanced across available nodes.
Before you begin
System Requirements to verify and prerequisites to follow before installing Customizer Lite:
- System Requirements
- Docker
- Docker Compose
- Setting up a reverse proxy
System Requirements
For Customizer Lite (with a single Docker instance), a single virtual machine (4 CPU, 2.x GHZ,
8GB memory, and at least 100GB disk) with a x86_64 Linux OS. Use two (or more) of these virtual
machines to configure for Customizer Lite HA deployment.
Customizer Lite has been validated to run on RHEL 7.6/CentOS 7.6 using: Docker 17.03.x configured
with devicemapper storage Docker-Compose version 1.22 .
Docker
Follow steps 1 through 3 in the following topic: Deploying an HA Kubernetes platform to install the same Docker configuration on each of the virtual machines.
Docker Compose
The following will install the latest stable version on RHEL 7.6/CentOS 7.6. (Customizer
docker-compose script requires a minimum version 1.22). Repeat on each of the virtual machines so
that they all have the same docker-compose configuration.
- Additional packages
sudo yum install epel-release
- Install python pip:
sudo yum install python-pip
- Install docker-compose
sudo pip install docker-compose
- Upgrade python packages on Cent0s for docker-compose
- Test docker-compose installation on each of the virtual machines
Setting up a reverse proxy
Customizer Lite requires the use of a reverse proxy. The following is an example using NGINX
.
Refer to steps 4 through 10 in the following topic: Configuring the NGINX proxy server for Customizer.
About this task
The following steps provide the instructions for installing Customizer Lite and the configuration
changes required for Connections to enable Customizer.
If deploying Customizer Lite, the files can be installed on the local storage of the single
virtual machine.
If deploying Customizer Lite HA with multiple Docker nodes, place the files on a shared volume
(such as NFS for Linux or SMB for Windows) so that all Docker nodes can mount and share the same
Docker images and service configuration files. Mount the shared volume into the file system of each
virtual machine before making the changes outlined below. Steps 1-7 below should only need to be
performed once if directories are shared and mounted properly on each Docker node; step 8 is
repeated on all nodes to start the services.
Procedure
-
Copy the Customizer Lite archive onto the required storage volume (either local or shared).
-
Extract the archive into the storage location, for example into a directory named
customizerLite. There should be four subdirectories as shown
below:
- customizations: this is where the customization files are placed. It will contain three files
initially. These are required by Customizer and should not be deleted or moved.
- data: this is where applications settings get saved
- images: directory for the three Customizer images
- scripts: scripts for deploying, starting and updating Customizer
-
Set the read permissions on the customizerLite
directory chmod -R 700 Compose.
-
In the customizerLite directory, change the permissions
of the following directories.
- chmod -R 005 customizations
- chmod -R 005 data
- chmod -R 007 data/settings
-
From the scripts directory run the "setupImages.sh" script. This will load all three of the
Customizer images. To run the script:
./setupImages.sh -dr Docker_registry -u Your_user_name -p Your_password -st cs_lite
Where:
- Docker_registry is your Docker Registry (include port if applicable)
- Your_user_name is your Docker user name (provide dummy value if not required)
- Your_password is your Docker password (provide dummy value if not used)
- cs_lite param.This does not change. It must be cs_lite.
-
Run docker images and make note of the full repository name for each of the
mw-proxy, appregistry-serviceand
appregistry-client. Edit the scripts/ .env file.
Set the three environmental variables as appropriate (do not specify http):
- CONNECTIONS_URL= my_installation_of_connections.com
- NGINX_URL=my_installation_of_nginx.com
For each of the following sections replace the _IMAGE values with the corresponding repository
name, For example:
- MW_PROXY_IMAGE=myDockerRegistry:5000/connections/mw-proxy
- APPREGISTRY_SERVICE_IMAGE=myDockerRegistry:5000/connections/appregistry-service
- APPREGISTRY_CLIENT_IMAGE=myDockerRegistry:5000/connections/appregistry-client
If deploying Customizer Lite HA, the interval at which the
appregistry-service nodes check and sync their customization data cache can be
modified by changing the value of this parameter (specified in milliseconds, default = 30 seconds):
APPREGISTRY_SERVICE_CACHE_CHECK=30000
-
NOTE: (Customizer Lite HA only) Edit the scripts/docker-compose.yml
file and in the mw-proxy environment section add the following line to ensure
that mw-proxy service accesses the appregistry-service via
load balancer using this URL: APPREGISTRY_SERVICE_PROXY: ‘http://${NGINX_URL}’
The environment section should then look similar to
this:
environment:
USE_SSL: 'true'
MW_REVERSE_PROXY: '${NGINX_URL}'
CUSTOMIZER_INTERSERVICE_HOST: '${CONNECTIONS_URL}'
CUSTOMIZER_INTERSERVICE_PORT: '443'
IS_PRIVATE_CLOUD: 'true'
REDIS_INSTALLED: 'false'
LOCAL_DIRECTORY: '/mnt'
ROARR_LOG: '${MW_PROXY_LOGGING_ENABLED}'
APPREGISTRY_SERVICE_PROXY: 'http://${NGINX_URL}'
-
Assuming all the prerequisites are in place and Connections configured, Customizer can be
started with either the command "docker-compose up" or running the script "update-all-services.sh".
The latter will return the command prompt, the former will output logs to the terminal.
"Docker-compose up" must be run from within the scripts directory. This step is repeated on each of
the Docker nodes to start the services on each.