Using Deep Security Agent with iptables
When Deep Security Agent 10.1 or earlier was installed on Linux, it disabled the iptables service to avoid firewall conflicts unless you added a configuration file that prevented that change. However, the iptables service is used for more than just firewall (for example, Docker manages iptables rules as part of its normal operation), so disabling it sometimes had negative consequences.
With Deep Security Agent 10.2 and higher, the functionality around iptables has changed. Deep Security Agent no longer disables iptables. (If iptables is enabled, it stays enabled after the agent installation. If iptables is disabled, it stays disabled.) However, if the iptables service is running, Deep Security Agent requires certain iptables rules, as described below.
Rules required by Deep Security Agent
If iptables is enabled on the computer where Deep Security Agent is being installed, iptables may require additional rules. By default, these rules are added when Deep Security Agent starts up and removed when the agent is stopped or uninstalled. Alternatively, you can Prevent Deep Security Agent from automatically adding iptables rules and add them manually instead:
- Allow incoming traffic on port 4118. This is required when the agent uses manager-initiated or bidirectional communication. (For more information, see Agent-manager communication.)
- Allow incoming traffic on port 4122. This is required when the agent is acting as a relay, so that the relay can distribute software updates. (For more information, see Distribute security and software updates with relays.)
These are the default port numbers - yours may be different. For a complete list of ports used in Workload Security, see Port numbers.
You can prevent Deep Security Agent from modifying iptables if you would rather add the required rules manually. To prevent the automatic modification of iptables, create the following file on the computers where you plan to install Deep Security Agent: