DSC 共享集群+备库的搭建

一、前期准备工作

1.1 在 Linux 中新增网卡操作（可选）

1、ifconfig 查看新增网卡名称
2、复制一份 ifcfg-ens33 文件
3、查看新增网卡的 uuidgen
4、修改新增网卡配置文件的 UUID、NAME 和 DEVICE 名称
5、重启网络服务

1.2 修改主机名（可选）

目的是为了更好的区分各台虚拟机所执行的功能并易于区分
1、编辑/etc/hostname 文件
2、将原有内容删除
3、输入需要更改的主机名
4、保存文件
5、重启虚拟机后生效（也可以直接在控制台输入 bash 但是不建议这么做）

二、两节点 DSC 集群的搭建

2.1 实验环境

注：本次实验环境均使用一块网卡

2.2 添加网卡（可选）

见一、在 Linux 中新增网卡操作（可选）

2.3 添加共享硬盘

CentOS 7-DSC01（原机）
CentOS 7-DSC02（克隆机）

1、在原机上新增一块 50GB 的硬盘
2、在克隆机上添加原机 50GB 的硬盘
| HostName | External IP | Intranet IP | Instance Name|
| :--: | :--: | :--: | :--: | :--: |
|Node0 |192.168.233.100 | 192.168.223.100 | DCS0|
|Node1 | 192.168.233.110 | 192.168.223.110 | DSC1|
|Standby | 192.168.223.6 | 192.168.223.6 | DW|
3、在原机的 vmx 文件中添加

disk.locking="FALSE"
disk.SharedBus="Virtual"
disk.enableUUID="TRUE"
disk.shared="TRUE"

（否则不能同时启动原机和克隆机）
4、在原机和克隆机使用 fdisk -l 验证是否添加成功

[root@Node0 ~]# fdisk -l

[root@Node1 ~]# fdisk -l

2.4 共享硬盘分区

Node0 或 Node1 节点执行

使用 parted 命令进行分区（主分区个数不限制，大小可以超过 2T----区别于 fdisk）

[root@Node0 ~]# parted
GNU Parted 3.1
Using /dev/sda
Welcome to GNU Parted! Type 'help' to view a list of commands.
(parted) select /dev/sdb
Using /dev/sdb
(parted) mklabel gpt
(parted) mkpart primary 0 2G
(parted) mkpart primary 2G 4G
(parted) mkpart primary 4G 14G
(parted) mkpart primary 14G 24G
(parted) mkpart primary 24G -1
(parted) quit
[root@Node0 ~]# parted /dev/sdb print

[root@Node0 ~]# blkid

2.5 创建磁盘链接

Node0 和 Node1 节点都需执行
配置相同

[root@Node0 ~]# cd /etc/udev/rules.d
[root@Node0 rules.d]# vim 88-dm-asmdevices.rules
[root@Node0 rules.d]# cat 88-dm-asmdevices.rules
KERNEL=="sdb1",SUBSYSTEM=="block",OWNER="dmdba",GROUP="dinstall",SYMLINK+="
DM/dmasm_vote",MODE="0660"
KERNEL=="sdb2",SUBSYSTEM=="block",OWNER="dmdba",GROUP="dinstall",
SYMLINK+="DM/dmasm_dcr",MODE="0660"
KERNEL=="sdb3",SUBSYSTEM=="block",OWNER="dmdba",GROUP="dinstall",
SYMLINK+="DM/dmasm_log",MODE="0660"
KERNEL=="sdb4",SUBSYSTEM=="block",OWNER="dmdba",GROUP="dinstall",
SYMLINK+="DM/dmasm_arch",MODE="0660"
KERNEL=="sdb5",SUBSYSTEM=="block",OWNER="dmdba",GROUP="dinstall",
SYMLINK+="DM/dmasm_data",MODE="0660"

2.6 刷新磁盘信息

Node0 和 Node1 节点都需执行
配置相同

[root@Node0 rules.d]# /sbin/udevadm control --reload
[root@Node0 rules.d]# /sbin/udevadm trigger --action change
[root@Node0 rules.d]# /sbin/udevadm trigger --action add
[root@Node0 rules.d]# chown -R dmdba:dinstall /dev/DM

