The knot-resolver documentation only appears to support Linux. Any plans to
make
the resolver available, or useful on systems other than Linux? While I could
put Linux in a VM, or BE, and use the resolver from there. It wouldn't really
be
very effective. Any insight to using this on any of the BSD's would be
greatly
appreciated (I'm currently using unbound along side knot authoritative).
Thanks!
--Chris
Hi,
I’ve stumbled across knot-resolver because I have an issue with my current DNS solution.
What is the best way to block a large number of domains.
I’ve trying to work with the below by it’s not functioning
Part of /etc/knot-resolver/kresd.conf
-- Domain Blocking
policy.add(
policy.rpz(policy.DENY_MSG('domain blocked by your IT department'),'/etc/knot-resolver/blacklist.rpz', true))
policy.add (
policy.rpz(policy.DENY, '/etc/knot-resolver/blacklist.rpz'))
/etc/knot-resolver/backlist.rpz
007bets.com,
Rrds,
Mike
Hi,
I’ve stumbled across knot-resolver because I have an issue with my current DNS solution.
What is the best way to block a large number of domains.
I’ve trying to work with the below by it’s not functioning
Part of /etc/knot-resolver/kresd.conf
-- Domain Blocking
policy.add(
policy.rpz(policy.DENY_MSG('domain blocked by your IT department'),'/etc/knot-resolver/blacklist.rpz', true))
policy.add (
policy.rpz(policy.DENY, '/etc/knot-resolver/blacklist.rpz'))
/etc/knot-resolver/backlist.rpz
007bets.com,
Rrds,
Mike
Hello,
recently we upgraded from 5.1 to 5.2 few servers (CentOS7 and Raspbian) and
all seems to be working fine, but I can see on (yes I know, not
recommend and supported) rasbian issue with metrics and in the log I can
issue this error:
Nov 13 10:21:33 dns-cache-2 kresd[15232]: map() error while connecting to
control socket /run/knot-resolver/control/H#003: socket:connect: No such
file or directory (ignoring this socket)
Nov 13 10:21:33 dns-cache-2 kresd[15232]: map() error while connecting to
control socket /run/knot-resolver/control/H: socket:connect: No such file
or directory (ignoring this socket)
Nov 13 10:21:33 dns-cache-2 kresd[15232]: map() error while connecting to
control socket /run/knot-resolver/control/: socket:connect: Connection
refused (ignoring this socket)
the error is triggered by opening "ip:8453/metrics" and it show almost
empty response:
# TYPE resolver_latency histogram
resolver_latency_count 0.000000
resolver_latency_sum 0.000000
root@dns-cache-2:/# apt -qq list knot* --installed
knot-resolver-module-http/unknown,now 5.2.0-1 all [installed]
knot-resolver-release/unknown,now 1.7-1 all [installed]
knot-resolver/unknown,now 5.2.0-1 armhf [installed]
root@dns-cache-2:/#
root@dns-cache-2:/# uname -a
Linux dns-cache-2 4.19.97-v7l+ #1294 SMP Thu Jan 30 13:21:14 GMT 2020
armv7l GNU/Linux
Could it be a bug or something related to mix armv7l and armhf? I know of
the limitation as we discussed this setup on raspberry some time ago. I
just want to share my experience as it can be useful and if you will have
any tip I will appreciate it.
Dear Knot Resolver users,
Knot Resolver 5.2.0 has been released!
One of the notable features is a new DNS-over-HTTTPS implementation
which is more scalable and stable than the old one. It also has less
dependencies and simpler configuration.
Another new feature is experimental eXpress Data Path (XDP) support for
UDP. With support from both the network card and the kernel, it can
provide superior performance and lower latency for UDP answers.
Some of the improvements and bugfixes required a few backward
incompatible changes, mainly regarding control sockets or module API.
Please refer to our upgrading guide for details:
https://knot-resolver.readthedocs.io/en/v5.2.0/upgrading.html#to-5-2
Improvements
------------
- doh2: add native C module for DNS-over-HTTPS (#600, !997)
- xdp: add server-side XDP support for higher UDP performance (#533,
!1083)
- lower default EDNS buffer size to 1232 bytes (#538, #300, !920);
see https://dnsflagday.net/2020/
- net: split the EDNS buffer size into upstream and downstream (!1026)
- lua-http doh: answer to /dns-query endpoint as well as /doh (!1069)
- improve resiliency against UDP fragmentation attacks (disable PMTUD)
(!1061)
- ta_update: warn if there are differences between statically configured
keys and upstream (#251, !1051)
- human readable output in interactive mode was improved (!1020)
- doc: generate info page (!1079)
- packaging: improve sysusers and tmpfiles support (!1080)
Bugfixes
--------
- avoid an assert() error in stash_rrset() (!1072)
- fix emergency cache locking bug introduced in 5.1.3 (!1078)
- migrate map() command to control sockets; fix systemd integration
(!1000)
- fix crash when sending back errors over control socket (!1000)
- fix SERVFAIL while processing forwarded CNAME to a sibling zone (#614,
!1070)
Incompatible changes
--------------------
- see upgrading guide:
https://knot-resolver.readthedocs.io/en/v5.2.0/upgrading.html#to-5-2
- minor changes in module API
- control socket API commands have to be terminated by "\n"
- graphite: default prefix now contains instance identifier (!1000)
- build: meson >= 0.49 is required (!1082)
- planned changes in future versions:
https://knot-resolver.readthedocs.io/en/v5.2.0/upgrading.html#upcoming-chan…
Full changelog:
https://gitlab.nic.cz/knot/knot-resolver/raw/v5.2.0/NEWS
Sources:
https://secure.nic.cz/files/knot-resolver/knot-resolver-5.2.0.tar.xz
GPG signature:
https://secure.nic.cz/files/knot-resolver/knot-resolver-5.2.0.tar.xz.asc
Documentation:
https://knot-resolver.readthedocs.io/en/v5.2.0/
--
Tomas Krizek
PGP: 4A8B A48C 2AED 933B D495 C509 A1FB A5F7 EF8C 4869
Hello team,
most probably we hit the issue !1070 "fix SERVFAIL in *FORWARD modes with
certain CNAME setup" and I would like to ask when we can expect release of
knot-resolver 5.2.0 (with the fix) or if there will be "hot fix" in 5.1.x?
