Here's the over-riding symptom set:
1. Everything was fine until we switched connectivity providers over the
weekend.
2. Mail seems to be arriving normally at the server.
3. Mail arriving for a username expanded from an alias in /etc/aliases is
delivered to the appropriate mailbox (e.g., users who have married and
changed their names since joining the company, users whose name, according
to our firstinitial+lastname convention, is longer than 8 characters).
However, it is not *marked* as delivered, and it is delivered again and
again, every twenty minutes or so.
4. Mail arriving for a username that is *not* expanded from an alias is
marked queued, and not delivered.
The particulars.
Running sendmail 8.9.3 on a freeBSD 2.2.2 system. (I know, I know -- but
since I *know* I'm in over my head, I've been reluctant to try and
implement major changes on a system I don't understand well.)
Our connectivity situation is a little complicated, and I'm afraid it may
be relevant.
Our DNS is hosted by Provider A. Until this week, Provider A had an MX
Record which pointed our mail services to Provider B. We had a leased
line from Provider B. The mail server was on our gateway box, doing
network address translation and IP filtering, connected directly to a
router leased from Provider B, with a single externally-visible IP.
We now have a new leased-line from Provider C. We connect to Provider C
through a Virtual Private Network, with a (Raptor) firewall on Provider
C's end. We lease from them a router and a VSU. The VSU has an
externally-visible IP address; the system I administer (or try to) no
longer has an externally visible IP address. The server is still
providing NAT, but is no longer doing IP filtering (or rather, it's
passing all in/out).
The VSU was shipped configured to talk to our server via subnet
192.168.0.*. (One of the reserved blocks of IPs perRFC 1198). We were
already using 192.168.0.* as the interface for our private network, so I
had to move our private network to subnet 192.168.1.* -- I don't see how
this could be affecting sendmail, but, whatever.
I didn't make any changes to sendmail.cf or any mail configuration file,
other than to change the IP address for pop & smtp to the new subnet. I
don't think that's likely to be the problem, since our users are able to
send mail normally, and those users who are getting their mail delivered
are able to get their pop mail normally. But then I don't know what
*would* be likely to be the problem, since I've never seen anything like
this before.
Anybody ever see anything like this? Anybody have any thoughts?
Please feel free to respond directly to my personal address, d...@mwmw.com.
--
- oh no, you've just read mail from doug = d...@radix.net - get yr pathos
- www.pathetic-caverns.com -- books, flicks, tunes, etc. = reviews
- www.fecklessbeast.com -- angst, guilt, fear, betrayal! = guitar pop
It may provide a crucial clue.
--
Andrzej (Andrew) A. Filip mailto:an...@bigfoot.com
Warsaw, Poland
Feel free to correct the posting if YOU think it is required.
Check your sendmail logfiles-- if the mail has been turned over to the LDA,
it's probably not sendmail which has the problem.
Check the ownership of the mbox files in your mail spool (probably,
/var/spool/mail), and make sure that each file is owned by the right user. If
this is wrong, it's possible that sendmail and/or your local mailer is no
longer setuid.
> Please feel free to respond directly to my personal address, d...@mwmw.com.
Nope. Read the newsgroups you post to.
-Chuck
Chuck Swiger | ch...@codefab.com | All your packets are belong to us.
-------------+-------------------+-----------------------------------
She said, you've taken me for granted because I please you. -P Simon