重新加载使之生效后会在/dev 目录下生成一个 DM 的文件夹

[root@Node0 ~]# cd /dev/DM/
[root@Node0 DM]# ll

2.7 集群搭建

2.7.1 安装达梦数据库

Node0 和 Node1 节点都需执行
配置相同

[root@Node0 ~]# groupadd dinstall
[root@Node0 ~]# useradd -g dinstall -m -s /bin/bash dmdba
[root@Node0 ~]# passwd dmdba
[root@Node0 ~]# vim /etc/security/limits.conf
dmdba hard nofile 65536
dmdba soft nofile 65536
dmdba hard stack 32768
dmdba soft stack 16384
[root@Node0 ~]# mount -o loop /opt/dm8_20230104_x86_rh6_64.iso /mnt
[root@Node0 ~]# mkdir /dm8
[root@Node0 ~]# chown -R dmdba:dinstall /dm8
[root@Node0 ~]# systemctl status firewalld
[root@Node0 ~]# systemctl stop firewalld
[root@Node0 ~]# systemctl disable firewalld
[root@Node0 ~]# vim /etc/selinux/config
SELINUX=disabled
[root@Node0 ~]# su - dmdba
[dmdba@Node0 ~]$ cd /mnt
[dmdba@Node0 mnt]$ ./DMInstall.bin -i
[root@Node0 ~]# /dm8/script/root/root_installer.sh

2.7.2 文件配置

2.7.2.1 创建 dmdcr_cfg.ini

Node0 和 Node1 节点都需执行
配置相同

该配置文件记录了集群环境全局信息、集群组信息（CSS、ASM 和 DB）、以及组内节点信息

[dmdba@Node0 conf]$ vim dmdcr_cfg.ini
[dmdba@Node0 conf]$ cat dmdcr_cfg.ini
DCR_N_GRP = 3
DCR_VTD_PATH = /dev/DM/dmasm_vote
DCR_OGUID = 20230612
[GRP]
DCR_GRP_TYPE = CSS
DCR_GRP_NAME = GRP_CSS
DCR_GRP_N_EP = 2
DCR_GRP_DSKCHK_CNT = 65
[GRP_CSS]
DCR_EP_NAME = CSS0
DCR_EP_HOST = 192.168.223.100
DCR_EP_PORT = 11286
[GRP_CSS]
DCR_EP_NAME = CSS1
DCR_EP_HOST = 192.168.223.110
DCR_EP_PORT = 11287
[GRP]
DCR_GRP_TYPE = ASM
DCR_GRP_NAME = GRP_ASM
DCR_GRP_N_EP = 2
DCR_GRP_DSKCHK_CNT = 61
[GRP_ASM]
DCR_EP_NAME = ASM0
DCR_EP_SHM_KEY = 2424
DCR_EP_SHM_SIZE = 1024
DCR_EP_HOST = 192.168.223.100
DCR_EP_PORT = 11276
DCR_EP_ASM_LOAD_PATH = /dev/DM
[GRP_ASM]
DCR_EP_NAME = ASM1
DCR_EP_SHM_KEY = 2425
DCR_EP_SHM_SIZE = 1024
DCR_EP_HOST = 192.168.223.110
DCR_EP_PORT = 11277
DCR_EP_ASM_LOAD_PATH = /dev/DM
[GRP]
DCR_GRP_TYPE = DB
DCR_GRP_NAME = GRP_DSC
DCR_GRP_N_EP = 2
DCR_GRP_DSKCHK_CNT = 20
[GRP_DSC]
DCR_EP_NAME = DSC0
DCR_EP_SEQNO = 0
DCR_EP_PORT = 5236
DCR_CHECK_PORT = 11256
[GRP_DSC]
DCR_EP_NAME = DSC1
DCR_EP_SEQNO = 1
DCR_EP_PORT = 5236
DCR_CHECK_PORT = 11257

