My problem: My radius servers that are linked to mysql databases are locking (requires start of service)every day.
Now I am trying to understand if there is a mysql problem that is impacting on the radius service in some way.

I have set up the slow query log. I would like some clarification from someone on the findings. Is there something I can do to investigate these slow responses

# User@Host: x[x] @ [. . .]
# Query_time: 4 Lock_time: 0 Rows_sent: 16208 Rows_examined: 32416
SELECT `radacct`.`RadAcctId`, `radacct`.`ts` FROM `radacct` `radacct` ORDER BY `radacct`.`ts`;
/usr/libexec/mysqld, Version: 4.1.22-log. started with:
Tcp port: 3306 Unix socket: /var/lib/mysql/mysql.sock
Time Id Command Argument
# Time: 081111 8:05:10
# User@Host: x[x] @ [. . .
# Query_time: 11 Lock_time: 0 Rows_sent: 16245 Rows_examined: 48767
use radius;
SELECT `ts` ,`UserName` ,`AcctQoS` ,SUM(`AcctInputOctets` ) ,SUM(`AcctInputPackets` ) FROM `radacct` GROUP BY `ts` ,`UserName` ,`AcctQoS`;
# Time: 081111 8:05:38
# User@Host: x[x] @ [. . .]
# Query_time: 2 Lock_time: 0 Rows_sent: 20546 Rows_examined: 20546
SELECT `radpostauth`.`id` FROM `radpostauth`;
# Time: 081111 8:06:12
# User@Host: x[x] @ [. . . ]
# Query_time: 11 Lock_time: 0 Rows_sent: 0 Rows_examined: 0
SELECT `ts` ,`UserName` ,`AcctQoS` ,SUM(`AcctInputOctets` ) ,SUM(`AcctInputPackets` ) FROM `radacct` GROUP BY `ts` ,`UserName` ,`AcctQoS`;
# Time: 081111 8:06:38
# User@Host: x[x] @ [. . . ]
# Query_time: 5 Lock_time: 0 Rows_sent: 16277 Rows_examined: 32554
SELECT `radacct`.`RadAcctId`, `radacct`.`ts` FROM `radacct` `radacct` ORDER BY `radacct`.`ts`;
# Time: 081111 8:10:15


I have also run the TUNING PRIMER and have some results from this..

What exactly are the log msgs re 'locking'? Does it say the db is hung ?
BTW, those queries you've shown don't seem to use a WHERE clause, so you're selecting a lot of rows, which will only get bigger/slower as time goes by.
You need to SELECT by userid or something of that ilk. Needs to be an indexed column.