Oracle 主从切换脚本

Hehuyi_In2024-03-14 18:59

一、 切换前预检查

1. dg_precheck_main_v1.4.sh

bash 复制代码 
#!/bin/bash

#**********************************************************************************
# Author: Hehuyi_In
# Date: 2022年06月16日
# FileName: dg_precheck_main_v1.4.sh
# 
# For sys user, execute the script directly. 
# For other sysdba privileged users, you need to manually synchronize the password file to standby server in versions below 12.2.
# The script cannot be executed on the Linux 5 os.
#
# Example: 
# ./dg_precheck_main_v1.4.sh &> dg_precheck.log
# cat dg_precheck.log | grep --color=no -E "Warning|Error" ,or less -R dg_precheck.log
#
#
# Description:
# 2022-06-16 v_1.0 just get and print the infomation we want
# 2022-06-17 v_1.1 auto check the database role
# 2022-06-22 v_1.2 adjust the output format
# 2022-06-28 v_1.3 add compare function, compare two nodes at a time
# 2022-06-28 split to two scripts dg_precheck_main_v1.0.sh and dg_precheck_func_v1.0.sh
# 2022-07-06 dg_precheck_main_v1.1 check os info
# 2022-07-12 dg_precheck_main_v1.2 adjust the output format
# 2022-07-13 dg_precheck_main_v1.3 adjust os info check function
# 2022-10-19 dg_precheck_main_v1.4 add checking items: dblink,standby_file_management,flashback_on,db_flashback_retention_target,db_recovery_file_dest,db_recovery_file_dest_size 
#**********************************************************************************

source ./dg_precheck_func_v1.4.sh

# TNS列表
TNS_LIST=(test_dg test)

# 连接信息
DBUSER='sys/"xxxxx"'


# --------------------------- 检查并设置TNS_LIST中各tnsname对应数据库角色 ----------------------------------
auto_check_db_role

# 字典定义
declare -A PRIMARY_TNS_DIC
declare -A STANDBY_TNS_DIC


echo -e "--------------------------------------------------------------------------------"
echo -e "- 各节点主要参数检查                                                "
echo -e "--------------------------------------------------------------------------------\n"

# 内存参数 及 compatible 检查和对比
PARAMS_LIST=(sga_max_size sga_target pga_aggregate_target shared_pool_size db_cache_size streams_pool_size compatible standby_file_management)

for PARAM in ${PARAMS_LIST[@]};
do

PRIMARY_TNS_DIC+=([$PARAM]=`get_params_value $PRIMARY_TNS $PARAM`)
STANDBY_TNS_DIC+=([$PARAM]=`get_params_value $STANDBY_TNS $PARAM`)

# 主从库间对比参数是否相等
compare_value ${PRIMARY_TNS_DIC[$PARAM]} ${STANDBY_TNS_DIC[$PARAM]} $PARAM

done

# 获取其他参数值
PARAMS_LIST=(log_archive_max_processes db_create_file_dest db_file_name_convert log_file_name_convert fal_client fal_server log_archive_config db_recovery_file_dest db_recovery_file_dest_size db_flashback_retention_target)

for PARAM in ${PARAMS_LIST[@]};
do

PRIMARY_TNS_DIC+=([$PARAM]=`get_params_value $PRIMARY_TNS $PARAM`)
STANDBY_TNS_DIC+=([$PARAM]=`get_params_value $STANDBY_TNS $PARAM`)

done

echo -e "\n"

echo -e "--------------------------------------------------------------------------------"
echo -e "- 各节点主要文件数检查                                                          "
echo -e "--------------------------------------------------------------------------------\n"

PRIMARY_TNS_DIC+=([redo_count]=`get_file_count $PRIMARY_TNS "LOG"`)
PRIMARY_TNS_DIC+=([standby_count]=`get_file_count $PRIMARY_TNS "STANDBY_LOG"`)
PRIMARY_TNS_DIC+=([tempfile_count]=`get_file_count $PRIMARY_TNS "TEMPFILE"`)
STANDBY_TNS_DIC+=([redo_count]=`get_file_count $STANDBY_TNS "LOG"`)
STANDBY_TNS_DIC+=([standby_count]=`get_file_count $STANDBY_TNS "STANDBY_LOG"`)
STANDBY_TNS_DIC+=([tempfile_count]=`get_file_count $STANDBY_TNS "TEMPFILE"`)

PRIMARY_TNS_DIC+=([datafile_online_count]=`get_datafile_online_count $PRIMARY_TNS`)
STANDBY_TNS_DIC+=([datafile_online_count]=`get_datafile_online_count $STANDBY_TNS`)

compare_value ${PRIMARY_TNS_DIC[redo_count]} ${STANDBY_TNS_DIC[redo_count]} "redo_count"
compare_value ${PRIMARY_TNS_DIC[standby_count]} ${STANDBY_TNS_DIC[standby_count]} "standby_count"
compare_value ${PRIMARY_TNS_DIC[tempfile_count]} ${STANDBY_TNS_DIC[tempfile_count]} "tempfile_count"
compare_value ${PRIMARY_TNS_DIC[datafile_online_count]} ${STANDBY_TNS_DIC[datafile_online_count]} "datafile_online_count"

echo -e "\n"

echo -e "--------------------------------------------------------------------------------"
echo -e "- 从节点redo日志状态检查                                                        "
echo -e "--------------------------------------------------------------------------------\n"