2.7.2.2 创建 dmdcr.ini

Node0 和 Node1 节点都需执行
配置不同

该配置文件记录了当前节点序列号以及 DCR 磁盘路径

Node0

[dmdba@Node0 conf]$ vim dmdcr.ini
[dmdba@Node0 conf]$ cat dmdcr.ini
DMDCR_PATH = /dev/DM/dmasm_dcr
DMDCR_MAL_PATH = /dm8/conf/dmasvrmal.ini
DMDCR_SEQNO = 0
DMDCR_ASM_RESTART_INTERVAL = 0
DMDCR_ASM_STARTUP_CMD = /dm8/bin/dmasmsvr dcr_ini = /dm8/conf/dmdcr.ini
DMDCR_DB_RESTART_INTERVAL = 0
DMDCR_DB_STARTUP_CMD=/dm8/bin/dmserver path=/dm8/conf/prod0_config/dm.ini
dcr_ini=/dm8/conf/dmdcr.ini

Node1

[dmdba@Node1 conf]$ vim dmdcr.ini
[dmdba@Node1 conf]$ cat dmdcr.ini
DMDCR_PATH = /dev/DM/dmasm_dcr
DMDCR_MAL_PATH = /dm8/conf/dmasvrmal.ini
DMDCR_SEQNO = 1
DMDCR_ASM_RESTART_INTERVAL = 0
DMDCR_ASM_STARTUP_CMD = /dm8/bin/dmasmsvr dcr_ini = /dm8/conf/dmdcr.ini
DMDCR_DB_RESTART_INTERVAL = 0
DMDCR_DB_STARTUP_CMD=/dm8/bin/dmserver path=/dm8/conf/prod0_config/dm.ini
dcr_ini=/dm8/conf/dmdcr.ini

2.7.2.3 创建 dmasvrmal.ini

该配置文件用来进行 asm 集群内部网络通讯

Node0 和 Node1 节点都需执行
配置相同

[dmdba@Node0 ~]$ cd /dm8/conf/
[dmdba@Node0 conf]$ cat dmasvrmal.ini
[MAL_INST1]
MAL_INST_NAME = ASM0
MAL_HOST = 192.168.223.100
MAL_PORT = 5244
[MAL_INST2]
MAL_INST_NAME = ASM1
MAL_HOST = 192.168.223.110
MAL_PORT = 5244

2.7.2.4 创建 dmcssm.ini

Standby 节点执行

[dmdba@Standby ~]$ cd /dm8/conf
[dmdba@Standby conf]$ cat dmcssm.ini
CSSM_OGUID=20230612
CSSM_CSS_IP=192.168.223.100:11286
CSSM_CSS_IP=192.168.223.110:11287
CSSM_LOG_PATH=/dm8/log
CSSM_LOG_FILE_SIZE=32
CSSM_LOG_SPACE_LIMIT=0

2.7.3 创建 dcr、vote 和 asm 磁盘

Node0 或 Node1 节点执行

[root@Node0 ~]# cd /dm8/bin
[root@Node0 bin]# ./dmasmcmd
DMASMCMD V8
ASM>listdisks /dev/DM
[/dev/DM/dmasm_vote]: Normal disk
[/dev/DM/dmasm_log]: Normal disk
[/dev/DM/dmasm_data]: Normal disk
[/dev/DM/dmasm_arch]: Normal disk
[/dev/DM/dmasm_dcr]: Normal disk
Used time: 1.312(ms).
ASM>create dcrdisk '/dev/DM/dmasm_dcr' 'DCR'
[TRACE]The ASM initialize dcrdisk /dev/DM/dmasm_dcr to name DMASMDCR
Used time: 1.928(ms).
ASM>create votedisk '/dev/DM/dmasm_vote' 'VOTE'
[TRACE]The ASM initialize votedisk /dev/DM/dmasm_vote to name DMASMVOTE
Used time: 11.233(ms).
ASM>create asmdisk '/dev/DM/dmasm_data' 'DATA'
[TRACE]The ASM initialize asmdisk /dev/DM/dmasm_data to name DMASMDATA
Used time: 12.769(ms).
ASM>create asmdisk '/dev/DM/dmasm_log' 'LOG'
[TRACE]The ASM initialize asmdisk /dev/DM/dmasm_log to name DMASMLOG
Used time: 14.482(ms).
ASM>create asmdisk '/dev/DM/dmasm_arch' 'ARCH'

