Please follow the upstream github issue. I am also affected by this bug, which temporarily disrupts ipv6 connectivity when a RA is received by the monitored interface. Some (streaming) clients handle those recurring outtakes of ipv6 connectivity not very well, longtime connections staying longer than RA intervals may also get closed. I could solve my situation by downgrading from 10.2.2 to also stable 10.1.0-r1 and add a in older releases not necessary "nogateway" option, since default route is handled by pppd. I added this here if someone is in search for a quick fix. Reproducible: Always
If you can give a list of commits which fix the issue when applied to 10.2.2, we can look at backporting them.
(In reply to Sam James from comment #1) > If you can give a list of commits which fix the issue when applied to > 10.2.2, we can look at backporting them. https://github.com/NetworkConfiguration/dhcpcd/commit/664952372f0965da16ef24abe079bea1ad87a166 Already testing a local build with this patch for regressions, like to give it a day and will report.
Revised and simplified my setup a little and give dhcpcd control over default route as well and everything is running smoothly. In case a (kernel based) default route already exists, dhcpcd will setup a redundant, but more specific (+ mtu) default-route with a higher metric -> useful behavior for HA setups. So far, everything seems ok.
The bug has been closed via the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=6945571fce4bf7557248af07f6d2e22a7ea56efb commit 6945571fce4bf7557248af07f6d2e22a7ea56efb Author: Sam James <sam@gentoo.org> AuthorDate: 2025-04-19 00:10:50 +0000 Commit: Sam James <sam@gentoo.org> CommitDate: 2025-04-19 00:10:50 +0000 net-misc/dhcpcd: backport IPv6 fix Closes: https://bugs.gentoo.org/953895 Signed-off-by: Sam James <sam@gentoo.org> net-misc/dhcpcd/dhcpcd-10.2.2-r1.ebuild | 173 ++++++++++++++ net-misc/dhcpcd/files/dhcpcd-10.2.2-ipv6.patch | 312 +++++++++++++++++++++++++ 2 files changed, 485 insertions(+)
