Hello,
We're running Radiator 2.16.1 under BSD/OS 4.1 and I've had some
problems with Radiator starting out of cron. I'll explain why we're
starting Radiator out of cron in a bit, but first the details of the
behavior.
When starting out of cron, Radiator fails with the following error:
Bad arg length for Socket::unpack_sockaddr_in, length is 0,
should be 16 at /usr/local/bin/radiusd line 322.
The relevant snippet of code is:
314 # Find out if we are running under inetd.
315 # If that is so, then stdin is a socket, adn we wont need
316 # to open a socket for that port
317 my ($stdin_sockaddr, $stdin_port, $stdin_addr);
318 $stdin_sockaddr = getsockname(STDIN);
319 ($stdin_port, $stdin_addr)
320 = Socket::unpack_sockaddr_in($stdin_sockaddr)
321 if defined $stdin_sockaddr;
322
I found that starting radiator with stdin redirected from /dev/null
is an effective workaround for the problem.
I mention this in the hopes that either this behavior can either be
modified or documented in the next release...
As for why I'm starting radiator out of cron, we are authentication
and logging out of a database. I run the database maintenance scripts
out of cron. Initially, I did not shut down the radius server when I
was doing this maintenance and auth/logging requests were still sent
to the radius server (with poor results). I decided that the best
solution would be to shutdown the radius server before doing database
maintenance and then restarting it afterward. This is when I
discovered the above problem.
--karl
-------------------------------------------------------------------------
Karl Friesen Sr. Systems/Network Administrator
316-263-7963 Southwind Internet Access, Inc.
a OneMain.com company
-------------------------------------------------------------------------
===
Archive at http://www.starport.net/~radiator/
Announcements on [EMAIL PROTECTED]
To unsubscribe, email '[EMAIL PROTECTED]' with
'unsubscribe radiator' in the body of the message.
