Dnsmasq[2757]: reply error is SERVFAIL

[Wiggum127]

Expected Behaviour:

Running bare metal version 6 on Raspbian (Bullseye). Noticed the DNS service being very unstable and dropping every so often.

Looking at the admin webui there was an errors:

Connection error (208.67.222.222#53): TCP connection failed (Operation in progress)
Connection error (208.67.220.220#53): TCP connection failed (Operation in progress)

The IP addresses are of the forward DNS servers of OpenDNS. Have been using these on Pihole for many years without any issues. Checking the connection with DNS and service of OpeDNS did show any problem. To debug the situation I changed my smartphone IP settings to use OpenDNS directly. Never gave any issue, while my other devices were suffering.

Looking at the pihole.log there are repeating / recurring errors:

Feb 24 00:13:26 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 00:49:23 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 00:49:31 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 01:09:28 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 02:07:55 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 02:07:55 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 02:07:55 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 02:07:55 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 02:07:57 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 02:07:57 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 02:08:00 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 02:08:00 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 02:08:00 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 02:08:00 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 02:08:00 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 14:31:32 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 14:31:37 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 14:31:42 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 14:31:47 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 14:31:52 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 14:31:57 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 14:32:02 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 14:32:07 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 14:32:12 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 14:32:17 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 14:32:22 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 14:32:27 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 14:32:33 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 14:32:38 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 14:32:43 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 14:32:48 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 14:32:53 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 14:32:58 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 16:01:03 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 16:37:43 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 16:37:43 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 16:37:43 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 16:54:11 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 17:14:19 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 17:29:57 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 17:30:01 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 17:30:34 dnsmasq[2757]: reply error is SERVFAIL
Feb 24 17:30:34 dnsmasq[2757]: reply error is 1
Feb 24 22:05:40 dnsmasq[2757]: reply error is SERVFAIL

I could not find any information what this error means and how to debug or fix this.

Actual Behaviour:

Stable DNS resolution.

Debug Token: https://tricorder.pi-hole.net/PoaH7bVd/

Thanks in advance for your help

[jfb]

Have you tried different upstream DNS servers with Pi-hole to see if that resolves the problem?

V6 uses a new dnsmasq version that can send DNS queries in mixed case, and some upstream servers object to that.

[Wiggum127]

Thanks for your suggestion.

Should mixed cases be an issue, there should be entries in the Pihole log files with mentioning "Case mismatch".

Doing a grep of all files pihole.* and FTL.* looking with -i "case" I found 0 entries back.

So, no, this doesn't seem to be the issue.

[radinsky]

I'm seeing the same issue - CONNECTION_ERROR TCP connection failed (Operation in progress)

[jfb]

@Wiggum127

this doesn't seem to be the issue.

Did you try using different upstream DNS servers?