注：如果误操作（比如说执行了两次创建 dcr 磁盘命令）使用 dd 命令擦除磁盘信息
例子：

[root@Node0 bin]# dd if=/dev/zero of=/dev/DM/dmasm_vote bs=1024k count=1
[root@Node0 bin]# dd if=/dev/zero of=/dev/DM/dmasm_dcr bs=1024k count=1

2.7.4 初始化 dcr 和 vote 磁盘

Node0 或 Node1 节点执行

[root@Node0 ~]# cd /dm8/bin
[root@Node0 bin]# ./dmasmcmd
ASM>init dcrdisk '/dev/DM/dmasm_dcr' from '/dm8/conf/dmdcr_cfg.ini' identified by
'dmdsc'
ASM>init votedisk '/dev/DM/dmasm_vote' from '/dm8/conf/dmdcr_cfg.ini'
ASM>listdisks /dev/DM

2.7.5 注册 css 服务并启动

Node0 和 Node1 节点都需执行
步骤相同

Node0

[root@Node0 root]# ./dm_service_installer.sh -t dmcss -p DSC0 -dcr_ini
/dm8/conf/dmdcr.ini
[root@Node0 bin]# ./DmCSSServiceDSC0 start

Node1

[root@Node1 root]# ./dm_service_installer.sh -t dmcss -p DSC1 -dcr_ini
/dm8/conf/dmdcr.ini
[root@Node1 bin]# ./DmCSSServiceDSC1 start

2.7.6 注册 asm 服务并启动

Node0 和 Node1 节点都需执行
步骤相同

Node0

[root@Node0 root]# ./dm_service_installer.sh -t dmasmsvr -p DSC0 -y DmCSSServiceDSC0
-dcr_ini /dm8/conf/dmdcr.ini
[root@Node0 bin]# ./DmASMSvrServiceDSC0 start

Node1

[root@Node1 root]# ./dm_service_installer.sh -t dmasmsvr -p DSC1 -y DmCSSServiceDSC1
-dcr_ini /dm8/conf/dmdcr.ini
[root@Node1 bin]# ./DmASMSvrServiceDSC1 start

2.7.7 创建 asm 磁盘组

Node0 或 Node1 节点执行

[dmdba@Node0 ~]$ cd /dm8/bin
[dmdba@Node0 bin]$ ./dmasmtool dcr_ini=/dm8/conf/dmdcr.ini
DMASMTOOL V8
ASM>create diskgroup 'DGDATA01' asmdisk '/dev/DM/dmasm_data'
Used time: 25.426(ms).
ASM>create diskgroup 'DGLOG01' asmdisk '/dev/DM/dmasm_log'
Used time: 29.763(ms).
ASM>create diskgroup 'DGARCH01' asmdisk '/dev/DM/dmasm_arch'
Used time: 20.525(ms).

2.7.8 初始化数据库（创建实例）

Node0 或 Node1 节点执行

[dmdba@Node0 ~]$ cd /dm8/conf/
[dmdba@Node0 conf]$ cat dminit.ini
db_name = DSC
system_path = +DGDATA01
system = +DGDATA01/DM/system.dbf
system_size = 128
roll = +DGDATA01/DM/roll.dbf
roll_size = 128
main = +DGDATA01/DM/main.dbf
main_size = 128
ctl_path = +DGDATA01/DM/dm.ctl
ctl_size = 8
log_size = 256
dcr_path = /dev/DM/dmasm_dcr
dcr_seqno = 0
auto_overwrite = 1
[DSC0]
config_path = /dm8/conf/DSC0_config
port_num = 5236
mal_host = 192.168.223.100
mal_port = 9340
log_path = +DGLOG01/DSC0/log01.log
log_path = +DGLOG01/DSC0/log02.log
[DSC1]
config_path = /dm8/conf/DSC1_config
port_num = 5236
mal_host = 192.168.223.110
mal_port = 9341
log_path = +DGLOG01/DSC1/log01.log
log_path = +DGLOG01/DSC1/log02.log