# 只有从库需要检查
STANDBY_TNS_DIC+=([redo_status_count]=`get_redo_status_count $STANDBY_TNS`)
check_redo_status_count ${STANDBY_TNS_DIC[db_create_file_dest]} ${STANDBY_TNS_DIC[log_file_name_convert]} ${STANDBY_TNS_DIC[redo_status_count]} 
echo -e "\n"


echo -e "--------------------------------------------------------------------------------"
echo -e "- 闪回开启及保留时间检查                                                        "
echo -e "--------------------------------------------------------------------------------\n"

# --------------------------- flashback_on 与 db_flashback_retention_target -------------------------------

PRIMARY_TNS_DIC+=([flashback_on]=`get_flashback_on $PRIMARY_TNS`)
echo -e "- primary database                                                           "
compare_value ${PRIMARY_TNS_DIC[flashback_on]} "NO" "flashback_on"
compare_value ${PRIMARY_TNS_DIC[db_flashback_retention_target]} "4320" "db_flashback_retention_target"

STANDBY_TNS_DIC+=([flashback_on]=`get_flashback_on $STANDBY_TNS`)
echo -e "- standby database                                                                  "
compare_value ${STANDBY_TNS_DIC[flashback_on]} "YES" "flashback_on"
compare_value ${STANDBY_TNS_DIC[db_flashback_retention_target]} "4320" "db_flashback_retention_target"

echo -e "\n"

echo -e "--------------------------------------------------------------------------------"
echo -e "- FRA区设置检查                                                                   "
echo -e "--------------------------------------------------------------------------------\n"

# --------------------------- db_recovery_file_dest 与 db_recovery_file_dest_size -------------------------------
# 主从库设置是否相同
compare_value ${PRIMARY_TNS_DIC[db_recovery_file_dest]} ${STANDBY_TNS_DIC[db_recovery_file_dest]} "db_recovery_file_dest"
compare_value ${PRIMARY_TNS_DIC[db_recovery_file_dest_size]} ${STANDBY_TNS_DIC[db_recovery_file_dest_size]} "db_recovery_file_dest_size"

echo -e "\n"


echo -e "--------------------------------------------------------------------------------"
echo -e "- 各节点主从相关参数检查                                                        "
echo -e "--------------------------------------------------------------------------------\n"

# --------------------------- fal_client 与 fal_server ---------------------------

# fal_server的值应该包括对方，因为${PRIMARY_TNS_DIC[fal_server]}返回值包含逗号，因此需要在传参时加""，避免参数传入函数后被截断

# 主节点参数值
include_value  $PRIMARY_TNS "fal_server" "${PRIMARY_TNS_DIC[fal_server]}" $STANDBY_TNS
# 从节点参数值
include_value  $STANDBY_TNS "fal_server" "${STANDBY_TNS_DIC[fal_server]}" $PRIMARY_TNS

# fal_client的值应该包含自己
# 主节点参数值
include_value  $PRIMARY_TNS "fal_client" "${PRIMARY_TNS_DIC[fal_client]}" $PRIMARY_TNS
# 从节点参数值
include_value  $STANDBY_TNS "fal_client" "${STANDBY_TNS_DIC[fal_client]}" $STANDBY_TNS


# --------------------------- log_archive_config -------------------------------
# 主节点参数值
include_value $PRIMARY_TNS "log_archive_config" ${PRIMARY_TNS_DIC[log_archive_config]} $STANDBY_TNS
# 从节点参数值
include_value $STANDBY_TNS "log_archive_config" ${STANDBY_TNS_DIC[log_archive_config]} $PRIMARY_TNS

# --------------------------- log_archive_max_processes -------------------------------
check_log_archive_max_processes $PRIMARY_TNS ${PRIMARY_TNS_DIC[log_archive_max_processes]} 
check_log_archive_max_processes $STANDBY_TNS ${STANDBY_TNS_DIC[log_archive_max_processes]} 

# --------------------------- log_archive_dest_n 与 log_archive_dest_state_n -------------------------------

check_archive_dest $PRIMARY_TNS
check_archive_dest $STANDBY_TNS

check_archive_dest_state $PRIMARY_TNS
check_archive_dest_state $STANDBY_TNS

# --------------------------- OMF 与 db_file_name_convert,log_file_name_convert -------------------------------

check_omf_and_convert_params $PRIMARY_TNS ${PRIMARY_TNS_DIC[db_create_file_dest]} ${PRIMARY_TNS_DIC[db_file_name_convert]} ${PRIMARY_TNS_DIC[log_file_name_convert]}
check_omf_and_convert_params $STANDBY_TNS ${STANDBY_TNS_DIC[db_create_file_dest]} ${STANDBY_TNS_DIC[db_file_name_convert]} ${STANDBY_TNS_DIC[log_file_name_convert]}

echo -e "\n"


echo -e "--------------------------------------------------------------------------------"
echo -e "- 从节点延迟 MRP进程与触发器检查                                                "
echo -e "--------------------------------------------------------------------------------\n"

check_db_lag $STANDBY_TNS "transport lag"
check_db_lag $STANDBY_TNS "apply lag"

check_mrp_process $STANDBY_TNS

check_mrp_trigger $PRIMARY_TNS
check_mrp_trigger $STANDBY_TNS
echo -e "\n"

echo -e "--------------------------------------------------------------------------------"
echo -e "- dblink检查                                                                   "
echo -e "--------------------------------------------------------------------------------\n"

