Oracle various log cleaning

Oracle various log cleaning (general oracle software directory is about 10G)


generally advance the $ORACLE_BASE directory, check which directory takes up more space, and go deeper into it, locate the directory that occupies a large amount, determine the file type and clean it up. As follows:
    [oracle@test-for-lihb ~]$ cat /etc/redhat-release 
    Red Hat Enterprise Linux Server release 6.4 (Santiago)
    [oracle@test-for-lihb ~]$ cd $ORACLE_BASE
    [oracle@test-for -lihb /opt/app/oracle]$ du -sm *|sort -n
    1 admin
    1 afiedt.buf
    1 checkpoints
    3 cfgtoollogs
    10 flash_recovery_area
    1377 diag
    8501 product


1 Archive file
 path:
    (SQL> archive log list Get archive from Archive destination path)
 method:
    delete archive via rman
    [oracle@test-for-lihb ~]$ rman target/
    RMAN> delete archivelog all completed before 'sysdate-1'; ("1" corresponds to one day, if you want to delete the archive log 6 hours ago, change it to 0.25)
    After the archive is deleted, if there is a corresponding backup strategy, you need to restart the full ready.

2 Alert alert log file
 path:
    /alert_.log (Get the log file location from SQL> show parameter background_dump_dest/core_dump_dest/user_dump_dest)
 Method:
    1) Clear alert_.log > alert_.log (suggestion: tail -100000 alert_.log > alert_.log_bak20170601, keep 10w rows of data and delete it)
    2) The system will automatically generate a new alert_.log
    3) find . -mtime +30 -name "*.trc/trm" | wc -l View trc/ one month ago Number of trm files
    4) find . -mtime +30 -name "*.trc/trm" | xargs rm -rf delete trc/trm files one month ago

3 Listen log file
 path:
    $ORACLE_HOME/network/log/listener.log (The location of the log file is obtained from the Listener Log File in the listener name of $ lsnrctl status $)
 If it is a log file
    , method 1:
        1) lsnrctl set log_status off
        2) Delete or remove the listener.log (recommendation: tail -100000 listener.log > listener.log_bak20170601, keep 10w rows of data and delete it)
        3) lsnrctl set log_status on At this time, the system regenerates the listener. Log 
    method 2:
        1) Directly clear the log file: >listener.log (recommendation: tail -100000 listener.log > listener.log_bak20170601, keep 10w rows of data and then delete it)
 If it is an xml file
    method:
        keep log.xml, others can be deleted, For example: find . -mtime +30 -name "log_*.xml" | xargs rm -rf 
        At the same time, the log files in the trace directory at the same level can be cleaned up:
        cd ../trace
        tail -100000 listener.log > listener.log_bak20170601
        >listener .log

4 OEM log file
 path:
    $ORACLE_HOME/oc4j/j2ee/OC4J_DBConsole__/log/http-web-access.log or em-application.log
 Method:
    Check the OEM status: $ emctl status If dbconsole
    is not running,
    directly remove or delete http-web-access.log/em-application.log.
    If it is running,
    stop dbconsole (emctl stop dbconsole),
    remove or delete http-web-access.log/em-application.log,
    restart dbconsole (emctl start dbconsole), and
    check the oem status (emctl start dbconsole).

5 CRS log file
 path:
    11g: grid user: $ORACLE_HOME/log/$HOSTNAME/crsd/crsdOUT.log (by default, 10 historical records will be kept, each 10M, and the file name is crsd.l01/02/03... )
    10g: grid user: $ORA_CRS_HOME/log/$HOSTNAME/crsd/
 Method:
    Clean up historical crsd.l0* files, keep crsdOUT.log and crsd.log in
    other same-level directories agent/cssd/gipcd/mdnsd/ctssd/ohasd The cleaning method is similar to the crsd directory

6 Audit file
 path:
    (SQL> show parameter audit_file_dest view)
 Method:
    Clear audit files older than 7 days: find . -mtime +7 -name "*.aud" | xargs rm -rf