Cyber Services / All lessons / CentOS / Linux Administration / Managing Networking

Managing Networking

Module 21 of 33 · 9 min read · Free to read

MODULES ↓ Command blocks have a COPY button

Mastering `nmcli` and `ip` for Robust Server Connectivity

A CentOS 8.5 server's utility is intrinsically tied to its network connectivity. Beyond basic reachability, a robust, securely configured network interface is paramount for delivering services, ensuring remote manageability, and integrating into complex infrastructures. This module delves into the sophisticated networking stack of CentOS 8.5, focusing on the `NetworkManager` service and its indispensable command-line interface, `nmcli`, alongside the powerful `ip` utility. We'll explore core concepts, advanced configuration, and expert troubleshooting techniques.

CentOS 8.5 and NetworkManager: A Synergistic Relationship

On modern RHEL/CentOS systems, NetworkManager is the authoritative service for managing network connections. It provides a dynamic, event-driven approach to networking, handling everything from wired Ethernet to Wi-Fi and VPNs. While `ifcfg` files in `/etc/sysconfig/network-scripts/` still exist and are read by NetworkManager, direct manipulation with `nmcli` is the recommended, most powerful, and future-proof method for configuring network interfaces persistently.

Foundational Network Concepts: The TCP/IP Model Perspective

Understanding the underlying principles of networking, particularly the TCP/IP model, provides crucial context for effective configuration and troubleshooting. While not a direct implementation of the OSI model, the TCP/IP model simplifies network communication into four logical layers:

Application Layer

(HTTP, FTP, DNS) - Where user applications and services interact with the network. Examples: Web browsers requesting pages, SSH clients connecting.

Transport Layer

(TCP, UDP) - Manages end-to-end communication, data segmentation, flow control, and reliability (TCP) or speed (UDP). Establishes connections for applications.

Internet Layer

(IP, ICMP) - Handles logical addressing (IP addresses) and routing of packets across different networks. This is where your server finds its way across the internet.

Network Access Layer

(Ethernet, Wi-Fi, MAC addresses) - Deals with physical transmission of data frames on the local network segment. Includes physical cabling/wireless and MAC addresses.

Essential Network Addressing Parameters (Internet Layer)

At the Internet Layer, a server requires specific parameters to participate in network communication:

IP Address (IPv4/IPv6)

The unique numerical label assigned to each device. E.g., `192.168.1.100` (IPv4) or `2001:0db8::1` (IPv6). It's the server's identity on the logical network.

Subnet Mask (or Prefix Length)

Defines which part of an IP address refers to the network and which part refers to the host. In CIDR notation (CentOS standard), this is `/24` (equivalent to `255.255.255.0`), or `/64` for IPv6. Crucial for determining local versus remote communication.

Default Gateway

The IP address of the router on the local network segment. All traffic destined for external networks (e.g., the internet) is forwarded to this gateway. Without it, your server is isolated from the world.

DNS Server(s)

Resolves human-readable hostnames (e.g., `www.redhat.com`) into numerical IP addresses. Without correctly configured DNS, your server cannot access resources by name, severely limiting its functionality.

Inspecting the Network State with `ip` and `nmcli`

Before making any changes, it's vital to understand the current network configuration. The `ip` utility provides raw kernel-level network information, while `nmcli` gives you a NetworkManager-centric view.

`ip` Utility: The Low-Level View

The `ip` command is the modern replacement for `ifconfig` (which is deprecated). It's granular and powerful.

# Show all network interfaces and their IP addresses
$ ip address show

# Show only UP interfaces
$ ip address show up

# Show routing table
$ ip route show

# Show link layer information (MAC addresses, link state)
$ ip link show

Example `ip address show` Output Deconstruction:

1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: enp0s3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000
    link/ether 08:00:27:a1:b2:c3 brd ff:ff:ff:ff:ff:ff
    inet 192.168.1.105/24 brd 192.168.1.255 scope global dynamic noprefixroute enp0s3
       valid_lft 85651sec preferred_lft 85651sec
    inet6 fe80::a00:27ff:fea1:b2c3/64 scope link 
       valid_lft forever preferred_lft forever

Key Interpretations

  • lo: The loopback interface. `127.0.0.1` is essential for local service communication.
  • enp0s3: This is the hardware-assigned name for the physical Ethernet interface. The `UP` and `LOWER_UP` flags indicate the interface is active and connected.
  • link/ether 08:00:27:a1:b2:c3: The unique MAC address of the network card.
  • inet 192.168.1.105/24: The server's IPv4 address and its network prefix. `dynamic` indicates DHCP assignment.
  • inet6 fe80::a00:27ff:fea1:b2c3/64: The link-local IPv6 address, automatically assigned.

