CentOS 7设备ARP回复错误MAC地址问题求助

Hey there, let’s tackle this frustrating ARP conflict issue you’re hitting with your CentOS server (10.150.39.5). When a single IP responds to ARP requests with two different MAC addresses, network devices get confused—they’ll flip between the two addresses, which is exactly why your communication keeps getting interrupted. Let’s walk through the most likely causes and fixes step by step.

First: Verify Your Server’s Network Interfaces #

The most common culprit here is accidentally assigning the same IP to two different network interfaces on your server. Let’s check:

• Run this command to list all active interfaces and their IPs:

  ip addr show
  

  Look for any duplicate entries of 10.150.39.5 across physical NICs, virtual NICs (like those from Docker/KVM), or bonded interfaces. If you find two interfaces with the same IP, that’s the problem.
• If you’re using NIC bonding, double-check the configuration to make sure the member interfaces aren’t assigned their own IPs (only the bond device should have the IP). You can inspect bond settings with:

  cat /proc/net/bonding/bond0
  

  (Replace bond0 with your actual bond interface name if different.)

Check for ARP Configuration Anomalies #

Sometimes misconfigured ARP settings or stale cache can cause unexpected responses:

• View your server’s local ARP cache to spot odd entries:

  arp -n
  

• Flush the ARP cache to clear any stale entries:

  ip neigh flush all
  

• Check your network interface configuration files (in /etc/sysconfig/network-scripts/) for unusual ARP-related parameters like ARPCHECK or ARP_ANNOUNCE that might be forcing multiple responses.

Rule Out Virtualization/Container Conflicts #

If your server is running Docker, KVM, or other virtualization tools, a virtual network might be accidentally using the same IP as your physical NIC:

• For Docker, list your networks and inspect their IP ranges to check for conflicts:

  docker network ls
  docker inspect <network-name>
  

  Look for any networks that include 10.150.39.5 in their subnet or have assigned that IP to a container.
• For KVM or other hypervisors, check the network settings of your virtual machines—ensure none are bridged to the physical network and using the same server IP.

Validate Network-Level Issues #

While the problem seems to originate from your server, it’s worth ruling out external factors:

• From another machine on the same subnet, run an arping to confirm you’re getting two MAC responses:

  arping -I <your-local-interface> 10.150.39.5
  

  This will show you all MACs responding to the ARP request for that IP.
• Check if any firewall or security tools (like IDS/IPS) on the server are modifying ARP packets. Inspect your iptables rules for unusual entries in the mangle or nat tables:

  iptables -t nat -L
  iptables -t mangle -L
  

Fix Steps to Resolve the Conflict #

Once you’ve identified the root cause:

• If duplicate IPs on interfaces: Remove the extra IP from one interface, then restart your network service. For CentOS 7 and earlier:

  systemctl restart network
  

  For CentOS 8+:

  nmcli networking restart
  

• If virtualization/container conflict: Adjust the virtual network’s subnet to avoid overlapping with your physical network, or reassign the conflicting IP to a different address.
• After making changes, re-run the arping test to confirm only one MAC responds to ARP requests for 10.150.39.5.

内容的提问来源于stack exchange，提问作者JohnDoe