From: Jim Gettys <jg@freedesktop.org>
To: Michael Richardson <mcr@sandelman.ca>
Cc: Kerry Lynn <kerlyn@ieee.org>,
"cerowrt-devel@lists.bufferbloat.net"
<cerowrt-devel@lists.bufferbloat.net>
Subject: Re: [Cerowrt-devel] Strange behavior with macbook and DNS
Date: Wed, 24 Oct 2012 21:18:33 -0400 [thread overview]
Message-ID: <50889369.1070705@freedesktop.org> (raw)
In-Reply-To: <1081.1351123608@sandelman.ca>
On 10/24/2012 08:06 PM, Michael Richardson wrote:
> Dave Taht <dave.taht@gmail.com> wrote:
> >> I think something is up with the Avahi reflector, but haven't had
> >> time to look into it.
>
> DT> This is an issue with avahi's mdns tying names to mac addresses,
> DT> and this being a routed, rather than bridged, network. It's
> DT> actually a flaw in mdns's design which I hope will be worked on
> DT> in the new dnsext working group at the ietf.
>
> So, basically, each time the mac goes from wired to wireless, the
> reflector sees the name, says, "hey, that name is in use over *THERE*",
> and declares a conflict.
>
> If tied to mac address, then I'd expect at most two names (one for the
> wired mac, one for the wireless mac), which is also suboptimal, but
> understandable. SO I think you are saying that in addition to this
> protocol problem, there is also some bug in the reflector causing all
> previous names to be taken?
Dunno. Kerry Lynn is probably in a better position to say.
>
> (I'm asking mostly so that I can make sure that mdnsext can capture this
> as part of the problem statement. BTW: who will be in Atlanta?)
We have another problem to capture in mdnsext.
Multiple CeroWrt routers can mesh; so there can be loops, and mdns has
problems with that: it will forward things it's already heard. I
observe Avahi in a hard loop (and probably flooding the network with
lots of messages).
So Avahi basically loops in that case forwarding announcements and
generating names.
Kerry and I got together a few weeks back and discussed this. I don't
know if he's had time to set up some routers to test and diagnose
exactly what is happening yet or not (I suspect not, that he's probably
been busy with the startup of the working group).
So this is a concrete example of the problems homenet faces with
multiple routers in the home. Naive people (like my family) want their
printer discovery to "just work".
- Jim
>
>
>
> _______________________________________________
> Cerowrt-devel mailing list
> Cerowrt-devel@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/cerowrt-devel
prev parent reply other threads:[~2012-10-25 1:18 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-10-24 13:04 Shannon Kendrick
2012-10-24 13:14 ` William Katsak
2012-10-24 15:50 ` Dave Taht
2012-10-25 0:06 ` Michael Richardson
2012-10-25 1:18 ` Jim Gettys [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
List information: https://lists.bufferbloat.net/postorius/lists/cerowrt-devel.lists.bufferbloat.net/
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=50889369.1070705@freedesktop.org \
--to=jg@freedesktop.org \
--cc=cerowrt-devel@lists.bufferbloat.net \
--cc=kerlyn@ieee.org \
--cc=mcr@sandelman.ca \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox