"Dan Fandrich" <dan_at_coneharvesters.com> wrote:

> On Wed, Mar 28, 2007 at 06:32:09AM +0200, Gisle Vanem wrote:
>> Trying to connect to a non-existant host, gives me these verbose
>> messages with the tftp and ftp protocols:
>>
>> >curl -v ftp://10.0.0.45
>> * Couldn't find host 10.0.0.45 in the _netrc file, using defaults
>> * About to connect() to 10.0.0.45 port 21 (#0)
>> * Trying 10.0.0.45... connected
>> * Connected to 10.0.0.45 (10.0.0.45) port 21 (#0)
>
> And you're sure there are no ftp server running on that host? Could
> there be a transparent proxy in the way? The tftp case might be a bit
> strange because there's no such thing as a "connection" with UDP, but
> I don't see why you'd get the above for ftp unless socket 21 was being
> listened to.

No host on that address. Running tcpdump along-side shows only
ARP-Requests.

But switching off my Win-XP firewall, the ftp-test
doesn't show the "connected" part (verifyconnect()).
In the case of tftp, maybe it's regarded as "connected" when the
bind() returns?

--gv
Received on 2007-03-28