Database Monitoring

Question

Hi guys,
I am pretty new into Oracle Database Administration. I am currently tasked with database monitoring and I would like to ask you guys what parameters or aspects of database we have to constantly monitor? I know for sure we have to monitor the harddisk space, tablespace free space and latch. However, apart from those parameters, anything else I have to pay attention to? And what are the threshold for each parameter? For example, what is the good percentage of freespace to determine when to expand a tablespace?
I would really appreciate your view as I am completely clueless. I have checked some books on this matter but most of them jump straight to the finetuning without revealing any detail on the monitoring parameter, threshold, etc.
My situation is worse since I am the only DBA and the only one competent in it (*sigh*).
Thank you.
Richard Siburian

Answer 1

just study following and try to grab
#Current working directory
cwd=`pwd`
#DBA to whome mail will be sent
DBALIST="[email protected]";
while true
do
clear
echo "**************************************************************************"
echo "* MENU *"
echo "**************************************************************************"
echo "* ( 1) Check status of Listener !! *"
echo "* ( 2) Check status of all Database whose entry in oratab *"
echo "* ( 3) Display database uptime (days and hours) *"
echo "* ( 4) Check the ORA- Error in alert log and mail it to dba *"
echo "* ( 5) Displays information on all database sessions. *"
echo "* ( *) Press any other key to Go back to main menu *"
echo "**************************************************************************"
echo "Enter Your Choice from Menu between 1 and 6 :-"
read choice
case $choice in
1)
echo "checking for the listiner"
listiner_up=`ps -ef | grep LISTENER | grep -v grep |awk '{print$9}'|wc -l`     
if [ $listiner_up -ge 2 ]
then
     echo "listiner is up, Have a Great time......"
while true
do
echo " Do you want to Stop the Listiner? (y or n) :"
read CONFIRM1
case $CONFIRM1 in
y|Y|YES|yes|Yes)
lsnrctl stop;
callme;
break ;;
n|N|no|NO|No)
echo Aborting - you entered $CONFIRM
callme;
break
*) echo Please enter only y or n
esac
done
     else
     echo "Listiner is Down, ."
while true
do
echo " Do you want to Start the Listiner? (y or n) :"
read CONFIRM1
case $CONFIRM1 in
y|Y|YES|yes|Yes)
lsnrctl start;
callme;
break ;;
n|N|no|NO|No)
echo Aborting - you entered $CONFIRM
callme;
break
*) echo Please enter only y or n
esac
done
fi
2)
#Script By DCP
#Check status of all Database whose entry in oratab
clear;
if [ -f /etc/oratab ]
then
ORATAB=/etc/oratab
elif [ -f /var/opt/oracle/oratab ]
then
ORATAB=/var/opt/oracle/oratab
else
echo "Unable to locate "
fi
echo "`date` "
echo "Oracle Database(s) Status `hostname` "
db=`egrep -i ":Y|:N" $ORATAB | cut -d":" -f1 | grep -v "\#" | grep -v "\*"`
for i in $db ; do
database_up=`ps -ef | grep ora_smon_$i|grep -v grep |awk '{print$8}'|wc -l`
if [ $database_up -eq 1 ]
then
     echo "Oracle Instance - $i : UP"
else
echo "Oracle Instance - $i : Down"
fi
done
while true
do
echo "Do you want to Continue or not Please confirm ? (y or n) :"
read CONFIRM
case $CONFIRM in
y|Y|YES|yes|Yes) break ;;
n|N|no|NO|No)
echo Aborting - you entered $CONFIRM
exit
*) echo Please enter only y or n
esac
done
echo You entered $CONFIRM. Continuing ...;;
3)
#Script By DCP
#Display database uptime (days and hours)
clear;
echo "---------------------------------------------------"
echo "`date` "
echo "Oracle Database(s) Status `hostname` "
echo "---------------------------------------------------"
Live=`ps -ef | grep ora_smon | grep -v grep |awk '{print$8}'|cut -d "_" -f2-3 >$cwd/live.txt`
db=`egrep -i "smon" $cwd/live.txt |cut -d "_" -f2`
for i in $db ; do
database_up=`ps -ef | grep ora_smon_$i|grep -v grep |awk '{print$8}'|wc -l`
if [ $database_up -eq 1 ]
then
     echo "Oracle Instance - $i : UP"
fi
echo "---------------------------------------------------"
done
echo "Enter the Database name:- "
read ORACLE_SID
db=`ps -ef | grep ora_smon_$ORACLE_SID|grep -v grep |awk '{print$8}'|cut -d "_" -f3`
if [ "$db" = "$ORACLE_SID" ]
then
export ORACLE_SID
sqlplus -s "/ as sysdba" <<EOF
set feed off
set linesize 280
set pagesize 200
select SYSDATE-logon_time "Days", (SYSDATE-logon_time)*24 "Hours"
from sys.v_\$session
where program like '%PMON%';
exit;
EOF
else
echo "Enter the Valid Name or DB is shudown"
fi
while true
do
