On Mon, 5 Sep 2005, Nigel Metheringham wrote:
>
> Personally I would tend to do this with a shared mailstore (maildirs or
> spool of some form) but per-instance queue/hints/logs/odds directories.
That's what we do, and the machines are generally reliable enough for our
purposes. However because we depend on the DNS to spread the load, a
failure is quite visible to users. This is the main cause of trouble; we
almost never lose data from failure of a mail queueing machine because
they are RAIDed etc. Some kind of lower-level load balancing (LVS, BigIP,
etc.) can fix the failure-visibility problem but that does not imply you
have to make the whole system tightly coupled.
Tony.
--
<fanf@???> <dot@???>
http://dotat.at/ ${sg{\N${sg{\
N\}{([^N]*)(.)(.)(.*)}{\$1\$3\$2\$1\$3\n\$2\$3\$4\$3\n\$3\$2\$4}}\
\N}{([^N]*)(.)(.)(.*)}{\$1\$3\$2\$1\$3\n\$2\$3\$4\$3\n\$3\$2\$4}}
