最近几天都研究SHELL脚本，为了方便对公司的Oracle运维，简化管理，学习一些SHELL脚本是非常有必要的，通过书本和网上的一些资料，整理出了一些比较精典的脚本，都是经过清自测试可行的，放上来共大家分享。
# 监控Oracle监听状态(chk_lsnr_stat.sh)
# ======================================================================================
# 监控Oracle监听器状态，发现状态异常启动监听，并发送邮件通知管理员，如果启动监听失败，发送邮件
# 通知管理员。
# ======================================================================================
#! /bin/bash
. /home/oracle/.bash_profile
tempfile=$ORACLE_BASE/admin/$ORACLE_SID/tempfile.lis
su - oracle -c "lsnrctl status" > /dev/null
if [ $? != '0' ]; then
echo "" >> $tempfile
echo "======================================================" >> $tempfile
echo "`date +%D-%T`" >> $tempfile
su - oracle -c "lsnrctl start" >> $tempfile
if [ $? = '0' ]; then
cat $tempfile | mail dba@163.com -s "The Listener Shutdown,and Restarted Success"
else
cat $tempfile | mail dba@163.com -s "The Listener Shutdown,and Restarted Failed"
fi
fi
----------------------------------------------------------------------------------------
# 监控Oracle实例状态(chk_inst_stat.sh)
# =====================================================================================
# 监控Oracle实例是否打开，实例打开时，数据库是否可用，当实例关闭，数据库不可用时发送告警邮件
# 通知管理员(判断时除开+ASM这个特殊实例)
# =====================================================================================
#! /bin/bash
ORATAB=/etc/oratab
tempfile=/home/oracle/tempfile.lis
db=`cat $ORATAB |egrep -i ":Y|:N"|cut -d ":" -f 1|grep -v "^+"`
pslist="`ps -ef | grep pmon|grep -v grep`"
mark=n
dbstat=`su - oracle << EOF
sqlplus -s /nolog
conn / as sysdba
set feedback off heading off pagesize 0
select status from v\\$instance;
exit
EOF`
for db_name in $db; do
echo "$pslist" | grep "ora_pmon_$db_name" > /dev/null 2>&1
if ［ $? = "0" ］; then
if [ $dbstat != "OPEN" ];then
mark=y
break
fi
else
mark=y
break
fi
done
if [ $mark != 'n' ];then
echo '>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>' >> $tempfile
echo "SERVER: $HOSTNAME" >> $tempfile
echo "`date +%D-%T`" >> $tempfile
echo 'WARN!!! Oracle Database Unavailable' >> $tempfile
echo "Maybe The Following Reasons: The Instance or Database is not OPEN" >>$tempfile
echo | mail -s
"Oracle Database Abnormal" dba@163.com < $tempfile
rm -f $tempfile
fi
------------------------------------------------------------------------------------------
# 监控归档目录空间（chk_arc_space.sh）
#=====================================================================================
# 将日志目录空间控制在200M，当容量大于200M时，将最旧日志打包复制到其它目录，并删除之
# 直到日志目录空间容量小于200M为止。
#=====================================================================================
#! /bin/bash
ARC_DIR=/disk01/tbs03
BAK_DIR=/opt/arcbackup
limit=200
capacity()
{
du -sh $ARC_DIR|awk -F " " '{print $1}'|tr -d M
}
oldlog()
{
ls -l|sort -k6|sed '1d'|head -1|awk -F " " '{print $9}'
}
cd $ARC_DIR
if [ `capacity` -gt $limit ]; then
echo ""|mail -s "The Archivelog Directory is Over Than $limit"\
dba@163.com
fi
while [ `capacity` -gt $limit ]
do
file=`oldlog`
tar -czf $file.`date +%Y%m%d%H%M`.tar.gz $file
cp $file.`date +%Y%m%d%H%M`.tar.gz $BAK_DIR/
rm -rf $ARC_DIR/$file $ARC_DIR/$file.`date +%Y%m%d%H%M`.tar.gz
done
# 监控警告文件错误信息(chk_alert_info.sh)
#================================================================================
# 固定时间间隔内检查alert_$ORACLE_SID.log文件中是否包含ORA-开头的错误信息，如果存在
# 则将其以邮件的形式通知管理员。
#================================================================================
#! /bin/bash
. /home/oracle/.bash_profile
cd $ORACLE_BASE/admin/$ORACLE_SID/bdump/
mv alert_$ORACLE_SID.log alert_temp.log
touch alert_$ORACLE_SID.log
cat alert_temp.log >> alert.$ORACLE_SID.hist
grep ORA- alert_temp.log > alert.err
if [ `cat alert.err|wc -l` -gt 0 ];then
mail -s "ORACLE ALERT ERROR" dba@163.com < alert.err
fi
rm -rf alert.err
rm -rf alert_temp.log
# 监控磁盘空间利用率(chk_disk_space)
# ====================================================================
# 判断/dev开头的磁盘或分区空间利用率，当磁盘空间利用率超过90%则发送邮件通知
# 管理员
# ====================================================================
#! /bin/bash
limit=90%
tempfile=chk_disk_space.tmp
mark=n
diskusage()
{