echo "Do you want to Continue or not Please confirm ? (y or n) :"
read CONFIRM
case $CONFIRM in
y|Y|YES|yes|Yes) break ;;
n|N|no|NO|No)
echo Aborting - you entered $CONFIRM
clear
exit
echo Please enter only y or n
esac
done
echo You entered $CONFIRM. Continuing ...
4)
#Script By DCP
# Check the ORA- Error in alert log and mail it to dba
clear;
echo "---------------------------------------------------"
echo "`date` "
echo "Oracle Database(s) Status `hostname` "
echo "---------------------------------------------------"
Live=`ps -ef | grep ora_smon | grep -v grep |awk '{print$8}'|cut -d "_" -f2-3 >$cwd/live.txt`
db=`egrep -i "smon" $cwd/live.txt |cut -d "_" -f2`
for i in $db ; do
database_up=`ps -ef | grep ora_smon_$i|grep -v grep |awk '{print$8}'|wc -l`
if [ $database_up -eq 1 ]
then
     echo "Oracle Instance - $i : UP"
fi
echo "---------------------------------------------------"
done
echo "---------------------------------------------------------------------------"
echo Alert \!!! This part is hardcoded, PLZ Modify it , just change Value of DLOC
echo "---------------------------------------------------------------------------"
sleep 2;
DLOC=/u01/oracle/product/10.2.0/test/test/bdump
while true
do
echo "Enter the database name: "
read SID
echo "Checking for alert log of $SID database"
case $SID in
$SID)
cd $DLOC
if [ -f alert_${SID}.log ]
then
echo Moving....................
     mv /$DLOC/alert_${SID}.log /$DLOC/alert_work.log
     echo Touching................
touch /$DLOC/alert_${SID}.log
cat /$DLOC/alert_work.log >> /$DLOC/alert_${SID}.hist
echo greping.....................
grep ORA- /$DLOC/alert_work.log > /$DLOC/alert.err
if [ `cat /$DLOC/alert.err|wc -l` -gt 0 ]
then
mail -s "${SID} ORACLE ALERT ERRORS" $DBALIST < /$DLOC/alert.err
     echo Sending Mail.................
     else
     echo "No Error from last ......Nothing to Mail to DBA...."
fi
rm -f /$DLOC/alert.err
rm -f /$DLOC/alert_work.log
while true
do
echo "Do you want to Continue or not Please confirm ? (y or n) :"
read CONFIRM
case $CONFIRM in
y|Y|YES|yes|Yes) break ;;
n|N|no|NO|No)
echo Aborting - you entered $CONFIRM
clear
exit
echo Please enter only y or n
esac
done
echo You entered $CONFIRM. Continuing ...
else
echo " "
echo "No alert Log file found......Nothing to Mail to DBA...."
echo " "     
callme;
fi
break ;;
n|N|no|NO|No)
echo Aborting - you entered $CONFIRM
exit
*) echo Please enter only y or n
esac
done
5)
#Script By DCP
#Displays information on all database sessions
clear;
echo "---------------------------------------------------"
echo "`date` "
echo "Oracle Database(s) Status `hostname` "
echo "---------------------------------------------------"
Live=`ps -ef | grep ora_smon | grep -v grep |awk '{print$8}'|cut -d "_" -f2-3 >$cwd/live.txt`
db=`egrep -i "smon" $cwd/live.txt |cut -d "_" -f2`
for i in $db ; do
database_up=`ps -ef | grep ora_smon_$i|grep -v grep |awk '{print$8}'|wc -l`
if [ $database_up -eq 1 ]
then
     echo "Oracle Instance - $i : UP"
fi
echo "---------------------------------------------------"
done
echo "Enter the Database name:- "
read ORACLE_SID
db=`ps -ef | grep ora_smon_$ORACLE_SID|grep -v grep |awk '{print$8}'|cut -d "_" -f3`
if [ "$db" = "$ORACLE_SID" ]
then
export ORACLE_SID
sqlplus -s "/ as sysdba" <<EOF
spool $cwd/curses.log
#set feed off
#set heading off
set line 180
SET LINESIZE 500
SET PAGESIZE 1000
SET VERIFY OFF
COLUMN username FORMAT A15
COLUMN machine FORMAT A25
COLUMN logon_time FORMAT A20
SELECT NVL(a.username, '(oracle)') AS username,
a.osuser,
c.value ,
a.status,
a.module,
a.machine,
a.program,
TO_CHAR(a.logon_Time,'DD-MON-YYYY HH24:MI:SS') AS logon_time
FROM v\$session a,
v\$sesstat c,
v\$statname d
WHERE a.sid = c.sid
AND c.statistic#
= d.statistic#
AND d.name = DECODE(UPPER('value'), 'READS', 'session logical reads',
'EXECS', 'execute count',
'CPU', 'CPU used by this session',
'CPU used by this session')
AND USERNAME IS NOT NULL
ORDER BY c.value DESC;
SET PAGESIZE 14
spool off;
exit;
EOF
if [ `cat $cwd/curses.log|wc -l` -gt 0 ]
then
cat $cwd/curses.log
else
echo "Oppppppssss....No Active user session"
fi
else
echo "Enter the Valid Name or DB is shudown"
fi
while true
do
echo "Do you want to Continue or not Please confirm ? (y or n) :"
read CONFIRM
case $CONFIRM in
y|Y|YES|yes|Yes) break ;;
n|N|no|NO|No)
echo Aborting - you entered $CONFIRM
clear
exit
echo Please enter only y or n
esac
done
echo You entered $CONFIRM. Continuing ...
#hi
clear
exit
esac
done