It was indeed. By mistake I had set the alternative accounting start query to an empty string. Something that I noticed when I looked better at the following piece of log: rlm_sql (sql): Couldn't insert SQL accounting START record - ORA-00001: unique constraint (RADUSER.RADACCT_IDX2) violated radius_xlat: '' radius_xlat: '/usr/local/var/log/radius/sqltrace.sql' rlm_sql (sql): Released sql socket id: 4 where you can see the radius_xlat tracing an empty string. By the way, probably something could be done in this respect. An idea could be that if the alternate query is empty (either a mistake or a wanter behaviour) the module keeps the return value of the primary query. Thanks for your attention :) Giuseppe On Thu, 2005-09-29 at 14:39 -0400, Alan DeKok wrote:
Giuseppe <gaaw@interfree.it> wrote:
This Accounting Response is unexpected to me because RFC 2866 says: "If the RADIUS accounting server is unable to successfully record the accounting packet it MUST NOT send an Accounting-Response acknowledgment to the client".
It appears to be an issue with the SQL module:
rlm_sql_oracle: execute query failed in sql_query: ORA-00001: unique constraint (RADUSER.RADACCT_IDX2) violated rlm_sql (sql): failed after re-connect rlm_sql (sql): Couldn't insert SQL accounting START record - ORA-00001: unique constraint (RADUSER.RADACCT_IDX2) violated ... rlm_sql (sql): Released sql socket id: 4 modcall[accounting]: module "sql" returns ok for request 0
If something fails, the SQL module shouldn't return "ok".
Alan DeKok. - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html