Martin-Éric RACINE

I created the following hook: case "$reason" in BOUND6|REBIND6|RENEW6|DELEGATED6) if command -v logger >/dev/null 2>&1; then logger -p daemon.info -t "dhcpcd" "Info: Logged IPv6 event. Reason: $reason $new_delegated_dhcp6_prefix" fi ;;...

Jan 30 12:46:53 p8h61 dhcpcd[595]: enp4s0: Router Advertisement from fe80::bb Jan 30 12:46:53 p8h61 dhcpcd[595]: enp4s0: no global addresses for default route Jan 30 12:46:54 p8h61 dhcpcd[595]: enp4s0: ADV 2001:14ba:a020:aa00::/56...

The above routing table answers that one. There are two routes for the same prefix: one added by dhcp and one by the kernel (presumably in response to router advert).

Has there been any progress on this?

Here, commit 5d372e8ac63a53aa1038655e91970d9d85fed277 seemingly resolves the issue of IPv6 randomly dropping completely whenever valid_lft=preferred_lft and PD expires.

Returning to this bug, Debian ships the following NTP implementations (plus systemd-timesyncd, for which we already have a separate hook), with each configuration found as follows: - **chrony**: `/etc/chrony/chrony.conf` -...

The key problem is this: Feb 02 10:01:31 p8h61 dhcpcd[588]: enp4s0: no global addresses for default route $ ip -6 r ::1 dev lo proto kernel metric 256 pref medium...

Same result before and after: 1: lo: mtu 65536 state UNKNOWN qlen 1000 inet6 ::1/128 scope host valid_lft forever preferred_lft forever 2: enp4s0: mtu 1500 state UP qlen 1000 inet6...

I'm still scratching my head over this one. I'm also baffled by the carrier loss at the beginning. This, too, happens systematically when dhcpcd is first launched upon (re)booting the...

I think that I've figured it out: When PD is used, dhcpcd needs to set 2 flags on the delegating interface: echo 1 > /proc/sys/net/ipv6/conf/enp4s0/forwarding echo 2 > /proc/sys/net/ipv6/conf/enp4s0/accept_ra As...