Script for getting mail if database is down

Question

Hi Friends,
OS Version : IBM AIX 5,2
Oracle version : 9.2.0.7
I am executing the following script for getting mail alert if database is down. some how the script is not working
check_stat=`ps -ef|grep ${ORACLE_SID}|grep pmon|wc -l`;
oracle_num=`expr $check_stat`
if [ $oracle_num -lt 1 ]
then
exit 0
fi
# Test to see if Oracle is accepting connections
$ORACLE_HOME/bin/sqlplus -s "/as sysdba" > /tmp/check_$ORACLE_SID.ora
select name from v$database;
exit
# If not, exit and e-mail . . .
check_stat=`cat /tmp/check_$ORACLE_SID.ora|grep -i error|wc -l`;
oracle_num=`expr $check_stat`
if [ $oracle_num -ne 0 ]
then
mailx -s "$ORACLE_SID is down!" [email protected] < /tmp/check_$ORACLE_SID.ora
exit 16
fi
I am saving this as .sh file and executing at command prompt. It is just hanging, but not throwing any error.
I would like to know if there is anything to be modified in the script or please provide me any such script. Thanks in advance

Answer 1

HI there.
I have a script I use that works really well. It sends out an email only if the database is down and also reads an ini file to process a blackout period and a priority level of the database... High priority databases are monitored every 5 minutes and Medium priority databases every hour.
There are two scripts, the shell script and the .ini file and I have two cron entries...
Check script:_
#!/bin/ksh
# check_oracle_status.sh
# Script to check if Oracle db's are up and running.
# Script is passed a priority field and reads check_oracle_status.ini
# to determine which db's to check. If db is down an email is sent.
# Priority Levels:
# H - Checks db's with "H"igh Priority every 5 minutes (cron)
# M - Checks db's with "M"edium Priority every hour (cron)
# L - db's with "L"ow Priority currently not checked
# Script Change History:
# ======================
# October 29th, 2009 - Initial Creation
# Set environment
export SCRIPTHOME=/opt/oracle/admin/scripts
export INIFILE=$SCRIPTHOME/check_oracle_status.ini
export PRIORITY=$1
. $HOME/.profile
db=`grep -i ":$PRIORITY" $INIFILE | cut -d":" -f1`
check_database()
sqlplus <<! > $SCRIPTHOME/check.out
/ as sysdba
select * from dba_data_files;
exit
grep ORA- $SCRIPTHOME/check.out > $SCRIPTHOME/error.out
if (( $? )); then
echo ""
else
mailx -s "Oracle instance $i is currently UNAVAILABLE" +<email address>+ < $SCRIPTHOME/error.out
fi
for i in $db ; do
fields=`grep $i $INIFILE | awk -F':' '{ total = total + NF }; END {print total}'`
export ORACLE_SID=$i
if [ $fields -gt 2 ]; then
BLACKOUT_START=`grep -i "$ORACLE_SID" $INIFILE | cut -d":" -f3`
BLACKOUT_END=`grep -i "$ORACLE_SID" $INIFILE | cut -d":" -f4`
CURRENT_HOUR=`date +%H`
CHECK_BASE=YES
if [ $BLACKOUT_START -gt $BLACKOUT_END ]; then
(( $CURRENT_HOUR >= $BLACKOUT_START || $CURRENT_HOUR <= $BLACKOUT_END )) && CHECK_BASE=
else
(( $CURRENT_HOUR >= $BLACKOUT_START && $CURRENT_HOUR <= $BLACKOUT_END )) && CHECK_BASE=
fi
if [ -n "$CHECK_BASE" ]; then
check_database
fi
else
check_database
fi
done
rm $SCRIPTHOME/check.out $SCRIPTHOME/error.out
INI File:_
oracle1:L
oracle2:M:17:08
oracle3:M
oracle5:M:17:08
oracle6:H
oracle7:M:17:08
oracle8:M
oracle9:M
Where oracle1,2,3 etc is your sid
L M and H your priority level
17 is blackout start (5 PM)
08 is blackout end (8 AM)
Note: Blackout is just a start hour and an end hour and must contain both or none and my script can only process one blackout per database. I guess if you
needed a second blackout you could add another line with different times for that sid
Cron entries:_
# Check Oracle Status
# The check_oracle_status.sh script monitors "H"igh priority databases every 5 minutes
# and "M"edium priority databases every hour
0,5,10,15,20,25,30,35,40,45,50,55 * * * * /opt/oracle/admin/scripts/check_oracle_status.sh H > /dev/null 2>&1
0 * * * * /opt/oracle/admin/scripts/check_oracle_status.sh M > /dev/null 2>&1
Not sure if you require blackouts or priority levels but this setup works great at our site.
Hope this helps.