DNS service not starting automatically

[Newton_tricks]

Expected Behaviour:

Raspberry Pi 4 with Raspberry Pi OS
Also using local DNS proxy cloudflared

Actual Behaviour:

Pi-hole doesn’t start DNS service automatically when restarting, have to restart DNS service manually to make it work.

Debug Token:

https://tricorder.pi-hole.net/jr21si26xs

Found this previous post:
https://discourse.pi-hole.net/t/dns-resolver-not-starting-on-restart/
But it didn’t help.

[DL6ER]

@Newton_tricks

maybe pihole can automatically run “restart DNS resolver” everytime when startup?

No, Pi-hole starts up on system boot. A following restart wouldn't make much sense.
Maybe it's an issue with the order of your network interfaces coming up.
What does /var/log/pihole.log and /var/log/pihole-FTL.log show?

There is the DELAY_STARTUP option for FTL to, well, do exactly what it says. It may be a solution for you. A debug token would still be very helpful so we can check if your interface is configured in a strange way or something similar.

@Newton_tricks

It’s even a good idea to run the restarting every 9 hours let’s say.

No, because you would loose your entire DNS cache at this point.

[Newton_tricks]

Thanks for the reply, updated debug token.

[yubiuser]

I can't spot any obvious errors in your debug log.

---

Did you manually restart Pi-hole twice within a minute?

   Mar  6 07:32:50 dnsmasq[971]: exiting on receipt of SIGTERM
   Mar  6 07:32:50 dnsmasq[22331]: started, version pi-hole-2.84 cachesize 10000
   Mar  6 07:32:50 dnsmasq[22331]: DNS service limited to local subnets
   Mar  6 07:32:50 dnsmasq[22331]: compile time options: IPv6 GNU-getopt no-DBus no-UBus no-i18n IDN DHCP DHCPv6 Lua TFTP no-conntrack ipset auth cryptohash DNSSEC loop-detect inotify dumpfile
   Mar  6 07:32:50 dnsmasq[22331]: DNSSEC validation enabled
   Mar  6 07:32:50 dnsmasq[22331]: configured with trust anchor for <root> keytag 20326
   Mar  6 07:32:50 dnsmasq[22331]: using only locally-known addresses for domain use-application-dns.net
   Mar  6 07:32:50 dnsmasq[22331]: using nameserver ::1#5053
   Mar  6 07:32:50 dnsmasq[22331]: using nameserver 127.0.0.1#5053
   Mar  6 07:32:50 dnsmasq[22331]: read /etc/hosts - 5 addresses
   Mar  6 07:32:50 dnsmasq[22331]: read /etc/pihole/custom.list - 0 addresses
   Mar  6 07:32:50 dnsmasq[22331]: read /etc/pihole/local.list - 2 addresses
   Mar  6 07:33:11 dnsmasq[22331]: exiting on receipt of SIGTERM
   Mar  6 07:33:11 dnsmasq[22461]: started, version pi-hole-2.84 cachesize 10000

Side note:
Your router distributes it's own and another IP as DNS server via DHCP. But it's not Pi-holes IP. Is this on purpose?

      dns-server: 192.168.2.1
      dns-server: 207.164.234.193

[Newton_tricks]

I restarted pi-hole again but not twice within a minute this time, and I have updated the debug token.

My router's DNS resolver is not pihole because I just use pihole for myself mostly.

[yubiuser]

@Newton_tricks

My router's DNS resolver is not pihole because I just use pihole for myself mostly.

Your Pi-hole has not received a single query, are you sure you use Pi-hole at all?

   [2021-03-07 13:34:01.703 695M] Imported 0 queries from the long-term database
   [2021-03-07 13:34:01.703 695M]  -> Total DNS queries: 0
   [2021-03-07 13:34:01.703 695M]  -> Cached DNS queries: 0
   [2021-03-07 13:34:01.703 695M]  -> Forwarded DNS queries: 0
   [2021-03-07 13:34:01.704 695M]  -> Blocked DNS queries: 0
   [2021-03-07 13:34:01.704 695M]  -> Unknown DNS queries: 0
   [2021-03-07 13:34:01.704 695M]  -> Unique domains: 0
   [2021-03-07 13:34:01.704 695M]  -> Unique clients: 0
   [2021-03-07 13:34:01.704 695M]  -> Known forward destinations: 0

Please run

nslookup pi.hole

nslookup pi.hole 192.168.2.56

from the device that you think is using Pi-hole

---

Are you blocking port 53 in your network?

*** [ DIAGNOSING ]: Operating system
[i] dig return code:  10
[i] dig response:  dig: couldn't get address for 'ns1.pi-hole.net': failure

[Newton_tricks]

I am using pihole on my phone, so I can't run nslookup. However, this time I restarted pihole and using my phone trying to make a few DNS request, then run pihole -d and updated the debug token.

The interesting thing is, under pihole DNS settings, if I set "Listen on all interfaces" then the problem occurs, but if I set "Listen only on interface eth0" then there won't be any problem at all...

[yubiuser]

@Newton_tricks

I am using pihole on my phone, so I can't run nslookup. However, this time I restarted pihole and using my phone trying to make a few DNS request, then run pihole -d and updated the debug token.

Please run echo ">stats >quit" | nc localhost 4711 on your Pi-hole.

@Newton_tricks

The interesting thing is, under pihole DNS settings, if I set "Listen on all interfaces" then the problem occurs, but if I set "Listen only on interface eth0" then there won't be any problem at all..

Did you try

@DL6ER

There is the DELAY_STARTUP option for FTL to, well, do exactly what it says. It may be a solution for you. A debug token would still be very helpful so we can check if your interface is configured in a strange way or something similar.

[Coro]

I just responded to this German discussion:

Pihole arbeitet nach Neukonfiguration nicht mehr richtig

Hallo zusammen, eigentlich wollte ich gestern nur OpenVPN auf WireGuard umstellen, aber irgendwie habe ich meine Pihole Config versaut. Diese lief jetzt einige Monate fehlerfrei. Was ist das Problem: Sobald der Raspi neu gestartet wird, bekomme ich an den Clients kein Internet und Namensauflösung (mittels nslookup) mehr. Es funktioniert erst wieder, nachdem ich am Pihole "Restart DNS Resolver" angeklickt habe. (eventuell ein Bug?) Ich habe alles probiert aber nach einem Neustart geht es nich…

Maybe it is related. Note that there isn't a solution yet over there. They describe as well that Pi-hole starts but doesn't reply to queries until it is restarted manually.