hello everyone,

I have been monitoring the mqueue (mailq) on my sendmail server for
the past couple of hours and I am seeing some strange issues.

If I run the mailq command I see that some domains receive a host
map: lookup (domain.com) defferred message. see below.

k1DE2Hbm032097- 13244 Mon Feb 13 09:02 <Heather.Dalton [at] phh.com>
(host map: lookup (mortgagefamily.com): deferred)

<sue.sunstein [at] mortgagefamily.com>

The strange part is if I run the queue Id it goes thru fine:

sendmail -v -q Ik1DE2Hbm032097

It seems sendmail doesn't have a resolve issue. If I run:

echo '/map host mortgagefamily.com' | sendmail -bt -d8.16_res.options =
12c7, HasWildcardMX = 0
ADDRESS TEST MODE (ruleset 3 NOT automatically invoked)
Enter <ruleset> <address>
> map_lookup: host (mortgagefamily.com) dns_getcanonname(mortgagefamily.com, trymx=1)
dns_getcanonname: trying mortgagefamily.com. (A)
YES
dns_getcanonname: mortgagefamily.com
returns mortgagefamily.com. (0)

everything looks ok.

nslookup -query=mx mortgagefamily.com

returns that it's ok.


It's not just this domain. But random domains.

I have been looking thru the newsgroups and faq, but I don't see
anything that resembles what I am seeing on my server.

Any suggestions?

Thanks,
sunckell

If anyone is remotely interested. I believe I solve my problem...


Problem:
1.spuratic resolution failures with domains that are properly
configured in DNS.
a. in mailq I would see (host map: lookup (domain.com) :
defferred)
b. I could run the mesgid manually and it would go through.
- ie. sendmail -v -q I$mesgid
2. Resolv.conf is correctly configured.
3. O ResolverOptions=WorkAroundBrokenAAAA did not fix the issue.

Solution:
1. In my scenerio, it's a Red Hat ES 3 server with a custom built
sendmail 8.13.3 binary.
2. The DNS servers reside on a different server.
3. RHN updated : bind, bind-libs, bind-utils
4. I found that the sendmail binary needs to be built with the
same version of resolver libraries that are on the server.

I rebuilt/install a new sendmail binary, built with the proper
resolver libraries

and so far so good...


Hope this helps someone...

Chad