# --------------------------- dblink -------------------------------
# 检查主库即可
check_dblink $PRIMARY_TNS
echo -e "\n"

2. dg_precheck_func_v1.4.sh

bash 复制代码 
#!/bin/bash

#**********************************************************************************
# Author: Hehuyi_In
# Date: 2022年06月16日
# FileName: dg_precheck_func_v1.4.sh
# 
# Description:
# 2022-06-16 v_1.0 just get and print the infomation we want
# 2022-06-17 v_1.1 auto check the database role
# 2022-06-22 v_1.2 adjust the output format
# 2022-06-28 v_1.3 add compare function, compare two nodes at a time
# 2022-06-28 split to two scripts dg_precheck_main_v1.0.sh and dg_precheck_func_v1.0.sh
# 2022-07-06 dg_precheck_func_v1.1 check os info
# 2022-07-12 dg_precheck_func_v1.2 adjust the output format
# 2022-07-13 dg_precheck_func_v1.3 adjust os info check function
# 2022-10-19 dg_precheck_func_v1.4 add checking items: dblink,standby_file_management,flashback_on,db_flashback_retention_target,db_recovery_file_dest,db_recovery_file_dest_size 
#**********************************************************************************

# 颜色定义
# 红色
ERROR_COLOR='\e[1;31m'
# 绿色
NARMAL_COLOR='\e[1;32m'
# 黄色
WARNING_COLOR='\e[1;33m'
# 重置
RESET_COLOR='\e[0m'

# ------------------------------------------------------------------------
# 检查TNS_LIST中各tnsname对应数据库角色
auto_check_db_role() {
    for TNS in ${TNS_LIST[@]}; do
        DB_ROLE=$(
            sqlplus -s $DBUSER@$TNS as sysdba <<EOF

set heading off feedback off pagesize 0 verify off echo off
select database_role from v\$database;
EOF
        )
        # 主库，则赋值给 PRIMARY_TNS
        if [[ "$DB_ROLE" = "PRIMARY" ]]; then
            PRIMARY_TNS=$TNS
            # 从库，则赋值给 STANDBY_TNS
        elif [[ "$DB_ROLE" = "PHYSICAL STANDBY" ]]; then
            STANDBY_TNS=$TNS
        fi        
    done

    if [[ "$PRIMARY_TNS" = "" ]]; then
       echo -e "${ERROR_COLOR} $(date "+%Y-%m-%d %H:%M:%S") | Error! There is no primary database tns ${RESET_COLOR}"
       exit 1
    elif [[ "$STANDBY_TNS" = "" ]]; then
       echo -e "${ERROR_COLOR} $(date "+%Y-%m-%d %H:%M:%S") | Error! There is no standby database tns ${RESET_COLOR}"
       exit 1
    fi
}

# ----------------------------------获取操作系统信息--------------------------------------
# 参数1：TNS连接串

# get_cpu_info() {
#     sqlplus -s $DBUSER@$1 as sysdba <<EOF
# !cat /proc/cpuinfo | grep "processor" | wc -l
# EOF
# }

# get_memory_info() {
#     sqlplus -s $DBUSER@$1 as sysdba <<EOF
# !free -h | grep 'Mem:' | awk '{print \$2}'
# EOF
# }

# get_disk_info() {
#     sqlplus -s $DBUSER@$1 as sysdba <<EOF
# !df -m | grep -v 'Filesystem' | awk '{sum+=\$2} END {print sum}'
# EOF
# }

# 获取Oracle参数值
# 参数1：TNS连接串，参数2：待检查参数名（不区分大小写）

get_params_value() {
    if [[ "$1" != "" && "$2" != "" ]]; then
        sqlplus -s $DBUSER@$1 as sysdba <<EOF
set heading off feedback off pagesize 0 verify off echo off
select value from v\$parameter where upper(name)=upper('$2');
EOF
    else
        echo -e "${ERROR_COLOR} $(date "+%Y-%m-%d %H:%M:%S") | Error! get_params_value() parameter 1 and 2 can't be null ${RESET_COLOR}"
        exit 1
    fi
}

# log_archive_max_processes 参数检查
# 参数1：TNS连接串，参数2：待检查参数值

check_log_archive_max_processes() {
    if [[ "$1" != "" && "$2" != "" ]]; then
        if [[ "$2" -ge 4 ]]; then
            printf "${NARMAL_COLOR}%-10s | Success | log_archive_max_processes >= 4${RESET_COLOR}\n" $1
        else
            printf "${WARNING_COLOR}%-10s | Warning | log_archive_max_processes < 4${RESET_COLOR}\n" $1
        fi
    else
        echo "${ERROR_COLOR} $(date "+%Y-%m-%d %H:%M:%S") | Error! check_log_archive_max_processes() parameter 1 and 2 can't be null ${RESET_COLOR}"
        exit 1
    fi
}