注：数据库实例端口问题（dmdcr_cfg.ini 配置文件中的数据库实例端口要和初始化的实例
端口保持一致），两个端口不同会导致后续在配置实时备库时 dsc 集群的一个节点的数据库
服务启动失败，而且连接数据库服务时使用的是 dmdcr_cfg.ini 配置文件中的实例端口

[dmdba@Node0 bin]$ ./dminit control=/dm8/conf/dminit.ini

初始化后会在 conf 目录下生成指定的两个文件夹

[dmdba@Node0 bin]$ cd /dm8/conf/
[dmdba@Node0 conf]$ ll

将其中一个目录发送到另外一个节点

[dmdba@Node0 conf]$ scp -r DSC1_config/ 192.168.223.110:/dm8/conf/

2.7.9 注册 dmserver 服务并启动

Node0 和 Node1 节点都需执行
步骤相同

Node0

[root@Node0 root]# ./dm_service_installer.sh -t dmserver -y DmASMSvrServiceDSC0 -p
DSC0 -dm_ini /dm8/conf/DSC0_config/dm.ini -dcr_ini /dm8/conf/dmdcr.ini
[dmdba@Node0 bin]$ ./DmServiceDSC0 start

Node0

[root@Node1 root]# ./dm_service_installer.sh -t dmserver -y DmASMSvrServiceDSC1 -p
DSC1 -dm_ini /dm8/conf/DSC1_config/dm.ini -dcr_ini /dm8/conf/dmdcr.ini
[dmdba@Node1 bin]$ ./DmServiceDSC1 start

2.7.10 注册 cssm 监视器服务（可选）

一般通过前台启动，便于观察各集群的状态信息

Node0、Node1 或 Standby 节点执行

[root@Standby root]# ./dm_service_installer.sh -t dmcssm -cssm_ini /dm8/conf/dmcssm.ini
-p Standby

2.7.11 启动 cssm 监视器（前台方式）

Standby 节点执行

[root@Standby bin]# ./dmcssm ini_path=/dm8/conf/dmcssm.ini

输入 show 命令查看 DSC 集群启动状态
CSS 集群启动正常

ASM 集群启动正常

DB 集群启动正常

2.7.12 验证数据同步

[dmdba@Node0 bin]$ ./disql sysdba/SYSDBA
SQL> select * from v$dsc_ep_info;

节点一创建表空间并新建表最后插入数据，节点二验证查询数据
问题一：在 DSC 中新建表空间时数据文件会如何存放？

SQL> create tablespace dsc_tb datafile '/dev/DM/dmasm_data/dsc.dbf' size 100;
create tablespace dsc_tb datafile '/dev/DM/dmasm_data/dsc.dbf' size 100;
[-1512]:Not support local path when create tablespace files on DSC system.
used time: 61.655(ms). Execute id is 0.

创建表空间

SQL> create tablespace dsc_tb datafile 'dsc.dbf' size 100;
executed successfully
used time: 213.875(ms). Execute id is 58702.

创建用户并指定表空间

SQL> create user dmtest identified by dameng123 default tablespace dsc_tb;
SQL> grant dba to dmtest;
SQL> conn dmtest/dameng123;
SQL> create table student(
2 id int primary key,
3 name varchar(20) not null,
4 age int not null);
SQL> insert into student values(1,'zhangsan',23);
SQL> insert into student values(2,'xiaoming',22);

SQL> select * from student;
SQL> commit;

注：不提交节点二查不到数据

[dmdba@Node1 bin]$ ./disql dmtest/dameng123

三、搭建实时备库

3.1 环境准备

◼ 操作系统：Centos-7
◼ IP 地址：192.168.223.6
注：搭建前需要关闭 DSC 集群的数据库服务

[dmdba@Node0 bin]$ ./DmServiceDSC0 stop
[dmdba@Node1 bin]$ ./DmServiceDSC1 stop

3.2 初始化数据库实例并注册服务

注：安装达梦数据库见 2.7.1 安装达梦数据库

[dmdba@Standby bin]$ ./dminit path=/dm8/data db_name=DSC instance_name=DW
[root@Standby root]# ./dm_service_installer.sh -t dmserver -p Standby -dm_ini
/dm8/data/DSC/dm.ini

