From: "Jev" <[EMAIL PROTECTED]>
> Anyway, my problem :)
> We are running a web-mail application that was written in-house. There is
> some horrible JS that refreshes in the users browsers every 60 seconds
> (in a small frame on the page) that check mail, with a few people logged
> onto the site the imap servers gets a lot of different checking
> mail.
>
> This would eventually get too much and I get an entry in
> /var/log/messages from inetd reading:
>
> inetd[430]: imap/tcp server failing (looping), service terminated
>
>
> So, is this a known issue? I have seen reference that others have used
> this same version under heavy loads. What should I do to track down and
> locate the problem? What other information can I provide?
>
Yes, this is a known issue with any service that uses inetd. The problem is
that too many connections are requested of that service in a 60 second
interval. You need to specify the maximum connections/second after
wait/nowait in inetd.conf in order to solve this problem.
imap stream tcp nowait.200 cyrus /usr/sbin/tcpd
/usr/local/cyrus/bin/imapd
Scot