# 检查log_archive_dest_n参数，是否有设置DELAY关键字（延迟从库）
# 参数1：TNS连接串

check_archive_dest() {
    if [[ "$1" != "" ]]; then
            COUNT=$(
        sqlplus -s $DBUSER@$1 as sysdba <<EOF
set heading off feedback off pagesize 0 verify off echo off
select count(*) FROM v\$parameter where name like 'log_archive_dest%' and upper(value) like '%DELAY%';
EOF
       )
        if [[ "$COUNT" -eq 0 ]]; then
            printf "${NARMAL_COLOR}%-10s | Success | There is no DELAY attribute in log_archive_dest_n${RESET_COLOR}\n" $1
        else
            printf "${WARNING_COLOR}%-10s | Warning | There is DELAY attribute in log_archive_dest_n${RESET_COLOR}\n" $1
        fi
    else
        echo -e "${ERROR_COLOR} $(date "+%Y-%m-%d %H:%M:%S") | Error! check_archive_dest() parameter 1 can't be null ${RESET_COLOR}"
        exit 1
    fi
}

# 检查log_archive_dest_state_n参数，是否有设置DEFER
# 参数1：TNS连接串

check_archive_dest_state() {
    if [[ "$1" != "" ]]; then
            COUNT=$(
        sqlplus -s $DBUSER@$1 as sysdba <<EOF
set heading off feedback off pagesize 0 verify off echo off
select count(*) FROM v\$parameter where name like 'log_archive_dest_state%' and upper(value)='DEFER';
EOF
       )
        if [[ "$COUNT" -eq 0 ]]; then
            printf "${NARMAL_COLOR}%-10s | Success | There is no DEFER value in log_archive_dest_state_n${RESET_COLOR}\n" $1 
        else
            printf "${WARNING_COLOR}%-10s | Warning | There is DEFER value in log_archive_dest_state_n${RESET_COLOR}\n" $1
        fi
    else
        echo -e "${ERROR_COLOR} $(date "+%Y-%m-%d %H:%M:%S") | Error! check_archive_dest_state() parameter 1 can't be null ${RESET_COLOR}"
        exit 1
    fi
}

# OMF与convert参数检查
# 参数1：TNS连接串；参数2：db_create_file_dest参数值；参数3：db_file_name_convert参数值；参数4：log_file_name_convert参数值

check_omf_and_convert_params() {
    if [[ "$2" != ""  ]]; then
             printf "${NARMAL_COLOR}%-10s | Success | db_create_file_dest parameter was set${RESET_COLOR}\n" $1
    else
        if [[ "$3" != "" && "$4" != "" ]]; then
            printf "${NARMAL_COLOR}%-10s | Success | dbfile and logfile convert parameters were set${RESET_COLOR}\n" $1
        else
            printf "${WARNING_COLOR}%-10s | Warning | db_create_file_dest,dbfile and logfile convert parameters were set${RESET_COLOR}\n" $1
        fi
    fi
}

# 获取各类文件数 
# 参数1：TNS连接串，参数2：待查询视图名（V$LOG V$STANDBY_LOG V$TEMPFILE）

get_file_count() {
    if [[ "$1" != "" ]]; then
        sqlplus -s $DBUSER@$1 as sysdba <<EOF
set heading off feedback off pagesize 0 verify off echo off
SELECT count(*) FROM V\$$2;
EOF
    else
        echo -e "${ERROR_COLOR} $(date "+%Y-%m-%d %H:%M:%S") | Error! get_file_count() parameter 1 and 2 can't be null ${RESET_COLOR}"
        exit 1
    fi
}

# 获取指定状态的redo文件数
# 参数1：TNS连接串

get_redo_status_count() {
    if [[ "$1" != "" ]]; then
        sqlplus -s $DBUSER@$1 as sysdba <<EOF
set heading off feedback off pagesize 0 verify off echo off
SELECT count(*) FROM V\$LOG WHERE STATUS NOT IN ('UNUSED', 'CLEARING','CLEARING_CURRENT');
EOF
    else
        echo -e "${ERROR_COLOR} $(date "+%Y-%m-%d %H:%M:%S") | Error! get_redo_status_count() parameter 1 can't be null ${RESET_COLOR}"
        exit 1
    fi
}

# 检查指定状态的redo文件数
# 参数1：db_create_file_dest参数值；参数2：log_file_name_convert参数值；参数3：${STANDBY_TNS_DIC[redo_status_count]} 

check_redo_status_count() {
    # 如果设置了OMF或者LOG_FILE_NAME_CONVERT，则可以跳过此步骤
    if [[ "$1" != "" || "$2" != "" ]]; then
            printf "${NARMAL_COLOR}%-30s | %-8s | There is no need to check this item${RESET_COLOR}\n" redo_status_count Success
    else
        # 否则，指定状态的redo文件数应该为0
        if [[ "$3" -eq 0 ]]; then
            printf "${WARNING_COLOR}%-30s | %-8s | redo_status_count value is 0${RESET_COLOR}\n" redo_status_count Warning
        else
            # 否则，发出告警
            printf "${WARNING_COLOR}%-30s | %-8s | redo_status_count value is not 0${RESET_COLOR}\n" $1 Warning
        fi
    fi
}

# 获取online状态的数据文件数量
# 参数1：TNS连接串

get_datafile_online_count() {
    if [[ "$1" != "" ]]; then
        sqlplus -s $DBUSER@$1 as sysdba <<EOF
set heading off feedback off pagesize 0 verify off echo off
SELECT count(*) FROM V\$DATAFILE WHERE STATUS='ONLINE';
EOF
    else
        echo -e "${ERROR_COLOR} $(date "+%Y-%m-%d %H:%M:%S") | Error! get_datafile_online_count() parameter 1 can't be null ${RESET_COLOR}"
        exit 1
    fi
}

