> > Our radiusd (Merit 2.4.16, running via inetd) seems to be aborting
> > frequently with an exit status (as reported by inetd) of 0xb...according
> > to the manpage, this is nonsensical...is inetd munging the exit code; if
> > so, what error is actually being reported? Nothing is logged to
> > .../raddb/logfile or syslog.
> I was not aware that inetd(8) presented exit codes in hexadecimal (I seem
> to remember it being decimal exit codes) so this is one source of confusion.
>
> If the reported (are you sure?) 0xb exit code is from Merit RADIUS (which
> seems odd since they are all negative -- don't ask...) it may be that some
> one has taken to converting the -11 (socket BIND error) code to +11 (0xb)
> but that seems very unlikely. Although you could check to see if some one
> else is holding the 1645 UDP port when you try to start the Merit server.
> What is the platform? What is the operating system? Is this inetd(8) or
> xinetd(8)? Is the Merit server statically linked or dynamically linked?
<smacks self> I managed to misread "exit signal 0xb" as being an exit
code...seems it is segfaulting after all. This is inetd(8), Linux
1.2.13/i586, dynamically linked. I checked the manpage for exit codes,
and pondered that it might be an error binding to the socket...but since
it's running from inetd (and processing authentication requests with no
(apparent) problem), it seemed pretty unlikely.
It came from an unmodified 2.4.16 distribution, so I haven't a clue why
it would segfault...I'm compiling 2.4.20 as I type; I'll see if that
works out better...
// Matt Zimmerman Chief of System Management NetRail, Inc.
// mdz@netrail.net sales@netrail.net
// (703) 524-4800 [voice] (703) 524-4802 [data] (703) 534-5033 [fax]