BIND 10 #2796: Add a counter for queries with RD=1
BIND 10 Development
do-not-reply at isc.org
Wed Jun 12 20:14:44 UTC 2013
#2796: Add a counter for queries with RD=1
-------------------------------------+-------------------------------------
Reporter: y-aharen | Owner:
Type: enhancement | jinmei
Priority: medium | Status:
Component: b10-auth | reviewing
Keywords: | Milestone:
Sensitive: 0 | Sprint-20130625
Sub-Project: DNS | Resolution:
Estimated Difficulty: 4 | CVSS Scoring:
Total Hours: 0 | Defect Severity: N/A
| Feature Depending on Ticket:
| Add Hours to Ticket: 0
| Internal?: 0
-------------------------------------+-------------------------------------
Comment (by jinmei):
Replying to [comment:17 y-aharen]:
> > Hmm, this explanation doesn't make much sense to me...
> > {{{
> > <para>
> > RD bit is not specific to queries (requests of opcode 0), but
> > <quote>qryrecursion</quote> is limited to queries because RD bit
is
> > meaningless for the other opcodes and they are passed to another
> > BIND 10 module such as <command>b10-ddns</command>.
> > </para>
> > }}}
> >
> > First, what's the source (RFC, etc) of the claim "RD bit is
> > meaningless for the other opcodes"? It's also somewhat inconsistent
> > with the first sentence "RD bit is not specific to queries".
> > Secondly, if we can say we don't count things related to non query
> > requests because they are forwarded to different processes, then why
> > do we have per opcode counters in Auth or generic requests/* counters?
> Well, there is no specification that claims it. However, RFC 1996
(Notify)
> does not mention recursing and some implementations such as NSD,
PowerDNS
> and BIND 10 don't care RD bit while processing Notify requests.
Moreover,
> RFC 2136 (Update) states that Update message uses only QR flag bit and
the
> other flags are explicitly zero or ignored if non-zero.
> I meant "RD bit is not specific to queries" as "RD bit can be set in the
> request regardless of opcode." If it is misleading I'll revise it.
>
> The latter is not a good reason; I'll remove it after the former
concludes
> in this ticket.
>
> If you don't agree with the explanation and if it should also be counted
> requests with RD=1 regardless of opcode, I'd like to add another item
> 'request.rd' for requests with RD=1 and keep 'qryrecursion' as is.
I'm not disagreeing with anything except for the wording. Maybe I
should suggest something, for example:
{{{
<para>
The <quote>qryrecursion</quote> counter is limited to queries
(requests of opcode 0) even though the RD bit is not specific
to queries. In practice, this bit is generally just ignored for
other types of requests, while DNS servers behave differently
for queries depending on this bit. It is also known that
some authoritative-only servers receive a non negligible
number of queries with the RD bit being set, so it would be
of particular interest to have a specific counters for such
requests.
</para>
}}}
if this (more or less) matches your reasoning, please just adopt it
and merge. If it's different from your intent, we can discuss it further.
--
Ticket URL: <http://bind10.isc.org/ticket/2796#comment:18>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development
More information about the bind10-tickets
mailing list