# 获取数据库闪回功能开启状态
# 参数1：TNS连接串
get_flashback_on() {
    if [[ "$1" != "" ]]; then
        sqlplus -s $DBUSER@$1 as sysdba <<EOF
set heading off feedback off pagesize 0 verify off echo off
SELECT FLASHBACK_ON FROM V\$DATABASE;
EOF
    else
        echo -e "${ERROR_COLOR} $(date "+%Y-%m-%d %H:%M:%S") | Error! get_flashback_on() parameter 1 can't be null ${RESET_COLOR}"
        exit 1
    fi
}

# mrp触发器检查
# 参数1：TNS连接串

check_mrp_trigger() {
    if [[ "$1" != "" ]]; then
        COUNT=$(
            sqlplus -s $DBUSER@$1 as sysdba <<EOF
set heading off feedback off pagesize 0 verify off echo off
select count(*) FROM dba_triggers where trigger_name in ('AUTO_START_STANDBY_MRP','AUTO_START_STANDBY_MRP_PDB');
EOF
       )
        if [[ "$COUNT" -eq 0 ]]; then
            printf "${ERROR_COLOR}%-10s | %-8s | There is no MRP trigger${RESET_COLOR}\n" $1 Error
        else
            printf "${NARMAL_COLOR}%-10s | %-8s | MRP trigger was found${RESET_COLOR}\n" $1 Success
        fi
    else
        echo -e "${ERROR_COLOR}$(date "+%Y-%m-%d %H:%M:%S") | Error! check_mrp_trigger() parameter 1 can't be null ${RESET_COLOR}"
        exit 1
    fi
}

# mrp进程检查
# 参数1：TNS连接串

check_mrp_process() {
    if [[ "$1" != "" ]]; then
        COUNT=$(
            sqlplus -s $DBUSER@$1 as sysdba <<EOF
set heading off feedback off pagesize 0 verify off echo off
select count(*) FROM v\$managed_standby where process like 'MRP%';
EOF
        )
        if [[ "$COUNT" -eq 0 ]]; then
            printf "${ERROR_COLOR}%-10s | %-8s | There is no MRP process${RESET_COLOR}\n" $1 Error
        else
            printf "${NARMAL_COLOR}%-10s | %-8s | MRP process was found${RESET_COLOR}\n" $1 Success
        fi

    else
        echo -e "${ERROR_COLOR} $(date "+%Y-%m-%d %H:%M:%S") | Error! check_mrp_process() parameter 1 can't be null ${RESET_COLOR}"
        exit 1
    fi
}

# 检查从库延迟（transport lag,apply lag ）
# 参数1：TNS连接串，参数2：待检查的lag类型

check_db_lag(){
if [[ "$1" != "" && "$2" != "" ]]	
then
DB_LAG=`sqlplus -s $DBUSER@$1 as sysdba << EOF 

set heading off feedback off pagesize 0 verify off echo off
select (extract(second from to_dsinterval(value)) + 
       extract(minute from to_dsinterval(value)) * 60 +
       extract(hour from to_dsinterval(value)) * 60 * 60 +
       extract(day from to_dsinterval(value)) * 60 * 60 * 24) as retvalue
from v\\$dataguard_stats 
where name = '$2';

EOF`

#去除左侧空格
DB_LAG=`eval echo $DB_LAG`

if [[ "$DB_LAG" -gt 30 ]]
then
    printf "${WARNING_COLOR}%-10s | %-8s | the $2 is too large($DB_LAG seconds)${RESET_COLOR}\n" $1 Warning
else 
    printf "${NARMAL_COLOR}%-10s | %-8s | the $2 is $DB_LAG second(s)${RESET_COLOR}\n" $1 Success
fi

else
    echo -e "${ERROR_COLOR} $(date "+%Y-%m-%d %H:%M:%S") | Error! check_db_lag() parameter 1 and 2 can't be null ${RESET_COLOR}"
    exit 1  
fi
}

# dblink检查
# 参数1：TNS连接串

check_dblink() {
    if [[ "$1" != "" ]]; then
        COUNT=$(
            sqlplus -s $DBUSER@$1 as sysdba <<EOF
set heading off feedback off pagesize 0 verify off echo off
select count(*) FROM dba_db_links;
EOF
       )
        if [[ "$COUNT" -eq 0 ]]; then
            printf "${NARMAL_COLOR}%-10s | %-8s | There is no dblink in our database${RESET_COLOR}\n" $1 Success
        else
            printf "${WARNING_COLOR}%-10s | %-8s | dblink was found,please check the tns file and firewall${RESET_COLOR}\n" $1 Warning
        fi
    else
        echo -e "${ERROR_COLOR}$(date "+%Y-%m-%d %H:%M:%S") | Error! check_mrp_trigger() parameter 1 can't be null ${RESET_COLOR}"
        exit 1
    fi
}

# 对比两个值是否相等
# 参数1：待比较值1，参数2：待比较值2
compare_value() {

    if [[ "$1" = "$2" ]]; then
        printf "${NARMAL_COLOR}%-30s | %-8s | The value is $1 ${RESET_COLOR}\n" $3 Success
    else
        printf "${WARNING_COLOR}%-30s | %-8s | The value 1 is: $1,but value 2 is: $2 ${RESET_COLOR}\n" $3 Warning
    fi
}


