-
-
Notifications
You must be signed in to change notification settings - Fork 211
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Pi-Hole does not appear to be forwarding all requests to unbound upstream #2235
Comments
How does the corresponding log line of
|
This is all I get when I
|
I am also having this issue with cloudflared as an upstream. I've rolled back to 5.18 for now. |
You set
|
Maybe I'm in the wrong place but I was getting no successful DNS resolutions whatsoever after upgrading to 6.0, neither internal nor external hosts were resolving. I do have a wildcard DNS configured for internal stuff, but I wasn't really able to test the bounds of the failure. My priority was to restore everything from backups and get everything to stop complaining, restore internet access. If/when I have a couple of hours I can try to do an upgrade on a test system and see if I can get better data. I use unbound as my upstream, that's the reason I subscribed to this issue... It certainly looked this morning like PiHole was failing to forward to Unbound. |
Please open a new issue and post a debug log |
Thanks @yubiuser, clearing out the DNS domain restores forwarding of my home.lan domain to unbound. However, there seems to be a disconnect between the comments in dnsmasq.conf vs what is described in the web UI: ![]()
In my case, I am not using Pi-hole's DHCP server but I did have a domain specified so I was led to believe that setting this acted more like a search domain feature rather than telling Pi-hole to be authoritative for the domain. Incidentally, I think the upgrade to Pi-hole v6 brought this over from my v5.x setup but I can't be sure. This is what I had in my
So maybe this is a documentation issue? Or maybe something was inadvertently changed with regard to this setting. |
Hmm, started an entirely new LXC with PiHole Unbound on it, upgraded to 6.0 and every instance of unbound on my network crashed instantly and simultaneously. Can reproduce this 100% of the time: restore from backup, turn on the 6.0 machine, they all suddenly go offline and must be restored from backup, restarts don't help. I don't see anything in the obvious logs and unfortunately to dig deeper I'd need to do this in a safer manner, and I'm frankly not equipped to do any more isolation at this time. Looks like to thoroughly test this I would need to set up a separate router and a few test systems. Well outside the time I have to sink into this. I guess my issue is different than OP, but I can't submit a ticket as the bug is far too invasive to analyze in my environment. |
I have a very similar issue.
debug token is: https://tricorder.pi-hole.net/PqCtgDTt/ |
Queries for Pi-hole's 'domain' are never forwarded upstream. Change the setting in Settings > DNS |
I would agree with others on two counts:
I also use a .myname.lan domain hosted on bind servers in my lan, and put pihole between certain clients and those servers to provide ad blocking. But I also need clients behind the piholes to be able to resolve my internal *.myname.lan addresses. Having this silently intercept all traffic for any *.lan domain out of the box is going to break a lot of peoples networks. |
Transferred to FTL repo as changes in |
The default domain is With Hence, Pi-hole enforces this " Do you think there is an error in this logic (I don't think so)? But I see we should extend the documentation. |
# The DNS domain used by your Pi-hole.
#
# This DNS domain in purely local. FTL may answer queries from its local cache and
# configuration but *never* forwards any requests upstream *unless* you have
# configured a dns.revServer exactly for this domain. In the latter case, all queries
# for this domain are sent exclusively to this server (including reverse lookups).
#
# For DHCP, this has two effects; firstly it causes the DHCP server to return the
# domain to any hosts which request it, and secondly it sets the domain which it is
# legal for DHCP-configured hosts to claim. The intention is to constrain hostnames so
# that an untrusted host on the LAN cannot advertise its name via DHCP as e.g.
# "google.com" and capture traffic not meant for it. If no domain suffix is specified,
# then any DHCP hostname with a domain part (ie with a period) will be disallowed and
# logged. If a domain is specified, then hostnames with a domain part are allowed,
# provided the domain part matches the suffix. In addition, when a suffix is set then
# hostnames without a domain part have the suffix added as an optional domain part.
# For instance, we can set domain=mylab.com and have a machine whose DHCP hostname is
# "laptop". The IP address for that machine is available both as "laptop" and
# "laptop.mylab.com".
#
# You can disable setting a domain by setting this option to an empty string.
#
# Possible values are:
# <any valid domain>
domain = "lan" This would be my proposal - any suggestions for changes/improvements? |
I think the web interface text for this option also needs clarification |
The web interface text is automatically generated from the description in |
Only in All Settings, not on the DNS page |
Seems it is still something wrong with the upgrade process, After upgrading to 6.0.x, domain "lan" was added, there was no domain entry in setupVars.conf before. Also "bogus-priv" was set, which was disabled before. |
This is somewhat expected. Not setting a value in |
I guess this is now considered 'fixed' after changing the documentation? If not, please excuse this oversight, and simply reopen the issue |
Versions
Core version is v6.0 (Latest: v6.0)
Web version is v6.0 (Latest: v6.0)
FTL version is v6.0 (Latest: v6.0)
Platform
Expected behavior
Pi-Hole is configured to use my unbound server running on localhost (port 5335) as a forwarder (the only one configured in Pi-Hole). Unbound contains the following stanza:
This allows me to create any subdomain on the enterprise.home.lan domain such as scrypted.enterprise.home.lan and have it resolve to my server at 192.168.8.15. For example:
Since upgrading to Pi-Hole 6, Pi-Hole no longer seems to forward requests to unbound and instead returns an NXDOMAIN result.
If I query Pi-Hole for the same domain name (which should forward the query to unbound):
NXDOMAIN is returned instead of an IP address that unbound would be returning.
Here is how this is represented in the query log immediately after restarting the DNS resolver (i.e., cache flushed but query status indicates it was somehow served from cache):
And from the FTL log:
However, if I query for
something.web.devel
(which also provides a similar wildcard feature in my unbound config), Pi-Hole correctly forwards that to unbound and resolves it:Actual behavior / bug
Pi-Hole does not appear to be forwarding all requests to upstream DNS (in my case, unbound running on localhost#5335).
Steps to reproduce
See above, not really sure how to reproduce this from scratch.
Debug Token
Screenshots
Additional context
I upgraded from Pi-Hole v5.18.4. This configuration worked as expected then but no longer does following the upgrade to v6.0.
The text was updated successfully, but these errors were encountered: