mysql failover issue

[andy]

Hi,

I have seen these errors when primary radius server goes offline and the secondary takes over:
All requests are coming via Redback SE400:

Wed Sep 13 07:51:34 2006 : Error: Dropping conflicting packet from client redback:1812 - ID: 1 due to unfinished 
request 1051
285
Wed Sep 13 07:51:35 2006 : Error: Dropping conflicting packet from client redback:1812 - ID: 2 due to unfinished 
request 1051
286


This happens on both servers and after an hour it seemed to sort itself out.

Setup is as follows:

2xfreeradius server
2xmysql cluster API nodes

each FR server is pointed to a different node, so when we take one node offline, the secondary defined freeradius 
server should kick in and talk to its API.

When the primary radius server came back after its mysql API was brought back after a planned outage, the primary and 
secondary radius servers got confused. They both started to drop connections with the above errors and no users could 
authenticate.

Is this issue a known issue, is there a fix, has anyone got any further info on when this would happen.
Restarting the primary radius server fixed the issue however this solution is meant to be a resiliant and redundant 
solution capable of working through either radius or mysql node failures.

any info or assistance would be helpful for my RFO here.

cheers

 -- 
andy    andy at shady.org
-----------------------------------------------
Never argue with an idiot. They drag you down 
to their level, then beat you with experience.
-----------------------------------------------