Server requirements
We recommend that you install a control panel on a fault-tolerant virtual or dedicated server that supports hypervisor virtualization or OVZ container virtualization.
Hardware requirements
Websites hosted on a server consumes the most part of its resources. Please note that websites with databases and mailboxes consume more server resources.
| Component | Requirements for the control panel | Requirements for the control panel and up to 100 sites | Requirements for the control panel and more than 100 sites |
|---|---|---|---|
| CPU | 1 GHz | 2.4 GHz | from 3 GHz |
| Cores | 1 | 16 | from 24 |
| RAM | 4 Gb | 16 Gb | from 24 Gb |
| Disk space | 60 Gb | 120 Gb | from 200 Gb |
Allocate all available disk space for the root filesystem.
You may change the partition if needed. Please note that ispmanager by default locates user data into /var/www (user disk quotas are set on the partition where user data are located), and the control panel is installed into /usr/local/mgr5/.
Software requirements
Supported operating systems (amd64)
- AlmaLinux 8 (recommended);
- CentOS 7;
- CloudLinux OS SHARED — installation is available only through the integration module.
We strongly recommend a clean install of ISPmanager. You should install a minimal version of the operating system, all the services that ISPmanager requires will be installed automatically during software installation.
Firewall settings
Open the following ports to allow incoming and outgoing connections:
- 20/tcp — to send data through FTP;
- 22/tcp — for remote access to a server through SSH;
- 110, 143, 993, 995, 587, 465, 25/tcp — to send and receive emails;
- 80, 443/tcp — to handle user requests to websites on the server with ISPmanager;
- 1500/tcp —to access the ISPmanager interface;
- 53/tcp, 53/udp — for domain name servers;
- 3306/tcp — for the database server.
Network
When configuring the network between cluster nodes please note that:
- cluster nodes must locate in the same data-center;
- use two physical network interfaces on every cluster node. Connect the first interface in a public network, the second interface should be connected to an internal one with cluster nodes only.
This will enhance cluster reliability and avoid network issues during migration.