https://bugs.exim.org/show_bug.cgi?id=1141
--- Comment #25 from Dave <exim-cc0@???> ---
(In reply to Jeremy Harris from comment #24)
> Can you confirm there were two distinct physical interfaces being selected
> by the
> two different IP addresses (and not a single interface that happens to have
> two
> addresses assigned?
No, in this case it is a single physical interface. IPs are assigned to
individual aliases (the live case is complex, so IPs assigned to bond0.343:1,
bond0.343:2, bond0.343:3 etc), so they're not all placed on a single interface
name.
Admittedly the behaviour expected is the alias matching the IP address, which
is usually the case. If the base physical interface is sometimes selected, this
should probably be documented as undefined behaviour.
Setting connection_max_messages = 1 works as expected with regards to the
source IP address/interface alias used, as do the majority of cases. Very
occassionally, the "wrong" transport gets chosen, usually under high load.
It appears that the proposed change in
https://git.exim.org/exim.git/commit/9d38e6c3845871c8774cd6494163b1625235080a
appears to work with regards to fixing this.
--
You are receiving this mail because:
You are on the CC list for the bug.
--
## subscription configuration (requires account):
##
https://lists.exim.org/mailman3/postorius/lists/exim-dev.lists.exim.org/
## unsubscribe (doesn't require an account):
## exim-dev-unsubscribe@???
## Exim details at
http://www.exim.org/
## Please use the Wiki with this list -
http://wiki.exim.org/
