What constitute "default auditing" under traditional auditing?

From the Oracle Documentation:

"When you use Database Configuration Assistant (DBCA) to create a new database, Oracle Database configures the database to audit the most commonly used security-relevant SQL statements and privileges"

and

"If you manually create a database, then you should run the secconf.sql script to apply the default audit settings to your database"

Oracle Database audits the following privileges by default:

ALTER ANY PROCEDURE

CREATE ANY LIBRARY

DROP ANY TABLE

ALTER ANY TABLE

CREATE ANY PROCEDURE

DROP PROFILE

ALTER DATABASE

CREATE ANY TABLE

DROP USER

ALTER PROFILE

CREATE EXTERNAL JOB

EXEMPT ACCESS POLICY

ALTER SYSTEM

CREATE PUBLIC DATABASE LINK

GRANT ANY OBJECT PRIVILEGE

ALTER USER

CREATE SESSION

GRANT ANY PRIVILEGE

AUDIT SYSTEM

CREATE USER

GRANT ANY ROLE

CREATE ANY JOB

DROP ANY PROCEDURE

Oracle Database audits the following SQL shortcuts by default:

ROLE

SYSTEM AUDIT

PUBLIC SYNONYM

DATABASE LINK

PROFILE

SYSTEM GRANT

Remember that secconf.sql turns on audting regardless of your database using Unified Auditing or Traditional Auditing.

How?

When executed manually, the DBA is prompted for which type of auditing style that should be considered:

sqlplus / as sysdba @secconf.sql

Session altered.


Profile altered.

Do you wish to configure 11g style Audit Configuration OR
Do you wish to configure 12c Unified Audit Policies?
Enter RDBMS_11G for former or RDBMS_UNIAUD for latter
Enter value for 1: RDBMS_11G
old   7:   USER_CHOICE := '&1';
new   7:   USER_CHOICE := 'RDBMS_11G';

What I have found, is that if you intend to harden your Traditional Auditing policies by executing the script in a database where the Unified Auditing polices exist but is disabled, or a database running in "Mixed mode" auditing, the script will fail:

DECLARE
*
ERROR at line 1:
ORA-46358: Audit policy ORA_ACCOUNT_MGMT already exists.
ORA-06512: at line 9

A simple workaround in such a case is to simply comment out the code that is irrelevant to your desired type of auditing, and rerun the script.

You can disable default auditing if you wish, see the section "Disabling and Enabling Default Audit Settings"

To check whether or not default auditing has been actived, you can query the view DBA_PRIV_AUDIT_OPTS, which describes current system privileges being audited across the system and by user. If the column USERNAME is NULL, you have introduced system-wide auditing.

How to convert stored outlines to use SQL Plan Baselines

If you have been using stored outlines for plan stability in versions prior to Oracle 11g, you should migrate them to use SQL Plan Baselines instead. Stored outlines is, according to oracle, "a legacy technique for providing plan stability".

My database had a number of stored outlines:

SELECT OWNER, CATEGORY, USED, MIGRATED,COUNT(*) 
FROM   DBA_OUTLINES
GROUP BY OWNER,CATEGORY,USED,MIGRATED
ORDER BY MIGRATED;

OWNER	CATEGORY	USED	MIGRATED	COUNT(*)
USER1	DEFAULT	USED	MIGRATED	3
USER2	DEFAULT	USED	MIGRATED	1
USER3	DEFAULT	USED	NOT-MIGRATED	7
USER1	DEFAULT	USED	NOT-MIGRATED	7
USER4	DEFAULT	USED	NOT-MIGRATED	1
USER2	DEFAULT	USED	NOT-MIGRATED	36

I created the following pl/sql to convert them to SQL Plan Baselines using the package DBMS_SPM.MIGRATE_STORED_OUTLINE:

SET TRIMSPOOL ON
SET LINES 200
SET PAGES 200
ALTER SESSION SET NLS_LANGUAGE='AMERICAN'; <-- to get English messages during execution
SPOOL CONVERT_OUTLINES.LOG
SET SERVEROUTPUT ON
DECLARE
  L_CLOB            CLOB; -- will display the resulting report

