Bernd Wurst writes:
Okay. Which component cares about connecting to the secondary MX (or any otheravailable IP address) when TCP connection to the primary fails? As I understood, the connection to the secondary MX should take place immediately after the second try on the first MX fails on TCP layer. But only one connection failure is logged.
Looking at the code, only the last connection failure gets logged. If the first IP address tried fails, but Courier connects to the next IP, the connections goes through, and nothing gets logged (and nobody would care, I think).
We have 0.65.2 on our production system. When I read the changelog, I stumbledupon the following changelog-entry: 2011-01-22 Gordon Messmer <[email protected]> * rfc1035/rfc1035mksocket.c (rfc1035_mksocket): Set IPV6_V6ONLY socket option to OFF for IPv6 sockets, if the system default is on. Could it be that this fixes my issues? Is this change included in 0.65.3?
No, it's not. If you're affected by this, you would not be able to establish IPv4 connections at all, ever. This change undoes a default setting in some kernel configurations that prevents IPv6 sockets from connecting to IPv4 addresses. On systems with IPv6 support, Courier always creates IPv6 sockets, and expects to be able to connect to IPv4 IP addresses, as part of the IPv6 specification.
This works under the default kernel configuration, but for some reason Debian's kernel is configured with this switched of by default, and this change explicitly enables IPV6←→4 interoperability.
pgppbtdaPGanY.pgp
Description: PGP signature
------------------------------------------------------------------------------ Colocation vs. Managed Hosting A question and answer guide to determining the best fit for your organization - today and in the future. http://p.sf.net/sfu/internap-sfd2d
_______________________________________________ courier-users mailing list [email protected] Unsubscribe: https://lists.sourceforge.net/lists/listinfo/courier-users
