Hello, I'm running Pi-hole on a Raspberry Pi 2, running debian 12 (bookworm) and utilising unbound. I've been running deb 12 for number of months now. Despite the old HW, the raspi and pihole generally run just fine. Until recently, my setup has been running fine almost all of the time. I don't appear to be constrained by the CPU and memory based on my current use case.
Previously, on occasion my pihole would stop responding to DNS queries. The raspi that it's hosted on would also become unresponsive. It was infrequent and I wasn't bothered to reboot it from time to time. The fact that the raspi become unresponsive hampered collection of diagnostic info.
Recently however, pihole seems to stop responding to queries every other day. The symptoms are not consistent. Sometimes it will only be some domains that are impacted and other times it will be all queries. One difference now is that I am able to interact with the raspi via SSH. Status commands indicate that the pihole and unbound are working normally, but dns queries from my host and locally on the raspi timeout. The issue will manifest at unpredictable times of the day. Power cylce or reset from the console resolves the issue.
As I am able to access the raspi, I was able to generate a diagnostic output. This may not be entirely useful at the moment, as the pihole currently seems to be working perfectly well.
Scanning all your interfaces for DHCP servers and IPv6 routers
Timeout: 6 seconds
Error while sending Router Solicitation on eth0: Network unreachable
* Received 300 bytes from 192.168.18.1 @ eth0
Offered IP address: 192.168.18.162
Server IP address: 192.168.18.1
Relay-agent IP address: N/A
BOOTP server: (empty)
BOOTP file: (empty)
DHCP options:
Message type: DHCPOFFER (2)
lease-time: 86400 ( 1d )
server-identifier: 192.168.18.1
--- end of options ---
Received 1 DHCP (IPv4) and 0 RA (IPv6) answers on eth0
In terms of the error messages in the diagnostics at the time I was experiencing the DNS timeouts (of which there were plenty), I didn't capture the terminal output when I was running that unfortunately. As these things go, I haven't experienced the same symptoms again since posting about the issue.
Was hoping to be ready to capture exactly what is required as soon as the issue arises again.
You don't need terminal output for Pi-hole diagnosis. It is a web UI panel, showing operational messages that occurred since pihole-FTL was (re)started.
It is strange that DNS servers are absent from your router's DHCP offers, but by itself, that shouldn't contribute to your issue, especially if your devices do talk directly to Pi-hole for DNS.
Your debug log shows that you have enabled Pi-hole's Conditional Forwarding.
If your router would use Pi-hole as its upstream DNS server (in addition to already telling its clients to use Pi-hole as local DNS server), then CF could close a partial DNS loop, which may result in a max concurrent or rate limit warning.
Either of those may render Pi-hole's DNS resolution unavailable for a short period, either for all clients or for only the rate limited ones.
If those would be causing your intermittent issues, you would see respective messages in Pi-hole diagnosis.
I have experienced one device that gets rate limited fairly regularly; for exceeding the 1000 queries per minute limit set. I have had a few other devices that have reported as rate limited from time to time. However, the times these errors are reflected in the diagnostics doesn’t corelate with the instances where I’ve lost DNS resolution entirely.
Best I can tell, the other errors reported weren’t related to this issue.
In terms of DNS on my router, I use my ISP DNS servers. I’ve elected to do that, as I don’t need to contain ads for the router. I also wanted to retain the ability to manually set any hosts DNS server to ,y router IP; to bypass the pi-hole.
After having this issue manifest several times over a few days, I haven’t had the issue again for a few days. I suppose, I’ll need to wait for the issue to come up again, to tshoot further
Pihole appears to be running flawlessly for several days now. Not sure what to make of it. If it give and issue again in the next 21 days, I'll update this topic.