Grant,

Thanks for the response.  Recall that this was in response to my question about
sendmail dying with the following unusual error message, on a busy server:

Mar 25 07:07:58 firewall sendmail[12796]: NOQUEUE: SYSERR(root): opendaemonsocket: cannot bind: Address already in use
Mar 25 07:07:58 firewall sendmail[12796]: problem creating SMTP socket
Mar 25 07:07:58 firewall sendmail[12796]: NOQUEUE: SYSERR(root): opendaemonsocket: server SMTP socket wedged: exiting

Whoever heard of a wedged SMTP socket??

> How hard is the CPU running?  If it is not running very hard (less than 50% 
> or so on top) you may be able to exchange some additional CPU overhead for 
> not having to have the daemon running all of the time.  I run sendmail from 
> inetd, and although spawning the processes takes some CPU overhead, any 
> changes I make to settings are immediate (I do not have to restart 
> sendmail) and any intentionally nasty messages that try to crash sendmail 
> with buffer overflows do not work, instead, they just crash that one message.

I don't know what the CPU usage is exactly but I suspect it's greater than 1.0
during the busy times.

Actually sendmail used to run from inetd, which is the way I prefer it, but this
was quite disastrous compared to having sendmail run standalone as a daemon.
inetd refused connections a lot more than sendmail did, and I would have to
send a SIGHUP to inetd to get it to start accepting connections again.

It seems to be more efficient to run it as a daemon, for a busy server.

If anyone knows, I still would like to know what would make sendmail generate
that error...

--
Jim Ockers (ockers@ockers.net)                     Ask me about Linux!
Contact info: please see http://www.ockers.net/

Fight Spam! Join CAUCE (Coalition Against Unsolicited Commercial Email)
at http://www.cauce.org/ .