How many instances of dhcp6c are expected?

  • I have a single WAN, single LAN setup. My ISP (comcast) utilizes DHCP6-PD for IPv6.

    I notice that there are two instances of dhcp6c. I was poking around the system and wanted to see the .conf files and noticed that there were two instances. Is this expected?

    The system otherwise works fine but there was a reboot of the cable modem on Oct 11. This used to trigger the complete loss of IPv6 connectivity for me until I used the "Use IPv4 connectivity as parent interface" setting. Perhaps this setting triggers a new instance of dhcp6c without cleaning up the old one first?

    Or is this much ado about nothing?  ;)

    $ ps auxwww|grep -i dhcp6c
    root   25926  0.0  0.0  8296  1744  ??  S     4:45PM   0:00.00 sh -c ps auxwww|grep -i dhcp6c 2>&1
    root   26245  0.0  0.0  9068  1504  ??  S     4:45PM   0:00.00 grep -i dhcp6c
    root   35420  0.0  0.0  5780  1496  ??  Is    5Oct13   0:02.90 /usr/local/sbin/dhcp6c -d -c /var/etc/dhcp6c_wan.conf -p /var/run/ re0
    root   84377  0.0  0.0  5780  1524  ??  INs  11Oct13   0:01.25 /usr/local/sbin/dhcp6c -d -c /var/etc/dhcp6c_wan.conf -p /var/run/ re0

    /var/run/ has contents 84377.

  • I detected the same issue this morning! My PPPoE session to my provider (Belgacom, Belgium) reached it maximum time and restarted. IPv6 is provided vi DHCPv6 and the old /64 was still configured and distributed across the LAN (yes, Belgacom provides dynamic IPv6!).

    I noticed that I had 4(!) dhcp6c processes. I killed them, restarted my PPPoE session and IPv6 was back in business with a new range.
    Any idea how to fix this? I'm now waiting the next pppoe restart to see if the problem will occur again.


  • My PPPoE0 session was renewed today and guess what? I lost again my IPv6 connection.
    There was 2 "dhcp6c" processes and the /var/run/ contained the PID of one of them.
    I checked the /etc/inc/*.inc files for some information and saw that the dhcp6c client is killed by looking its PID in the file above
    => Could it be the issue?

    Anybody had seen the same behavior?


  • It appears this may still be a bug in 2.2.2  My PPPoE sessions being renewed seem to completely break DHCP6 prefix delegation with my ISP.  The WAN link comes up and can be used to communicate but routing for the delegated subnet is completely broken.

    I was able to confirm that killing the processes and re-saving the wan config as suggested on the documentation page here resolved the issue.

Log in to reply