Tony Finch wrote:
>> So, can you tell me why sesame delivers to mx.ymmv.de with IPv4 most of
>> time, but there are times it keeps using IPv6.
> No :-) I should probably turn up the logging. It's notable that deliveries
> switch between the transports infrequently - it'll stick to one for rarely
You mean transport layer (that would be tcp, then) or Exim-transport? If
the latter, why do you have different transports?
> less than a day at a time. But I don't know what this implies.
Here's a little grep'n'cut output (when protocol switched):
2005-11-01 06:00:49 [131.111.8.41]:1798
2005-11-18 19:00:41 [2001:630:200:8080:20e:cff:fe09:a58a]:2332
2005-11-19 21:02:19 [131.111.8.41]:1578
2005-11-20 03:09:46 [2001:630:200:8080:20e:cff:fe09:a58a]:1327
2005-11-21 04:22:00 [131.111.8.41]:3762
2005-11-26 14:10:48 [2001:630:200:8080:20e:cff:fe09:a58a]:4157
2005-11-27 16:29:35 [131.111.8.41]:3519
2005-11-27 19:41:25 [2001:630:200:8080:20e:cff:fe09:a58a]:1932
2005-11-28 21:06:12 [131.111.8.41]:4603
2005-11-30 13:11:53 [2001:630:200:8080:20e:cff:fe09:a58a]:3585
> (Hmm, this might be easier to diagnose if we had different host names for
> 4 and 6.)
I though about having only AAAA for the highest priority mx, but there
might be MTAs that barf about that, you never know.
>> 2005-11-15 11:57:44 1EbySR-0005JO-8D sesame.csx.cam.ac.uk
>> [2001:630:200:8080:20e:cff:fe09:a58a] Connection timed out
> I wonder what caused that...
network problem? I have only a tunnel to sixxs, but the connectivity is
very good. Then again, I don't use ipv6 much so I'm not very likely to
notice outages.
