Blogger traces router outage to IPv6 prefix delegation bug after ISP reconnect
A developer running a custom Linux-based router diagnosed an issue where restarting pppd (the ISP dial-up daemon) caused Google and other IPv6 sites to become unreachable, even though DNS resolution still worked. Investigation showed the router's stack of networkd, kea-dhcp4, blocky and pppd was individually functioning, but a stale or duplicated IPv6 prefix delegation left client devices holding addresses from an old, now-invalid prefix.
GoKawiil's interpretation of the reporting above, not reported fact.
The case illustrates how IPv6 delegated prefixes can silently go stale across reconnects in ways that are hard to detect, since DNS continues to work while routing quietly fails. This suggests that home-router software built from independent, well-tested components can still produce subtle interoperability bugs at their seams, a risk that may affect other DIY router setups using similar daemon combinations.
- Restarting the PPP daemon on a custom router caused IPv6 connectivity to break while DNS kept working.
- The failure stemmed from stale IPv6 prefix delegation rather than a bug in any single daemon.
- The case highlights a subtle class of networking failures that can arise when combining independently reliable Linux networking tools.
Source: savyasacheejha.com — Savyasachee Jha, 2026-11-15
Published there as: “IPv6 delegation, an unusually subtle failure mode”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.