------- You are receiving this mail because: -------
You are on the CC list for the bug.
http://bugs.exim.org/show_bug.cgi?id=1405
Graeme Fowler <graeme@???> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |graeme@???
Status|NEW |RESOLVED
Resolution| |INVALID
--- Comment #2 from Graeme Fowler <graeme@???> 2013-10-25 10:49:21 ---
Firstly, use
/sbin/ip address list
(or the short form /sbin/ip add ls)
to determine all your local IP addresses.
Exim will bind the whatever addresses are on the system if told to; by default
it will send from the one that is used as the source when the system itself is
making outbound connections.
If I recall correctly, most kernel IP stacks will default to the numerically
'lowest' address as their outgoing one if several are available. You should be
able to see which one by using
/sbin/ip route list
(or the short form /sbin/ip ro ls)
This is from a machine with 30 or so 'secondary' addresses (it's doing a lot of
websites)
[root@web ~]# ip ro ls
aaa.bbb.ccc.0/24 dev eth0 proto kernel scope link src aaa.bbb.ccc.89
169.254.0.0/16 dev eth0 scope link metric 1002
default via aaa.bbb.ccc.1 dev eth0
You can see there that the default gateway is the aaa.bbb.ccc.1 address on the
same range; the range is reached via device eth0 with a source address of
aaa.bbb.ccc.89.
The behaviour you're seeing implies two equal 'weight' addresses; i.e. neither
are configured as 'secondary'. It's perfectly normal.
If you want Exim to utilise a fixed address for outgoing connections, see the
docs for the SMTP transport and look for the 'interface' option.
--
Configure bugmail:
http://bugs.exim.org/userprefs.cgi?tab=email
