Note that if you set up sendmail on a local box, make sure it either handballs stuff off to a smarthost straight away, or you have stuff set up to check for sendmail getting upset and just queuing messages for ever. Normally you'd notice breakages on your mail server - but your web server might be a different matter entirely.

I've already dropped chris a mail about the issue. We run exim on our main mailservers with the configuration such that it verifies envelope addresses as much as it can (checks the domain exists for remote addresses). This is likely to cause speed issues, and for mass mailing or quick mailing from internal hosts it's better to have a different copy set up to handle the mail (it can even be on the same server, just on a different port with a different config, assuming the mail send class being used allows you to specify an alternate port). There are other config parameters to tweak if large amounts of mail are going to churn through it to optimise for speed and parallel delivery (if you queue mail and then send every 15 minutes, you can take advantage of the fact exim will batch mails if it detects they can go down the same SMTP connection). Where it's a forking mail server, if you are going to hit it lots, there's options to configure the limit on accepted connections/forked copies to stop the mail server grinding to a halt. Matt