In message <[EMAIL PROTECTED]>, Scott writes : > George, > > I was recently monitoring some testing of my own radius server the other day. > > I'm using the CVS release (not a daily snapshot) from 2004-12-15 on RedHat > 7.2. While briefly testing my setup, I noticed there was a reject delay for > failed authentications. I didn't time anything, but your comment about two > seconds plus the config paramater sounds accurate. However, I have not > experienced any excessive delays like the 31 second ballpark you mention.
The 31-second delay wasn't obvious, because most RADIUS clients will retry after a couple of seconds. A packet trace will reveal that no reply is sent to the first request, but the reject reply is sent immediately after the retry. Using 'radclient' with a really long timeout revealed the 31-second delay. > If you're not making any progress with this, I'd suggest investigating the > source code, to see where this is handled. I would then, either enable extra > > debugging (if logging is already present for this section), or add some debug > > messages of your own (including the current time+millisecond timestamp). I've sidestepped the immediate problem by turning off the reject_delay, but I'll try to dig through the source if I can find time. (I'm wary of debugging modes, as I've already seen that -X hides the problem). -- George C. Kaplan [EMAIL PROTECTED] Communication & Network Services 510-643-0496 University of California at Berkeley - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
