Yesterday I had a problem with sendmail. I have a dialup IP connection (PPP)
and sendmail will defer any mail sent while I'm not on-line. Then, when I
connect to the net, /etc/ppp/ip-up.local will run "/usr/sbin/sendmail -q" to
purge the queue.
But yesterday there was a problem with the ISP's nameserver. Sendmail put out
the request to resolve the hostname on the mail which was pending and it never
got a reply nor a timeout. Forgetting that I had that process running, I
disconnected from the ISP to try again later. Normally, if a transfer is in
process and I accidentally take the link down, it will timeout and just hold
the message for the next time the link comes up.
This time, however, there was a process which was trying to resolve the
hostname which was zombified. That kept any subsequently sent mail from even
getting properly into the queue ("mailq" would only give the message ID and
say "(no control file)"). I tried taking down the sendmail process running on
the system, but that wasn't the one which was problematic, so it had no
effect.
The only way I came up with was to reboot the system, which did work, but I
don't like that solution. Surely, there has to be a way to fix this problem
w/o a reboot to clean out the zombie...
Anyone have similar experience?
-Michael
--
"They that can give up essential liberty to obtain a little temporary
safety deserve neither liberty nor safety."
-- Benjamin Franklin, 1759
--
PLEASE read the Red Hat FAQ, Tips, Errata and the MAILING LIST ARCHIVES!
http://www.redhat.com/RedHat-FAQ /RedHat-Errata /RedHat-Tips /mailing-lists
To unsubscribe: mail [EMAIL PROTECTED] with
"unsubscribe" as the Subject.