3.3 脱机备份 DSC 集群

Node0 或 Node1 节点执行

注：备份前提是 asm 服务要打开，而 asm 服务要打开必须要启动 css 服务

[dmdba@Node0 bin]$ ./dmrman dcr_ini=/dm8/conf/dmdcr.ini
dmrman V8
RMAN> backup database '/dm8/conf/DSC0_config/dm.ini' full backupset
'/dm8/dmbak/BACKUP_FILE'
RMAN> check backupset '/dm8/dmbak/BACKUP_FILE';

3.4 备机还原

Standby 节点执行

将 Node0 的备份文件发送到备机，并使用 dmrman 工具进行还原
先在备份节点上创建 dmbak 目录（dmdba 用户）

[dmdba@localhost dm8]$ mkdir dmbak
[dmdba@Node0 bin]$ scp -r /dm8/dmbak/BACKUP_FILE/
dmdba@192.168.223.6:/dm8/dmbak/BACKUP_FILE/
[dmdba@Standby bin]$ ./dmrman
RMAN> check backupset '/dm8/dmbak/BACKUP_FILE' ;
RMAN> restore database '/dm8/data/DSC/dm.ini' from backupset
'/dm8/dmbak/BACKUP_FILE';
RMAN> recover database '/dm8/data/DSC/dm.ini' from backupset
'/dm8/dmbak/BACKUP_FILE';
RMAN> recover database '/dm8/data/DSC/dm.ini' update DB_MAGIC;

3.5 修改 dm.ini 文件

Node0、Node1 和备库节点都需执行
配置相同

修改如下参数项使所有节点保持一致即可

ALTER_MODE_STATUS = 0
ENABLE_OFFLINE_TS = 2
MAL_INI = 1
ARCH_INI = 1

3.6 修改/创建 dmmal.ini 文件

Node0、Node1 和备库节点都需执行
配置相同

在集群的 dmmal.ini 文件添加备机节点的信息即可，将此文件发送到备机节点

[dmdba@Node0 DSC0_config]$ vim dmmal.ini
[dmdba@Node0 DSC0_config]$ cat dmmal.ini
MAL_CHECK_INTERVAL=87
MAL_CONN_FAIL_INTERVAL=180
MAL_SYS_BUF_SIZE=600
MAL_BUF_SIZE=300
MAL_VPOOL_SIZE=500
MAL_COMPRESS_LEVEL=0
[MAL_INST0]
 MAL_INST_NAME = DSC0
 MAL_HOST = 192.168.223.100
 MAL_PORT = 9340
 MAL_INST_HOST = 192.168.223.100
 MAL_INST_PORT = 5236
 MAL_DW_PORT = 52141
 MAL_INST_DW_PORT = 5276
[MAL_INST1]
 MAL_INST_NAME = DSC1
 MAL_HOST = 192.168.223.110
 MAL_PORT = 9340
 MAL_INST_HOST = 192.168.223.110
 MAL_INST_PORT = 5236
 MAL_DW_PORT = 52141
 MAL_INST_DW_PORT = 5276
[MAL_INST2]
 MAL_INST_NAME = DW
 MAL_HOST = 192.168.223.6
 MAL_PORT = 9340
 MAL_INST_HOST = 192.168.223.6
 MAL_INST_PORT = 5236
 MAL_DW_PORT = 52141
MAL_INST_DW_PORT = 5276
[dmdba@Node0 DSC0_config]$ scp dmmal.ini dmdba@192.168.223.6:/dm8/data/DSC/
[dmdba@Node0 DSC0_config]$ scp dmmal.ini
dmdba@192.168.223.6:/dm8/conf/DSC1_config

