1.数据库备份

1）逻辑备份脚本参考

此脚本需要修改以下内容：

username="SYSDBA"
password="Hn@dameng123"
address="127.0.0.1:5136"
# 备份并发数
parallel=2
# 保存天数
retention=5
# 要备份的目录未知
base_path="/dm/bak/dexp"

脚本内容参考

#!/bin/bash

set -e

# 00 02 * * * /bin/bash /dm/bak/dexp/backup.sh
##### 公共配置 #####
username="SYSDBA"
password="Hn@dameng123"
address="127.0.0.1:5136"
# 备份并发数
parallel=2
# 保存天数
retention=5
# 要备份的目录未知
base_path="/dm/bak/dexp"

########################


format_time=`date +%Y%m%d%H`
backup_path="${base_path}/backup"
log_path="${base_path}/logs"
log_file="${log_path}/${format_time}.log"

test ! -d "${log_path}" && mkdir -p "${log_path}"
test ! -d "${backup_path}" && mkdir -p "${backup_path}"

log() {
    local level=$1
    shift
    local message="$*"
    local timestamp=$(date '+%Y-%m-%d %H:%M:%S')
    echo "[${timestamp}] [${level}] ${message}" | tee -a "${log_file}"
}

error_exit() {
    log "ERROR" "$1"
    exit 1
}

set_env() {
    log "INFO" "Setting up DM environment variables..."
    dm_pid=`pgrep -f dmserver`
    if [ -z "$dm_pid" ]; then
        log "WARN" "dmserver is not running, try to find it by ps"
        dp_pid=`ps -ef | grep dmserver | grep -v grep | awk '{print $2}'`
        if [ -z "$dp_pid" ]; then
            log "ERROR" "dmserver is not running"
            error_exit "dmserver is not running"
        fi
    fi
    log "INFO" "Found dmserver process: PID=$dm_pid"

    dm_home=$(pwdx "$dm_pid" 2>/dev/null | awk '{print $2}')
    if [ -z "$dm_home" ]; then
        # 如果pwdx失败，尝试其他方法
        dm_home=$(lsof -p "$dm_pid" 2>/dev/null | grep cwd | awk '{print $9}')
        if [ -z "$dm_home" ]; then
            error_exit "Cannot determine DM_HOME from process $dm_pid"
        fi
    fi
    log "INFO" "DM_HOME determined as: $dm_home"
    if [ `grep -c "$dm_home" /etc/profile` -eq 0 ]; then
        echo "export DM_HOME=$dm_home" >> /etc/profile
        echo "export PATH=\$DM_HOME:\$PATH" >> /etc/profile
        echo "export LD_LIBRARY_PATH=\$DM_HOME:\$LD_LIBRARY_PATH" >> /etc/profile
        source /etc/profile
    else
        log "INFO" "DM_HOME is already set in /etc/profile"
        source /etc/profile
    fi
}

validate_connection() {
    log "INFO" "Validating database connection..."
    id_code=`disql $username/'"'$password'"'@$address -e "select id_code;"`
    if [[ $id_code == *"id_code"* ]]; then
        log "INFO" "Connected to the database successfully."
    else
        log "ERROR" "Failed to connect to the database."
        error_exit "Failed to connect to the database."
    fi
}

# 按文件数量删除备份文件
delete_backup_by_num() {
  # 备份文件目标目录
  dst_dir=$1
  # 目标保留天数
  save_days=$2
  c_num=0
  # 获取当前保留的备份文件数量
  current_save_days=$(ls -t "$dst_dir" | wc -l)
  log "INFO" "目标目录: $dst_dir"
  log "INFO" "当前保留的文件列表: "
  ls -t "$dst_dir"
  log "INFO" "目标保留天数: $save_days"
  log "INFO" "当前保留的备份文件数量: $current_save_days"
  if [ "$save_days" -lt "$current_save_days" ]; then
    sort_backup_files=$(ls -t $dst_dir)
    log "INFO" "$sort_backup_files"
    for f in $sort_backup_files; do
      if [ "$c_num" -lt "$save_days" ]; then
        ((c_num += 1))
      else
        rm -rf "${dst_dir}/${f}"
        if [ $? -ne 0 ]; then
          return
        fi
        log "INFO" "${dst_dir}/${f} 删除完成"
      fi
    done
  else
    log "INFO" "当前备份文件数为: $current_save_days, 目标保留文件数: $save_days, 不需要删除"
  fi
}

