Checking Radius user test:
Sent: 2 Good: 1 Bad: 0 T/O:0 Out: 1 Avg: 1234
Sent: 3 Good: 2 Bad: 0 T/O:0 Out: 1 Avg: 1320
Sent: 4 Good: 3 Bad: 0 T/O:0 Out: 1 Avg: 1447
Sent: 5 Good: 4 Bad: 0 T/O:0 Out: 1 Avg: 1566
Sent: 7 Good: 5 Bad: 0 T/O:0 Out: 2 Avg: 1671
Sent: 8 Good: 6 Bad: 0 T/O:0 Out: 2 Avg: 1778
Sent: 9 Good: 7 Bad: 0 T/O:0 Out: 2 Avg: 1870
Sent: 10 Good: 8 Bad: 0 T/O:0 Out: 2 Avg: 1956
Sent: 11 Good: 9 Bad: 0 T/O:0 Out: 2 Avg: 2027
Sent: 12 Good: 11 Bad: 0 T/O:0 Out: 1 Avg: 2119
Sent: 13 Good: 12 Bad: 0 T/O:0 Out: 1 Avg: 2140
Sent: 14 Good: 13 Bad: 0 T/O:0 Out: 1 Avg: 2142
Sent: 15 Good: 14 Bad: 0 T/O:0 Out: 1 Avg: 2127
Sent: 16 Good: 16 Bad: 0 T/O:0 Out: 0 Avg: 2052
Sent: 17 Good: 17 Bad: 0 T/O:0 Out: 0 Avg: 2001
Sent: 18 Good: 18 Bad: 0 T/O:0 Out: 0 Avg: 1959
Sent: 19 Good: 19 Bad: 0 T/O:0 Out: 0 Avg: 1916
At the Radius -x15 debug screen the response time for each of those requests
was between 1100-1299, the times do not correspond.
When I add on the 8th/9th/etc.. chassis and run radlogin it just times out
on every request. When I switch all the chassis off Radius NT it will
continue displaying debug/auth. information for awhile 10-20 seconds. It
seems like it's getting behind and it's processing the requests but by the
time it processes the request the NAS has given up and sent a new request
and it ignore's the last ACK???
I have run this on multiple different PCs, SQL Servers and run radlogin
directly from the Radius/SQL server and the response times always seem to be
between 1000-2000. Lasts results ran against a Double Pentium II/300, 512MB
Ram, RadiusNT, SQL and Radlogin all running on same PC I think that should
eliminate network performance. Also, run as service and without accounting
(currently still running without accounting).
SQL Performance Monitor Stats:
Cache Hit Response 97 97 97 97
IO Tran/sec .67 1.78 .66 18.6
IO Page read/sec .0 .3 .0 11.65
IO SP w/s .0 .14 .0 3.3
user 20 20 20 20