3.7 创建 dmarch.ini 文件

Node0、Node1 和备库节点都需执行
配置不同

Node0

[dmdba@Node0 DSC0_config]$ cat dmarch.ini
ARCH_WAIT_APPLY = 0
ARCH_LOCAL_SHARE = 1
ARCH_LOCAL_SHARE_CHECK=0
[ARCHIVE_LOCAL1]
ARCH_TYPE = LOCAL
ARCH_DEST = +DGARCH01/DSC0/ARCH
ARCH_FILE_SIZE = 2048
ARCH_SPACE_LIMIT = 102400
[ARCHIVE_REMOTE1]
ARCH_TYPE = REMOTE
ARCH_DEST = DSC1
ARCH_INCOMING_PATH = +DGARCH01/DSC1/ARCH
ARCH_FILE_SIZE = 2048
ARCH_SPACE_LIMIT = 102400
[ARCHIVE_REALTIME1]
ARCH_TYPE = REALTIME
ARCH_DEST = DW

Node1

[dmdba@Node1 DSC1_config]$ cat dmarch.ini
ARCH_WAIT_APPLY = 0
ARCH_LOCAL_SHARE = 1
ARCH_LOCAL_SHARE_CHECK=0
[ARCHIVE_LOCAL1]
ARCH_TYPE = LOCAL
ARCH_DEST = +DGARCH01/DSC1/ARCH
ARCH_FILE_SIZE = 2048
ARCH_SPACE_LIMIT = 102400
[ARCHIVE_REMOTE1]
ARCH_TYPE = REMOTE
ARCH_DEST = DSC0
ARCH_INCOMING_PATH = +DGARCH01/DSC0/ARCH
ARCH_FILE_SIZE = 2048
ARCH_SPACE_LIMIT = 102400
[ARCHIVE_REALTIME1]
ARCH_TYPE = REALTIME
ARCH_DEST = DW

Standby

[dmdba@Standby DSC]$ cat dmarch.ini
[ARCHIVE_REALTIME]
ARCH_TYPE = REALTIME
ARCH_DEST = DSC0/DSC1
[ARCHIVE_LOCAL1]
ARCH_TYPE = LOCAL
ARCH_DEST = /dm8/dmarch
ARCH_FILE_SIZE = 2048
ARCH_SPACE_LIMIT = 204800

3.8 创建 dmwatcher.ini 文件并注册服务

Node0、Node1 和备库节点都需执行
配置不同

Node0

[dmdba@Node0 DSC0_config]$ vim dmwatcher.ini
[dmdba@Node0 DSC0_config]$ cat dmwatcher.ini
[GRP1]
DW_TYPE = GLOBAL
DW_MODE = MANUAL
DW_ERROR_TIME = 120
INST_RECOVER_TIME = 60
INST_ERROR_TIME = 120
INST_OGUID = 453331
INST_INI = /dm8/conf/DSC0_config/dm.ini
DCR_INI= /dm8/conf/dmdcr.ini
INST_STARTUP_CMD = /dm8/bin/DmServiceDSC0 start
INST_AUTO_RESTART = 0
RLOG_SEND_THRESHOLD = 0
RLOG_APPLY_THRESHOLD = 0

注册服务

[root@Node0 root]# ./dm_service_installer.sh -t dmwatcher -p DSC0 -watcher_ini
/dm8/conf/DSC0_config/dmwatcher.ini

Node1

[dmdba@Node1 DSC1_config]$ vim dmwatcher.ini
[dmdba@Node1 DSC1_config]$ cat dmwatcher.ini
[GRP1]
DW_TYPE = GLOBAL
DW_MODE = MANUAL
DW_ERROR_TIME = 120
INST_RECOVER_TIME = 60
INST_ERROR_TIME = 120
INST_OGUID = 453331
INST_INI = /dm8/conf/DSC1_config/dm.ini
DCR_INI= /dm8/conf/dmdcr.ini
INST_STARTUP_CMD = /dm8/bin/DmServiceDSC1 start
INST_AUTO_RESTART = 0
RLOG_SEND_THRESHOLD = 0
RLOG_APPLY_THRESHOLD = 0

