SUMMARY: sendmail.mx fails with in.named, 4.1.3C on Classic

From: Joel L. Seber ... CH210 (JLS2013@tntech.edu)
Date: Fri May 06 1994 - 16:45:20 CDT


My original query (sorry this one took so long):

>System setup: Sun SparcClassic, 24MB, mailhost, NIS server configured with
> B=-b for DNS, SUNOS 4.1.3C; NFS, UDP, LOCKD patches applied;
> running a cacheing-only DNS server on my NIS master
>
>PROBLEM: Sendmail.mx refuses to function correctly with in.named running
>
>I have recently been trying to make mail work outside my NIS domain using
>DNS. I have RTFM and followed to the letter the procedure for setting up
>DNS as described in the manual and the O'Reilly DNS & BIND book. Whenever I
>rename sendmail.mx as sendmail, kill and restart the sendmail daemon on the
>NIS server and attempt to mail locally or remotely, nothing happens. No
>error message, nothing. ps -aux says in.named, sendmail -bd -q15m, etc. is
>running, but no mail. If I replace the original sendmail from a saved copy,
>local mail works again and I can mail to other machines outside my domain,
>but not to MX-record based mail destinations (I do realize that sendmail.mx
>is required for that).
>
>I did in fact copy /usr/lib/sendmail.main.cf on my NIS server, and
>/usr/lib/sendmail.subsidiary.cf on my clients. I changed the NIS copy very
>little, putting in our domain name and a major relay host (a campus VAX
>system on Internet) as suggested by the manual and by my Sun rep.
>I also have /etc/{resolv.conf,named.boot,named.ca} installed on my NIS
>server. As noted above, I am only a cacheing server with no authority
>delegated to my domain.
>
>(one side note that may/may not matter: nslookup does NOT function for me
>directly; I have to do 'nslookup -l <a local name server>' to get it to work
>at all. I downloaded the 'dig' program from internet and it works fine
>locally. I add this because it may be symptomatic of a DNS failure, although
>clients such as ftp, telnet, etc. seem to be able to call the resolver and get
>their info OK...)
>
>I have scanned the sun-managers summaries through WAIS.QUAKE.COM, and
>contacted my local Sun technical support person. He sent me some examples
>from a site that seems to work for them. I will try them. In the meantime,
>I would really like to hear from anyone who had had to deal with this and
>resolved their problems. If I have done something stupid, I need to know.
>I would really like to summarize this for others going through this, too.
>ANY SUGGESTIONS ARE WELCOME, NO MATTER HOW TRIVIAL THEY MAY SEEM! Example
>files are also most welcome. If you send an example file, please tell me if
>you do NOT want it included in a summary if it helps solve the problem!
>
>If I have not included enough information here for a diagnosis, let me know
>and I will send what you need.
>
>Thanks in advance,
>
>Joel L. Seber | Dry humor is wasted around here.
>SUN Workstation Laboratory Manager |
>Center for Manufacturing Research | -Joel L. Seber
> and Technology Utilization |
>Tennessee Technological University | recursive, adj.
>Cookeville, TN 38505 | See 'recursive'
> |
>jls2013@tntech.edu |
>

Well, folks, I have to admit that it was my own less-than-perfect
understanding of DNS that was the culprit with most of this problem. It
seems that I had mis-configured my named.boot file, so sendmail.mx had
bad info on my cacheing servers' setup. As for 'nslookup', my setup is such
that I had to point to our campus nameserver as our primary in
'resolv.conf', and it works fine now, too.

Thanks especially to Rick Weisner of Sun for sending sample files, as well
as some suggestions that pointed me in the right direction.

Several others sent interesting suggestions that I will try to one- or
two-line summarize here also:

Make sure you have "B=-b" set in /var/yp/Makefile, touch "hosts.time", and
run 'make'.

Try telnet to port 25 and see if it still works; if so, try 'sendmail -bt'
and a local address.

Check /etc/nsswitch.conf and add/modify it with a line that reads:
       hosts: nis dns [NOTFOUND=return] files

Watch wildcards in DNS records - at least one had problems with this

and a couple of long posts that I won't include here. If anyone wants the
entire text of all the replies, I still have it and can forward it.

Thanks much to the following:

rick.weisner@east.sun.com Rick Weisner
davec@cs.ust.hk Dave Curado
crane@cadlab.ece.drexel.edu Richard Crane
blymn@mulga.awadi.com.AU Brett Lymn
poc@usb.ve Patrick O'Callaghan
feldt@phyast.nhn.uoknor.edu Andy Feldt
andrew@sdsu.edu Andrew Scherpbier
julik@borlas.com Resident Hacker

Joel L. Seber | Dry humor is wasted around here.
SUN Workstation Laboratory Manager | -Joel L. Seber
Center for Manufacturing Research |
      and Technology Utilization | recursive, adj.
Tennessee Technological University | See 'recursive'
Cookeville, TN 38505 |
                                        | What if there were no hypothetical
jls2013@tntech.edu | situations? -Steven Wright



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:09:00 CDT