[quote user="Peter Strömblad"]If Mercury can't locate the user accounts, it can't decide whether or not to accept inbound mail. Maybe it shouldn't crash as you say, just say service unavailable on all ports.[/quote]

That's all I'm asking for.

In our configuration we have Mercury setup to authenticate POP against a Novell Netware 6.5 server.  Our user's mailboxes reside on this same server, also, we have a 'secondary' mail queue on this server which is where Pmail puts it's local outgoing mail.  The primary mail queue resides locally on the Mercury server.

The problem is that when the Novell server crashes or becomes unstable (crap!) it causes Mercury to crash too.  It would be nice if Mercury could fail gracefully and allow incoming SMTP connections to continue and resume proper services when Netware recovers.

If Mercury can't locate the user accounts, it can't decide whether or not to accept inbound mail. Maybe it shouldn't crash as you say, just say service unavailable on all ports.

The problem is that when the Novell server crashes or becomes unstable

(crap!) it causes Mercury to crash too.  It would be nice if Mercury

could fail gracefully and allow incoming SMTP connections to continue

and resume proper services when Netware recovers.

I'm using a really old v3.2 Netware host with everything  in Mercury/32 except the mailboxes on the Windows system and my Mercury/32 does not crash when it loses the connection to the Netware host. 

That said, the Netware host is like a rock!  It never crashes, it's been running for years without problem.  The only time I've ever had it go down is when the power goes out longer than the UPS battery can handle.  I do have a daemon (Status Petr Jaklin) running that shuts down MercuryS though if I ever lose the connection to the Netware host so that I do not inadvertently lose mail.  The mail goes to an off site backup MX host.

Program Status is daemon module for Mercury/32.  http://www.3net.cz/software

It can test specified functions of network and server with Mercury.

When any problem is detected, daemon sends warning email to the configured

addresses (it is possible to create "shortened" version of message, e.g. for mobile phones).

Daemon configuration can be changed remotely.

It tests:
1. Computers availability by ICMP Echo protocol (PING).
2. Mercury server memory availability.
3. Free disk space on specified drives.
4. Availability of these drives (it suspends receiving of mails if disk would be lost)