Oracle11g Tracefile Housekeeping by ADRCI (Automatic Diagnostic Repository Command Interpreter)－DanBrother的部落格

ADRCI (Automatic Diagnostic Repository Command Interpreter) :

[Introduction]
Housekeeping tasks are automatically picked up by MMON (Memory Monitor) backround process.

SHORTP_POLICY(short term) defaults to 720 hours (30 days) => automatically clean trace & core dump files

LONGP_POLICY(long term) defaults to 8760 hours (365 days) => automatically clean Incidents and Health Monitor warnings.
SQL> sho parameter diagnostic_dest
NAME TYPE VALUE
------------------------------------ -------------------- ------------------------------
diagnostic_dest string /oracle

[Usage ]
# adrci
ADRCI: Release 11.2.0.2.0 - Production on Tue Oct 4 15:55:58 2011

No ADR base is set

adrci> set base /oracle
adrci> show base
ADR base is "/oracle"

adrci> show homes
ADR Homes:
diag/rdbms/testdb/TESTDB1
diag/rdbms/testdb2/TESTDB21
diag/rdbms/testdb3/TESTDB31

adrci> set home TESTDB

adrci> show control

ADR Home = /oracle/diag/rdbms/testdb/TESTDB1:
*************************************************************************
ADRID SHORTP_POLICY LONGP_POLICY LAST_MOD_TIME LAST_AUTOPRG_TIME LAST_MANUPRG_TIME ADRDIR_VERSION ADRSCHM_VERSION ADRSCHMV_SUMMARY ADRALERT_VERSION CREATE_TIME
-------------------- -------------------- -------------------- ---------------------------------------- ---------------------------------------- ---------------------------------------- -------------------- -------------------- -------------------- -------------------- ----------------------------------------
45261763 720 8760 2011-08-18 16:19:23.458764 +08:00 1 2 80 1 2011-08-18 16:19:23.458764 +08:00
1 rows fetched

Manually purging all trace files (.log) older than 6 days (8640 minutes) by:
adrci> purge -age 8640 -type TRACE

Manually purging all alert files (.xml) older than 6 days (8640 minutes) by:
adrci> purge -age 8640 -type ALERT

adrci> help purge

Purpose: Purge the diagnostic data in the current ADR home. If no
option is specified, the default purging policy will be used.

Options:
[-i id1 | id1 id2]: Users can input a single incident ID, or a
range of incidents to purge.

    [-age ]: Users can specify the purging policy either to all
    the diagnostic data or the specified type. The data older than
    ago will be purged

Examples:
    purge
    purge -i 123 456
    purge -age 60 -type incident

Without specifying purging type, all types will be purged.
adrci> purge -age 8640

Changing SHORTP_POLICY & LONGP_POLICY values (in hours) by:
adrci> set control (SHORTP_POLICY = 360 LONGP_POLICY = 1440)

adrci> show control

ADR Home = /oracle/diag/rdbms/testdb/TESTDB1:
*************************************************************************
ADRID SHORTP_POLICY LONGP_POLICY LAST_MOD_TIME LAST_AUTOPRG_TIME LAST_MANUPRG_TIME ADRDIR_VERSION ADRSCHM_VERSION ADRSCHMV_SUMMARY ADRALERT_VERSION CREATE_TIME
-------------------- -------------------- -------------------- ---------------------------------------- ---------------------------------------- ---------------------------------------- -------------------- -------------------- -------------------- -------------------- ----------------------------------------
45261763 360 1440 2011-10-04 15:13:15.646709 +08:00 2011-10-03 00:13:43.774374 +08:00 2011-10-04 15:13:23.590306 +08:00 1 2 80 1 2011-08-31 15:10:54.197781 +08:00
1 rows fetched

** For log files such as listener tracelog & listener scan1 tracelog, the housekeeping method is different by using Linux logrotate program:

1.
Login as root
cd /etc/logrotate.d

2.
Create a file named ora_housekeeping and type the following commands:

# listener tracelog
/grid/orabase/diag/tnslsnr/RAC1/listener/trace/listener.log {
weekly
compress
rotate 4
notifempty
missingok
copytruncate
nocreate
}

# listener scan1 tracelog
/grid/grid11203/log/diag/tnslsnr/RAC1/listener_scan1/trace/listener_scan1.log {
weekly
compress
rotate 4
notifempty
missingok
copytruncate
nocreate
}

** For ASM RDBMS audit log & trace files housekeeping method using crontab:
EXPORT ORACLE_SID=+ASM
sqlplus / as sysasm
SQL> sho parameter audit_file_dest
NAME TYPE VALUE
--------------------- ------------ ----------------------------------------------
audit_file_dest string /grid/grid11203/rdbms/audit

crontab -e
>>
# Clear 30 days old ASM audit log files daily at 7AM
0 7 * * * find /grid/grid11203/rdbms/audit -mtime +30 -name "*.aud" -exec rm -rf {} \; > /dev/null 2>&1

0 7 * * * find /grid/grid11203/rdbms/log -mtime +30 -name "*.trc" -exec rm -rf {} \; > /dev/null 2>&1

See also Auditing Using Operating System Audit Trail:
http://danbrother.pixnet.net/blog/post/40366278

[Reference]
[1] http://gavinsoorma.com/2010/09/purging-trace-and-dump-files-with-11g-adrci/
[2] http://www.morganslibrary.com/reference/adr.html
[3] http://itdavid.blogspot.tw/2011/02/manage-oracle-11gr2-asm-and-rdbms-audit.html