`nmcli` Utility: The NetworkManager Interface

`nmcli` provides a consistent way to interact with NetworkManager. It manages "connections," which are profiles containing configuration details for a network interface.

# List all NetworkManager connections (profiles)
$ nmcli connection show

# Show active connections only
$ nmcli connection show --active

# Show device status (physical NICs)
$ nmcli device status

# Show detailed information for a specific connection
$ nmcli connection show enp0s3

Configuring Network Interfaces with `nmcli` (Static IP)

For CentOS servers, a static IP address is almost always preferred for predictability, DNS records, and firewall rules. The following advanced workflow demonstrates configuring a static IPv4 address.

Workflow: Transitioning from DHCP to Static IP

This process ensures your changes are persistent across reboots and are managed correctly by NetworkManager.

  1. Identify the Target Connection:
    $ nmcli connection show
    NAME    UUID                                  TYPE      DEVICE
    enp0s3  1234abcd-5678-9ef0-1234-56789abcdef0  ethernet  enp0s3
    
    We'll use `enp0s3` as our example connection name.
  2. Modify IPv4 Configuration to Static:

    Set the IP address, gateway, and DNS servers. Crucially, set the `ipv4.method` to `manual`.

    # Set a static IPv4 address with CIDR prefix
    $ sudo nmcli connection modify enp0s3 ipv4.addresses 192.168.1.200/24
    
    # Set the default gateway for IPv4
    $ sudo nmcli connection modify enp0s3 ipv4.gateway 192.168.1.1
    
    # Set the primary and secondary DNS servers
    $ sudo nmcli connection modify enp0s3 ipv4.dns "8.8.8.8,1.1.1.1"
    
    # Crucially, change the method from 'auto' (DHCP) to 'manual' (static)
    $ sudo nmcli connection modify enp0s3 ipv4.method manual
    
    # Disable IPv6 if not needed (optional, but common on dedicated IPv4 servers)
    $ sudo nmcli connection modify enp0s3 ipv6.method ignore
    
  3. Apply and Activate Changes:

    The changes are saved to the connection profile, but not yet active. You must restart the NetworkManager connection.

    # Bring the connection down, then back up to apply changes
    $ sudo nmcli connection down enp0s3 && sudo nmcli connection up enp0s3
    

    Caution: Remote Access!

    When running `nmcli connection down` remotely via SSH, you will immediately lose connectivity. Always execute `nmcli connection up` as a chained command (using `&&`) or ensure you have console access. A safer alternative for remote reconfigurations is `nmcli connection reload enp0s3` if the changes are minor and don't require a full interface restart, or `nmcli device reapply enp0s3`.

  4. Verify the New Configuration:
    $ ip address show enp0s3
    $ ip route show
    $ cat /etc/resolv.conf # Verify DNS servers
    

Lab: Adding a New Network Interface and Configuring it Statically

Scenario: Your CentOS server has a second network card (e.g., `enp0s8`) that needs to be configured for a specific internal network segment. We'll add a new NetworkManager connection for it.

  1. Identify the New Device:
    $ ip link show # Look for a device that is UP but has no IP, e.g., enp0s8
    $ nmcli device status # Verify no connection is associated with enp0s8
    
    If `enp0s8` isn't listed, ensure it's enabled in your VM settings or physical server hardware.
  2. Create a New NetworkManager Connection:

    Create a new connection profile explicitly for `enp0s8`.

