- Date: Fri, 12 Mar 1999 12:04:38 +0000 (GMT)
On Fri, 12 Mar 1999, Dave Wilson wrote:
> While a five minute delay on a single mail is usually acceptable, several
er, the five minutes I quote is the propagation time through the machine.
> five minute delays on every post to a mailing list really starts to hurt
> those list managers that work their way through a list sequentially (and
> I can only speculate what it does to those that insist on simultaneously
> forking a process for every recipient).
ah, you mean the time taken for the sender to time-out when trying to talk to
the customer's machine; I guess if you're running a list expander, you might
tweak your TCP connection-attempt timeouts to avoid this?
> It's a good solution, but this LISTSERV maintainer for one would be a lot
> happier if the reachable MTA were the primary MX host, and the MTA was either
> configured to forward to the customer, or run off a different zone file.
Good point. However, this might mean getting our customers to change their end
slightly, and if they can't fix their mail machine, chances are their clue
level isn't high enough to add additional mail domains for receiving mail.
# P Mansfield, Senior SysAdmin PSINet UK Ltd, +44-1223-577611 fax:~577600 #
# You can wait for a bus-error all day and then three come along all at once #
# registered linux user 112301 - http://counter.li.org #