Hi Bill,
sorry for the late answer, I haven't forgotten you and I'm very happy
for the answer. Unfortunately I wasn't at the customer again to check
out the tipps!
I'll tell you after checking out.
Wish you the best
Fabian
On 21 Oct 2016, at 19:31, Bill Cole wrote:
On 19 Oct 2016, at 4:28, Fabian Blechschmidt wrote:
Good morning everyone,
I have the problem, that mail mate tries to use my virtual network
card, therefore having a wrong IP address when trying to access a
mail server, which blocks the connection (or the packet doesn't even
reach it)
I have Parallels and VirtualBox installed. Both using a couple of IP
ranges to do their stuff. Neither of them is using 19.2.168.42.*
My routing table - if I interpret that correctly, says, that
connections to 192.168.42.146 (which is the mail server), should rund
through default connection.
No, it says (in part) this:
default 192.168.42.254 UGSc 777 0
en0
169.254 link#4 UCS 1 0
en0
192.168.42 link#4 UCS 30 0
en0
192.168.42.146/32 link#4 UCS 1 0
en0
192.168.42.254/32 link#4 UCS 2 0
en0
192.168.42.254 0:e:38:38:ed:ff UHLWIir 778 92
en0 1196
192.168.42.255 link#4 UHLWbI 1 425
en0
Which, at first glance, suggests that 192.168.42.146 is an IP address
assigned to your physical ethernet interface: a LOCAL address.
However, for that I would expect to see a host route for that IP via
lo0 with the flags "UHLWIi", which isn't present. Yet, the only other
"UCS" route for a /32 net is for your default gateway, clearly a
working device with a Cisco MAC address. I'm a bit confused as to what
is going on here, but I don't think this can be a working config.
What address do you think your Ethernet interface should have?
en0 is my cable lan device. This is what I meant with default. Default
device, not route. So en0 is the normal card, with a normal cable on a
"normal" internet ISP router, so should have public DNS, and just access
to internet.
I understand, that with the domain "my.customer.ads" we don't know
yet which IP address the server has. I assume (but might be wrong
here) that we than simply use the default route and change the device
if needed after resolving the domain.
I am unable to parse that paragraph. I'm sure whatever you mean is
important to your interpretation of this problem, but it does not
makes sense in English.
Side issue: note that "my.customer.ads" has MX and A records in public
DNS, but they are bogus. The .ads gTLD is a Google project and no real
domains exist under it yet. I think you are using that name as a
placeholder here, but maybe not.
So in short: Whatever happens inside of MailMate leads to a wrong
network device to be used when contacting the server.
This is extraordinarily unlikely. It is technically possible for an
application to specify what address and/or interface it uses for a
specific connection, but for a client like MailMate there is
absolutely no reason to do so. Normal client app behavior when setting
up a TCP connection is to ask the OS to resolve a name to an IP
address then ask the OS to open a connection to that IP address on a
specific port. The OS typically determines the best local IP,
ephemeral local port, and routing for a connection, NOT the app.
Any idea what to do or how to debug? Tell me if I can help. I assume
I'm here until friday.
Fix your network config. If it does not seem wrong to you, the output
of these two commands might help illuminate what's going on:
networksetup -getinfo Ethernet
ifconfig -av
_______________________________________________
mailmate mailing list
[email protected]
https://lists.freron.com/listinfo/mailmate
--
Fabian Blechschmidt
Tel: +49 30 419 932 55
Handy: +49 176 666 55 256
_______________________________________________
mailmate mailing list
[email protected]
https://lists.freron.com/listinfo/mailmate
