Did you build a new klernel first - you should _always_ build a
	new kernel before a new userland (modulus occasional new versions
	of make, config etc, but those should be built with the old
	includes, then the new kernel built & booted, then the userland
	built with 'make build')

		David/abs		david@mono.org

.---- I've been too drunk to love ----.-- I've been too drunk to remember -.
|          too drunk to care          |    the hell of the night before    |
|  looked like death, felt like hell  |   I've been drinking myself blind  |
`------ been the worse for wear ------'-- and still I'll drink some more --'

On Mon, 17 Mar 1997, William O Ferry wrote:

>     Just yesterday I did a make build on my /usr/src, sup'ed a few
> minutes beforehand.  Everything went pretty smoothly, but after
> rebooting with the new binaries installed, any inetd connections failed,
> and I started getting the following syslogs:
> 
> Mar 16 14:49:41 WarpDrive inetd[117]: select: Bad file descriptor
> Mar 16 14:50:12 WarpDrive last message repeated 31 times
> Mar 16 14:52:13 WarpDrive last message repeated 120 times
> Mar 16 15:02:14 WarpDrive last message repeated 599 times
> Mar 16 15:12:15 WarpDrive last message repeated 599 times
> Mar 16 15:22:15 WarpDrive last message repeated 598 times
> Mar 16 15:32:16 WarpDrive last message repeated 598 times
> 
> I reinstalled the 1.2 inetd, and this problem went away.  At first I
> figured it might be a problem that my inetd.conf was still using tcpd. 
> But I removed all occurrances of tcpd in the file and set it back to
> "normal", and I still get these syslogs.  All incoming connections to
> the machine just hang when this happens.
> 
>     The problem doesn't occur immediately on boot, it sometimes takes up
> to an hour before it starts complaining.  Traffic to inetd via localhost
> seems to aggrivate the problem.  I also occasionally got the syslog:
> 
> Mar 16 16:33:58 WarpDrive inetd[621]: warning: can't get client address:
> Socket is not connected
> 
> with localhost connections.
> 
>     Has anybody else seen problems like this?  The new inetd doesn't
> want to compile without -DLIBWRAP, so I can't see if it's in that part
> of the code or not without doing some more hacking.
> 
>     I'd appreciate any help / suggestions, thanks in advance.
> 
>                                                           Will Ferry
> 
> --------------------------------------------------------------------------
> William O Ferry <woferry@CMU.EDU> | finger:  woferry@WarpDrive.RES.CMU.EDU
> talk:  finger for online status   | http://warpdrive.res.cmu.edu/~woferry/
> --------------------------------------------------------------------------
>