I have been trying to run a list with Auto-Delete on Manual and no renewal
policy. This was because I didn't trust automatic deletion: when we changed
to LISTSERV last year, those who trusted the staff to move their lists got a
configuration that allowed spam, and instantly deleted anyone who rejected
it.

We're now using 14.3. My list has accumulated a bunch of suspect NOMAIL
subscriptions, added so people can send to list from a different address
than the one at which they receive list email. (We're an edited list, but
everyone but new, untrusted subscribers is an editor.) I've also got a
running problem with remapped subscriptions: I get error reports but cannot
match them to subscribed addresses.

I am hoping to use both active and passive probing to identify the bad
addresses. The manual tells me that I can't get passive probes (to find
those remapped subscribers) until I make them subject to renewal. I don't
want to risk that until I understand renewal and prove it works. My test
setup has Renewal= 2-daily,Delay(1),Probe but everyone on the list is set to
NORENEW except my own subscriptions, so I get retests every couple of days
and nobody else will be bothered until I change that to a realistic value
and make them renewable.

The problem is that one of my test subscriptions is an old, abandoned email
address. LISTSERV reports that it's being probed, and I know email to that
address will bounce, yet my daily error log does not report the bad address,
and nothing suggests it will ever be deleted. (I switched Auto-Delete to
Semi-Auto because the manual suggests I might get better error reports.) I
don't know whether there would be a different report if a Probe 2 was sent,
but I haven't seen one and the timing suggests it's not happening.

I tried sending a reply to a probe for one of my working addresses, in hopes
of triggering a bounce report, but that just got handled as a bad command
email to LISTSERV (no commands recognized). Is that because a reply doesn't
look enough like a bounce, or because it goes to the wrong address?

How is an active probe supposed to detect a bad address? Are there any
addressing considerations I should be checking?

Hal Keen