Starting around 8am this morning when my LG TV was switched on my PiHole logs started filling up with this
message: Oct 23 08:08:36 dnsmasq-dhcp[53204]: DHCPREBIND(eth0) << MAC address of TV >>
I noticed this around 1:45pm this afternoon, and resolved it by restarting DNS via the frontend. Approx 3.7 million DHCPREBIND messages were recorded in the logs in this time:
$ grep REBIND pihole.log | wc -l
3786219
Given that a DNS restart fixed the problem, does that indicate this problem is likely on the PiHole side rather than the TV misbehaving?
I'm trying to upload the debug log via the frontend but pihole.log is 237Mb it'll take quite a while.
If you don't set quiet-dhcp6 in any custom config file (it isn't enabled by default), this message is expected. I'm not sure how to interpret that the restart of the DNS server fixed this. It may either be that the embedded DHCP server got stuck (I just checked the code and don't see how this could happen) or it was the LG TV that was having issues. The short interruption in DHCPv6 offers caused by the restart might have been enough to fix this.
Did this happen for the first time or have you seen this more often? Is there a reliable way for you to reproduce this behavior somehow?
We haven't implemented the DHCP(v6) features ourselves, it comes from the embedded dnsmasq. If you want to get this analysed any further by the original developers, the best way would be contacting them on their mailing list at mailto:dnsmasq-discuss@lists.thekelleys.org.uk