Re: Temporary failure in name resolution
>> That I can believe (and isn't a bad option, IMO),
>> but the discussion was about broken DNS proxies/servers.
> Well, I assume that's your term for the modem/router being leased
> by the OP from att, and similar.
Not really. It was my attempt at reproducing the approximate
description of some vague potential sources of problems seen earlier in
the thread.
> I would assume that, rather like BT (British Telecom) and their
> HomeHubs for domestic customers, they lock in the addresses of their
> own DNS servers for the reasons outlined above.
That would make sense, indeed: it could even be a perfectly fine piece
of software, just purposefully misconfigured like you describe.
> But I don't recognise your statement "most home routers used `dnsmasq`",
> because AIUI dnsmasq implies "It can serve the names of local machines
> which are not in the global DNS" (Wikipedia), and none of my routers has
IIRC while "it can" this depends on the configuration being used.
My statement was based on my memory of the contents of the "big&ugly GPL
tarball" that you could usually get from the somewhat complying
companies (after they were reminded of the terms of the GPL license
used in the software inside their boxes).
[ To me, this is actually one of the great successes of the GPL, since
it spurred the whole development of OpenWRT. ]
>> That's not even needed: just tell the DHCP clients to use the ISP's
>> DNS servers.
> That seems to be the general solution. But I'm not sure if that's
> possible with certain devices, like "smart" TVs that can browse
> the web.
It should work under the condition that the DHCP client uses the info
provided by DHCP. I'm sure there are devices braindead enough to
hardcode their own list of DNS servers (IIUC it's somewhat common to
have hardcoded Google-DNS servers in some Android devices).
Stefan
Reply to: