Standby Database on Oracle 10g Standard Edition ?

Question

Hi All,
Is it possible to do Oracle Stand By Database on Oracle 10g SE ?
I read somewhere in this group that we can do standby database even in SE by writing a script that copy the archive log to the secondary server.
Is this approach reliable enough to be done in production ?
Does anybody have sample of the script ?
Thank you for your help,
xtanto

Answer 1

Hi,
Well, I think I'd beter put the scripts right there. Remember, they have been tested under MY environment, using MY releases, MY O.Ses. It's up to you to check whether the're valid for you. They are provided as is, for sample.
There are 4 files:
. generic.sh : you can duplicate this file in order to set up as many standby as needed, etc. It calls other scripts in order to:
. archivemove.sh : Get the archived redo logs to the standby host
. recover.sh : Synch the standby
. getrecid.sql : get the maximum progess point on the Manual Standby (used by archivemove.sh)
These scripts are used from the standby host. Remember to throughly check it before relying on it for production.
generic.sh
#!/bin/sh
# Be sure environment variable are set. If not, then it might fail!
# These environment variables are those for the Manual Physical Standby host
export ORACLE_HOME=/logical/oracle/Ora9i
export ORACLE_BASE=/logical/oracle
export ORACLE_STANDBY=tns_alias
export ORACLE_STANDBY_SYSDBA_PASS=change_on_install
export PATH=$ORACLE_HOME/bin:$PATH
export SOURCE_HOST=primary_host
export SOURCE_DRIVE=/primary/absolute/path/to/archived/redo/logs
export LOCAL_ARC_PARTH=/path/to/logical/archive/dest
# Check the date command usage depending on the platform
dateexec=`date "+%Y-%m-%d-%H-%M"`
# copy archived redo logs from main database
archivemove.sh > $dateexec.generic.log
# recover/sync the Manual Standby Database
recover.sh >> $dateexec.generic.log
archivemove.sh
#!/bin/sh
echo ----------------------------------------------------------------
echo ----------------------------------------------------------------
echo Get what log has last been applied to: $ORACLE_STANDBY
echo ----------------------------------------------------------------
sqlplus /nolog @getrecid.sql $ORACLE_STANDBY
echo ----------------------------------------------------------------
maxval=`tail -1 recid.log`
echo maxval=$maxval
rm recid.log
echo ----------------------------------------------------------------
# Check source drive to see what we're missing locally (source = primary)
for filename in `remsh $SOURCE_HOST 'ls $SOURCE_DRIVE' | sort`
do
     # get archive number.
     # WARNING here I'm based on MY archived redo log name format! Put yours for the cut
     filename_parsed=`echo $filename | cut -c12-16`
     # Check if the number is after the last one applied to standby
     if [ $filename_parsed -gt $maxval ]
     then
          # grab it!
          echo $filename
          rcp $SOURCE_HOST:$SOURCE_DRIVE/$filename $LOCAL_ARC_PARTH
     fi
done
echo ----------------------------------------------------------------
echo Removing old files
echo ----------------------------------------------------------------
# Check in local directory
for filename in `ls $LOCAL_ARC_PATH | sort`
do
     # WARNING again about filename format
     filename_parsed=`echo $filename | cut -c12-16`
     # Check the arc number...
     if [ $filename_parsed -lt `expr $maxval - 15` ]
     then
          # Delete it!
          echo $filename
          rm -f $LOCAL_ARC_PATH/$filename
     fi
done
echo ----------------------------------------------------------------
echo end archivemove.sh
echo ----------------------------------------------------------------
recover.sh
#!/bin/sh
echo ----------------------------------------------------------------
echo Traitement de la base $ORACLE_STANDBY
echo ----------------------------------------------------------------
sqlplus /nolog << EOF
connect sys/$ORACLE_STANDBY_SYSDBA_PASS@$ORACLE_STANDBY as sysdba
SELECT MAX(RECID) "Log id now" FROM V\$LOG_HISTORY;
RECOVER AUTOMATIC DATABASE UNTIL CANCEL USING BACKUP CONTROLFILE
CANCEL
SELECT MAX(RECID) "Log id after recover" FROM V\$LOG_HISTORY;
exit;
EOF
echo ----------------------------------------------------------------
echo End of recovery process
echo ----------------------------------------------------------------
getrecid.sql
connect sys/change_on_install@&1 as sysdba
SET HEAD OFF FEEDBACK OFF VERIFY OFF TERMOUT ON ECHO OFF TRIMSPOOL ON SERVEROUTPUT OFF
SPOOL recid.log
SELECT MAX(RECID) FROM V$LOG_HISTORY;
SPOOL OFF
exitHTH building your own scripts.
Yoann.