  CURSOR C1 IS
    SELECT OWNER,NAME
    FROM DBA_OUTLINES
    WHERE MIGRATED = 'NOT-MIGRATED';

    C1_REC C1%ROWTYPE;

 BEGIN
    DBMS_OUTPUT.ENABLE( 1000000 );

    IF NOT C1%ISOPEN THEN
       OPEN C1;
    END IF;
 
     LOOP
       FETCH C1 INTO C1_REC;
       EXIT WHEN C1%NOTFOUND;
       DBMS_OUTPUT.PUT_LINE('Now converting: ' || C1_REC.OWNER || '.' || C1_REC.NAME);
       L_CLOB := DBMS_SPM.MIGRATE_STORED_OUTLINE( ATTRIBUTE_NAME=>'OUTLINE_NAME', ATTRIBUTE_VALUE=>C1_REC.NAME, FIXED=>'NO');
       DBMS_OUTPUT.PUT_LINE( L_CLOB );
     END LOOP;
    CLOSE C1;
END;
/
EXIT

The resulting log files shows that several of the stored outlines could not be converted:

Now converting: USER2.SYS_OUTLINE_11021513055564321
-------------------------------------------------------------------------------

               Migrate Stored Outline to SQL Plan Baseline

Report
-------------------------------------------------------------------------------

Summary:
--------

Number of stored outlines to be migrated: 1
Stored outlines migrated successfully: 0
Stored outlines failed to be migrated: 1

Summary of Reasons for failure:
-------------------------------

Number of invalid stored outlines: 1

Details on stored outlines not migrated or name changes during migration:
-------------------------------------------------------------------------

* Notes on name change:
* New SQL plan baselines are assigned the same names as their original stored 
* outlines. If a stored outline has the same name as an existing
* SQL plan baseline, a system generated name is used for the new
* SQL plan baseline.

I then checked with the developers. It turned out that the outlines that didn't convert properly were remnants from the database when it was totally different laid out - the schemas had by now diverged and I could simply ignore these obsolete outlines.

So the last step was simply to generate a drop-script for the non-migrated outlines and then execute these:

SELECT 'DROP OUTLINE ' || NAME || ';'
FROM DBA_OUTLINES
WHERE MIGRATED = 'NOT-MIGRATED';

Login storm against database caused exhausted library cache

One of our databases experienced massive contention in the shared pool, in form of wait events alerted as "library cache locks".

The database was very small indeed, so my natural instinct was to throw some more memory at the virtual host, and rearrange the memory parameters.

This turned out to be a misconception; the resources were sufficient for the instance to work properly.

The problem was caused by an incorrect password configuration on the application server.

What we could observe was:

A totally exhausted shared pool, caused by "library cache lock"

The SQL that seemed to be repeatedly executed was

SELECT /*+ connect_by_filtering */
          privilege#, LEVEL
      FROM sysauth$
CONNECT BY grantee# = PRIOR privilege# AND privilege# > 0
START WITH grantee# = :1 AND privilege# > 0;


SELECT privilege#
  FROM sysauth$
WHERE (grantee# = :1 OR grantee# = 1) AND privilege# > 0;

The V$EVENT_NAME view showed that the wait event was accompanied by the additional information found in the columns parameter1 through parameter3, which turned out to be helpful further on:

select  name, wait_class,parameter1,parameter2,parameter3
from v$event_name
where wait_class = 'Concurrency'
and name = 'library cache lock';

NAME	WAIT_CLASS	PARAMETER1	PARAMETER2	PARAMETER3
library cache lock	Concurrency	handle address	lock address	100*mode+namespace

Further research showed that the problem was due to a built-in delay between failed login attempts in Oracle 11g:

"The 'library cache lock' wait is seen due to the fact that the account status gets updated due to incorrect login.
To prevent password guessing attack, there's a sleep() in the code when incorrect login attempts exceed count of 3.
And because of this sleep() you see a wait on library cache, as the process is yet to release the lock."

In release 11.1.0.7, patch 7715339 was released to remove this delay.

