I have a couple CGI scripts that send confirmation emails using sendmail. overnight, the messages stopped sending. They seem to be stuck in the /var/spool/clientmqueue directory. I haven't changed any settings and cannot figure out the problem. I did try to send an email from a shell prompt and that didn't work either.

My internet connection seems to be up, as I can access the web server (apache). I can ftp and ssh. nslookups work so I believe my DNS is functional.

Any ideas on where to look to debug this problem?

/var/log/maillog

Sometimes if there has been a network hiccup sendmail may miss a mesage then everything queues up behind it. Just restarting sendmail may cause it to begin sending again:

Do a "tail -f /var/log/maillog" in one screen and in another do
"service sendmail restart"

This should either start sendmail and begin sending messages (you'll see them as queued in the log) or it will give you errors which hopefully will tell you why it isn't queueing.

Its always possible that you've been blacklisted by spamfilters if you didn't restrict relaying and some spammer was using your system for spam.

Thank you.. I will try that. If by chance I have been blacklisted.....any solution aside from getting a new IP?

You can actually contact the people that do the blacklisting and request they remove you. I haven't needed to do that (knock on wood) but I know it can be done. A Google search will likely tell you who to contact.

Prior to doing that though you'd want to make sure you had fixed whatever let the spammer use you or they'll do it again. I'm sure there's some threshold at which the blacklisting folks will quit removing you. Also contacting those people only removes you from their spam lists. If any organization does their own blacklist instead of subscribing to the common ones you'd have to contact them. That would only be necessary if you found it was someone to whom you need to send email. There'd be no reason to bother if it was just someone the spammer had hit using your server.

A quick Google search led to the following page for one of these:
http://dsbl.org/faq-listed

No luck...the maillog shows the messages as "stat=queued". The restart updated the maillog with an alias rebuild message and daemon starting, but no error messages.

I am using SMART_HOST option to forward to my isp's smtp. Not sure if they changed any settings on their end.

Any other ideas on how to fix this?

Nothing in the log complained about any of the messages?

It may be your ISP is denying relating.

Do:
telnet <isp's mailhost> 25

Where isp's mailhost is the name or IP of the host you are doing your smtp to?

The above will open the smtp port (port 25) on the mail host.

The link here discusses how to test.
http://72.14.209.104/search?q=cache:...ient=firefox-a

Essentially you can send a mail with this direct connection instead of going through your sendmail config.

At the end of your telnet session hit Ctrl-[ then you'll see a telnet prompt then type "close" and hit return.

Thank you for that link. It will help in the future.

My problem has been solved. How? I don't know. There where no errors in the maillong just messages going into the queue. I could access the smtp domain through telnet using the ip address, but would get name resolution failures when trying to use the domain name.

In short, I believe my ISP was having DNS issues, but if they weren't I have no idea how the problem was solved. Thanks for the help.