instead of removing the server for the entire life of the process, why
not back it off in expotentially increasing times (ala TCP/IP) and then
remove it if it goes over a threshhold AND remove the oldest if the
table of bad servers goes over a threshold size (to keep the list of bad
servers constrained)
Marc Langer wrote:
> Hi Philip,
>
> WishList items 85 and 109 seem to be similar. The possible solutions
> given so far:
>
> - remove a server from the list within a single Exim process when a
> connection to it fails
>
> - implement a retry database for database lookups
>
> Other solutions could be:
>
> - introduce an option "randomize_database_servers", sort of round robin
> as used in DNS
>
> - remember response times of database servers and use the fastest
> (should also be a special configure option)
>
> When using database lookups for delivering mails it is important to
> provide load sharing and failover methods, in my point of view.
>
> Thanks,
>
> Marc
>
> --
>
> ## List details at http://www.exim.org/mailman/listinfo/exim-users Exim details at http://www.exim.org/ ##
>
--
=================================================================
= Management is doing things right; leadership is doing the =
= right things. - Peter Drucker =
=_______________________________________________________________=
= http://www.sun.com/service/sunps/jdc/javacenter.pdf =
= www.sun.com | www.javasoft.com | http://wwws.sun.com/sunone =
=================================================================