example of affected domain: www.action.foundation.total
Thank you.
These two patches modify Knot Resolver's build process to check for
Texinfo utilities during configuration and, if they're found, to build
and install documentation in Info format (the standard for GNU
distributions such as Guix[0]) as well as HTML. (If the Texinfo
utilities aren't found, the build works as before.)
The first patch implements these changes and additionally updates the
manual to list Texinfo as an optional dependency.
The second patch isn't required but genericizes a few references
within the manual to documentation in specific formats, in the
interest of readability. Note I've left unchanged the "build-html-doc"
ref-ID to help ensure external links to the documentation continue to
function.
[0] https://guix.gnu.org/
----------
With these patches applied you'll find makeinfo generates a large
number of warnings of the form
knotresolver.texi:11811: warning: @anchor should not appear in @deffn
These warnings are harmless but result from the way Breathe processes
C declarations. It generates output with this structure (here for the
"kr_request_selected" macro):
<desc_signature ids="c.kr_request_selected" is_multiline="True">
<desc_signature_line add_permalink="True" xml:space="preserve">
<target ids="resolve_8h_1a9ba8c6cef807cdf580e3249675a2589c"
names="resolve_8h_1a9ba8c6cef807cdf580e3249675a2589c"></target>
<desc_name xml:space="preserve">kr_request_selected</desc_name>
<desc_parameterlist xml:space="preserve">
<desc_parameter noemph="True" xml:space="preserve"><emphasis>req</emphasis></desc_parameter>
</desc_parameterlist>
</desc_signature_line>
</desc_signature>
The issue is the "target" node's being placed within
desc_signature_line. I assume this is done for convenience when
generating HTML, but it also leads to this Texinfo output:
@deffn {Define} @anchor{lib resolve_8h_1a9ba8c6cef807cdf580e3249675a2589c}@anchor{25f}kr_request_selected (req)
which is invalid, since (as the warning says) @anchor is not supposed
to appear within @deffn.
Sphinx's Texinfo writer doesn't account for this, since there's no way
(that I can see) for normal ReST input to generate a "target" tag in
this position. However it's not clear Breathe is at fault here either,
since I can't find anything that defines where target is allowed to
appear or which tags desc_signature_line is meant to contain.
A straightforward approach to fixing this is implemented by this patch
to Sphinx, which causes its Texinfo writer to reorder target nodes
embedded within function descriptions:
diff --git a/sphinx/writers/texinfo.py b/sphinx/writers/texinfo.py
index 5ad2831dd..bd76acdc3 100644
--- a/sphinx/writers/texinfo.py
+++ b/sphinx/writers/texinfo.py
@@ -1383,6 +1383,11 @@ class TexinfoTranslator(SphinxTranslator):
if objtype != 'describe':
for id in node.get('ids'):
self.add_anchor(id, node)
+ # embedded anchors need to go in front
+ for n in node.traverse(nodes.target, include_self=False):
+ if isinstance(n, nodes.target):
+ n.walkabout(self)
+ n.parent.remove(n)
# use the full name of the objtype for the category
try:
domain = self.builder.env.get_domain(node.parent['domain'])
However, whether this is the correct approach or if a fix would better
be applied to Breathe isn't clear to me. Engaging with the community
of either project means signing up for services I'd rather not use, so
there's little I can do to carry this forward. Perhaps someone else is
interested?
---
Simon South
simon(a)simonsouth.net
Simon South (2):
doc: generate Info manual
doc: use non-format-specific references to documentation
doc/build.rst | 11 +++++++----
doc/meson.build | 7 +++++++
scripts/install-doc-info.sh | 8 ++++++++
scripts/make-doc.sh | 6 ++++++
4 files changed, 28 insertions(+), 4 deletions(-)
create mode 100644 scripts/install-doc-info.sh
--
2.28.0
Hi,
I am using knot-resolver with default configuration in arch linux. I am not
able to resolve some domains (like costco.ca). It works fine with other
resolvers like cloudflare/google. Is there way to debug this?
Thanks,
Bala