Hey Scott,
inetd gets upset with life if you make too many connections to it
too fast. It defaults to 40 connections per minutes - after that it
decides it is looping and stops responding to that service. The idea
(presumably) being to sacrfice one service to protect the rest.
The simplest solution is just to up the limit in /etc/inetd.conf:
shell stream tcp nowait.120 root /usr/sbin/tcpd in.rshd
^^^^- a big number. see inetd(8) for more.
Assuming this really is your problem I'd also be inclined to take another
look at your script. If your script is spawning enough rsh's to loop
inetd then it prolly has problems of it's own. ;)
The other thing you might want to think about is xinetd - an inetd
replacement. But there's no real reason to go replacing major system
components...
M.
On Thu, 28 Oct 1999, Scott L. McCracken wrote:
> We just upgraded to Red Hat Linux v6.0 Kernel 2.2.5-15. We are running a
> shell script using rsh from another UNIX platform, and we are getting a
> "Connection refused" error during the processing. The problem doesn't
> happen in a consistent manner. Sometimes we will get through half of the
> script, sometimes more other times less. We can't start up the script
> again until we stop and re-start the inetd daemon. The error message in
> /var/log/messages is "inetd[368]: shell/tcp server failing(looping),
> service terminated".
>
> Any assistance would be appreciated,
> Scott