# 待判断参数值中是否包含给定的子字符串
# 参数1：TNS连接串；参数2：待判断参数名；参数3：待判断参数值；参数3：子字符串
# 例如：include_value $PRIMARY_TNS "fal_server" "${PRIMARY_TNS_DIC[fal_server]}" $STANDBY_TNS

include_value(){
    # 过滤结果
    cnt=`echo $3 | grep $4 | wc -l`

    if [[ "$cnt" -gt 0 ]]; then
        printf "${NARMAL_COLOR}%-10s | %-8s| %-20s | The value is $3,include $4 ${RESET_COLOR}\n" $1 Success $2
    else
        printf "${WARNING_COLOR}%-10s | %-8s| %-20s | The value is $3,doesn't include $4 ${RESET_COLOR}\n" $1 Warning $2
    fi
}

3. 执行效果

二、 执行切换

1. dg_switchover_v1.4.sh

bash 复制代码 
#!/bin/bash

#**********************************************************************************
# Author: Hehuyi_In
# Date: 2022年06月10日
# FileName: dg_switchover_v1.4.sh
# 
# For sys user, execute the script directly. 
# For other sysdba privileged users, you need to manually synchronize the password file to standby server and chown file mode to 640 in versions below 12.2.
#
# Example: ./dg_switchover_v1.4.sh or ./dg_switchover_v1.4.sh &>> dg_switchover.log 
# Description: 
# 2022-06-10 v_1.0 only oracle dataguard switchover
# 2022-06-10 v_1.1 check instance status and db role after switchover.
# 2022-06-10 v_1.2 check transport lag,apply lag,switchover_status before switchover.
# 2022-06-13 v_1.3 add logging for main steps 
# 2022-06-14 v_1.4 auto check and select the target primary and standby database.
#**********************************************************************************

# 连接信息
DBUSER='sys/xxxxx'

# TNS列表，越靠前的被选为目标主库的概率较高
TNS_LIST=(mrptest mrptest_dg mrpuat)

# 列表元素个数检查，目前仅支持一主一从或一主两从
if [[ ${#TNS_LIST[@]} -lt 2 || ${#TNS_LIST[@]} -gt 3 ]]
then
	echo "`date "+%Y-%m-%d %H:%M:%S"` | Error! The number of standby db is less than 1 or greater than 2."
fi

# ------------------------------------------------------------------------
# 判断 TNS_LIST 中每个连接串角色，并设置各切换目标参数
# 若为主库，则作为TARGET_STANDBY_TNS
# 若为从库，则检查lag。如果lag均为0，则作为TARGET_PRIMARY_TNS，剩余的从库作为OUTLOOKING_STANDBY_TNS
# 如果所有从库lag都大于0，则报错退出。

auto_set_target_db_tns(){
	
	for TNS in ${TNS_LIST[@]}
	do
	    # 获取TNS对应db角色
		DB_ROLE=`check_db_role $TNS`

		# 若为主库，则赋值给TARGET_STANDBY_TNS
		if [[ "$DB_ROLE" = "PRIMARY" ]]
		then
			TARGET_STANDBY_TNS=$TNS			
		# 若为从库
		elif [[ "$DB_ROLE" = "PHYSICAL STANDBY" ]]
		then
			    # 如果TARGET_PRIMARY_TNS不为空，说明已经设置过了目标主库，直接设置旁观从库即可
				if [[ "$TARGET_PRIMARY_TNS" != "" ]]
				then
					OUTLOOKING_STANDBY_TNS=$TNS				
				else 
				    # 若TARGET_PRIMARY_TNS为空，则需要再检查lag
					DB_TRAN_LAG=`check_db_lag $TNS "transport lag"`
					DB_APPLY_LAG=`check_db_lag $TNS "apply lag"`
			
				    # 如果lag均为0，则赋值给TARGET_PRIMARY_TNS
					if [[ "$DB_TRAN_LAG" -eq 0 && "$DB_APPLY_LAG" -eq 0 ]]
					then
						TARGET_PRIMARY_TNS=$TNS
					fi	
				fi
		fi
	done

    # 设置后检查，若TARGET_PRIMARY_TNS或TARGET_STANDBY_TNS为空，直接报错退出
	if [[ "$TARGET_PRIMARY_TNS" = "" || "$TARGET_STANDBY_TNS" = "" ]]	
    then
		echo "`date "+%Y-%m-%d %H:%M:%S"` | Error! TARGET_PRIMARY_TNS and TARGET_STANDBY_TNS can't be null."
  		exit 1  		
	fi

}

# ------------------------------------------------------------------------
# 检查从库延迟（transport lag,apply lag ）
# 参数1为db对应的tnsname，参数2为待检查的lag
check_db_lag(){
if [[ "$1" != "" && "$2" != "" ]]	
then
sqlplus -s $DBUSER@$1 as sysdba << EOF 
set heading off feedback off pagesize 0 verify off echo off
select (extract(second from to_dsinterval(value)) + 
       extract(minute from to_dsinterval(value)) * 60 +
       extract(hour from to_dsinterval(value)) * 60 * 60 +
       extract(day from to_dsinterval(value)) * 60 * 60 * 24) as retvalue
from v\$dataguard_stats 
where name = '$2';
EOF

else
    echo "`date "+%Y-%m-%d %H:%M:%S"` | Error! parameter 1,2 can't be null."
    exit 1  
fi
}

