Problem sending mail after IP/hostname change

UNIX Solaris
1

Hello,
I had to replace my Solaris 10 workstation with the other one because of hardware issues. I need to configure the new WS with the same configuration. It looked like it was using DHCP before and I was able to send mail with mailx command. Since old HW used static IP I have disabled DHCP and modified
/etc/inet/hosts
/etc/inet/ipnodes
/etc/nsswitch.conf
/etc/resolv.conf
/etc/defaultrouter
/etc/nodename
/etc/hostname.eri0

to handle the new configuration of a static IP and now I'm getting mail delivery subsystem warnings: Deferred: Connection timed out with foo.com

Could anybody give me a hint how to troubleshoot this? Obviously It was working before and problem is not in foo.com :frowning: when I do nslookup foo.com (foo is not my real domain) it comes back with IP address. Any ideas?

Thanks,
PN

2

Check the connectivity from the workstation to your mailserver.

ping mailserver.foo.com
telnet mailserver.foo.com 25
3

"when I do nslookup foo.com (foo is not my real domain) it comes back with IP address"
Does the reverse work? whats the output of sendmail -q -v

4

Thanks, you have a good point. I did check ping foo.com and telnet foo.com 25 timed out BUT if I do telnet mail.foo.com 25 I get response! Is there any configuration file where mail server prefix is configured?

Thanks,
PN

5

Hmm.. Sorry, I'm not sure that I understood you correctly. What does reverse mean in this case?

That probably times out. The only output I got was:

WARNING: local host name (wayossws55) is not qualified; see cf/README: WHO AM I?

What should I bee looking for?

Thank You,
Pavel.

6

The quick (wrong) way:
Edit /etc/mail/sendmail.cf and replace the line that starts with "DS" with "DSmail.foo.com". Then restart sendmail.

The good way:
You'll have to edit the .mc file and rebuild the .cf file.
Configure sendmail as a smart host

Alternatively, I think you could just set the MX record of your domain to the correct mailserver. (But this would only allow you to send to @foo.com addresses).

7

Thank you! I tried the good way and the generated sendmail.cf file was much smaller then the original Solaris one so I went for a "wrong" one. The good news is - it seems to improve the situation, I'm not getting any more timeouts. The bad news is I'm getting the following report:

... while talking to mail.foo.com.:
>>> HELO wayossws55.
<<< 501 5.5.4 Invalid Address
554 5.0.0 Service unavailable

It looks like mailserver does not like my WS name. Is there any way to trace a good scenario? I have another WS that seems to be sending the e-mails ok.

Regards,
PN

8

Can you compare /etc/hosts on both WS?
Also, compare the sendmail configs...

9

Thank you for your help! It is now working! The problem was located in the /etc/hosts. While trying to eliminate "unable to qualify my own domain name" error messages I have changed record in /etc/hosts for my WS IP. For whatever reason this change cause problems just now while it was actually done some time ago.
Thanks again!
PN.