In release 11.2.X, the DBA must set an event to remove the delay, as follows:

alter system set events '28401 trace name context forever, level 1'; 

According to Oracle, the purpose of the built-sleep is to make it harder to succeed in a "password guessing attack", particularly in cases where FAILED_LOGIN_ATTEMPTS is set to UNLIMITED. Oracle Development is pointing out that disabling the sleep-function is not recommended. A better solution is to set the FAILED_LOGIN_ATTEMPTS to a reasonable value.
When the number of failed login attempts for a session hits the limit, the account will be locked. Subsequent logon attempts with incorrect password will then be rejected immediately without any contention in the library cache.

See Bug 15882590 : 'LIBRARY CACHE LOCK' DURING WRONG PASSWORD LOGON ATTEMPTS on My Oracle Support (MOS) for further information.

How to use the case statement in a shell script, with case-insensitive input

Use the pipe (|) character for each possible case:

echo "TABLESPACE REPORT FOR $ORACLE_SID"

stty echo
   echo "All tablespaces or specific tablespace? [A|S]"
   read answer
stty echo

case $answer in
A|a) echo "All tablespaces selected...please wait...";
    getFiles A;;
S|s) echo "Which tablespace?";
    read wt;
    getFiles $wt;;
*) echo "Only A or S are supported parameters.";
    exit 1;;
esac

How to find the SID, serial# and their operating system process ID from the database

A simple statement to find the SID, serial# and their operating system process ID from the database:

SET LINES 300 PAGES 300
SET TRIMSPOOL ON
SPOOL SQL
COL "Os pid" FORMAT A10
COL MACHINE  FORMAT A30
COL "SQL Text" FORMAT A100 WRA
COL PROGRAM  FORMAT A20
COL USERNAME FORMAT A12

SELECT
  S.SID,
  S.SERIAL#,
  S.USERNAME,
  S.STATUS,
  P.SPID "Os pid",
  S.MACHINE,
  CONCAT(SUBSTR(A.SQL_TEXT,1,60), '...(output truncated)')  "SQL text"
FROM V$SESSION S,
          V$SQLAREA A,
          V$PROCESS P
WHERE A.ADDRESS = S.SQL_ADDRESS
AND S.PADDR=P.ADDR
ORDER BY 1
/

Example output:

SID          SERIAL# USERNAME     Os pid     MACHINE          SQL text
---------- ---------- ------------ ---------- -------------  ------------------------------------------------------------------------------------
         4          3              6357052    myserver       insert into obj$(owner#,name,namespace,obj#,type#,ctime,mtim...(output truncated)
        18       7035 SYS          21496004   myserver       call DBMS_AQADM_SYS.REGISTER_DRIVER (  )...(output truncated)
       107      62143 SYS          23527460   myserver       SELECT   S.SID,   S.SERIAL#,   S.USERNAME,   P.SPID "Os pid"...(output truncated)
       303      27191 SYS          16777372   myserver       analyze table scott.man_to_stage validate structure cascade o...(output truncated)
       393          1              8716332    myserver       insert into obj$(owner#,name,namespace,obj#,type#,ctime,mtim...(output truncated)
       402       3125 BATCHUSR     18612294   myserver       call mypackage.start_dorg (  )...(output truncated)
       490          1              8257548    myserver       insert into obj$(owner#,name,namespace,obj#,type#,ctime,mtim...(output truncated)
       499      25461 SYS          13762730   myserver       call DBMS_AQADM_SYS.REGISTER_DRIVER (  )...(output truncated)
       503       2635 SYS          11272234   myserver       call DBMS_AQADM_SYS.REGISTER_DRIVER (  )...(output truncated)
       593      12091 BATCHUSR     19726590   myserver       call myprocedure.read_queue (  )...(output truncated)
       600       9193 SYS          21102724   COMP\PC1       select * from dba_locks...(output truncated)
       694      44601 BATCHUSR     20840656   myserver       call mypackage.check_job_consistency(  )...(output truncated)
       696        819 SYS          15269968   myserver       call DBMS_AQADM_SYS.REGISTER_DRIVER (  )...(output truncated)
       700       1135 BATCHUSR     21364890   myserver       DECLARE job BINARY_INTEGER := :job;  next_date TIMESTAMP WIT...(output truncated)

What is the meaning of the "10G 11G" value in DBA_USERS.PASSWORD_VERSIONS?