    $ sudo nmcli connection add type ethernet con-name "Internal Network" ifname enp0s8 \
        ipv4.addresses 172.16.10.10/24 \
        ipv4.method manual \
        autoconnect yes
    
    • `con-name "Internal Network"`: A human-readable name for the connection.
    • `ifname enp0s8`: Binds this connection to the physical `enp0s8` device.
    • `ipv4.addresses 172.16.10.10/24`: The static IP and prefix.
    • `ipv4.method manual`: Ensures it's static.
    • `autoconnect yes`: Tells NetworkManager to bring this connection up automatically on boot.
  3. Configure DNS for the New Connection (Optional, but good practice):

    If this internal network requires its own DNS, or if you want to override system-wide DNS for this interface:

    $ sudo nmcli connection modify "Internal Network" ipv4.dns "172.16.10.1,172.16.10.2"
    
  4. Bring the New Connection Up:
    $ sudo nmcli connection up "Internal Network"
    
  5. Verify:
    $ ip address show enp0s8
    $ nmcli connection show --active
    

Lab: Modifying an Existing Network Interface

Scenario: You need to change the gateway of your primary interface (`enp0s3`) and add an additional IP alias.

  1. Modify the Gateway:
    $ sudo nmcli connection modify enp0s3 ipv4.gateway 192.168.1.254
    
  2. Add an IP Alias (Secondary IP Address):

    NetworkManager allows multiple IP addresses on a single interface. Append the new address to the existing ones.

    $ sudo nmcli connection modify enp0s3 +ipv4.addresses 192.168.1.201/24
    
    The `+` before `ipv4.addresses` appends the new address without overwriting existing ones.
  3. Apply and Verify:
    $ sudo nmcli connection up enp0s3 # This will re-apply the modified connection
    $ ip address show enp0s3 # Verify both 192.168.1.200 and 192.168.1.201 are present
    $ ip route show # Verify the new gateway
    

Lab: Deleting a Network Interface Connection

Scenario: The internal network connection (`Internal Network`) you created earlier is no longer needed.

  1. Deactivate the Connection (Optional, but good practice):
    $ sudo nmcli connection down "Internal Network"
    
  2. Delete the Connection Profile:
    $ sudo nmcli connection delete "Internal Network"
    
    This removes the configuration profile. The physical device `enp0s8` will remain, but it will no longer have an associated NetworkManager connection to bring it up automatically.
  3. Verify:
    $ nmcli connection show
    $ ip address show enp0s8 # Should no longer have an IP from the deleted connection
    

Advanced Network Troubleshooting with `ping`, `traceroute`, `dig`, and `nmtui`

Effective troubleshooting requires a systematic approach, moving from basic connectivity to higher-level services.

`ping` (ICMP Echo)

Purpose: Tests basic IP-level reachability and latency. Operates at the Internet Layer.

Usage: `ping 8.8.8.8` (test internet gateway/public IP), `ping ` (test local router).

Interpretation: `Destination Host Unreachable` suggests incorrect IP/subnet/gateway. `Request timeout` indicates a route problem or firewall block.

`traceroute` (or `tracepath`)

Purpose: Maps the path (hops) a packet takes to reach a destination. Helps identify where connectivity breaks down.

Usage: `traceroute google.com` or `traceroute 8.8.8.8`.

Interpretation: Asterisks (`*`) indicate a router that isn't responding to ICMP (often by design) or a point of failure.

`dig` (Domain Information Groper)

Purpose: Queries DNS servers for domain name resolution. Operates at the Application Layer (DNS).

Usage: `dig google.com`, `dig @8.8.8.8 example.com` (specify DNS server).

Interpretation: No `ANSWER SECTION` or `SERVFAIL` usually means DNS configuration is wrong or DNS servers are unreachable.

`nmtui` (NetworkManager Text User Interface)

Purpose: A graphical (curses-based) interface for NetworkManager. Excellent for quick visual checks or when `nmcli` syntax is tricky.

Usage: Just type `sudo nmtui`.

Interpretation: Provides a user-friendly way to see and modify connections, activate/deactivate interfaces, and troubleshoot visually.

Troubleshooting Flow:
1. `ip address show` and `ip route show`: Verify IP, subnet, gateway are correct. 2. `ping `: Can you reach your local router? 3. `ping 8.8.8.8`: Can you reach the internet (via IP)? 4. `dig google.com`: Can you resolve domain names? 5. `traceroute google.com`: If ping fails to internet, where is the path breaking? 6. Check `firewalld` with `sudo firewall-cmd --list-all`. 7. Examine `/var/log/messages` or `journalctl -u NetworkManager.service` for clues.
Modules (33) ← CentOS / Linux Administration