Kristina Pfaff-Harris <[EMAIL PROTECTED]> wrote: > A little more info. (I haven't removed those kill lines from session.c > yet.) I had put in a couple of debug statements to tell me how checkrad > is called, and which thing it is calling (finger, snmp, telnet, etc).
What's more important is which PID was used for checkrad. That will tell you if the TERM signal is going to the right place. > When the one server crashed, the NAS sent the request to the alternate > server, and it crashed that one as well. Maybe the 16-character username? > Or is that a coincidence? It's a coincidence. I'm still amazed that sending a TERM signal to a child PID which is *not* radiusd will cause radiusd to intercept the signal, and die. > I see where it does the "killing all active servers" thing in radiusd.c, > but am not sure where to look to find out exactly what "bad thing" the > child did to cause this. Find out which PID is used for checkrad. If the server is sending a TERM signal to another PID, then that's a bug. If the server IS sending the TERM signal to checkrad, but a radiusd process is getting it, that's incredibly bad. Alan DeKok. - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
