Hello to everyone.
Yesterday I installed freeradius-0.9.3 on a Solaris 2.7 production
environment. I use sql module and oracle as my database to store
accounting. I first try to store the accounting data in the database
and afterwards I write them in files with the detail module. The
problem is that sometimes I get data that fail to be inserted in my
database (wrong size, stupid error) so I see the following in
radius.log:
...
Thu Apr 1 15:55:13 2004 : Error: rlm_sql_oracle: execute query failed
in sql_query: ORA-01401: inserted value too large for column
Thu Apr 1 15:55:13 2004 : Auth: Login OK: [EMAIL PROTECTED] (from
client fworks port 352 cli 2221038953)
Thu Apr 1 15:55:14 2004 : Error: rlm_sql_oracle: execute query failed
in sql_query: ORA-01401: inserted value too large for column
Thu Apr 1 15:55:14 2004 : Error: rlm_sql (sql1): failed after re-connect
Thu Apr 1 15:55:14 2004 : Error: rlm_sql: Couldn't insert SQL
accounting STOP record - ORA-01401: inserted value too large for column
Thu Apr 1 15:55:18 2004 : Error: rlm_sql_oracle: execute query failed
in sql_query: ORA-01401: inserted value too large for column
Thu Apr 1 15:55:18 2004 : Error: rlm_sql_oracle: execute query failed
in sql_query: ORA-01401: inserted value too large for column
Thu Apr 1 15:55:18 2004 : Error: rlm_sql (sql1): failed after re-connect
Thu Apr 1 15:55:18 2004 : Error: rlm_sql: Couldn't insert SQL
accounting STOP record - ORA-01401: inserted value too large for column
Thu Apr 1 15:55:21 2004 : Auth: Login OK: [EMAIL PROTECTED] (from
client fworks port 0 cli 2104131605)
...
The error is very clear and I found the problem data. As I can
understand,
when a query fails, the sql module will try to execute the alternative
query (if it exists) and if this also fails it reconnects to the db
and tries one more time. Is this correct?
Anyway, the radius server keeps servicing other requests and after a
few such failures, the whole server stops functioning!
Is this a correct behaviour? Shouldn't the server keep working servicing
other requests? Is there some sort of threshold of errors after which
the server stops functioning? Should we consider that a bug? (at first
sight it seems this way)
My configuration regarding accounting follows below.
# Pre-accounting. Decide which accounting type to use.
#
preacct {
preprocess
#
# Look for IPASS-style 'realm/', and if not found, look for
# '@realm', and decide whether or not to proxy, based on
# that.
#
# Accounting requests are generally proxied to the same
# home server as authentication requests.
# realmslash
suffix
#
# Read the 'acct_users' file
files
}
#
# Accounting. Log the accounting data.
#
accounting {
#
# Ensure that we have a semi-unique identifier for every
# request, and many NAS boxes are broken.
acct_unique
#
# Create a 'detail'ed log of the packets.
# Note that accounting requests which are proxied
# are also logged in the detail file.
# detail
# daily
# unix # wtmp file
#
# For Simultaneous-Use tracking.
#
# Due to packet losses in the network, the data here
# may be incorrect. There's little we can do about it.
# radutmp
# sradutmp
# Return an address to the IP Pool when we see a stop record.
# main_pool
# see acct_users file
Acct-Type SQL1{
sql1
detail
}
}
The acct_users file:
DEFAULT Acct-Type := SQL1
Thank you in advance.
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
