[ On Thursday, July 19, 2001 at 12:03:58 (+0100), Mark S. Guz wrote: ]
> Subject: Re: [Exim] Re: RBL Domains
>
> Given that exim becomes unusable when it can't reach one of the servers
> in its rbl domain list
> I do think that a failure message would help diagnose problems quickly.
You really really cannot do any such thing. The the process of
resolving DNS names is much more complicated and involved process.
If Philip makes Exim complain when DNS_AGAIN or DNS_FAIL re returned
then your logs will fill up with totally misleading and usless junk
messages. At any given time there are a very very large number of
nameservers that'll trigger such a problem. Given the way the resolver
chooses which nameserver to query you've got between a 100% and 1/N
chance of trying to contact a "duff" nameserver (where N is the number
of nameservers for a given domain that has at least one "duff"
nameserver).
You do not want that.
The real question is what you mean when you say "exim becomes unusable".
What do you mean?
Did Exim simply stop handling incoming mail efficiently? I.e. did each
incoming connection take many seconds to do anything?
If so then maybe it would be sensible to log a warning about DNS issues
IFF a DNS_AGAIN or DNS_FAIL error is encountered on an RBL lookup, and
IFF there was a significant delay between the time the query was issued
and the error was returned. That's about as far as it's sane to go....
Generally though when a problem like this occurs with my mailer I simply
try connecting to it manually from afar, turn on full debugging traces,
and see where it "hangs".....
--
Greg A. Woods
+1 416 218-0098 VE3TCP <gwoods@???> <woods@???>
Planix, Inc. <woods@???>; Secrets of the Weird <woods@???>
