I have exim4 (exim4-daemon-heavy) running on etch. This machine has only a few users who send directly. One of them has an ISP who seems to have a broken DNS setup.
The symptom I see is that when this user connects to the server (port 25 or 587) there is a large delay before the 220 banner is shown e.g.: 220 bryanek.chrissearle.org ESMTP Exim 4.63 Tue, 17 Jun 2008 17:25:02 +0200 This can be a delay of between 20 to 60 secs and is at times enough for the sending client to time out before sending either HELO or EHLO. I have narrowed this down to the fact that exim (as far as I can tell) is doing first a reverse lookup from the IP the user is connecting from followed by a forward lookup for the resulting hostname - and this lookup fails - no server found. This is the delay that is then causing the 220 banner not to show in a timely fashion. I've tried setting exim4's host_lookup_order=byaddr:bydns (by default it is set to dns first) and adding the host to /etc/hosts - but this didn't help. What I would like best is for the ISP to fix it. If not - then I would like preferably for a given IP range or if not possible then for all users not to perform this lookup prior to sending the 220 header - but I can't seem to get the correct exim4 config for skipping this check in particular. Is it possible to whitelist a range or turn off the check? -- Chris Searle -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