# ------------------------------------------------------------------------

# 检查切换前switchover_status
# 参数1为db对应的tnsname，参数2为待检查的状态
check_switchover_status(){
if [[ "$1" != "" && "$2" != "" ]]	
then
STATUS=`sqlplus -s $DBUSER@$1 as sysdba << EOF 
set heading off feedback off pagesize 0 verify off echo off
select switchover_status from v\\$database;
EOF`

if [[ "$STATUS" = $2 || "$STATUS" = "SESSIONS ACTIVE" ]]
then
    echo "`date "+%Y-%m-%d %H:%M:%S"` | Success! switchover_status of $1 is $STATUS."
else 
    echo "`date "+%Y-%m-%d %H:%M:%S"` | Error! switchover_status of $1 is $STATUS (should be $2 or SESSIONS ACTIVE)."
    exit 1
fi

else
    echo "`date "+%Y-%m-%d %H:%M:%S"` | Error! parameter 1 and 2 can't be null."
    exit 1  
fi
}

# ------------------------------------------------------------------------
# 检查db角色，参数1为db对应的tnsname

check_db_role(){
if [[ "$1" != "" ]]	
then
sqlplus -s $DBUSER@$1 as sysdba << EOF 

set heading off feedback off pagesize 0 verify off echo off
select database_role from v\$database;
EOF

else
    echo "`date "+%Y-%m-%d %H:%M:%S"` | Error! parameter 1 should be tnsname."
    exit 1  
fi
}

# ------------------------------------------------------------------------
# 检查切换后实例状态
# 参数1为db对应的tnsname

check_instance_status(){
if [[ "$1" != "" ]]	
then
STATUS=`sqlplus -s $DBUSER@$1 as sysdba << EOF 

set heading off feedback off pagesize 0 verify off echo off
select status from v\\$instance;

EOF`

if [[ "$STATUS" != "OPEN" ]]
then
    echo "`date "+%Y-%m-%d %H:%M:%S"` | Error! instance status of $1 is $STATUS after switchover."
    exit 1
else 
    echo "`date "+%Y-%m-%d %H:%M:%S"` | Success! instance status of $1 is $STATUS after switchover."
fi

else
    echo "`date "+%Y-%m-%d %H:%M:%S"` | Error! parameter 1 can't be null."
    exit 1  
fi
}

# ------------------------------- main() -----------------------------------------

echo "------------------------------------------------------------------------"
echo "`date "+%Y-%m-%d %H:%M:%S"` | Precheck before switchover..."

# 判断 TNS_LIST 中每个连接串角色，并设置各切换目标参数
auto_set_target_db_tns

# 设置后各目标参数值
echo "`date "+%Y-%m-%d %H:%M:%S"` | TARGET_PRIMARY_TNS: $TARGET_PRIMARY_TNS"
echo "`date "+%Y-%m-%d %H:%M:%S"` | TARGET_STANDBY_TNS: $TARGET_STANDBY_TNS"
echo "`date "+%Y-%m-%d %H:%M:%S"` | OUTLOOKING_STANDBY_TNS: $OUTLOOKING_STANDBY_TNS"

# 切换前switchover_status检查
check_switchover_status $TARGET_STANDBY_TNS "TO STANDBY"

# 目标从库（原主库）执行：将原主库转切为从库
echo "`date "+%Y-%m-%d %H:%M:%S"` | Begin switchover $TARGET_STANDBY_TNS to standby database..."
    
sqlplus -s $DBUSER@$TARGET_STANDBY_TNS as sysdba << EOF 
alter database commit to switchover to standby with session shutdown;
shutdown immediate
EOF

echo "`date "+%Y-%m-%d %H:%M:%S"` | Restarting $TARGET_STANDBY_TNS database..."

# 重新启动目标从库
sqlplus -s $DBUSER@$TARGET_STANDBY_TNS as sysdba << EOF 
startup
EOF

# 切换后角色检查
DB_ROLE=`check_db_role $TARGET_STANDBY_TNS`

if [[ "$DB_ROLE" != "PHYSICAL STANDBY" ]]; 
then
    echo "`date "+%Y-%m-%d %H:%M:%S"` | Error! db role of $TARGET_STANDBY_TNS is $DB_ROLE(should be PHYSICAL STANDBY) after switchover."
    exit 1
else 
    echo "`date "+%Y-%m-%d %H:%M:%S"` | Success! db role of $TARGET_STANDBY_TNS is $DB_ROLE after switchover."
fi

# 切换后实例状态检查
check_instance_status $TARGET_STANDBY_TNS

echo "`date "+%Y-%m-%d %H:%M:%S"` | Finish switchover $TARGET_STANDBY_TNS to standby database."

# ------------------------------------------------------------------------

echo "`date "+%Y-%m-%d %H:%M:%S"` | Precheck before switchover $TARGET_PRIMARY_TNS to primary database..."

