Gert van Dijk [Tue, 2 Apr 2019 09:09:19 +0000 (11:09 +0200)]
docs: update pdnsutil 'set-nsec3' and NSEC3 narrow mode
* The 4-parameter quoted string argument in "pdnsutil set-nsec3" appears to
have a default value of "1 0 1 ab" according to the sources:
https://github.com/PowerDNS/pdns/blob/e62422ce0e86cfe959073ef061f07873ceff6be8/pdns/pdnsutil.cc#L2459
introduced with b8adb30dff5fc18b55e27580c07799d8fae1bafb (in 3.3+)
* The example command included smart quotes, which don't really work when
copy-pasted to a shell - changed to monospaced formatting.
* Include note about online signing requirement for "White Lies" / narrow
mode support and that zone transfers are denied in this mode.
* Mention RFC about "White Lies".
* Mention defaults on 'DNSSEC Modes of Operation'.
* Mention possible limits of ITERATIONS via 'max-nsec3-iterations'.
Gert van Dijk [Mon, 1 Apr 2019 19:54:07 +0000 (21:54 +0200)]
docs: Note that supermaster support is off by default since 4.2
With commit b80139777ebd2e0dc14cfd88d30eed80cfc04054 / PR #6021, it appears
that the behaviour of PowerDNS has changed to not act as a supermaster any
longer by default. Having to add a new setting that did not exist before,
to retain the old behaviour should be mentioned, I believe.
Gert van Dijk [Sun, 31 Mar 2019 18:51:55 +0000 (20:51 +0200)]
docs: fix formatting of 'rng' setting options
Was rendered as a field definition, because both:
* A blank line was omitted before starting the list items.
* Leading whitespace for each item (much like the commit that removes the
superfluous blockquotes).
Gert van Dijk [Sun, 31 Mar 2019 18:30:28 +0000 (20:30 +0200)]
docs: Align position of version annotation in settings
On two out of many other occurrences the versionadded/deprecated directive
was rendered above the parameters, but the vast majority had it below. This
change aligns the two outliers.
Gert van Dijk [Sun, 31 Mar 2019 17:37:08 +0000 (19:37 +0200)]
docs: Add warning on empty bind-dnssec-db for slave operation
I fell right into the pitfall of configuring a slave with the BIND
backend, serving presigned records, assuming it will serve the RRSIGs
just fine, but no, my domain went bogus. This was documented, but
not as clearly as I hoped for, this commit improves the documentation
regarding that.
Gert van Dijk [Sun, 31 Mar 2019 17:25:22 +0000 (19:25 +0200)]
docs: BIND backend - improve formatting of output status
Before this change it was hard to distinguish the three possible statuses
due to the styling of the inline-monospaced text without contrast in the
background for the current Sphinx theme.
Gert van Dijk [Sun, 31 Mar 2019 17:25:22 +0000 (19:25 +0200)]
docs: HTTP API tsigkeys example shows wrong body
Two errors fixed:
* The JSON body parameter key 'key' was provided twice. I think it should
have been the 'name' key.
* One value wasn't terminated with a double quote.
The HTTP code block parser warned about this in the sphinx-build output.
Gert van Dijk [Sun, 31 Mar 2019 17:25:22 +0000 (19:25 +0200)]
docs: Other minor fixes
* Make sentence actually part of the 'versionchanged' directive on the
settings page. Fixes sphinx-build warning.
* LDAP backend formatting changes so that ldap:// URIs don't get
linkified.
* Fix sphinx-build warning about under and over underlined sections.
Gert van Dijk [Sun, 31 Mar 2019 17:25:21 +0000 (19:25 +0200)]
docs: Fix formatting of Lua2 backend API
Without a blank line between the paragraphs in the descriptions of
RST description lists [1], the list items will not render as list items,
but as a long single paragraph.
Gert van Dijk [Sun, 31 Mar 2019 17:25:21 +0000 (19:25 +0200)]
docs: fix link in Lua backend
Two problems with the existing link:
* target was non-existent (anchor is different in Sphinx)
* nested markup for monospaced text on link is not possible, see
https://stackoverflow.com/q/4743845/1254292
Gert van Dijk [Sun, 31 Mar 2019 17:25:21 +0000 (19:25 +0200)]
docs: remove unintentional blockquotes in HTML
Having leading spaces in either:
* Lists
* RST directives, such as
.. toctree::
will lead to the listing being wrapped in an HTML <blockquote> element with
also the styling as such. This is probably unintentional and at least
inconsistent with other occurrences.
Keep, for each frontend and backend:
- the number of concurrent TCP connections
- the average number of queries per connection
- the average duration of a connection
dnsdist: Try reading from the TCP backend right away
Instead of waiting for the socket to be readable, as it might
already be, so we save a multiplexer trip, and prevent an issue
if we ever add a TLS layer between dnsdist and the backends.
dnsdist: Actually try to read before checking if the socket is readable
We need to because the TLS layer might already have data waiting
for us, while there might not be anything left on the OS-level
buffer associated to the socket.
If we don't ask the TLS layer, we might wait indefinitely for
something to arrive while the client has already sent everything,
and it's just waiting for us because the TLS record has been read.