networking
layer2
ai_generated
true
NDP:eth0上fe80::1的重复地址检测失败,地址已被00:1a:2b:3c:4d:5e使用
NDP: Duplicate Address Detection failed for fe80::1 on eth0, address already in use by 00:1a:2b:3c:4d:5e
ID: networking/ipv6-ndp-duplicate-address
79%修复率
86%置信度
1证据数
2024-03-05首次发现
版本兼容性
| 版本 | 状态 | 引入 | 弃用 | 备注 |
|---|---|---|---|---|
| Linux kernel 6.2.0 | active | — | — | — |
| FreeBSD 13.2 | active | — | — | — |
| Windows 11 IPv6 stack | active | — | — | — |
根因分析
IPv6重复地址检测(DAD)检测到链路本地地址fe80::1已被同一链路上的另一主机声明,阻止接口使用该地址。
English
IPv6 Duplicate Address Detection (DAD) detected that the link-local address fe80::1 is already claimed by another host on the same link, preventing the interface from using that address.
官方文档
https://datatracker.ietf.org/doc/html/rfc4862解决方案
-
Change the link-local address on one host by modifying the interface identifier: `ip link set dev eth0 address 02:00:00:00:00:01` (EUI-64 based), then reboot or restart networking.
-
Use a static link-local address with a unique suffix on each host: `ip -6 addr add fe80::1/64 dev eth0` but ensure no conflict by checking with `ndp -a`.
-
Reconfigure the network to use globally routable IPv6 addresses instead of link-local for communication, and disable DAD for link-local only: `sysctl -w net.ipv6.conf.eth0.dad_transmits=0`.
无效尝试
常见但无效的做法:
-
95% 失败
Disabling DAD leads to network instability and packet loss as two hosts respond to the same address.
-
90% 失败
Both hosts will conflict; only one can respond to neighbor solicitations, causing intermittent failures.
-
80% 失败
Flushing caches does not resolve the underlying address duplication; the conflict persists.