# 切换前switchover_status检查
check_switchover_status $TARGET_PRIMARY_TNS "TO PRIMARY"

# 目标主库（原从库）执行
# 将原从库切为主库

echo "`date "+%Y-%m-%d %H:%M:%S"` | Begin switchover $TARGET_PRIMARY_TNS to primary database..."

sqlplus -s $DBUSER@$TARGET_PRIMARY_TNS as sysdba << EOF 
alter database commit to switchover to primary with session shutdown;
ALTER DATABASE OPEN;
EOF

# 切换后角色检查
DB_ROLE=`check_db_role $TARGET_PRIMARY_TNS`

if [[ "$DB_ROLE" != "PRIMARY" ]]; 
then
    echo "`date "+%Y-%m-%d %H:%M:%S"` | Error! db role of $1 is $DB_ROLE(should be PRIMARY) after switchover."
    exit 1
else 
    echo "`date "+%Y-%m-%d %H:%M:%S"` | Success! db role of $1 is $DB_ROLE after switchover."
fi

# 切换后实例状态检查
check_instance_status $TARGET_PRIMARY_TNS

# 切换后从库lag检查
DB_TRAN_LAG=`check_db_lag $TARGET_STANDBY_TNS "transport lag"`
DB_APPLY_LAG=`check_db_lag $TARGET_STANDBY_TNS "apply lag"`

echo "`date "+%Y-%m-%d %H:%M:%S"` | After switchover: DB_TRAN_LAG is $DB_TRAN_LAG second(s), DB_APPLY_LAG is $DB_APPLY_LAG second(s)"

echo "`date "+%Y-%m-%d %H:%M:%S"` | Finish switchover $TARGET_PRIMARY_TNS to primary database."

# ------------------------------------------------------------------------
# 如果有旁观从库
if [ "$OUTLOOKING_STANDBY_TNS" != "" ]; then

echo "`date "+%Y-%m-%d %H:%M:%S"` | Begin onlooking standby database setting..."

# 目标主库调整参数，指向旁观从库，先清空再设置，否则有可能报参数冲突
sqlplus -s $DBUSER@$TARGET_PRIMARY_TNS as sysdba << EOF 
alter system set LOG_ARCHIVE_DEST_2='';
alter system set LOG_ARCHIVE_DEST_3='';
alter system set LOG_ARCHIVE_DEST_2='SERVICE=$TARGET_STANDBY_TNS ASYNC VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=$TARGET_STANDBY_TNS';
alter system set LOG_ARCHIVE_DEST_3='SERVICE=$OUTLOOKING_STANDBY_TNS ASYNC VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=$OUTLOOKING_STANDBY_TNS';
EOF

# 目标从库调整参数，断开到旁观从库的连接
sqlplus -s $DBUSER@$TARGET_STANDBY_TNS as sysdba << EOF 
alter system set LOG_ARCHIVE_DEST_3='';
alter system set LOG_ARCHIVE_DEST_2='SERVICE=$TARGET_PRIMARY_TNS ASYNC VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=$TARGET_PRIMARY_TNS';
EOF

# 旁观从库调整参数，指向新主库
sqlplus -s $DBUSER@$OUTLOOKING_STANDBY_TNS as sysdba  << EOF 
alter system set LOG_ARCHIVE_DEST_3='';
alter system set LOG_ARCHIVE_DEST_2='SERVICE=$TARGET_PRIMARY_TNS ASYNC VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=$TARGET_PRIMARY_TNS';
EOF

# 切换后旁观从库lag检查
DB_TRAN_LAG=`check_db_lag $OUTLOOKING_STANDBY_TNS "transport lag"`
DB_APPLY_LAG=`check_db_lag $OUTLOOKING_STANDBY_TNS "apply lag"`

echo "`date "+%Y-%m-%d %H:%M:%S"` | After setting: DB_TRAN_LAG is $DB_TRAN_LAG second(s), DB_APPLY_LAG is $DB_APPLY_LAG second(s)"

echo "`date "+%Y-%m-%d %H:%M:%S"` | Finish onlooking standby database setting."

else # 没有旁观从库
echo "`date "+%Y-%m-%d %H:%M:%S"` | There is no onlooking standby database."
fi
相关推荐
leegong2311115 小时前
Oracle、PostgreSQL该学哪一个？
数据库·postgresql·oracle
夜光小兔纸16 小时前
Oracle 普通用户连接hang住处理方法
运维·数据库·oracle
步、步、为营1 天前
解锁.NET配置魔法：打造强大的配置体系结构
数据库·oracle·.net
MasterNeverDown1 天前
解决 PostgreSQL 中创建 TimescaleDB 扩展的字符串错误
数据库·postgresql·oracle
limts1 天前
Oracle之开窗函数使用
数据库·oracle
张飞光2 天前
MongoDB 创建数据库
数据库·mongodb·oracle
码农丁丁2 天前
为什么数据库不应该使用外键
数据库·mysql·oracle·数据库设计·外键
青灯文案12 天前
Oracle 数据库常见字段类型大全及详细解析
数据库·oracle
雾里看山2 天前
【MySQL】数据库基础知识
数据库·笔记·mysql·oracle
码农研究僧2 天前
Oracle SQL: TRANSLATE 和 REGEXP_LIKE 的知识点详细分析
数据库·sql·oracle·translate·regexp_like
热门推荐
01Dell服务器升级ubuntu 22.04失败解决02Windows10安装PCL1.14.0及点云配准03半导体应用系统一些小知识收集(strip&wafer mapping,EAP&scada)04机器人持续学习基准LIBERO系列10——文件结构05xgboost: Why not implement distributed XGBoost on top of spark06密码学原理技术-第六章-introduction to pulibc-key cryptography072023美赛A题之Lotka-Volterra【完整思路+代码】08centos7 init.d 和system.d09(欧拉)openEuler系统添加网卡文件配置流程、(欧拉)openEuler系统手动配置ipv6地址流程、(欧拉)openEuler系统网络管理说明10BMC 虚拟i2c访问PCA9545(switch芯片)后面的设备,为什么找不到PCA9545?