Follow-up Comment #4, bug #14106 (project freeciv):

> > In the end, it appears that --Bind-meta= is a
> > more sensible thing to do in the case of a multihomed host

> Why is that? It seems to me that this only forces user to enter same IP
> twice; once for --bind and second time for --Bind-meta.

Not at all, as --bind= already implies --Bind-meta= on the same address
with my patch:

 +  if ((b_a = srvarg.meta_bind_addr != 0
 +       ? srvarg.meta_bind_addr
 +       : srvarg.bind_addr) != 0) {

> Actually, if they differ, doesn't client get wrong IP from metaserver (the
> one server used to connect metaserver, and not the one server is listening
> client connections from)?

This option allows for a multi-homed host to use a specific IPv4 address
when reporting to the metaserver, while still accepting connections on
any of its addresses.

In my case, e. g., the server has a NATed private (RFC 1918) IPv4
address from a cheaper ISP, which is the default route, and a real IPv4
address from another ISP (not to talk about of yet another RFC 1918
address for LAN.)  While I wish for civserver to accept connections on
any of these addresses, I need a specific IPv4 to be reported to the
metaserver.  Unless --Bind-meta= is used, the connection will be done
using the NATed address, and the address of the NAT will be in the
Metaserver's list.


Reply to this item at:


  Message sent via/by Gna!

Freeciv-dev mailing list

Reply via email to