When the value of DBA_USERS.PASSWORD_VERSIONS is shown as "10G 11G", it means that both old and new-style hash values are available for the user.

Note that instead of storing the hashed password values directly in the DBA_USERS table, they are from 11gR1 and onwards both stored in the table USER$, column PASSWORD for the 10G style hash value, and column SPARE4 for the 11G SHA-1 style hash value.

A NULL value indicates that the password has not been changed since the migration and the user still has the old case insensitive password.

In my query below, USER1 through USER4 have been migrated and will now take advantage of 11g password case-sensitivity, if enabled.

USER5 is still using 10G style case insensitive passwords.

USER6 is created after migration, and will also take advantage of the 11g password case-sensitivity, if available.

SELECT U.NAME "Name",
            NVL(REGEXP_REPLACE(U.SPARE4,'^.+', 'Password changed since migration',1,0), 'Password unchanged since migration') "Action undertaken",
            DU.PASSWORD_VERSIONS "Password Version"
FROM USER$ U LEFT OUTER JOIN DBA_USERS DU
ON U.NAME = DU.USERNAME
WHERE DU.USERNAME NOT IN (SELECT ROLE FROM DBA_ROLES)
ORDER BY NAME ASC;

Name	Action undertaken	Password Version
USER1	Password changed since migration	10G 11G
USER2	Password changed since migration	10G 11G
USER3	Password changed since migration	10G 11G
USER4	Password changed since migration	10G 11G
USER5	Password unchanged since migration	10g
USER6	Password changed since migration	11G

Users that are imported from an earlier release into an 11g database, will remain case-insensitive until the password is changed.


To generate "alter user identified by values" statements, use the following:

SELECT 'alter user ' || NAME || ' identified by values ' || '''' || SPARE4 ||';' || PASSWORD ||''';' 
FROM USER$ 
WHERE  NAME IN ('USER1','USER2');

Sources: Oracle Documentation

How to use trace event 10046

ALTER SESSION SET max_dump_file_size = unlimited;
ALTER SESSION SET tracefile_identifier = 'normal_run_with_trace_10046';
ALTER SESSION SET statistics_level = ALL;
ALTER SESSION SET events '10046 trace name context forever, level 12';

Your SQL statement(s) here

ALTER SYSTEM SET EVENTS '10046 trace name context off';
EXIT

Alternatively, use the oradebug utility:

select s.sid,s.serial#,p.spid
from v$session s, v$process p 
where sid = 29
and s.paddr = p.addr

Output:

SID	SERIAL#	SPID
29	9	24510662

Check that the process is indeed there:

prodserver1>ps -ef | grep 24510662
  ora11g 24510662        1   0 17:24:42      -  0:00 oraclemagr (LOCAL=NO)
prodserver1>proctree 24510662
  24510662    oraclemagr (LOCAL=NO)

Enable tracing on the operating system process, inside sqlplus:

SQL> oradebug setospid 24051998
Oracle pid: 420, Unix process pid: 24051998, image: oracle@ystu032ma
SQL> oradebug unlimit
SQL> oradebug event 10046 trace name context forever, level 12

When done, disable tracing:

SYS@magr SQL> oradebug setospid 24051998
Oracle pid: 420, Unix process pid: 24051998, image: oracle@ystu032ma

SQL> oradebug unlimit
SQL> oradebug event 10046 trace name context forever, level 12

You can now parse the trace file produced through tkprof:

tkprof mytracefile.trc mytracefile.out sys=yes waits=yes sort=exemis

For an extensive list of commands to be used with tkprof, simply type tkprof at the prompt.

An excellent note on 10046 trace event is called "How To Collect 10046 Trace (SQL_TRACE) Diagnostics for Performance Issues (Doc ID 376442.1)".

Another great note on tracing in general is "Tracing Enhancements Using DBMS_MONITOR (In 10g, 11g and Above) (Doc ID 293661.1)"

- both available on Oracle Supports site.