Mail Queue error

[David Crowe]

When I look into the mail queue using mailq command, I see a number of ": deferred>". Can anyone tell me what this actually means? Is this a DNS problem or a host name problem or an anti-relay problem?

[Steffi Jones_1]

Hi David,

you might have a message "stuck" and you can try the following steps to clear it up:

1) /sbin/init.d/sendmail stop
2) ps -ef | grep send
3) kill the process which is still holding the message
4) /sbin/init.d/sendmail start

Good luck,

Steffi Jones

[David Crowe]

I tried that and it didn't do anything. I'd really like to know what that error really means. When you check the mail.log, all the messages going to that particular domain seems to go directly into queue status.

[Unix Systems Admin]

This looks like a problem to do with DNS resolution. Your system is unable to resolve the MX for 123.com or the IP for 123.com. This can occur if you have a number of machines within a firewalled domain, as we have here, not allowing resolution of domains outside of local.

For my site I set up a main relay mail system which relays to the firewall mail system. All other UNIX servers then have a simple 'DSmailhost' entry in the sendmail.cf.

[Ossie de Jongh]

I agree with the others, but also want to ad something. I found its mainly caused by network problems (DNS's etc temporarely unavailable). Although the problem is temporarely, the queue doesnt retransmit as it is suppose to do. I got into the habbit of forcing delivery at least once per day by using "sendmail -q" and if I want to see what is happening "sendmail -q -v"

[David Crowe]

What's really strange is that I put in a static route in the mailertable to the domain that is giving me the most trouble and I still get the same error. Does sendmail do a MX lookup before sending email for a reason other than getting the destination IP? I am wondering about something to do with anti-relay?