We have some pppoe servers with radius-authentication. Sometimes, our customers tune their pppoe-clients to retry every second forever. For blocked (by billing) customers we offer so-called ‘gray’ pppoe-session (with ip 192.168.x.x), that have access only to a few corp sites. But some of them made errors in password, use invalid logins or do other mistakes. So our radius server takes a high load to process such useless frequent queries, while they (customers) sleeps or went away to job, and etc.
Is there any method to limit these negative pppoe-session creation requests?
There is nothing to do on PPPoE server, how do you know specific PPPoE connection is good or bad?
Educate your customer and create configuration guide for PPPoE client configuration to avoid bad username/password users.
Sometimes user’s accounts disabled for non-payment. It is a lot every begin of month usually. This causes explosive growth in the number of attempts:
89172209 Jan 2 03:15 radiator.log.1.bz2
13945926 Jan 1 03:13 radiator.log.2.bz2
6846862 Dec 31 03:13 radiator.log.3.bz2
7216602 Dec 30 03:13 radiator.log.4.bz2