Re: RADIUS accounting doesn't work properly

Jeff Woods ( jeff@delta.com )
Wed, 12 Mar 1997 12:38:14 -0500

At 11:34 AM 3/12/97 -0600, you wrote:

>We have racks running a variety of kernels (1.5.1d - 1.5.1f) and do not
>experience accounting problems. The only time we have a user that
>'hangs' is if we have cycled the rack and we get a duplicate key error
>the next time the rack happens to hit the same port/sessionid/type
>combination (seems like it s/b rare but it does happen).

OK. Well, we had a lot of trouble early on. Is it possible there's some
port/session/id combos that are out there frozen in limbo permanently, that
I'd need to remove to get them to work? How would I remove such entries if
they exist?

>We did have to add AcctInputPackets and AcctOutputPackets to the calls
>table once upon a time to get these fields but that was the only
>rack-specific thing we did.

What do you mean by this? I don't understand what you're saying we should
look for. It's the calls table I'm trying to get to work -- I don't see
anywher to manually add ANYTHING to it, much less what looks like a RADIUS
flag....
>>
>> I have an application called EMERALD which is RADIUS NT and and SQL
>> backender. In it, there is a table that should be able to reasonably
>> replicate what is in the WHODO output, based solely on RADIUS accounting
>> information.
>>
>> Recently, it has come to light on the Emerald mailing list (to which I have
>> cc'd this post) that many of your PowerRack clients using Emerald are
>> experiencing similar difficulties. One user wrote this on the mailing
list:
>>
>> >I have the same problems with my Computone
>> >Computone even told me that they think someone is going in and
>> >changing the racks setup. I don't think so but good try!
>>
>> I have determined that the accounting SEEMS to work properly, IF and ONLY
>> if you are telnetted into the Rack permanently, and do a WHODO once each
>> minute. My accounting was NOT working before, and is working now that I am
>> doing this perpetual telnetting.
>>
>> This shouldn't need to be the case. With so many of your users
>> complaining about this on the Emerald list, you probably ought to acquire a
>> copy of Emerald (they might even give you a comp copy good for a couple
>> dozen records if you ask nicely -- email daler@iea.com) and try to
>> replicate this and get it corrected in 1.5.1g you have forthcoming. The
>> NOPROMPT bug is the least of your worries at this time. Without
>> accounting, I'm dead in the water.
>>
>> Stevek, you have the password to get root access on our box. Please check
>> it out, and let me know what you think. If you have SQL server clients
>> installed there, I can give you a copy of the Emerald client to access our
>> database and see what I mean (for testing only, Dale).
>>
>> ----------------------------------------------------------
>> Emerald Mailing List listserver@emerald.iea.com