提取mysql的error信息并报警，可以按下面步骤：
##变量设置##
mark=`date +%Y%m%d%H%M`
error_file=/data/mysql/data/error.log
bak_file=/tmp/error_bak.log
report_file=/tmp/error_rport.log

###没有error.bak就建一个，提取error字段信息到 ${report_file(这边用>>是为了假设后面发邮件失败，需要保留error信息)；读取${error_file}追加到${bak_file}(相当于备份)；清空${error_file}
[ ! -f ${bak_file} ] && touch ${bak_file}
[ ! -f ${report_file} ] && touch ${report_file}
cat ${error_file} | grep "\[ERROR\]" >> ${report_file}&& cat ${error_file} >> ${bak_file}&& > ${error_file}

###检查 ${report_file大小，如果大于0（说明有error信息），则发邮件，然后清空 ${report_file}
logsize= `du -b ${report_file} |awk '{print $1}'`
[ $logsize -gt 0 ] &&/usr/local/bin/sendEmail ...&& > ${report_file}
上面其实也可以简化为[ -s ${logfile} ] &&/usr/local/bin/sendEmail ...&& > ${report_file}


实时提取Mysql的慢日志并报警，可以按下面步骤：
##变量设置##
mark=`date +%Y%m%d%H%M`
process=/tmp/showprocesslist.log
slow=/tmp/slow.log

##输出processlis到${process}。提取time>5的查询信息到${slow}，用>>是为了假设后面发邮件失败，需要保留slow信息
/usr/local/mysql/bin/mysql -usystem -p'itiaoling.com@123' -e "show full processlist;" | sed 's/\\t\|\\n/ /g' > ${process}
cat ${process} | awk '($6>=5) && ($6<10) && ($5=="Query") {printf $4" "$7" for "$6" seconds\nDetails:\n"$0}' >> ${slow}
cat ${process} | awk '($6>=60) && ($6<65) && ($5=="Query") {printf $4" "$7" for "$6" seconds\nDetails:\n"$0}' >> ${slow}

##检查${slow}，如果存在且非空，则发邮件，然后清空 ${slow}
[ -s ${slow} ] &&/usr/local/bin/sendEmail ...&& > ${slow}


监控备份，备份失败报警
##变量设置##
mark=`date +%Y%m%d%H%M`
check=/tmp/check.log
backfile=/backup/crm/ncrm_nobh_antdb_wechat_tiaoling$mark.sql.gz

(/usr/local/mysql/bin/mysqldump -uroot -pxx --default-character-set=utf8 --master-data=1 --single-transaction -R -B test || echo "backup not completed" > ${check}) | (gzip || echo "gzip not completed" > ${check}) > ${backfile}
[ -s ${check} ] &&/usr/local/bin/sendEmail ...&&> ${check}