注册服务

[root@Node1 root]# ./dm_service_installer.sh -t dmwatcher -p DSC1 -watcher_ini
/dm8/conf/DSC1_config/dmwatcher.ini

Standby

[dmdba@Standby DSC]$ vim dmwatcher.ini
[dmdba@Standby DSC]$ cat dmwatcher.ini
[GRP1]
DW_TYPE = GLOBAL
DW_MODE = MANUAL
DW_ERROR_TIME = 120
INST_RECOVER_TIME = 60
INST_ERROR_TIME = 120
INST_OGUID = 453331
INST_INI = /dm8/data/DSC/dm.ini
INST_AUTO_RESTART = 0
INST_STARTUP_CMD = /dm8/bin/dmserver
RLOG_SEND_THRESHOLD = 0
RLOG_APPLY_THRESHOLD = 0

注册服务

[root@Standby root]# ./dm_service_installer.sh -t dmwatcher -p DW -watcher_ini
/dm8/data/DSC/dmwatcher.ini

3.9 创建 dmmonitor.ini 文件

Node0、Node1 或 Standby 节点执行

[dmdba@Node0 DSC0_config]$ vim dmmonitor
[dmdba@Node0 DSC0_config]$ cat dmmonitor
MON_DW_CONFIRM = 1
MON_LOG_PATH = /dm8/dmmlog
MON_LOG_INTERVAL = 0
MON_LOG_FILE_SIZE = 32
MON_LOG_SPACE_LIMIT = 0
[GRP1]
MON_INST_OGUID = 453331
MON_DW_IP = 192.168.223.100:52141/192.168.223.110:52141
MON_DW_IP = 192.168.223.6:52141

3.10 启动数据库服务并修改参数（mount）

Node0、Node1 和备库节点都需执行
步骤相同

Node0

[dmdba@Node0 bin]$ ./DmServiceDSC0 start mount
[dmdba@Node0 bin]$ ./disql sysdba/SYSDBA
Server[LOCALHOST:5236]:mode is normal, state is mount
SQL> sp_set_oguid(453331) ;
DMSQL executed successfully
used time: 101.363(ms). Execute id is 0.
SQL> alter database primary ;
executed successfully
used time: 86.672(ms). Execute id is 0.

Node1

[dmdba@Node1 bin]$ ./DmServiceDSC1 start mount
[dmdba@Node1 bin]$ ./disql sysdba/SYSDBA
Server[LOCALHOST:5236]:mode is primary, state is mount
login used time : 24.262(ms)
disql V8
SQL> sp_set_oguid(453331) ;
sp_set_oguid(453331) ;
[-720]:Dmwatcher is active, or current configuration(ALTER_MODE_STATUS) not allowed to
alter database.
used time: 18.933(ms). Execute id is 0.

注：Node2 已自动变成主库，无需修改任何参数

Standby

[dmdba@Standby bin]$ ./DmServiceStandby start mount
[dmdba@Standby bin]$ ./disql sysdba/SYSDBA
Server[LOCALHOST:5236]:mode is normal, state is mount
SQL> sp_set_oguid(453331) ;
DMSQL executed successfully
used time: 8.690(ms). Execute id is 0.
SQL> alter database standby ;
executed successfully
used time: 9.400(ms). Execute id is 0.

3.11 启动守护进程

Node0、Node1 和备库节点都需执行
步骤相同

[dmdba@Node0 bin]$ ./DmWatcherServiceDSC0 start
[dmdba@Node1 bin]$ ./DmWatcherServiceDSC1 start
[dmdba@Standby bin]$ ./DmWatcherServiceDW start

3.12 在 cssm 监视器中查看信息

见 2.7.11 启动 cssm 监视器（前台方式）

3.13 启动 monitor 监视器

Node0 节点执行

[dmdba@Node0 bin]$ ./dmmonitor /dm8/conf/DSC0_config/dmmonitor.ini

输入 show 命令查看集群和备库信息，所有节点数据库状态已从 mount 状态变成 open 状态
代表实时备库搭建成功