df -h|grep -v Filesystem|sed '/\/dev\/mapper/N;s/\n//'|grep "^/dev"|awk -F " " '{print $5}'
}
for percent in `diskusage`
do
if [[ $percent > $limit]];then
mark=y
break
fi
done
if [ $mark != 'n' ];then
df -h > $tempfile
mail -s "Disk Usage Over than $limit on `hostname`" < $tempfile
rm -rf $tempfile
fi
# 监控表空间空闲表空间(chk_tbs_free.sh)
#==================================================================================
# 监控空闲表空间，当空闲表空间低于20%时，发送邮件通知管理员
#==================================================================================
#! /bin/bash
su - oracle > /dev/null << EOF
sqlplus -s /nolog
conn / as sysdba
set feedback off
set heading off
set verify off
set pagesize 0
set linesize 200
spool tbsfree.alert
select t.tablespace_name,f.free_space/t.total_space from
(select tablespace_name,sum(bytes) total_space from
dba_data_files group by tablespace_name) t,
(select tablespace_name,sum(bytes) free_space from
dba_free_space group by tablespace_name) f
where t.tablespace_name=f.tablespace_name and f.free_space/t.total_space < 0.20
/
spool off
exit
EOF
if [ `cat tbsfree.alert|wc -l` -gt 0 ];then
cat tbsfree.alert|mail -s "No Free Space in Oracle db" dba@163.com
rm -rf tbsfree.alert
fi
# 全库冷备份(full_cold_backup.sh)
# =================================================================================
# 数据库打开时，自动生成备份脚本。然后关闭数据库，对控制文件，数据文件，重做日志文件，
# 初始化参数文件及口令文件做冷备，完成后打开数据库。
# =================================================================================
#! /bin/bash
. /home/oracle/.bash_profile
backup_dir=/disk01/backup/coldbak
log_file=/disk01/backup/coldbak/cold_backup_$ORACLE_SID.log
echo 'Begin Cold Backup>>>>>>>>>>>>>>>>' >> $log_file
date >> $log_file
su - oracle > /dev/null << EOF
sqlplus -s /nolog
conn / as sysdba
set feedback off heading off pagesize 0 line 1000
spool file_copy_$ORACLE_SID.sh
select 'cp ' || name || ' $backup_dir/' from v$controlfile;
select 'cp' || file_name || ' $backup_dir/' from dba_data_files;
select 'cp' || member || ' $backup_dir/' from v$logfile;
spool off
shutdown immediate
! bash file_copy_$ORACLE_SID.sh
startup
exit
EOF
if [ -e $ORACLE_HOME/dbs/init$ORACLE_SID.ora ];then
cp $ORACLE_HOME/dbs/init$ORACLE_SID.ora $backup_dir/
fi
if [ -e $ORACLE_HOME/dbs/spfile$ORACLE_SID.ora ];then
cp $ORACLE_HOME/dbs/spfile$ORACLE_SID.ora $backup_dir/
fi
if [ -e $ORACLE_HOME/dbs/orapw$ORACLE_SID ];then
cp $ORACLE_HOME/dbs/orapw$ORACLE_SID $backup_dir/
fi
echo 'Cold Backup Finished>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>' >>$log_file
date >> $log_file
# RMAN备份SHELL脚本(rman_backup.sh)
#===========================================================================
# 实现指定级别的增量备份，由用户传入备份级别参数，如果不指参数则进行0级备份
#==========
=================================================================
#! /bin/bash
. /home/oracle/.bash_profile
if [ $1 ];then
backup_level=$1
else
backup_level=0
fi
backup_user=sys
backup_user_pw=oracle
#catalog_user=rman
#catalog_user_pw=rman
log_file=/home/oracle/rman_backup.log
echo 'Begining rman backup >>>>>>>>>>>>>>>>>>>>>>>>>>' >> $log_file
date >> $log_file
su - oracle >> $log_file << EOF
rman target $backup_user/$backup_user_pw
# catalog $catalog_user/$catalog_user_pw
backup incremental level = $backup_level database;
quit;
EOF
echo 'rman backup finished >>>>>>>>>>>>>>>>>>>>>>>>>' >> $log_file
date >> $log_file
# 逻辑备份SHELL脚本(schema_exp.sh)
#=========================================================================
# EXP对数据库schema对象进行备份，用户可以将需要备份的用户名做为参数传入SHELL脚本
#=========================================================================
#! /bin/bash
BAK_DIR=/disk01/backup/logical_bak/
log_file=/disk01/backup/logical_bak/user_full_bak.log
exp_par="userid=system/oracle buffer=10485760 owner=$1"
if [ $2 ];then
exp_par="$exp_par file=$2"
else
exp_par="$exp_par file="$BAK_DIR/$1_`date +%Y%m%d%H%M`.dmp""
fi
echo "Begining User $1 Export ---------------------" >> $log_file
echo "Export with following parameters: $exp_par" >> $log_file
date >> $log_file
su - oracle -c "exp $exp_par" >> $log_file 2>&1
echo "Backup Finished ---------------------" >> $log_file
date >> $log_file
