Re: USR Total Control

Brian Lube ( brian@mpinet.net )
Mon, 11 May 1998 09:19:52 -0400

At 08:30 PM 5/10/98 -0500, Michael Whisenant wrote:
>Questions regarding USR Total Control, HiPer ARCs, and HiPer DSPs; relating
>to their interaction with RadiusNT and Emerald. I have just added 2 230
>port configuration chassis into my main POP, the additional 460 ports are
>nice, yet I am running into a few strange problems.
>
>1. Has anyone gotten a mega chassis (say 10DSPs and 2 HiPerARCs) working
>comfortable with RadiusNT and Emerald?
>
Haven't tried with more than 6 DSP's at this point....

> 1.1 I have had to increase the lenght of the field for call ports to
>extend to 4 digits. The modems in slot 10 will report as 101 to 1024. If
>you have 11 DSPs then you would get multiple users logged to the same port,
>ie Slot 1 modem 11 is reported as 111 and Slot 11 modem 1 is also reported
>as 111. Then what?
>
The actual NASPort configuration is 256*<slot>+channel

> 1.2 I have stray accounting packets when I telnet into the HiPer Arcs that
>are constantly retransmitting. I remember Brian Lube had a solution, but
>it now escapes me. This excessive accounting packets should die after the
>max transmissions, yet they do not.
>
These were comming from the !root user. What I had to do to fix this was
to give a default value to the NASPort value in the calls table (I used
9999) so that the packets for the !root (or whatever you call your
administrative user) user are accepted by RadiusNT.

The only problem I have seen so far with the newest release is that the
HiPerARC is not releasing the admin user from the sessions table. I've got
over 10 'ghost' users in my list sessions that are all my admin user (and
they are not actually connected at the time). By default you should only
see two (the console and the one for when you are telnet'ed in). I have
been told that they are working on a fix for this, but it will not be until
the next full release of the HiPerARC (mid-late summer).

Other than that the only problem that I perceive is bringing the HiPerARC
down. If for any reason you want to work on that box you have to busy out
6-15 (depending on the number of DSP's) T1/PRI's. For us that means
bringing down a large portion of our pool. I was contemplating a dual ARC
solution, but this still brings down the DSPs that are being controlled by
the ARC you want to bring down. So far this is my only 'complaint' with
the newest version of the ARC/DSP. Hopefully this is something that they
will work to correct.

Brian Lube
senior technician
MPInet

http://www.mpinet.net