On 17 Apr 2005, at 00:42, Andrew Nimmo wrote:
>
> On 16 Apr 2005, at 22:09, Marc Haber wrote:
>
>> On Sat, 16 Apr 2005 17:52:57 +0200, Andrew Nimmo
>> <andrew.nimmo+exim.org@???> wrote:
>>> no (ok, 20 minutes or so...). note however, i continued testing and
>>> modified the firewall to not deny on 113 (because that was an obvious
>>> difference). a subsequent connect to mail.krwdigital.com responded
>>> with
>>> a 220 almost immediately, as expected. initial findings would suggest
>>> an ident-related issue, so i would consider the poster's exim config
>>> ident settings, first. i've not checked, but were there any related
>>> issues with 4.44?
>>
>> I don't remember any. Both servers probably run with Debian's default
>> configuration, which means that ident requests are always sent out.
>> This usually does not do any harm besides the usual 30 seconds delay
>> on incoming mail from sites with broken firewall setups. I have never
>> seen this causing remote sites to hang.
>>
>
> interesting. in that case it may be useful to find out which brand
> adsl router and its settings, etc, for comparison with the mx=1
> server. for example, i've seen various problems (tho not this one)
> with the linksys wag54g, with the 'firewall' active. note that the
> non-220 can also be seen via:
>
> http://www.checkdns.net/quickcheck.aspx?
> domain=krwdigital.com&detailed=1
>
to conclude this thread, the user:
* upgraded to exim-4.50, which made no difference
* upgraded the adsl router firmware, which made no difference
* resolved the issue by setting:
rfc1413_hosts = *
rfc1413_query_timeout = 0s
in the absence of more experimental information, we've assumed that
there is an issue with the adsl router being used when port 113 is
blocked at the remote site, resulting in no 220 response whilst waiting
indefinitely for the ident query reply.
regards
andrew
