Deploy PortSIP PBX HA for Ubuntu
This manual is for guide deploy the PortSIP HA on Ubuntu 20.04.
In the examples used in this manual, a total of three servers are used (PBX nodes are used to refer to servers in the following; the three PBX nodes or all PBX nodes refer to three servers, master node and node 1 refer to pbx01, node 2 refer to pbx02, and node 3 refer to pbx03).
To deploy the HA, requires to have two disk volumes. In this example the hard disk of each server is 98GB and created two volumes, one is 48G(/dev/sda) for Linux OS we called it as "system volume", another one 50G(/dev/sdb) for store the PBX data, we called it as "data volume".
You are free to decide the disk size, but the size must be the same with all servers.
In this example we deploy the HA on three servers which have the IP: 192.168.78.101, 192.168.78.102, 192.168.78.103, the VIP (virtual IP) used in this example is 192.168.78.90.
Please change the hostnames and IP addresses of the three node machines as follows:
The OS version is required Ubuntu 20.04.
Principle
PortSIP PBX HA The principle of the scheme is as follows:
Three PBX nodes are running at the same time to prevent split-brain during HA switching.
Use DRBD to synchronize data between nodes, including current active call information, recording files, logs, call records, and prompt voice files uploaded by users.
Use virtual IP (VIP) as the access entry point of the system.
Use Pacemaker to monitor the status between servers.
The primary node provides services. Once the primary node has a DOWN machine, Pacemaker will immediately drift the VIP to the standby node that takes over after detecting it. User service requests will be automatically routed to the standby node, and the standby node will continue to provide services.
The established call on the primary node will be automatically restored by the standby node.
Prerequisites
Unless otherwise specified, the operations mentioned below are only performed on the master node pbx01
Please change the host names of the three nodes to pbx01, pbx02 and pbx03
It is recommended to configure 4 cores CPU and 4G memory hardware. If it is lower than 2-cores CPU or 2G memory, ha handoff will be affected by hardware, and call recovery time will increase
Installation system
That you need to follow the same steps described in this section to install and configure the CentOS operating system on all three node servers.
Select language and keyboard type
Choose minimize installation
Network connections [default]
Configure Proxy [default]
Configure Ubuntu archive mirror [default]
Guided storage configuration [Default]
Storage configuration [default]
Note: /dev/sdb state is unused
Profile setup
Enter the information in the fields.
SSH Setup
Enable OpenSSH server
Featured Server Snaps [default]
Installing system
Start to install the OS and waiting...
Install complete
After installation is completed, reboot it.
Download installation resource package
Perform the following command only at the master node, that is, node pbx01.
Install ansible
Perform the following command only at the master node, that is, node pbx01.
Set ansible resource group
Perform the following command only at the master node, that is, node pbx01.
Resolve the hostname
Perform the following commands at all nodes: pbx01, pbx02, pbx03. Note: you need to replace the IP and hostname in the following command with your actual IP and hostname.
Set password-free login
In this example, pbx01, pbx02, and pbx03 refer to node 1, node 2, and node 3 respectively. The following command only needs to be executed on the primary node, that is, pbx01 and the certificate is generated according to the prompt:
Set up pbx01 password-free login (executed on pbx01 node):
Set up pbx02 password-free login (executed on pbx01 node):
Set up pbx03 password-free login (executed on pbx01 node):
Set variables
Please prepare the value for the below Variables.
Name | Type | Description |
pbx01_hostname | string | The hostname of node 1, in this case is pbx01 |
pbx02_hostname | string | The hostname of node 2, in this case is pbx02 |
pbx03_hostname | string | The hostname of node 2, in this case is pbx02 |
pbx01_private_ip | string | The private static IP of node 1, in this case is 192.168.78.101 |
pbx02_private_ip | string | The private static IP of node 2, in this case is 192.168.78.102 |
pbx03_private_ip | string | The private static IP of node 3, in this case is 192.168.78.103 |
vip | string | The virtual IP in this case is 192.168.78.90 |
pcs_hacluster_pwd | string | The HA cluster user password, in this case is 123456 |
pbx_image | string | PortSIP PBX docker image in this case is portsip/pbx:12.6 |
pbx_datapath_disk | string | The disk volume for PBX data store in this case is /dev/sdb |
pbx_datapath_size | string | The disk volume size for PBX data store,in this case is 49G(If use 50G maybe report error) For example if the disk volume size is 500G , suggest set as 490G. |
The following command is only executed on the master node, that is, node pbx01.
Pay attention to changing the information about pbx01, pbx02, pbx03, VIP, and pbx_datapath_size to your own real information.
Install resources
Execute the following command only on the master node, that is, node pbx01.
Restart
Execute the following command only on the master node, that is, node pbx01.
Start PCs
After restart, execute the following command only at the primary node, that is, node pbx01:
Initialize configuration resources
Only at the master node, that is, pbx01, execute the following commands (the execution process may be long, just wait patiently, and don't interrupt, restart or shut down in the middle of the process)
After the resource configuration is completed, you can use the browser to open it http://192.168.78.90 : 8888 or https://192.168.78.90 : 8887 to configure your PBX. Subsequent PBX management, configuration and access are carried out through virtual IP 192.168.78.90. The first step of the setup wizard is to fill in the virtual IP. You need to replace it with your actual virtual IP.
After entering the web management interface of PBX, under the menu, advanced > Settings > Advanced page, select enable call recovery and click the "apply" button.
Several common commands
Perform the commands from the /root/portsip-pbx-ha-guide
directory
View PBX status
Restart PBX resource
After the following command is executed, the PBX-related resources of the whole cluster will be restarted.
Upgrade PBX
Only at the master node, that is, pbx01, perform the following commands.
Download and update the resource package
Modify the config file
Edit the /root/portsip-pbx-ha-guide/vars.yml
to change the PortSIP PBX
docker image version.
Upgrade
Last updated