# 获取所有的schema
get_schemas() {
    echo `disql $username/'"'$password'"'@$address -e "select object_name from dba_objects where object_type='SCH';"   | awk '/^object_name/ && !found {found=1} found' | awk 'NF' | grep -v -E "object_name|---" | grep -v SYS`
}

# 获取指定schema下的所有表
get_tables() {
    local schema=$1
    echo `disql $username/'"'$password'"'@$address -e "SELECT TABLE_NAME FROM DBA_TABLES WHERE OWNER = '$schema';" | awk '/^TABLE_NAME/ && !found {found=1} found' | awk 'NF' | grep -v -E "TABLE_NAME|---"`
}

# 备份
backup() {
    set_env
    validate_connection
    schemas=$(get_schemas)
    for schema in $schemas; do
        backup_file="${backup_path}/${format_time}/${schema}.dmp"
        test ! -d "${backup_path}/${format_time}" && mkdir -p "${backup_path}/${format_time}"
        log "INFO" "backup $schema to $backup_file"
        dexp USERID=$username/'"'$password'"'@$address FILE=${schema}.dmp SCHEMAS=${schema} DIRECTORY=${backup_path}/${format_time} NOLOG=Y PARALLEL=${parallel} TABLE_PARALLEL=${parallel} COMPRESS=Y COMPRESS_LEVEL=4
        if [ $? -eq 0 ]; then
            log "INFO" "backup $schema success"
            rm -f ${backup_path}/${format_time}/dexp.log ${backup_path}/${format_time}/dexp_${schema}.log
        else
            mv ${backup_path}/${format_time}/dexp.log ${backup_path}/${format_time}/dexp_${schema}.log
            log "ERROR" "backup $schema failed"
        fi
    done
}

main() {
    # 设置环境变量
    set_env
    # 验证连接
    validate_connection
    # 清理历史备份
    delete_backup_by_num "${backup_path}" "${retention}"    
    # 备份
    backup
}

main

2）创建计划任务

将此脚本放到任意目录中执行，保存成backup.sh文件名

执行 crontab -e命令后，会进入编辑框，输入以下内容：

# 需要修改下备份脚本路径
00 02 * * * /bin/bash /dm/bak/dexp/backup.sh

2.数据库恢复

1）明确要恢复的库(schema)

直接去备份目录中，找到要恢复的库，位置在[基础目录]/backup/[年月日]/库名.dmp，例如：

/dm/bak/dexp/backup/2026020502/procurement_contract_gcbp.dmp

这里的procurement_contract_gcbp即为库名

2）恢复原库数据到其他库（重命名的新库）中

恢复之前，建议先把要恢复的库文件先拷贝到新的目录中

执行命令：

dimp USERID=SYSDBA/'"Hn@dameng123"':5136 FILE=implementation_engine_gcbp.dmp LOG=implementation_engine_gcbp.log REMAP_SCHEMA=implementation_engine_gcbp:import_20260107_implementation_engine_gcbp DIRECTORY=/dm/bak/dexp/recovery/260107

参数说明

• FILE 指定逻辑备份的备份文件，以dmp结尾
• LOG 指定恢复时的日志文件，随意设置即可
• REMAP_SCHEMA 设置要把原库恢复到目标库的规则，写法为：原库:新库
• DIRECTORY 指定备份文件的目录位置